Oct. 29, 2023
RE: File No. S7-04-23 You claim to support anonymous comments, yet you have red stars next to the email address and phone number forms. Surely you see how that is antithetical to anonymity? Your form mutilates peoples comments by removing all the line breaks, harming the intent of their feedback and making it harder for you to understand. Your website has a pitifully small maximum file size limit of 12MB total. It's the year 2023 guys, and I know you're funded with hundreds of millions of dollars, if not billions. Your rule also affects billions of dollars of congress. Pitiful. You should call up your colleagues over at the federal register and let them know that this rule concerning cryptocurrencies doesn't even show up if you search for "cryptocurrency." Specifically: https://federalregister.gov/documents/2023/08/30/2023-18667/safeguarding-advisory-client-assets-reopening-of-comment-period#addresses The first possible APA violation. Searching cryptocurrency, cryptocurrencies, virtual currency or virtual currencies in that box doesn’t bring up all the important and currently open for comment results. Saying you want comments, and then hiding the issues from people searching for them; misleading them that they don’t exist, is perhaps a violation of the APA. How is anyone going to have faith in your agency when you can't even follow your own rules properly? Here's the rule. 5 U.S. Code § 553 - Rule making U.S. Code Notes prev | next (a)This section applies, according to the provisions thereof, except to the extent that there is involved— (1)a military or foreign affairs function of the United States; or (2)a matter relating to agency management or personnel or to public property, loans, grants, benefits, or contracts. (b)General notice of proposed rule making shall be published in the Federal Register, unless persons subject thereto are named and either personally served or otherwise have actual notice thereof in accordance with law. The notice shall include— (1)a statement of the time, place, and nature of public rule making proceedings; (2)reference to the legal authority under which the rule is proposed; and (3)either the terms or substance of the proposed rule or a description of the subjects and issues involved. Except when notice or hearing is required by statute, this subsection does not apply— (A)to interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice; or (B)when the agency for good cause finds (and incorporates the finding and a brief statement of reasons therefor in the rules issued) that notice and public procedure thereon are impracticable, unnecessary, or contrary to the public interest. (c)After notice required by this section, the agency shall give interested persons an opportunity to participate in the rule making through submission of written data, views, or arguments with or without opportunity for oral presentation. After consideration of the relevant matter presented, the agency shall incorporate in the rules adopted a concise general statement of their basis and purpose. When rules are required by statute to be made on the record after opportunity for an agency hearing, sections 556 and 557 of this title apply instead of this subsection. (d)The required publication or service of a substantive rule shall be made not less than 30 days before its effective date, except— (1)a substantive rule which grants or recognizes an exemption or relieves a restriction; (2)interpretative rules and statements of policy; or (3)as otherwise provided by the agency for good cause found and published with the rule. (e)Each agency shall give an interested person the right to petition for the issuance, amendment, or repeal of a rule. (Pub. L. 89–554, Sept. 6, 1966, 80 Stat. 383.) Judicial case law has created further requirements of you, the SEC, but as you're getting paid to read this, and I'm not getting paid to write it, I don't really feel like looking for them right now. How is the pay over there by the way? You guys make lots more than the judicial branch does right? Does that create animosity when you are in front of the court, or do you usually just play pretend court with your own internal judges? Speaking of which, here is that ruling that found your internal courts unconstitutional. United States Court of Appeals for the Fifth Circuit No. 20-61007 George R. Jarkesy, Jr.; Patriot28, L.L.C., Petitioners, versus Securities and Exchange Commission, Respondent. Petition for Review of an Order of the United States Securities and Exchange Commission No. 3-15255 Before Davis, Elrod, and Oldham, Circuit Judges. Jennifer Walker Elrod, Circuit Judge: Congress has given the Securities and Exchange Commission substantial power to enforce the nation’s securities laws. It often acts as both prosecutor and judge, and its decisions have broad consequences for personal liberty and property. But the Constitution constrains the SEC’s powers by protecting individual rights and the prerogatives of the other branches of government. This case is about the nature and extent of those constraints in securities fraud cases in which the SEC seeks penalties. United States Court of Appeals Fifth Circuit FILED May 18, 2022 Lyle W. Cayce Clerk Case: 20-61007 Document: 00516323784 Page: 1 Date Filed: 05/18/2022 No. 20-61007 2 The SEC brought an enforcement action within the agency against Petitioners for securities fraud. An SEC administrative law judge adjudged Petitioners liable and ordered various remedies, and the SEC affirmed on appeal over several constitutional arguments that Petitioners raised. Petitioners raise those same arguments before this court. We hold that: (1) the SEC’s in-house adjudication of Petitioners’ case violated their Seventh Amendment right to a jury trial; (2) Congress unconstitutionally delegated legislative power to the SEC by failing to provide an intelligible principle by which the SEC would exercise the delegated power, in violation of Article I’s vesting of “all” legislative power in Congress; and (3) statutory removal restrictions on SEC ALJs violate the Take Care Clause of Article II. Because the agency proceedings below were unconstitutional, we GRANT the petition for review, VACATE the decision of the SEC, and REMAND for further proceedings consistent with this opinion. I. Petitioner Jarkesy established two hedge funds and selected Petitioner Patriot28 as the investment adviser. The funds brought in over 100 investors and held about $24 million in assets. In 2011, the SEC launched an investigation into Petitioners’ investing activities, and a couple of years later the SEC chose to bring an action within the agency, alleging that Petitioners (along with some former co-parties) committed fraud under the Securities Act, the Securities Exchange Act, and the Advisers Act. Specifically, the agency charged that Petitioners: (1) misrepresented who served as the prime broker and as the auditor; (2) misrepresented the funds’ investment parameters and safeguards; and (3) overvalued the funds’ assets to increase the fees that they could charge investors. Petitioners sued in the U.S. District Court for the District of Columbia to enjoin the agency proceedings, arguing that the proceedings infringed on Case: 20-61007 Document: 00516323784 Page: 2 Date Filed: 05/18/2022 No. 20-61007 3 various constitutional rights. But the district court, and later the U.S. Court of Appeals for the D.C. Circuit, refused to issue an injunction, deciding that the district court had no jurisdiction and that Petitioners had to continue with the agency proceedings and petition the court of appeals to review any adverse final order. See Jarkesy v. SEC, 48 F. Supp. 3d 32, 40 (D.D.C. 2014), aff’d, 803 F.3d 9, 12 (D.C. Cir. 2015). Petitioners’ proceedings moved forward. The ALJ held an evidentiary hearing and concluded that Petitioners committed securities fraud. Petitioners then sought review by the Commission. While their petition for Commission review was pending, the Supreme Court held that SEC ALJs had not been properly appointed under the Constitution. Lucia v. SEC, 138 S. Ct. 2044, 2054–55 (2018). In accordance with that decision, the SEC assigned Petitioners’ proceeding to an ALJ who was properly appointed. But Petitioners chose to waive their right to a new hearing and continued under their original petition to the Commission. The Commission affirmed that Petitioners committed various forms of securities fraud. It ordered Petitioners to cease and desist from committing further violations and to pay a civil penalty of $300,000, and it ordered Patriot28 to disgorge nearly $685,000 in ill-gotten gains. The Commission also barred Jarkesy from various securities industry activities: associating with brokers, dealers, and advisers; offering penny stocks; and serving as an officer or director of an advisory board or as an investment adviser. Critical to this case, the Commission rejected several constitutional arguments Petitioners raised. It determined that: (1) the ALJ was not biased against Petitioners; (2) the Commission did not inappropriately prejudge the case; (3) the Commission did not use unconstitutionally delegated legislative power—or violate Petitioners’ equal protection rights—when it decided to Case: 20-61007 Document: 00516323784 Page: 3 Date Filed: 05/18/2022 No. 20-61007 4 pursue the case within the agency instead of in an Article III court; (4) the removal restrictions on SEC ALJs did not violate Article II and separationof-powers principles; and (5) the proceedings did not violate Petitioners’ Seventh Amendment right to a jury trial. Petitioners then filed a petition for review in this court. II. Petitioners raise several constitutional challenges to the SEC enforcement proceedings.1 We agree with Petitioners that the proceedings suffered from three independent constitutional defects: (1) Petitioners were deprived of their constitutional right to a jury trial; (2) Congress unconstitutionally delegated legislative power to the SEC by failing to provide it with an intelligible principle by which to exercise the delegated power; and (3) statutory removal restrictions on SEC ALJs violate Article II. A. Petitioners challenge the agency’s rejection of their constitutional arguments. We review such issues de novo. See Emp. Sols. Staffing Grp. II, L.L.C. v. Off. of Chief Admin. Hearing Officer, 833 F.3d 480, 484 (5th Cir. 2016); Trinity Marine Prods., Inc. v. Chao, 512 F.3d 198, 201 (5th Cir. 2007). B. Petitioners argue that they were deprived of their Seventh Amendment right to a jury trial. The SEC responds that the legal interests at issue in this case vindicate distinctly public rights, and that Congress therefore appropriately allowed such actions to be brought in agency 1 Multiple amici have filed briefs with this court as well: the Cato Institute, Phillip Goldstein, Mark Cuban, Nelson Obus, and the New Civil Liberties Alliance. Each argues that the SEC proceedings exceeded constitutional limitations for reasons that Petitioners raise. Case: 20-61007 Document: 00516323784 Page: 4 Date Filed: 05/18/2022 No. 20-61007 5 proceedings without juries. We agree with Petitioners. The Seventh Amendment guarantees Petitioners a jury trial because the SEC’s enforcement action is akin to traditional actions at law to which the jury-trial right attaches. And Congress, or an agency acting pursuant to congressional authorization, cannot assign the adjudication of such claims to an agency because such claims do not concern public rights alone. 1. Thomas Jefferson identified the jury “as the only anchor, ever yet imagined by man, by which a government can be held to the principles of its constitution.” Letter from Thomas Jefferson to Thomas Paine (July 11, 1789), in The Papers of Thomas Jefferson 267 (Julian P. Boyd ed., 1958). And John Adams called trial by jury (along with popular elections) “the heart and lungs of liberty.” The Revolutionary Writings of John Adams 55 (C. Bradley Thompson ed., 2000); see also Jennifer W. Elrod, Is the Jury Still Out?: A Case for the Continued Viability of the American Jury, 44 Tex. Tech L. Rev. 303, 303–04 (2012) (explaining that the jury is “as central to the American conception of the consent of the governed as an elected legislature or the independent judiciary”). 2 2 Veneration of the jury as safeguard of liberty predates the American Founding. Our inherited English common-law tradition has long extolled the jury as an institution. William Blackstone said that trial by jury is “the glory of the English law” and “the most transcendent privilege which any subject can enjoy or wish for, that he cannot be affected, either in his property, his liberty, or his person, but by the unanimous consent of twelve of his neighbors and equals.” Mitchell v. Harmony, 54 U.S. 115, 142–43 (1851) (quoting 4 William Blackstone, Commentaries on the Laws of England 227–29 (Oxford, Clarendon Pr. 1992) (1765)); see also Jennifer W. Elrod, W(h)ither The Jury? The Diminishing Role of the Jury Trial in Our Legal System, 68 Wash. & Lee L. Rev. 3, 7 (2011). Indeed, King George III’s attempts to strip colonists of their right to trial by jury was one of the chief grievances aired against him and was a catalyst for declaring independence. The Declaration of Independence para. 20 (U.S. 1776). Case: 20-61007 Document: 00516323784 Page: 5 Date Filed: 05/18/2022 No. 20-61007 6 Civil juries in particular have long served as a critical check on government power. So precious were civil juries at the time of the Founding that the Constitution likely would not have been ratified absent assurance that the institution would be protected expressly by amendment. 2 The Debate on the Constitution 549, 551, 555, 560, 567 (Bernard Bailyn ed. 1993) (collecting various state ratification convention documents calling for the adoption of a civil jury trial amendment); The Federalist No. 83 (Alexander Hamilton) (“The objection to the plan of the convention, which has met with most success in this State [i.e., New York], and perhaps in several of the other States, is that relative to the want of a constitutional provision for the trial by jury in civil cases.”); Mercy Otis Warren, Observations on the Constitution (1788), in 2 The Debate on the Constitution 290 (Bernard Bailyn ed. 1993) (worrying that the unamended Constitution would lead to “[t]he abolition of trial by jury in civil causes”); Parsons v. Bedford, 28 U.S. (3 Pet.) 433, 446 (1830) (“One of the strongest objections originally taken against the constitution of the United States, was the want of an express provision securing the right of trial by jury in civil cases.”). 3 Trial by jury therefore is a “fundamental” component of our legal system “and remains one of our most vital barriers to governmental arbitrariness.” Reid v. Covert, 354 U.S. 1, 9–10 (1957). “Indeed, ‘[t]he right to trial by jury was probably the only one universally secured by the first American state constitutions . . . .’” Parklane Hosiery Co., Inc. v. Shore, 439 3 See also Kenneth Klein, The Validity of The Public Rights Doctrine in Light of the Historical Rationale of the Seventh Amendment, 21 Hastings Const. L.Q. 1013, 1015 (1994) (“At the time the Constitution was proposed, the people of the United States greatly distrusted government, and saw the absence of a guaranteed civil jury right as a reason, standing alone, to reject adoption of the Constitution; only by promising the Seventh Amendment did the Federalists secure adoption of the Constitution in several of the state ratification debates.”). Case: 20-61007 Document: 00516323784 Page: 6 Date Filed: 05/18/2022 No. 20-61007 7 U.S. 322, 341 (1979) (Rehnquist, J., dissenting) (quoting Leonard Levy, Legacy of Suppression: Freedom of Speech and Press in Early American History 281 (1960)). Because “[m]aintenance of the jury as a fact-finding body is of such importance and occupies so firm a place in our history and jurisprudence[,] . . . any seeming curtailment of the right to a jury trial should be scrutinized with the utmost care.” Dimick v. Schiedt, 293 U.S. 474, 486 (1935). The Seventh Amendment protects that right. It provides that “[i]n Suits at common law, where the value in controversy shall exceed twenty dollars, the right of trial by jury shall be preserved, and no fact tried by a jury, shall be otherwise reexamined in any Court of the United States, than according to the rules of the common law.” U.S. Const. amend. VII. The Supreme Court has interpreted “Suits at common law” to include all actions akin to those brought at common law as those actions were understood at the time of the Seventh Amendment’s adoption. Tull v. United States, 481 U.S. 412, 417 (1987). The term can include suits brought under a statute as long as the suit seeks common-law-like legal remedies. Id. at 418–19. And the Court has specifically held that, under this standard, the Seventh Amendment jury-trial right applies to suits brought under a statute seeking civil penalties. Id. at 418–24. That is not to say, however, that Congress may never assign adjudications to agency processes that exclude a jury. See Atlas Roofing Co. v. Occupational Safety & Health Rev. Comm’n, 430 U.S. 442, 455 (1977). “[W]hen Congress properly assigns a matter to adjudication in a non-Article III tribunal, the Seventh Amendment poses no independent bar to the adjudication of that action by a nonjury factfinder.” Oil States Energy Servs., LLC v. Greene’s Energy Grp., LLC, 138 S. Ct. 1365, 1379 (2018) (internal quotations omitted). Case: 20-61007 Document: 00516323784 Page: 7 Date Filed: 05/18/2022 No. 20-61007 8 Whether Congress may properly assign an action to administrative adjudication depends on whether the proceedings center on “public rights.” Atlas Roofing, 430 U.S. at 450. “[I]n cases in which ‘public rights’ are being litigated[,] e.g., cases in which the Government sues in its sovereign capacity to enforce public rights created by statutes within the power of Congress to enact[,] the Seventh Amendment does not prohibit Congress from assigning the factfinding function and initial adjudication to an administrative forum with which the jury would be incompatible.” Id. Describing proper assignments, the Supreme Court identified situations “where the Government is involved in its sovereign capacity under an otherwise valid statute creating enforceable public rights. Wholly private tort, contract, and property cases, [and] a vast range of other cases as well are not at all implicated.” Id. at 458. The Supreme Court refined the public-right concept as it relates to the Seventh Amendment in Granfinanciera, S.A. v. Nordberg, 492 U.S. 33 (1989). There, the Court clarified that Congress cannot circumvent the Seventh Amendment jury-trial right simply by passing a statute that assigns “traditional legal claims” to an administrative tribunal. Id. at 52. Public rights, the Court explained, arise when Congress passes a statute under its constitutional authority that creates a right so closely integrated with a comprehensive regulatory scheme that the right is appropriate for agency resolution. Id. at 54. The analysis thus moves in two stages. First, a court must determine whether an action’s claims arise “at common law” under the Seventh Amendment. See Tull, 481 U.S. at 417. Second, if the action involves common-law claims, a court must determine whether the Supreme Court’s public-rights cases nonetheless permit Congress to assign it to agency adjudication without a jury trial. See Granfinanciera, 492 U.S. at 54; Atlas Roofing, 430 U.S. at 455. Here, the relevant considerations include: Case: 20-61007 Document: 00516323784 Page: 8 Date Filed: 05/18/2022 No. 20-61007 9 (1) whether “Congress ‘creat[ed] a new cause of action, and remedies therefor, unknown to the common law,’ because traditional rights and remedies were inadequate to cope with a manifest public problem”; and (2) whether jury trials would “go far to dismantle the statutory scheme” or “impede swift resolution” of the claims created by statute. Granfinanciera, 492 U.S. at 60–63 (quoting Atlas Roofing, 430 U.S. at 454 n.11, 461 (first and second quotations)). 2. The rights that the SEC sought to vindicate in its enforcement action here arise “at common law” under the Seventh Amendment. Fraud prosecutions were regularly brought in English courts at common law. See 3 William Blackstone, Commentaries on the Laws of England *42 (explaining the common-law courts’ jurisdiction over “actions on the case which allege any falsity or fraud; all of which savour of a criminal nature, although the action is brought for a civil remedy; and make the defendant liable in strictness to pay a fine to the king, as well as damages to the injured party”). And even more pointedly, the Supreme Court has held that actions seeking civil penalties are akin to special types of actions in debt from early in our nation’s history which were distinctly legal claims. Tull, 481 U.S. at 418–19. Thus, “[a] civil penalty was a type of remedy at common law that could only be enforced in courts of law.” Id. at 422. Applying that principle, the Court in Tull held that the right to a jury trial applied to an action brought by an agency seeking civil penalties for violations of the Clean Water Act. Id. at 425. Likewise here, the actions the SEC brought seeking civil penalties under securities statutes are akin to those same traditional actions in debt. Under the Seventh Amendment, both as originally understood and as interpreted by the Supreme Court, the jury-trial right applies to the penalties action the SEC brought in this case. Case: 20-61007 Document: 00516323784 Page: 9 Date Filed: 05/18/2022 No. 20-61007 10 That conclusion harmonizes with the holdings of other courts applying Tull. The Seventh Circuit followed the Supreme Court’s lead in that case and has specifically said that when the SEC brings an enforcement action to obtain civil penalties under a statute, the subject of the action has the right to a jury trial. SEC v. Lipson, 278 F.3d 656, 662 (7th Cir. 2002) (“Because the SEC was seeking both legal and equitable relief (the former under the Insider Trading Sanctions Act, 15 U.S.C. § 78u–1, which (in subsection (a)(1)) authorizes the imposition of civil penalties for insider trading at the suit of the SEC[)] . . . [the defendant] was entitled to and received a jury trial.”); see also id. (explaining that another circuit was wrong to tacitly assume “that civil penalties in SEC cases are not a form of legal relief”4 ). Some district courts have applied Tull similarly. See, e.g., SEC v. Badian, 822 F. Supp. 2d 352, 365 (S.D.N.Y. 2011) (explaining that “whether the facts are such that the defendants can be subjected to a civil penalty . . . is a question for the jury, [and] the determination of the severity of the civil penalty to be imposed . . . is a question for the Court, once liability is established”); SEC v. Solow, 554 F. Supp. 2d 1356, 1367 (S.D. Fla. 2008) (applying Tull for the proposition that civil penalties are “legal, as opposed to equitable, in nature,” and that it therefore “was [the defendant’s] constitutional right to have a jury determine his liability, with [the court] thereafter determining the amount of penalty, if any”). Other elements of the action brought by the SEC against Petitioners are more equitable in nature, but that fact does not invalidate the jury-trial right that attaches because of the civil penalties sought. The Supreme Court has held that the Seventh Amendment applies to proceedings that involve a mix of legal and equitable claims—the facts relevant to the legal claims 4 The Seventh Circuit was referring to the Ninth Circuit’s opinion in SEC v. Clark, 915 F.2d 439, 442 (9th Cir. 1990). Clark did not address the issue whatsoever. Case: 20-61007 Document: 00516323784 Page: 10 Date Filed: 05/18/2022 No. 20-61007 11 should be adjudicated by a jury, even if those facts relate to equitable claims too. See Ross v. Bernhard, 396 U.S. 531, 537–38 (1970); see also Lipson, 278 F.3d at 662 (noting that the defendant was entitled to a jury trial because the SEC sought legal relief in the form of penalties, even though the SEC also sought equitable relief). Here, the SEC sought to ban Jarkesy from participation in securities industry activities and to require Patriot28 to disgorge ill-gotten gains—both equitable remedies. Even so, the penalty facet of the action suffices for the jury-trial right to apply to an adjudication of the underlying facts supporting fraud liability. 3. Next, the action the SEC brought against Petitioners is not the sort that may be properly assigned to agency adjudication under the public-rights doctrine. Securities fraud actions are not new actions unknown to the common law. Jury trials in securities fraud suits would not “dismantle the statutory scheme” addressing securities fraud or “impede swift resolution” of the SEC’s fraud prosecutions. And such suits are not uniquely suited for agency adjudication. Common-law courts have heard fraud actions for centuries, even actions brought by the government for fines. See Blackstone, supra at *42; see also Tull, 481 U.S. at 422 (“A civil penalty was a type of remedy at common law that could only be enforced in courts of law.”). Naturally, then, the securities statutes at play in this case created causes of action that reflect common-law fraud actions. The traditional elements of common-law fraud are (1) a knowing or reckless material misrepresentation, (2) that the tortfeasor intended to act on, and (3) that harmed the plaintiff. In re Deepwater Horizon, 857 F.3d 246, 249 (5th Cir. 2017). The statutes under which the SEC brought securities fraud actions use terms like “fraud” and “untrue statement[s] of material fact” to describe the prohibited conduct. Case: 20-61007 Document: 00516323784 Page: 11 Date Filed: 05/18/2022 No. 20-61007 12 See 15 U.S.C. §§ 77a–77aa, 78j(b), 80b-6. When “Congress uses terms that have accumulated settled meaning under . . . the common law, a court must infer, unless the statute otherwise dictates, that Congress means to incorporate the established meaning of these terms.” Nationwide Mut. Ins. Co. v. Darden, 503 U.S. 318, 322 (1992) (quoting Cmty. for Creative NonViolence v. Reid, 490 U.S. 730, 739 (1989)); see also Felix Frankfurter, Some Reflections on the Reading of Statutes, 47 Colum. L. Rev. 527, 537 (1947) (explaining that “if a word is obviously transplanted from another legal source, whether the common law or other legislation, it brings the old soil with it”). Accordingly, the Supreme Court has often looked to common-law principles to interpret fraud and misrepresentation under securities statutes. See, e.g, Omnicare, Inc. v. Laborers Dist. Council Indus. Pension Fund, 575 U.S. 175, 191 (2015) (considering the Restatement (Second) of Torts to determine whether material omissions are actionable under a securities statute); Dura Pharms., Inc. v. Broudo, 544 U.S. 336, 343–44 (2005) (relying on “the common-law roots of the securities fraud action” in “common-law deceit and misrepresentation actions” to interpret the statutory securities-fraud action); SEC v. Cap. Gains Rsch. Bureau, 375 U.S. 180, 192–95 (1963) (considering the principles of common-law fraud to determine the requirements of fraud under the Advisers Act). Thus, fraud actions under the securities statutes echo actions that historically have been available under the common law. Next, jury trials would not “go far to dismantle the statutory scheme” or “impede swift resolution” of the statutory claims. See Granfinanciera, 492 U.S. at 60–63. For one, the statutory scheme itself allows the SEC to bring enforcement actions either in-house or in Article III courts, where the jurytrial right would apply. See Dodd–Frank Act § 929P(a), 15 U.S.C. § 78u-2(a). If Congress has not prevented the SEC from bringing claims in Article III Case: 20-61007 Document: 00516323784 Page: 12 Date Filed: 05/18/2022 No. 20-61007 13 courts with juries as often as it sees fit to do so, and if the SEC has in fact brought many such actions to jury trial over the years,5 then it is difficult to see how jury trials could “dismantle the statutory scheme.” Congress could have purported to assign such proceedings solely to administrative tribunals, but it did not. And there also is no evidence that jury trials would impede swift resolution of the claims.6 In this case, for example, the SEC took seven years to dispose of Petitioners’ case and makes no argument that proceedings with a jury trial would have been less efficient. Relatedly, securities-fraud enforcement actions are not the sort that are uniquely suited for agency adjudication. Again, Congress has not limited the SEC’s ability to bring enforcement actions in Article III courts. Consider the statutory scheme in Atlas Roofing for contrast. The statutes in that case were new and somewhat unusual. They provided elaborate enforcement mechanisms for the sorts of claims that likely could not have been brought in legal actions before that point. See Atlas Roofing, 430 U.S. at 445 (describing how the statutes required factfinders to undertake detailed assessments of workplace safety conditions and to make unsafe-conditions findings even if no injury had occurred). But the federal courts have dealt with actions under 5 Indeed, the SEC regularly brings securities-fraud actions in Article III courts and adjudicates them through jury trials. See, e.g., SEC v. Fowler, 6 F.4th 255, 258–60 (2d Cir. 2021); SEC v. Johnston, 986 F.3d 63, 71 (1st Cir. 2021); SEC v. Life Partners Holdings, Inc., 854 F.3d 765, 772 (5th Cir. 2017); SEC v. Quan, 817 F.3d 583, 587 (8th Cir. 2016); SEC v. Miller, 808 F.3d 623, 626 (2d Cir. 2015); SEC v. Jasper, 678 F.3d 1116, 1119, 1121–22 (9th Cir. 2012); SEC v. Seghers, 298 F. App’x 319, 321 (5th Cir. 2008). 6 The dissenting opinion contends that these considerations are “not decisive” (that the SEC has for decades sued in Article III courts under securities statutes) or “not determinative” (that those same suits are not unique to agency adjudication). To disregard these facts is to ignore the Supreme Court’s explanation for what public rights are made of. And in any event, though the facts may not in isolation make up a private right, they together establish (along with the other considerations discussed above) that the right being vindicated here is a private right, not a public one. Case: 20-61007 Document: 00516323784 Page: 13 Date Filed: 05/18/2022 No. 20-61007 14 the securities statutes for many decades, and there is no reason to believe that such courts are suddenly incapable of continuing that work just because an agency may now share some of the workload. In fact, for the first decades of the SEC’s existence, securities-fraud actions against nonregistered parties could be brought only in Article III courts. Thomas Glassman, Ice Skating Uphill: Constitutional Challenges to SEC Administrative Proceedings, 16 J. Bus. & Sec. L. 47, 50–52 (2015).7 The SEC counters that the securities statutes are designed to protect the public at large, and that some circuits have identified SEC enforcement actions as vindicating rights on behalf of the public. Indeed, the SEC says, the statutes allow for enforcement proceedings based on theories broader than actions like fraud that existed at common law. Those facts do not convert the SEC’s action into one focused on public rights. Surely Congress believes that the securities statutes it passes serve the public interest and the U.S. economy overall, not just individual parties. Yet Congress cannot convert any sort of action into a “public right” simply by finding a public purpose for it and codifying it in federal statutory law. See Granfinanciera, 492 U.S. at 61 (explaining that “Congress cannot eliminate a party’s Seventh Amendment right to a jury trial merely by relabeling the cause of action to which it attaches and placing exclusive jurisdiction in an administrative agency or a specialized court of equity”). Purely private suits for securities fraud likely would have a similar public purpose—they too would serve to discourage and remedy fraudulent 7 Moreover, the Supreme Court has noted that agency adjudicators generally do not have special expertise to address structural constitutional claims—precisely the issues central to this case. Carr v. Saul, 141 S. Ct. 1352, 1360 (2021) (“[T]his Court has often observed that agency adjudications are generally ill suited to address structural constitutional challenges, which usually fall outside the adjudicators’ areas of technical expertise.”). Case: 20-61007 Document: 00516323784 Page: 14 Date Filed: 05/18/2022 No. 20-61007 15 behavior in securities markets. That does not mean such suits concern public rights at their core. Granted, some actions provided for by the securities statutes may be new and not rooted in any common-law corollary. The fact remains, though, that the enforcement action seeking penalties in this case was one for securities fraud, which is nothing new and nothing foreign to Article III tribunals and juries. That being so, Petitioners had the right for a jury to adjudicate the facts underlying any potential fraud liability that justifies penalties. And because those facts would potentially support not only the civil penalties sought by the SEC, but the injunctive remedies as well, Petitioners had a Seventh Amendment right to a jury trial for the liability-determination portion of their case. 4. The dissenting opinion cannot define a “public right” without using the term itself in the definition. That leads to a good bit of question-begging. It says at times that the “SEC’s enforcement action” is itself “a ‘public right’ because it is a case ‘in which the Government sues in its sovereign capacity to enforce public rights.” Post at 37. So the action is a public right because (1) the SEC is the government, and (2) it is vindicating a public right. And what is that public right being vindicated? The dissenting opinion does not say. In reality, the dissenting opinion’s rule is satisfied by the first step alone: The action is itself a “public right” because the SEC is the government. And the not-so-far-removed consequences that flow from that conclusion: When the federal government sues, no jury is required. This is perhaps a runner-up in the competition for the “Nine Most Terrifying Case: 20-61007 Document: 00516323784 Page: 15 Date Filed: 05/18/2022 No. 20-61007 16 Words in the English Language.”8 But fear not, the dissenting opinion’s proposal runs headlong into Granfinanciera: “Congress cannot eliminate a party’s Seventh Amendment right to a jury trial merely by relabeling the cause of action to which it attaches and placing exclusive jurisdiction in an administrative agency or a specialized court of equity” 492 U.S. at 61. With that limit in place, the dissenting opinion’s bright-line rule burns out. Congress cannot change the nature of a right, thereby circumventing the Seventh Amendment, by simply giving the keys to the SEC to do the vindicating. In this light, this approach treats the government’s involvement as a sufficient condition for converting “private rights” into public ones. But from 1856 to 1989, the government’s involvement in a suit was only a necessary condition, not a sufficient condition, for determining whether a suit vindicated public rights. See Granfinanciera, 492 U.S. at 65–66, 68–69 (Scalia, J., concurring in part) (referring to Murray’s Lessee v. Hoboken Land & Improvement Co., 18 U.S. (How.) 272, 283 (1856), and N. Pipeline Constr. Co. v. Marathon Pipeline Co., 458 U.S. 50, 68–69 (1982) (plurality op.)); cf. N. Pipeline Constr. Co., 458 U.S. at 69 n.23 (“It is thus clear that the presence of the United States as a proper party to the proceeding is a necessary but not sufficient means of distinguishing ‘private rights’ from ‘public rights.’”). Then Granfinanciera said that a dispute between two private parties could still vindicate “public rights,” such that the government was no longer a necessary condition for such suits. See 492 U.S. at 53–55. The dissenting opinion thus says that, after Granfinanciera, the government is no longer a necessary condition, but it is now a sufficient condition. That is at odds with Granfinanciera and does not follow from any of the Court’s previous 8 Cf. Ronald Reagan, Presidential News Conference (Aug. 12, 1986), https://www.presidency.ucsb.edu/documents/the-presidents-news-conference-957. Case: 20-61007 Document: 00516323784 Page: 16 Date Filed: 05/18/2022 No. 20-61007 17 decisions, which stressed that the government’s involvement alone does not convert a suit about private rights into one about public rights. The question is not just whether the government is a party, but also whether the right being vindicated is public or private, and how it is being vindicated. Tracing the roots of, and justification for, the public-rights doctrine, the Supreme Court has explained “that certain prerogatives were [historically] reserved to the political Branches of Government.” N. Pipeline Constr. Co., 458 U.S. at 67. Specifically, “[t]he public-rights doctrine is grounded in a historically recognized distinction between matters that could be conclusively determined by the Executive and Legislative Branches and matters that are ‘inherently . . . judicial.’” Id.. at 68 (quoting Ex parte Bakelite Corp., 279 U.S. 438, 458 (1929)). The inquiry is thus inherently historical. The dissenting opinion tries to avoid the history by again emphasizing that Granfinanciera dealt with private parties, not the government. But again, if the right being vindicated is a private one, it is not enough that the government is doing the suing. That means we must consider whether the form of the action—whether brought by the government or by a private entity—is historically judicial, or if it reflects the sorts of issues which courts of law did not traditionally decide. As discussed in Part II.B.2, history demonstrates that fraud claims like these are “traditional legal claims” that arose at common law. Even aside from post-Atlas Roofing refinements of the “public rights” doctrine, this fact, among others, distinguishes that case. In Atlas Roofing, OSHA empowered the government to pursue civil penalties and abatement orders whether or not any employees were “actually injured or killed as a result of the [unsafe working] condition.” 430 U.S. at 445; see also id. at 461 (“[Congress] created a new cause of action, and remedies therefor, unknown to the common law Case: 20-61007 Document: 00516323784 Page: 17 Date Filed: 05/18/2022 No. 20-61007 18 .. . . .”). The government’s right to relief was exclusively a creature of statute and was therefore distinctly public in nature. In contrast, fraud claims, including the securities-fraud claims here, are quintessentially about the redress of private harms. Indeed, the government alleges that Petitioners defrauded particular investors. Cf. 15 U.S.C. §§ 77q(a), 78j(b), 80b-6. As explained above, these fraud claims and civil penalties are analogous to traditional fraud claims at common law in a way that the “new” claims and remedies in Atlas Roofing were not. See Atlas Roofing, 430 U.S. at 461. That being so, Granfinanciera’s considerations about whether Congress created a new action unfamiliar to the common law, and whether jury trial rights are incompatible with the statutory scheme, are appropriate for us to address even if the suit involves the federal government. And as discussed above: (1) this type of action was commonplace at common law, (2) jury trial rights are consistent and compatible with the statutory scheme, and (3) such actions are commonly considered by federal courts with or without the federal government’s involvement. Thus, the agency proceedings below violated Petitioners’ Seventh Amendment rights, and the SEC’s decision must be vacated. C. Petitioners next argue that Congress unconstitutionally delegated legislative power to the SEC when it gave the SEC the unfettered authority to choose whether to bring enforcement actions in Article III courts or within the agency. Because Congress gave the SEC a significant legislative power Case: 20-61007 Document: 00516323784 Page: 18 Date Filed: 05/18/2022 No. 20-61007 19 by failing to provide it with an intelligible principle to guide its use of the delegated power, we agree with Petitioners.9 “We the People” are the fountainhead of all government power. Through the Constitution, the People delegated some of that power to the federal government so that it would protect rights and promote the common good. See The Federalist No. 10 (James Madison) (explaining that one of the defining features of a republic is “the delegation of the government .. . . to a small number of citizens elected by the rest”). But, in keeping with the Founding principles that (1) men are not angels, and (2) “[a]mbition must be made to counteract ambition,” see The Federalist No. 51 (James Madison), the People did not vest all governmental power in one person or entity. It separated the power among the legislative, executive, and judicial branches. See The Federalist No. 47 (James Madison) (“The accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointed, or elective, may justly be pronounced the very definition of tyranny.”). The legislative power is the greatest of these powers, and, of course, it was given to Congress. U.S. Const. art. I, § 1. The Constitution, in turn, provides strict rules to ensure that Congress exercises the legislative power in a way that comports with the People’s will. Every member of Congress is accountable to his or her constituents through regular popular elections. U.S. Const. art I, §§ 2, 3; id. amend. XVII, cl. 1. And a duly elected Congress may exercise the legislative power only through the assent of two separately constituted chambers 9 This is an alternative holding that provides ground for vacating the SEC’s judgment. “This circuit follows the rule that alternative holdings are binding precedent and not obiter dictum.” Texas v. United States, 809 F.3d 134, 178 n.158 (5th Cir. 2015) (quoting United States v. Potts, 644 F.3d 233, 237 n.3 (5th Cir. 2011)). Case: 20-61007 Document: 00516323784 Page: 19 Date Filed: 05/18/2022 No. 20-61007 20 (bicameralism) and the approval of the President (presentment). U.S. Const. art. I, § 7. This process, cumbersome though it may often seem to eager onlookers,10 ensures that the People can be heard and that their representatives have deliberated before the strong hand of the federal government raises to change the rights and responsibilities attendant to our public life. Cf. Rachel E. Barkow, Separation of Powers and the Criminal Law, 58 Stan. L. Rev. 989, 1017 (2006). (“[T]he Framers weighed the need for federal government efficiency against the potential for abuse and came out heavily in favor of limiting federal government power over crime.”). But that accountability evaporates if a person or entity other than Congress exercises legislative power. See Gundy v. United States, 139 S. Ct. 2116, 2134 (2019) (Gorsuch, J., dissenting) (“[B]y directing that legislating be done only by elected representatives in a public process, the Constitution sought to ensure that the lines of accountability would be clear: The sovereign people would know, without ambiguity, whom to hold accountable for the laws they would have to follow.”). Thus, sequestering that power within the halls of Congress was essential to the Framers. As John Locke— 10 Indeed, President Woodrow Wilson, the original instigator of the agency that became the SEC, believed agencies like that one could solve the “problem” of congressional gridlock and the burden of popular accountability. See Cochran v. SEC, 20 F.4th 194, 218 (5th Cir. 2021) (Oldham, J., concurring) (“Wilson’s ‘new constitution’ would ditch the Founders’ tripartite system and their checks and balances for a ‘more efficient separation of politics and administration, which w[ould] enable the bureaucracy to tend to the details of administering progress without being encumbered by the inefficiencies of politics.’” (quoting Ronald J. Pestritto, Woodrow Wilson and the Roots of Modern Liberalism 227 (2005))), cert. granted sub nom., SEC v. Cochran, 21-1239, 2022 WL 1528373 (U.S. May 16, 2022); see also id. (“Wilson’s goal was to completely separate ‘the province of constitutional law’ from ‘the province of administrative function.’” (quoting Philip Hamburger, Is Administrative Law Unlawful? 464 (2014))). Case: 20-61007 Document: 00516323784 Page: 20 Date Filed: 05/18/2022 No. 20-61007 21 a particularly influential thinker at the Founding—explained, not even the legislative branch itself may give the power away: The legislative cannot transfer the power of making laws to any other hands; for it being but a delegated power from the people, they who have it cannot pass it over to others. The people alone can appoint the form of the commonwealth, which is by constituting the legislative, and appointing in whose hands that shall be. And when the people have said we will submit to rules, and be governed by laws made by such men, and in such forms, nobody else can say other men shall make laws for them; nor can the people be bound by any laws but such as are enacted by those whom they have chosen and authorised to make laws for them. Id. at 2133–34 (quoting John Locke, The Second Treatise of Civil Government and a Letter Concerning Toleration § 141, p. 71 (1947)).11 Article I of the Constitution thus provides that “[a]ll legislative Powers herein granted shall be vested in a Congress of the United States.” U.S. Const. art. I, § 1 (emphasis added). In keeping with Founding conceptions of separation of powers,12 the Supreme Court has made clear that Congress cannot “delegate to the Courts, or to any other tribunals, powers which are strictly and exclusively legislative.” Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 42 (1825); see also A.L.A. Schechter Poultry Corp. v. 11 Locke’s perspective on the legislature’s delegation of its power was influential in the United States around the time of the framing of the Constitution. See Hamburger, supra at 384. 12 Principles of non-delegation had even taken hold in England before the American Founding. See Hamburger, supra at 381 (explaining that “even under [King] James I, the judges recognized that the king’s prerogative power came from his subjects—that he was exercising a power delegated by the people” and, as a result, he could not transfer the royal powers to anyone else); see also id. (“[P]arliamentary subdelegations were widely understood to be unlawful.”). Case: 20-61007 Document: 00516323784 Page: 21 Date Filed: 05/18/2022 No. 20-61007 22 United States, 295 U.S. 495, 529 (1935) (“Congress is not permitted to abdicate or to transfer to others the essential legislative functions with which it is thus vested.”). According to the Supreme Court’s more recent formulations of that longstanding rule,13 Congress may grant regulatory power to another entity only if it provides an “intelligible principle” by which the recipient of the power can exercise it. Mistretta v. United States, 488 U.S. 361, 372 (1989) (quoting J.W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 409 (1928)). The two questions we must address, then, are (1) whether Congress has delegated power to the agency that would be legislative power but-for an intelligible principle to guide its use and, if it has, (2) whether it has provided an intelligible principle such that the agency exercises only executive power.14 We first conclude that Congress has delegated to the SEC what would be legislative power absent a guiding intelligible principle. Government actions are “legislative” if they have “the purpose and effect of altering the legal rights, duties and relations of persons . . . outside the legislative branch.” INS v. Chadha, 462 U.S. 919, 952 (1983). The Supreme Court has noted that the power to assign disputes to agency adjudication is “peculiarly 13 Some contemporary academics have argued that the non-delegation doctrine lacks a sound historical basis. See Julian Davis Mortenson & Nicholas Bagley, Delegation at the Founding, 121 Colum. L. Rev. 277 (2021); but see Ilan Wurman, Nondelegation at the Founding, 130 Yale L.J. 1490 (2021) (arguing that the doctrine was present at the Founding); Philip Hamburger, Delegating or Divesting?, 115 Nw. U. L. Rev. Online 88 (2020) (similar). Of course, our role as an inferior court is to faithfully apply Supreme Court precedent, so we do not reach the proper historical scope of the non-delegation doctrine. See Morrow v. Meachum, 917 F.3d 870, 874 n.4 (5th Cir. 2019). 14 Adrian Vermeule, No, 93 Tex. L. Rev. 1547, 1558 (2015) (“[T]here is [no] delegation of legislative power at all so long as the legislature has supplied an ‘intelligible principle’ to guide the exercise of delegated discretion. Where there is such a principle, the delegatee is exercising executive power, not legislative power.” (emphasis and footnote omitted)). Case: 20-61007 Document: 00516323784 Page: 22 Date Filed: 05/18/2022 No. 20-61007 23 within the authority of the legislative department.” Oceanic Steam Navigation Co. v. Stranahan, 214 U.S. 320, 339 (1909).15 And, as discussed above, in some special circumstances Congress has the power to assign to agency adjudication matters traditionally at home in Article III courts. Atlas Roofing, 430 U.S. at 455. Through Dodd–Frank § 929P(a), Congress gave the SEC the power to bring securities fraud actions for monetary penalties within the agency instead of in an Article III court whenever the SEC in its unfettered discretion decides to do so. See 15 U.S.C. § 78u-2(a). Thus, it gave the SEC the ability to determine which subjects of its enforcement actions are entitled to Article III proceedings with a jury trial, and which are not. That was a delegation of legislative power. As the Court said in Crowell v. Benson, “the mode of determining” which cases are assigned to administrative tribunals “is completely within congressional control.” 285 U.S. 22, 50 (1932) (quoting Ex parte Bakelite Corp., 279 U.S. at 451). The SEC argues that by choosing whether to bring an action in an agency tribunal instead of in an Article III court it merely exercises a form of prosecutorial discretion—an executive, not legislative, power. That position reflects a misunderstanding of the nature of the delegated power. Congress did not, for example, merely give the SEC the power to decide whether to bring enforcement actions in the first place, or to choose where to bring a case among those district courts that might have proper jurisdiction. It instead effectively gave the SEC the power to decide which defendants should 15 Moreover, at the Virginia Ratifying Convention in 1788, then-delegate John Marshall suggested that it is proper to the legislative power to determine the expedience of assigning particular matters for jury trial. See John Marshall on the Fairness and Jurisdiction of the Federal Courts, in 2 The Debate on the Constitution 740 (Bernard Bailyn ed. 1993) (“The Legislature of Virginia does not give a trial by jury where it is not necessary. But gives it wherever it is thought expedient. The Federal Legislature will do so too, as it is formed on the same principles.”). Case: 20-61007 Document: 00516323784 Page: 23 Date Filed: 05/18/2022 No. 20-61007 24 receive certain legal processes (those accompanying Article III proceedings) and which should not. Such a decision—to assign certain actions to agency adjudication—is a power that Congress uniquely possesses. See id. Next, Congress did not provide the SEC with an intelligible principle by which to exercise that power. We recognize that the Supreme Court has not in the past several decades held that Congress failed to provide a requisite intelligible principle. Cf. Whitman v. Am. Trucking Ass’ns, Inc., 531 U.S. 457, 474–75 (2001) (cataloguing the various congressional directives that the Court has found to be “intelligible principle[s]”). But neither in the last eighty years has the Supreme Court considered the issue when Congress offered no guidance whatsoever. The last time it did consider such an openended delegation of legislative power, it concluded that Congress had acted unconstitutionally: In Panama Refining Co. v. Ryan, 293 U.S. 388, 405–06 (1935), the Court considered a statutory provision granting the President the authority to prohibit the transportation in interstate commerce of petroleum and related products. The Court scoured the statute for directives to guide the President’s use of that authority, but it found none. Id. at 414–20. It therefore explained: [I]n every case in which the question has been raised, the Court has recognized that there are limits of delegation which there is no constitutional authority to transcend. We think that section 9(c) goes beyond those limits. As to the transportation of oil production in excess of state permission, the Congress has declared no policy, has established no standard, has laid down no rule. Id. at 430. Congress’s grant of authority to the SEC here is similarly open-ended. Even the SEC agrees that Congress has given it exclusive authority and absolute discretion to decide whether to bring securities fraud enforcement Case: 20-61007 Document: 00516323784 Page: 24 Date Filed: 05/18/2022 No. 20-61007 25 actions within the agency instead of in an Article III court. Congress has said nothing at all indicating how the SEC should make that call in any given case. If the intelligible principle standard means anything, it must mean that a total absence of guidance is impermissible under the Constitution.16 See Gundy, 139 S. Ct. at 2123 (Kagan, J., plurality op.) (noting that “we would face a nondelegation question” if the statutory provision at issue had “grant[ed] the Attorney General plenary power to determine SORNA’s applicability to pre-Act offenders—to require them to register, or not, as she sees fit, and to change her policy for any reason and at any time” (emphasis added)).. We therefore vacate the SEC’s judgment on this ground as well. D. The SEC proceedings below suffered from another constitutional infirmity: the statutory removal restrictions for SEC ALJs are unconstitutional.17 SEC ALJs perform substantial executive functions. The President therefore must have sufficient control over the performance of their functions, and, by implication, he must be able to choose who holds the 16 As a member of this court aptly noted just last year, the fact that the modern administrative state is real and robust does not mean courts are never called to declare its limits. See Cochran, 20 F.4th at 222 (Oldham, J., concurring) (“If administrative agencies ‘are permitted gradually to extend their powers by encroachments—even petty encroachments—upon the fundamental rights, privileges and immunities of the people,’ the Court warned that ‘we shall in the end, while avoiding the fatal consequences of a supreme autocracy, become submerged by a multitude of minor invasions of personal rights, less destructive but no less violative of constitutional guaranties.’” (quoting Jones v. SEC, 298 U.S. 1, 24–25 (1936))). 17 Because we vacate the SEC’s judgment on various other grounds, we do not decide whether vacating would be the appropriate remedy based on this error alone. See Collins v. Yellen, 27 F.4th 1068, 1069 (5th Cir. 2022) (remanding to the district court to determine what remedy, if any, is appropriate in light of the Supreme Court’s holding that removal restrictions applicable to the Director of the Federal Housing Finance Agency were unconstitutional). Case: 20-61007 Document: 00516323784 Page: 25 Date Filed: 05/18/2022 No. 20-61007 26 positions. Two layers of for-cause protection impede that control; Supreme Court precedent forbids such impediment. Article II provides that the President must “take Care that the Laws be faithfully executed.” U.S. Const. art. II, § 3. The Supreme Court has held that this provision guarantees the President a certain degree of control over executive officers; the President must have adequate power over officers’ appointment and removal.18 Myers v. United States, 272 U.S. 52, 117 (1926). Only then can the People, to whom the President is directly accountable, vicariously exercise authority over high-ranking executive officials. Free Enterprise Fund v. Public Co. Accounting Oversight Bd., 561 U.S. 477, 498 (2010). Yet not all removal restrictions are constitutionally problematic. “Inferior officers” may retain some amount of for-cause protection from firing. See, e.g., Morrison v. Olson, 487 U.S. 654, 691–92 (1988). Likewise, even principal officers may retain for-cause protection when they act as part of an expert board. Seila Law LLC v. CFPB, 140 S. Ct. 2183, 2192 (2020). But a problem arises when both of those protections act in concert. In Free Enterprise Fund, the Supreme Court considered the constitutionality of two layers of for-cause protection for members of the Public Company Accounting Oversight Board (PCAOB). 561 U.S. at 492. The members of the board answered to the SEC Commissioners. But the SEC could remove them only for “willful violations of the [Sarbanes–Oxley] Act, Board rules, or the securities laws; willful abuse of authority; or unreasonable failure to enforce compliance—as determined in a formal Commission order, rendered on the record and after notice and an opportunity for a hearing.” Id. at 503. 18 Of course, the President’s authority over appointments derives from the Appointments Clause as well. See U.S. Const. art. II, § 2, cl. 2. Case: 20-61007 Document: 00516323784 Page: 26 Date Filed: 05/18/2022 No. 20-61007 27 On top of that, the President could only remove SEC Commissioners for “inefficiency, neglect of duty, or malfeasance in office.” Id. at 486–87, 502. The Supreme Court held that this extensive system insulating PCAOB members from removal deprived the President of the ability to adequately oversee the Board’s actions. Id. at 492, 496. The question here is whether SEC ALJs serve sufficiently important executive functions, and whether the restrictions on their removal are sufficiently onerous, that the President has lost the ability to take care that the laws are faithfully executed. Petitioners’ argument on this point is straightforward: SEC ALJs are inferior officers; they can only be removed by the SEC Commissioners if good cause is found by the Merits Systems Protection Board; SEC Commissioners and MSPB members can only be removed by the President for cause; so, SEC ALJs are insulated from the President by at least two layers of for-cause protection from removal, which is unconstitutional under Free Enterprise Fund. The SEC responds that this case is not like Free Enterprise Fund. First, it contends that SEC ALJs primarily serve an adjudicatory role. Second, it asserts that the for-cause protections for ALJs are not as stringent as those which applied to PCAOB members at the time of Free Enterprise Fund—or, at least, that this court should read the removal protections for ALJs that way to avoid constitutional problems. We agree with Petitioners and hold that the removal restrictions are unconstitutional. The Supreme Court decided in Lucia that SEC ALJs are “inferior officers” under the Appointments Clause because they have substantial authority within SEC enforcement actions. Lucia v. SEC, 138 S. Ct. 2044, 2053 (2018). And in Free Enterprise Fund it explained that the President must have adequate control over officers and how they carry out their functions. 561 U.S. at 492, 496. If principal officers cannot intervene in their inferior officers’ actions except in rare cases, the President lacks the Case: 20-61007 Document: 00516323784 Page: 27 Date Filed: 05/18/2022 No. 20-61007 28 control necessary to ensure that the laws are faithfully executed. So, if SEC ALJs are “inferior officers” of an executive agency, as the Supreme Court in Lucia indicated was the case at least for the purposes of the Appointments Clause, they are sufficiently important to executing the laws that the Constitution requires that the President be able to exercise authority over their functions. Specifically, SEC ALJs exercise considerable power over administrative case records by controlling the presentation and admission of evidence; they may punish contemptuous conduct; and often their decisions are final and binding. Lucia, 138 S. Ct. at 2053–54. But 5 U.S.C. § 7521(a) provides that SEC ALJs may be removed by the Commission “only for good cause established and determined by the Merit Systems Protection Board (MSPB) on the record after opportunity for hearing before the Board.” (Parenthetical not in original.) And the SEC Commissioners may only be removed by the President for good cause. The dissenting opinion’s response is all built on dicta from Free Enterprise Fund. There, in noting what issues the Court was leaving open, the Court identified characteristics that were true of ALJs that were not true of PCAOB members: “[U]nlike members of the [PCAOB], many” ALJs “perform adjudicative rather than enforcement or policymaking functions.” Free Enterprise Fund, 561 U.S. at 507 n.10. Far from “stat[ing]” that this “may justify multiple layers of removal protection,” post at 22, the Court merely identified that its decision does not resolve the issue presented here. In any event, the Court itself said in Myers that “quasi[-]judicial” executive officers must nonetheless be removable by the President “on the ground that the discretion regularly entrusted to that officer by statute has not been on Case: 20-61007 Document: 00516323784 Page: 28 Date Filed: 05/18/2022 No. 20-61007 29 the whole intelligently or wisely exercised.” 272 U.S. at 135.19 So even if ALJs’ functions are more adjudicative than PCAOB members, the fact remains that two layers of insulation impedes the President’s power to remove ALJs based on their exercise of the discretion granted to them.20 Finally, the SEC urges us to interpret the for-cause protections for ALJs to instead allow removal for essentially any reason. Even if we could do so (and the statutory language likely does not give us that flexibility), that 19 The dissenting opinion deems this proposition from Myers to be obiter dicta that the Court subsequently disregarded in Humphrey’s Executor v. United States, 295 U.S. 602, 626–28 (1935). Post at 54 n.113. But that itself is to disregard the Supreme Court’s more recent guidance, which fortifies the Court’s “landmark decision” in Myers and narrowed Humphrey’s Executor. See Seila Law, 140 S. Ct. at 2191–92, 2197–99 & n.2 (limiting the Humphrey’s Executor exception to Myers to cases involving “for-cause removal protections [given] to a multimember body of experts, balanced along partisan lines, that perform[] legislative and judicial functions and [are] said not to exercise any executive power,” while casting doubt on the existence of wholly non-executive, quasi-legislative or quasi-judicial agency powers altogether); see also City of Arlington v. F.C.C., 569 U.S. 290, 305 n.4 (2013) (noting that “[agency] activities take ‘legislative’ and ‘judicial’ forms, but they are exercises of—indeed, under our constitutional structure they must be exercises of—the ‘executive Power’” (citing U.S. Const. art. II, § 1, cl. 1)). 20 In the next breath, the dissenting position draws from a law review article that “[t]he ALJs’ role is similar to that of a federal judge.” Post at 52. It then concludes that they must be insulated from removal by the president to maintain their independence. But that analogy runs out under a little scrutiny. The SEC’s ALJs are not mere neutral arbiters of federal securities law; they are integral pieces within the SEC’s powerful enforcement apparatus. The ALJs report to the Commission itself and act under authority delegated by it. SEC Organization Chart (2020), https://www.sec.gov/about/secorg.pdf; 15 U.S.C. § 78d-1(a); 17 C.F.R. § 200.30-10. As the amicus brief by the Cato Institute points out, these administrative proceedings differ significantly from cases resolved in federal district courts and reviewed by federal courts of appeals. Cato Amicus Br. at 19–31. First, the Commission has ex parte discussions with the prosecutors to determine whether to pursue securities-fraud claims. Then the Commission itself decides what claims should be brought by the prosecutors. Only then do ALJs resolve the claims, which are then again reviewed by the Commission. Suffice it to say, even if ALJs have some of the same “tools of federal trial judges,” Lucia, 138 S. Ct. at 2053, they use those tools at the direction of and with the power delegated to them by the Commission. Case: 20-61007 Document: 00516323784 Page: 29 Date Filed: 05/18/2022 No. 20-61007 30 would not solve the Article II problem. As noted above, the MSPB is part of the mix as well. Furthermore, MSPB members “may be removed by the President only for inefficiency, neglect of duty, or malfeasance in office.” 5 U.S.C. § 1202(d). So, for an SEC ALJ to be removed, the MSPB must find good cause and the Commission must choose to act on that finding. And members of both the MSPB and the Commission have for-cause protection from removal by the President. Simply put, if the President wanted an SEC ALJ to be removed, at least two layers of for-cause protection stand in the President’s way. Thus, SEC ALJs are sufficiently insulated from removal that the President cannot take care that the laws are faithfully executed. The statutory removal restrictions are unconstitutional. III. In sum, we agree with Petitioners that the SEC proceedings below were unconstitutional. The SEC’s judgment should be vacated for at least two reasons: (1) Petitioners were deprived of their Seventh Amendment right to a civil jury; and (2) Congress unconstitutionally delegated legislative power to the SEC by failing to give the SEC an intelligible principle by which to exercise the delegated power. We also hold that the statutory removal restrictions for SEC ALJs are unconstitutional, though we do not address whether vacating would be appropriate based on that defect alone.21 We GRANT the petition for review, VACATE the decision of the SEC, and REMAND for further proceedings consistent with this opinion. 21 Petitioners also argue that the SEC violated their equal protection rights, and that its decision was infected with bias and violated their due process rights. Because we vacate the SEC’s decision on other grounds, we decline to reach these issues. Case: 20-61007 Document: 00516323784 Page: 30 Date Filed: 05/18/2022 No. 20-61007 31 W. Eugene Davis, Circuit Judge, dissenting: The majority holds that (1) administrative adjudication of the SEC’s enforcement action violated Petitioners’ Seventh Amendment right to a jury trial; (2) Congress unconstitutionally delegated an Article I legislative power to the executive branch when it gave the SEC the discretion to choose between bringing its enforcement action in an Article III court or before the agency without providing an intelligible principle to guide the SEC’s decision; and (3) the removal protections on SEC administrative law judges violate Article II’s requirement that the President “take Care that the Laws be faithfully executed.” I respectfully disagree with each of these conclusions. I. The majority holds that the Seventh Amendment grants Petitioners the right to a jury trial on the facts underlying the SEC’s enforcement action, and administrative adjudication without a jury violated that right. In reaching this conclusion, the majority correctly recognizes that a case involving “public rights” may be adjudicated in an agency proceeding without a jury notwithstanding the Seventh Amendment.1 But, the majority then erroneously concludes that the SEC’s enforcement action does not involve “public rights.” In my view, the majority misreads the Supreme Court’s decisions addressing what are and are not “public rights.” 1 See, e.g., Granfinanciera, S.A. v. Nordberg, 492 U.S. 33, 42 n.4 (1989) (“If a claim that is legal in nature asserts a ‘public right,’ . . . then the Seventh Amendment does not entitle the parties to a jury trial if Congress assigns its adjudication to an administrative agency or specialized court of equity. The Seventh Amendment protects a litigant’s right to a jury trial only if a cause of action is legal in nature and it involves a matter of ‘private right.’” (citation omitted)). Case: 20-61007 Document: 00516323784 Page: 31 Date Filed: 05/18/2022 No. 20-61007 32 A. As declared by Professors Wright and Miller, “A definitive statement by the Supreme Court regarding congressional authority in this context is found in Atlas Roofing v. Occupational Safety & Health Review Commission.”2 That case concerned the Occupational Safety and Health Act (“OSHA” or “the Act”), which created a new statutory duty on employers to avoid maintaining unsafe or unhealthy working conditions. OSHA also empowered the Federal Government, proceeding before an administrative agency without a jury, to impose civil penalties on those who violated the Act.3 Two employers who had been cited for violating the Act argued that a suit in a federal court by the Government seeking civil penalties for violation of a statute is classically a suit at common law for which the Seventh Amendment provides a right to a jury trial; therefore, Congress cannot deprive them of that right by simply assigning the function of adjudicating the Government’s right to civil penalties to an administrative forum where no jury is available.4 The Court, in a unanimous opinion, disagreed: At least in cases in which “public rights” are being litigated— e.g., cases in which the Government sues in its sovereign capacity to enforce public rights created by statutes within the power of Congress to enact—the Seventh Amendment does not prohibit Congress from assigning the factfinding function and initial adjudication to an administrative forum with which the jury would be incompatible. . . . This is the case even if the Seventh Amendment would have required a jury where the 2 9 Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 2302.2, at 59 (4th ed. 2020) (citing Atlas Roofing Co. v. Occupational Safety & Health Rev. Comm’n, 430 U.S. 442 (1977)) (italics added). 3 Atlas Roofing, 430 U.S. at 445. 4 Id. at 449–50. Case: 20-61007 Document: 00516323784 Page: 32 Date Filed: 05/18/2022 No. 20-61007 33 adjudication of those rights is assigned instead to a federal court of law instead of an administrative agency.5 Atlas Roofing drew its definition of “public rights” from, inter alia, Crowell v. Benson, which described “public rights” in slightly broader terms: matters “which arise between the Government and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments.”6 The Supreme Court has never retreated from its holding in Atlas Roofing. 7 In fact, the Court implicitly re-affirmed Atlas Roofing’s definition of “public rights” as recently as 2018, when it decided Oil States Energy Services, LLC v. Greene’s Energy Group, LLC. 8 That case involved the LeahySmith America Invents Act, which granted the Patent and Trademark Office (“PTO”) the power to reconsider a previously-issued patent via an administrative process called “inter partes review.”9 This was a departure from historical practice, which placed this function in Article III courts alone.10 The petitioner argued that inter partes review violated both Article 5 Id. at 450, 455 (emphasis added; paragraph break omitted); see also id. at 458 (“Our prior cases support administrative factfinding in only those situations involving ‘public rights,’ e.g., where the Government is involved in its sovereign capacity under an otherwise valid statute creating enforceable public rights.”). 6 Id. at 452 (quoting Crowell v. Benson, 285 U.S. 22, 50 (1932)) (emphasis added); see also id. at 456, 457, 460 (citing Crowell, 285 U.S. 22). 7 Gideon Mark, SEC and CFTC Administrative Proceedings, 19 U. Pa. J. Const. L. 45, 95 (2016). 8 138 S. Ct. 1365 (2018). 9 Id. at 1370–72. 10 Id. at 1384 (Gorsuch, J., dissenting) (“[F]rom the time it established the American patent system in 1790 until about 1980, Congress left the job of invalidating patents at the federal level to courts alone.”). Case: 20-61007 Document: 00516323784 Page: 33 Date Filed: 05/18/2022 No. 20-61007 34 III and the Seventh Amendment.11 The Court disagreed and explained that Congress has “significant latitude” to assign adjudication of “public rights” to non-Article III tribunals that do not use a jury.12 Moreover, the Court, quoting Crowell, defined “public rights” as “matters ‘which arise between the Government and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments.’”13 As mentioned, Atlas Roofing’s definition of “public rights” is a slightly narrower version of Crowell’s definition. Thus, when Oil States reaffirmed Crowell, it necessarily re-affirmed Atlas Roofing’s definition as well.14 Oil States is also significant because it held that historical practice is not determinative in matters governed by the public rights doctrine, as such matters “‘from their nature’ can be resolved in multiple ways.”15 Accordingly, the Court rejected the view that “because courts have traditionally adjudicated patent validity in this country, courts must forever continue to do so.”16 11 Id. at 1372. 12 Id. at 1373, 1379. 13 Id. at 1373 (quoting Crowell, 285 U.S. at 50). 14 Oil States did not purport to provide an exhaustive definition of “public rights,” and the opinion alludes to the possibility that, under certain circumstances, matters not involving the Government may also fall within the realm of “public rights.” See id. However, the Court did not need to address these other, “various formulations” of “public rights,” because inter partes review fell squarely within Crowell’s definition. See id. This court reached a similar conclusion in Austin v. Shalala, discussed below. 15 Id. at 1378 (quoting Ex parte Bakelite Corp., 279 U.S. 438, 451 (1929)). 16 Id.; see also id. (“That Congress chose the courts in the past does not foreclose its choice of the PTO today.”). Case: 20-61007 Document: 00516323784 Page: 34 Date Filed: 05/18/2022 No. 20-61007 35 Like Oil States, this court relied on Crowell to define “public rights” in Austin v. Shalala. 17 That case involved the Government’s action to recover overpayment of social security benefits via an administrative proceeding before the Social Security Administration.18 Austin rejected the plaintiff’s argument that the proceeding violated her Seventh Amendment right, explaining that “if Congress may employ an administrative body as a factfinder in imposing money penalties for the violation of federal laws”—as was done in Atlas Roofing and in the securities statutes at issue here—“it plainly may employ such a body to recover overpayments of government largess.”19 Consistent with the above cases, our sister circuits routinely hold that an enforcement action by the Government for violations of a federal statute or regulation is a “public right” that Congress may assign to an agency for adjudication without offending the Seventh Amendment.20 For example, the Eleventh Circuit relied solely on Atlas Roofing when it rejected a Seventh Amendment challenge to administrative adjudication of an SEC enforcement action and declared “it is well-established that the Seventh 17 994 F.2d 1170, 1177 (5th Cir. 1993). 18 Id. at 1173. 19 Id. at 1177-78 (citing Oceanic Steam Navigation Co. v. Stranahan, 412 U.S. 320, 339 (1909)). 20 See, e.g., Imperato v. SEC, 693 F. App’x 870, 876 (11th Cir. 2017) (unpublished) (administrative adjudication for violations of the Securities Exchange Act); Crude Co. v. FERC, 135 F.3d 1445, 1454–55 (Fed. Cir. 1998) (Mandatory Petroleum Allocation Regulations); Cavallari v. Office of Comptroller of Currency, 57 F.3d 137, 145 (2d Cir. 1995) (Financial Institutions Reform, Recovery and Enforcement Act); Sasser v. Adm’r EPA, 990 F.2d 127, 130 (4th Cir. 1993) (Clean Water Act). Case: 20-61007 Document: 00516323784 Page: 35 Date Filed: 05/18/2022 No. 20-61007 36 Amendment does not require a jury trial in administrative proceedings designed to adjudicate statutory ‘public rights.’”21 The SEC’s enforcement action satisfies Atlas Roofing’s definition of a “public right,” as well as the slightly broader definition set forth in Crowell and applied in Oil States and Austin. The broad congressional purpose of the securities laws is to “protect investors.”22 For example, the Securities Act of 1933 was “designed to provide investors with full disclosure of material information concerning public offerings of securities in commerce, to protect investors against fraud and, through the imposition of specified civil liabilities, to promote ethical standards of honesty and fair dealing.”23 The Dodd-Frank Act, which, inter alia, expanded the SEC’s authority to pursue civil penalties in administrative proceedings,24 was “intended to improve investor protection,” particularly in light of the Bernard Madoff Ponzi scheme.25 Other circuits have consistently recognized that “[w]hen the SEC sues to enforce the securities laws, it is vindicating public rights and furthering public interests, and therefore is acting in the United States’s 21 Imperato, 693 F. App’x at 876 (citing Atlas Roofing, 430 U.S. at 455–56). 22 Smallwood v. Pearl Brewing Co., 489 F.2d 579, 592 (5th Cir. 1974). 23 Ernst & Ernst v. Hochfelder, 425 U.S. 185, 195 (1976). In a similar vein, the Investment Advisers Act of 1940 seeks to “protect[] investors through the prophylaxis of disclosure,” in order to eliminate “the darkness and ignorance of commercial secrecy,” which “are the conditions upon which predatory practices best thrive.” SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 200 (1963). 24 Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111- 203, Sec. 929P, 124 Stat. 1376, 1862–64 (2010) (codified at 15 U.S.C. §§ 77h-1(g), 78u-2(a), 80a-9(d), 80b-3(i)). 25 Mark Jickling, Congressional Research Service, R41503 The Dodd-Frank Wall Street Reform and Consumer Protection Act: Title IX, Investor Protection at i (2010). Case: 20-61007 Document: 00516323784 Page: 36 Date Filed: 05/18/2022 No. 20-61007 37 sovereign capacity.”26 Thus, the SEC’s enforcement action is a “public right” because it is a case “in which the Government sues in its sovereign capacity to enforce public rights created by statutes within the power of Congress to enact.”27 It is also a matter “which arise[s] between the Government and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments.”28 Because the SEC’s enforcement action is a “public right,” the Seventh Amendment does not prohibit Congress from assigning its adjudication to an administrative forum that lacks a jury.29 As discussed below, the fact that the securities statutes at issue resemble (but are not identical to) common-law fraud does not change this result.30 It also makes 26 SEC v. Diversified, 378 F.3d 1219, 1224 (11th Cir. 2004), abrogated on other grounds by Kokesh v. SEC, 137 S. Ct. 1635 (2017); see also SEC v. Rind, 991 F.2d 1486, 1491 (9th Cir. 1993); United States v. Badger, 818 F.3d 563, 566 (10th Cir. 2016). 27 Atlas Roofing, 430 U.S. at 450. 28 Crowell, 285 U.S. at 22; Oil States, 138 S. Ct. at 1373; Austin, 994 F.2d at 1177. The majority asserts that “[t]he dissenting opinion cannot define a ‘public right’ without using the term itself in the definition.” First, I rely on definitions the Supreme Court has provided. Second, while Atlas Roofing does use “public rights” to define “public rights,” Crowell does not. Furthermore, Granfinanciera observed that Atlas Roofing “left the term ‘public rights’ undefined” and so looked to Crowell to fill in any perceived gap. Granfinanciera, 492 U.S. at 51 n.8; see also id. at 53 (noting that, under Atlas Roofing, a “public right” is simply “a statutory cause of action [that] inheres in, or lies against, the Federal Government in its sovereign capacity”). 29 Atlas Roofing, 430 U.S. at 450; Granfinanciera, 492 U.S. at 52–54; Oil States, 138 S. Ct. at 1379. 30 See Granfinanciera, 492 U.S. at 52 (“Congress may fashion causes of action that are closely analogous to common-law claims and place them beyond the ambit of the Seventh Amendment by assigning their resolution to a forum in which jury trials are unavailable” if the action involves “public rights.”). Case: 20-61007 Document: 00516323784 Page: 37 Date Filed: 05/18/2022 No. 20-61007 38 no difference that federal courts have decided claims under the securities statutes for decades.31 B. The majority’s conclusion that the SEC’s enforcement action is not a “public right” is based primarily on an erroneous reading of Granfinanciera, S.A. v. Nordberg.32 Specifically, the majority interprets that case as abrogating Atlas Roofing. Granfinanciera did nothing of the sort. In Granfinanciera, a bankruptcy trustee sued in bankruptcy court (where a jury was unavailable) to avoid allegedly fraudulent transfers the defendants had received from the debtor.33 The defendants argued that they were entitled to a jury trial under the Seventh Amendment.34 A key issue was whether the trustee’s claim involved “public” or “private” rights. The Court held that the action was a private right.35 Unlike Atlas Roofing, Granfinanciera did not involve a suit by or against the Federal Government. This distinction is important. In discussing what constitutes a “public right,” Granfinanciera, citing Atlas Roofing, recognized that “Congress may effectively supplant a common-law cause of action carrying with it a right to a jury trial with a statutory cause of action shorn of a jury trial right if that statutory cause of action inheres in, or lies 31 See Oil States, 138 S. Ct. at 1378 (“[W]e disagree with the dissent’s assumption that, because courts have traditionally adjudicated patent validity in this country, courts must forever continue to do so. Historical practice is not decisive . .. . [in] matters governed by the public-rights doctrine . . . . That Congress chose the courts in the past does not foreclose its choice of the PTO today.”) 32 492 U.S. 33. 33 Id. at 36. 34 Id. at 40. 35 Id. at 55, 64. Case: 20-61007 Document: 00516323784 Page: 38 Date Filed: 05/18/2022 No. 20-61007 39 against, the Federal Government in its sovereign capacity.”36 Granfinanciera then clarified that “the class of ‘public rights’ whose adjudication Congress may assign to administrative agencies . . . is more expansive than Atlas Roofing’s discussion suggests”;37 i.e., the “Government need not be a party for a case to revolve around ‘public rights’” provided certain other criteria are met.38 Nevertheless, and contrary to what is implied by the majority, Granfinanciera’s recognition that the public-rights doctrine can extend to cases where the Government is not a party in no way undermines or alters Atlas Roofing’s holding that a case where the Government sues in its sovereign capacity to enforce a statutory right is a case involving “public rights.”39 Because the bankruptcy trustee’s suit involved only private parties and not the Government, Granfinanciera’s analysis is solely concerned with whether the action was one of the “seemingly ‘private’ right[s]” that are 36 Granfinanciera, 492 U.S. at 53 (citing Atlas Roofing, 430 U.S. at 458) (emphasis added). 37 Id. at 53 (emphasis added). 38 Id. at 54 (citing Thomas v. Union Carbide Agric. Prods. Co., 473 U.S. 568, 586, 596–99 (1985)). 39 Granfinanciera itself makes this clear when it states: The crucial question, in cases not involving the Federal Government, is whether “Congress, acting for a valid legislative purpose pursuant to its constitutional powers under Article I, [has] create[d] a seemingly ‘private’ right that is so closely integrated into a public regulatory scheme as to be a matter appropriate for agency resolution with limited involvement by the Article III judiciary.” If a statutory right is not closely intertwined with a federal regulatory program Congress has power to enact, and if that right neither belongs to nor exists against the Federal Government, then it must be adjudicated by an Article III court. Id. at 54-55 (quoting Thomas, 473 U.S. at 593–94) (footnote omitted; emphasis added; bracketed alterations in original). Case: 20-61007 Document: 00516323784 Page: 39 Date Filed: 05/18/2022 No. 20-61007 40 within the reach of the public-rights doctrine. Thus, any considerations or requirements discussed in Granfinanciera that go beyond Atlas Roofing or Crowell apply only to cases not involving the Government. This understanding of Granfinanciera is supported by our subsequent decision in Austin, which stated: Although the definition is somewhat nebulous, at a minimum, suits involving public rights are those “which arise between the Government and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments.” Crowell v. Benson, 285 U.S. 22, 50, 52 S. Ct. 285, 292, 76 L.Ed. 598 (1932). Beyond that, certain other cases are said to involve public rights where Congress has created a “seemingly ‘private’ right that is so closely integrated into a public regulatory scheme as to be a matter appropriate for agency resolution with limited involvement by the Article III judiciary.” Granfinanciera, 492 U.S. at 54 . . . .40 Similarly, while Oil States acknowledged that Crowell did not provide the sole definition of what constitutes a “public right,” it did not discuss any of the other “formulations” because Crowell’s definition was met.41 The majority overlooks the fact that Granfinanciera’s expansion of the public-rights doctrine applies only when the Government is not a party to the case. As a result, the majority applies “considerations” that have no relevance here. For example, the majority, quoting Granfinanciera, states that “jury trials would not ‘go far to dismantle the statutory scheme’ or ‘impede swift resolution’ of statutory claims.” Again, Granfinanciera discussed these considerations in the context of a suit between private 40 Austin, 994 F.2d at 1177 (emphasis added). 41 Oil States, 138 S. Ct. at 1373. Case: 20-61007 Document: 00516323784 Page: 40 Date Filed: 05/18/2022 No. 20-61007 41 persons, not a case involving the Government acting in its sovereign capacity under an otherwise valid statute creating enforceable public rights. 42 Indeed, neither Austin nor Oil States, both of which were decided after Granfinanciera and which found public rights to exist, mentions these considerations.43 The majority also states that the securities statutes at issue created causes of action that “reflect” and “echo” common-law fraud. But this does not matter, because, as Granfinanciera itself recognized, the public-rights doctrine allows Congress to “fashion causes of action that are closely analogous to common-law claims and place them beyond the ambit of the Seventh Amendment by assigning their resolution to a forum in which jury trials are unavailable.”44 The majority asserts that Atlas Roofing is distinguishable from the SEC’s enforcement action because “OSHA empowered the government to pursue civil penalties regardless of whether any employe[e]s were ‘actually injured or killed as a result of the [unsafe working] condition.’”45 But the securities statutes share this feature: The SEC may impose civil penalties on 42 Granfinanciera, 492 U.S. at 61, 63. 43 The same goes for the out-of-circuit decisions cited in footnote 20 above. Atlas Roofing, in a footnote, does make a passing reference to “go far to dismantle the statutory scheme.” 430 U.S. at 454 n.11. But the Court was merely describing its reasoning in another bankruptcy case. Nothing in Atlas Roofing suggests that this consideration is relevant to whether Congress may assign the Government’s enforcement action to an administrative proceeding lacking a jury. 44 Granfinanciera, 492 U.S. at 52 (citations omitted); see also id. at 53 (“Congress may effectively supplant a common-law cause of action carrying with it a right to a jury trial with a statutory cause of action shorn of a jury trial right if that statutory cause of action inheres in, or lies against, the Federal Government in its sovereign capacity.” (citing Atlas Roofing, 430 U.S. at 458)); accord Crude Co., 135 F.3d at 1455 (“The public right at issue is not converted into a common law tort simply because the theory of liability underlying the enforcement action is analogous to a common law tort theory of vicarious liability.”). 45 Majority Op. at 17–18 (quoting Atlas Roofing, 430 U.S. at 445). Case: 20-61007 Document: 00516323784 Page: 41 Date Filed: 05/18/2022 No. 20-61007 42 a person who makes a material misrepresentation even if no harm resulted from the misrepresentation.46 The statutory cause of action created by the securities statutes is as “new” to the common law as the one created by OSHA.47 Relatedly, the majority harps on the fact that federal courts have dealt with actions under the securities statutes for decades. But Oil States makes clear that “[h]istorical practice is not decisive here.”48 “That Congress chose the courts in the past does not foreclose its choice of [an administrative adjudication] today.”49 The majority also states that “securities-fraud enforcement actions are not the sort that are uniquely suited for agency adjudication.” Again, this is not relevant. As Oil States explained, “the public-rights doctrine applies to matters ‘arising between the government and others, which from their nature 46 See 15 U.S.C. §§ 78u-2(c), 77h-1(g)(1), 80a-9(d)(3), 80b-3(i)(3). 47 Atlas Roofing recognized that, before (and after) OSHA, a person injured by an unsafe workplace condition may have an action at common law for negligence. See Atlas Roofing, 430 U.S. at 445. Through OSHA, specific safety standards were promulgated, and the Government could bring an enforcement action for a violation even if no one was harmed by the violation. Id. Similarly, before enactment of the securities statutes, an investor who was defrauded in the course of a securities transaction had a common-law action for fraud. Like OSHA, the securities statutes expressly prohibited certain conduct and empowered the SEC to bring an enforcement action for a violation, even if no one was actually harmed by the violation. 48 138 S. Ct. at 1378. 49 Id. Oil States likewise refutes the majority’s assertion that “[t]he inquiry is thus inherently historical.” I add that the majority’s support for this proposition consists of a concurring opinion in Granfinanciera and the plurality opinion in Northern Pipeline Construction Co. v. Marathon Pipeline Co., 458 U.S. 50 (1982) (plurality), which addressed whether a bankruptcy court may decide a breach of contract action between two private parties. Case: 20-61007 Document: 00516323784 Page: 42 Date Filed: 05/18/2022 No. 20-61007 43 do not require judicial determination and yet are susceptible of it.’”50 Indeed, “matters governed by the public-rights doctrine ‘from their nature’ can be resolved in multiple ways.”51 Finally, it should be emphasized that Tull v. United States52 does not control the outcome here. That case concerned the Government’s suit in district court seeking civil penalties and an injunction for violations of the Clean Water Act.53 Tull did not involve an administrative proceeding. Thus, while Tull concluded that the Government’s claim was analogous to a “Suit at common law” for Seventh Amendment purposes,54 the Court did not engage in the “quite distinct inquiry” into whether the claim was also a “public right” that Congress may assign to a non-Article III forum where juries are unavailable.55 Tull itself acknowledges in a footnote prior decisions “holding that the Seventh Amendment is not applicable to administrative proceedings,” making clear that it was not deciding whether the defendant would be entitled to a jury in an administrative adjudication.56 C. In summary, the SEC’s enforcement action against Petitioners for violations of the securities laws is a “public right” under Supreme Court precedent as well as our own. Accordingly, Congress could and did validly 50 Id. at 1373 (citing Crowell, 285 U.S. at 50) (emphasis added). 51 Id. at 1378 (quoting Ex parte Bakelite Corp., 279 U.S. at 451). 52 481 U.S. 412 (1987). 53 Id. at 414–15. 54 Id. at 425. 55 Granfinanciera, 492 U.S. at 42 n.4; accord Sasser, 990 F.2d at 130. 56 Tull, 481 U.S. at 418 n.4 (citing Atlas Roofing, 430 U.S. at 454; Pernell v. Southall Realty, 416 U.S. 363, 383 (1974)). Case: 20-61007 Document: 00516323784 Page: 43 Date Filed: 05/18/2022 No. 20-61007 44 assign adjudication of that action to an administrative forum where the Seventh Amendment does not require a jury. II. I also disagree with the majority’s alternative holding that Congress exceeded its power by giving the SEC the authority to choose to bring its enforcement action in either an agency proceeding without a jury or to a court with a jury. The majority reasons that giving the SEC this power without providing guidelines on the use of that power violates Article I by delegating its legislative authority to the agency. The majority’s position runs counter to Supreme Court precedent. As set forth below, by authorizing the SEC to bring enforcement actions either in federal court or in agency proceedings, Congress fulfilled its legislative duty. In support of its determination that Congress unconstitutionally delegated its authority to the SEC, the majority relies on Crowell v. Benson, wherein the Supreme Court explained that “the mode of determining” cases involving public rights “is completely within congressional control.”57 Crowell did not state that Congress cannot authorize that a case involving public rights may be determined in either of two ways. By passing DoddFrank § 929P(a), Congress established that SEC enforcement actions can be brought in Article III courts or in administrative proceedings. In doing so, Congress fulfilled its duty of controlling the mode of determining public rights cases asserted by the SEC. The majority maintains that because the SEC has “the power to decide which defendants should receive certain legal processes (those accompanying Article III proceedings) and which should not,” then such a 57 285 U.S. at 50 (quoting Ex parte Bakelite Corp., 279 U.S. at 451). Case: 20-61007 Document: 00516323784 Page: 44 Date Filed: 05/18/2022 No. 20-61007 45 decision falls under Congress’s legislative power. The Supreme Court’s decision in United States v. Batchelder58 demonstrates that the majority’s position on this issue is incorrect. In Batchelder, the issue presented was whether it was constitutional for Congress to allow the Government, when prosecuting a defendant, to choose between two criminal statutes that “provide[d] different penalties for essentially the same conduct.”59 The defendant had been convicted under the statute with the higher sentencing range, and the Court of Appeals determined that the delegation of authority to prosecutors to decide between the two statutes, and thus choose a higher sentencing range for identical conduct, was a violation of due process and the nondelegation doctrine.60 Specifically, the Court of Appeals determined that “such prosecutorial discretion could produce ‘unequal justice’” and that it might be “impermissibl[e] [to] delegate to the Executive Branch the Legislature’s responsibility to fix criminal penalties.”61 The Supreme Court disagreed. The Court explained that “[t]he provisions at issue plainly demarcate the range of penalties that prosecutors and judges may seek and impose.”62 The Court further stated: “In light of that specificity, the power that Congress has delegated to those officials is no broader than the authority they routinely exercise in enforcing the criminal laws.”63 The Court concluded: “Having informed the courts, prosecutors, 58 442 U.S. 114 (1979). 59 Id. at 116. 60 Id. at 123, 125–26. 61 Id. at 125–26. 62 Id. at 126. 63 Id. Case: 20-61007 Document: 00516323784 Page: 45 Date Filed: 05/18/2022 No. 20-61007 46 and defendants of the permissible punishment alternatives available under each Title, Congress has fulfilled its duty.”64 The Supreme Court has analogized agency enforcement decisions to prosecutorial discretion exercised in criminal cases.65 If the Government’s prosecutorial authority to decide between two criminal statutes that provide for different sentencing ranges for essentially the same conduct does not violate the nondelegation doctrine, then surely the SEC’s authority to decide between two forums that provide different legal processes does not violate the nondelegation doctrine. Thus, the SEC’s forum-selection authority is part and parcel of its prosecutorial authority.66 Although no other circuit court appears to have addressed the particular nondelegation issue presented in this case, a district court did so in Hill v. SEC. 67 Like the majority does here, the plaintiff in Hill relied on I.N.S. v. Chadha68 to assert that the SEC’s choice of forum is a legislative action because it “alter[s] the rights, duties, and legal relations of individuals.”69 Chadha addressed the question whether a provision in the Immigration and 64 Id. (citation omitted). 65 See Heckler v. Chaney, 470 U.S. 821, 832 (1985) (“[W]e recognize that an agency’s refusal to institute proceedings shares to some extent the characteristics of the decision of a prosecutor in the Executive Branch not to indict—a decision which has long been regarded as the special province of the Executive Branch . . . ..”). 66Cf. SEC v. Chenery Corp., 332 U.S. 194, 203 (1947) (“[T]he choice made between proceeding by general rule or by individual, ad hoc litigation is one that lies primarily in the informed discretion of the administrative agency.”) (citation omitted). 67 114 F. Supp. 3d 1297 (N.D. Ga. 2015) (holding that SEC’s forum-selection authority does not violate the nondelegation doctrine), vacated and remanded on other grounds, 825 F.3d 1236 (11th Cir. 2016). 68 462 U.S. 919 (1983). 69 Hill, 114 F. Supp. 3d at 1312 (quoting Chadha, 462 U.S. at 952). Case: 20-61007 Document: 00516323784 Page: 46 Date Filed: 05/18/2022 No. 20-61007 47 Nationality Act (INA) allowing one House of Congress to veto the Attorney General’s decision to allow a particular deportable alien to remain in the United States violated the Presentment Clauses and bicameral requirement of Article I.70 Specifically, it addressed whether Congress, after validly delegating authority to the Executive, can then alter or revoke that valid delegation of authority through the action of just one House. I agree with the district court in Hill that if Chadha’s definition of legislative action is interpreted broadly and out of context, then any SEC decision which affected a person’s legal rights—including charging decisions—would be legislative actions, which is contrary to the Supreme Court’s decision in Batchelder. 71 Chadha, one of the primary authorities the majority relies on, does not touch on any issue involved in this case. I agree with the persuasive and well-reasoned decision of the district court in Hill that “Congress has properly delegated power to the executive branch to make the forum choice for the underlying SEC enforcement action.”72 In sum, it is clear to me that Congress’s decision to give prosecutorial authority to the SEC to choose between an Article III court and an administrative proceeding for its enforcement actions does not violate the nondelegation doctrine. III. Finally, the majority concludes that the statutory removal restrictions applicable to SEC administrative law judges are unconstitutional because they violate Article II’s requirement that the President “take Care that the 70 462 U.S. at 923, 946. 71 Hill, 114 F. Supp. 3d at 1313. 72 Id. Case: 20-61007 Document: 00516323784 Page: 47 Date Filed: 05/18/2022 No. 20-61007 48 Laws be faithfully executed.” Specifically, the majority determines that SEC ALJs enjoy at least two layers of for-cause protection, and that such insulation from the President’s removal power is unconstitutional in light of the Supreme Court’s decisions in Free Enterprise Fund v. Public Company Accounting Oversight Board73 and Lucia v. SEC. 74 I disagree. Rather than support the majority’s conclusion, these cases explain why the SEC ALJs’ tenure protections are constitutional: ALJs perform an adjudicative function. Free Enterprise concerned the Public Company Accounting Oversight Board (“PCAOB”), which Congress created in 2002 to regulate the accounting industry.75 The PCAOB’s powers included promulgating standards, inspecting accounting firms, initiating formal investigations and disciplinary proceedings, and issuing sanctions. 76 In other words, PCAOB members were inferior officers who exercised “significant executive power.”77 The President could not remove the members of the PCAOB; rather, they could be removed by the Securities and Exchange Commission under certain, limited circumstances.78 Furthermore, SEC Commissioners cannot themselves be removed by the President except for inefficiency, neglect of duty, or malfeasance in office.79 While prior cases upheld restrictions on the President’s removal power that imposed one level of protected tenure, Free Enterprise held that these dual for-cause limitations on 73 561 U.S. 477 (2010). 74 138 S. Ct. 2044 (2018). 75 Id. at 484-85. 76 Id. at 485. 77 Id. at 514. 78 Id. at 486, 503. 79 Id. at 487. Case: 20-61007 Document: 00516323784 Page: 48 Date Filed: 05/18/2022 No. 20-61007 49 the removal of PCAOB members unconstitutionally impaired the President’s ability to ensure that the laws are faithfully executed, because “[n]either the President, nor anyone directly responsible to him, nor even an officer whose conduct he may review only for good cause, has full control over the [PCAOB].”80 Free Enterprise, however, “did not broadly declare all two-level forcause protections for inferior officers unconstitutional.”81 Furthermore, the Court expressly declined to address “that subset of independent agency employees who serve as administrative law judges.”82 The Court made two observations about ALJs that potentially distinguished them from the PCAOB: (1) whether ALJs are “Officers of the United States” was, at that time, a disputed question, and (2) “unlike members of the [PCAOB], many administrative law judges of course perform adjudicative rather than enforcement or policymaking functions or possess purely recommendatory powers.”83 The Supreme Court subsequently addressed the first observation in Lucia v. SEC. 84 There, the Court held that SEC ALJs are “inferior officers” within the meaning of the Appointments Clause in Article II.85 However, the Court again expressly declined to decide whether multiple layers of statutory removal restrictions on SEC ALJs violate Article II. 86 80 Id. at 496. 81 Decker Coal Co. v. Pehringer, 8 F.4th 1123, 1122 (9th Cir. 2021). 82 Free Enter. Fund, 516 U.S. at 507 n.10. 83 Id. (citations omitted; emphasis added). 84 138 S. Ct. 2044 (2018). 85 Id. at 2055. 86 Id. at 2051 & n.1. Case: 20-61007 Document: 00516323784 Page: 49 Date Filed: 05/18/2022 No. 20-61007 50 Thus, neither Free Enterprise nor Lucia decided the issue raised here: whether multiple layers of removal restrictions for SEC ALJs violate Article II. As the Ninth Circuit recently concluded, the question is open.87 It is important to recognize that the Constitution does not expressly prohibit removal protections for “Officers of the United States.”88 The concept that such protections may be unconstitutional is drawn from the fact that “Article II vests ‘[t]he executive Power . . . in a President of the United States of America,’ who must ‘take Care that the Laws be faithfully executed.’”89 The test is functional, not categorical: The analysis contained in our removal cases is designed not to define rigid categories of those officials who may or may not be removed at will by the President, but to ensure that Congress does not interfere with the President’s exercise of the “executive power” and his constitutionally appointed duty to “take care that the laws be faithfully executed” under Article II.90 Consistent with this standard, Free Enterprise thoroughly explained why two levels of removal protection for the PCAOB interfered with the executive power.91 The first step in the Court’s analysis focused on the fact that the PCAOB exercised “significant executive power”92 as it 87 See Decker Coal Co., 8 F.4th at 1122. 88 ERWIN CHEMERINSKY, CONSTITUTIONAL LAW § 4.2 (5th ed. 2015) (“No constitutional provision addresses the [President’s] removal power.”). 89 Free Enter. Fund, 561 U.S. at 483 (quoting U.S. CONST. , art. II §§ 1 & 3). 90 Morrison v. Olson, 487 U.S. 654, 689–90 (1988) (footnote omitted; emphasis added). 91 Free Enter. Fund, 561 U.S. at 495–96. 92 Id. at 514. Case: 20-61007 Document: 00516323784 Page: 50 Date Filed: 05/18/2022 No. 20-61007 51 “determine[d] the policy and enforce[d] the laws of the United States.”93 Then the Court explained how the PCAOB’s removal protections subverted the President’s ability to oversee this power.94 The point here is that the function performed by the officer is critical to the analysis—the Court did not simply conclude that because members of the PCAOB were “Officers of the United States” (which was undisputed)95 that dual for-cause protections were unconstitutional. Unlike the PCAOB members who determine policy and enforce laws, SEC ALJs perform solely adjudicative functions. As the Lucia Court stated, “an SEC ALJ exercises authority ‘comparable to’ that of a federal district judge conducting a bench trial.”96 Their powers include supervising discovery, issuing subpoenas, deciding motions, ruling on the admissibility of evidence, hearing and examining witnesses, generally regulating the course of the proceeding, and imposing sanctions for contemptuous conduct or procedural violations.97 After a hearing, the ALJ issues an initial decision that is subject to review by the Commission.98 Commentators have similarly observed that “SEC ALJs do not engage in enforcement or rulemaking”99 93 Id. at 484; see also id. at 508 (describing the PCAOB as “the regulator of first resort and the primary law enforcement authority for a vital sector of our economy”). 94 Id. at 498. 95 Id. at 506. 96 Lucia, 138 S. Ct. at 2049 (quoting Butz v. Economou, 438 U.S. 478, 513 (1978)). 97 Id. 98 Id. 99 Mark, supra, at 107. Case: 20-61007 Document: 00516323784 Page: 51 Date Filed: 05/18/2022 No. 20-61007 52 and proceedings before them are “analogous to that which would occur before a federal judge.”100 Free Enterprise stated, albeit in dicta, that the fact that an ALJ performs adjudicative rather than enforcement or policymaking functions may justify multiples layers of removal protection.101 I believe this to be the case. The ALJs’ role is similar to that of a federal judge;102 it is not central to the functioning of the Executive Branch for purposes of the Article II removal precedents.103 As the Southern District of New York concluded, invalidating the “good cause” removal restrictions enjoyed by SEC ALJs would only “undermine the ALJs’ clear adjudicatory role and their ability to ‘exercise[ ] .. . . independent judgment on the evidence before [them], free from pressures by the parties or other officials within the agency.’”104 In fact, the Ninth Circuit recently employed similar reasoning in Decker Coal Co. v. Pehringer, which held that two layers of removal protection for ALJs in the Department of Labor do not violate Article II.105 Like SEC ALJs, the ALJs in Decker Coal performed “a purely adjudicatory 100 David Zaring, Enforcement Discretion at the SEC, 94 Tex. L. Rev. 1155, 1166 (2016). 101 561 U.S. at 507 n.10. 102 Lucia, 138 S. Ct. at 2049. 103 Free Enter. Fund v. Public Co. Accounting Oversight Bd., 537 F.3d 667, 669 (D.C. Cir. 2008) (Kavanaugh, J., dissenting) (citing Morrison, 487 U.S. at 691–92). 104 Duka v. SEC, 103 F. Supp. 3d 382, 395–96 (S.D.N.Y. 2015), abrogated on other grounds by Tilton v. SEC, 824 F.3d 276 (2d Cir. 2016) (quoting Butz, 438 U.S. at 513–14). See also Mark, supra, at 102–08 (arguing that multiple layers of removal protection for SEC ALJs do not violate Article II); Zaring, supra, at 1191–95 (same). 105 Decker Coal Co., 8 F.4th at 1133. Case: 20-61007 Document: 00516323784 Page: 52 Date Filed: 05/18/2022 No. 20-61007 53 function.”106 The majority’s decision is in tension, if not direct conflict, with Decker Coal. Free Enterprise also noted that the exercise of “purely recommendatory powers” may justify multiple removal protections.107 When an SEC ALJ issues a decision in an enforcement proceeding, that decision is essentially a recommendation as the Commission can review it de novo.108 Even when the Commission declines review, the ALJ’s decision is “deemed the action of the Commission.”109 Furthermore, the Commission is not required to use an ALJ and may elect to preside over the enforcement action itself.110 This further supports the conclusion that the SEC ALJs’ removal protections do not interfere with the President’s executive power. The majority reasons that because Lucia determined that SEC ALJs are inferior officers under the Appointments Clause, “they are sufficiently important to executing the laws that the Constitution requires that the President be able to exercise authority over their functions,” and, consequently, multiple for-cause protections inhibit the President’s ability to take care that the laws be faithfully executed. But nowhere does the majority explain how the ALJs’ tenure protections interfere with the President’s ability to execute the laws. The majority does not mention Free Enterprise’s observation that the performance of “adjudicative rather than enforcement or policymaking functions” or “possess[ing] purely recommendatory powers” distinguishes ALJs from the PCAOB and may justify multiples 106 Id. 107 Free Enter. Fund, 561 U.S. at 507 n.10. 108 See Lucia, 138 S. Ct. at 2049 (citing 17 C.F.R. § 201.360(d)); 5 U.S.C. § 557(b). 109 Lucia, 138 S. Ct. at 2049 (quoting 15 U.S.C. § 78d-1(c)). 110 Id. (citing 17 C.F.R. § 201.110). Case: 20-61007 Document: 00516323784 Page: 53 Date Filed: 05/18/2022 No. 20-61007 54 layers of removal protection for ALJs.111 The majority does not mention that Lucia found SEC ALJs to be similar to a federal judge.112 The majority does not mention Decker Coal. Instead, the majority applies what is essentially a rigid, categorical standard, not the functional analysis required by the Supreme Court’s precedents.113 Accordingly, I disagree with the majority that multiple layers of removal protection for SEC ALJs violate Article II. Because SEC ALJs solely perform an adjudicative function, and because their powers are recommendatory, these removal restrictions do not interfere with the President’s ability to “take Care that the Laws be faithfully executed.” IV. I find no constitutional violations or any other errors with the administrative proceedings below. Accordingly, I would deny the petition for review. 111 561 U.S. at 507 n.10. 112 138 S. Ct. at 2049. 113 Morrison, 487 U.S. at 689–90. The majority also cites Myers v. United States, 272 U.S. 52, 135 (1926), for the proposition that quasi-judicial executive officers must be removable by the President. But that part of Myers is dicta, which is why the Court disregarded it in Humphrey’s Executor v. United States, 295 U.S. 602, 626–28 (1935). Case: 20-61007 Document: 00516323784 Page: 54 Date Filed: 05/18/2022 What affect will it have on the proposed rule that you have lost your last 4 of 5 cases in the supreme court? Is there actually a revolving door between the giant megacorps that you guys slap on the wrist with tiny fines, while totally savaging the small companies? Have you refuted this study? This is just one of many such studies: The SEC Revolving Door and Comment Letters? Michael Shen NUS Business School, National University of Singapore Samuel T. Tan School of Accountancy, Singapore Management University Abstract Government officials, advocacy groups, and the business press have raised concerns that former SEC employees may continue to influence the SEC after leaving the agency. Using hand-collected data on the characteristics of 1,384 lawyers who represented firms in responding to SEC comment letters between 2005 and 2016, we examine the impact of post-revolving SEC employees on the SEC comment letter process. Among other determinants, we find that older and larger firms with a history of litigation are more likely to hire former SEC lawyers over non-SEC lawyers. Relative to firms that involve only non-SEC lawyers, we find that firms that involve former SEC lawyers in responding to SEC comment letters negotiate to a greater extent with the SEC, and have a lower likelihood and number of amendment filings, after matching on lawyer, law firm, comment letter, and firm characteristics. Keywords: Revolving door, regulatory capture, SEC comment letters, external counsel JEL: G18, K22, M41, M48 ?We thank Marco Trombetta (editor), two anonymous reviewers, Minyue Dong (discussant), Ira Yeung (discussant) and other conference participants at the 2018 AAA Western Region Meeting, the 2018 AAA Annual Meeting, the 2019 SMU/NTU/NUS Accounting Research Conference, and the 2019 EAA Annual Congress; and workshop participants at American University, Hong Kong University of Science and Technology, and University of Technology Sydney for very helpful comments. Shen acknowledges financial support from National University of Singapore’s Start-Up Grant (A-0003869-00-00), and Tan acknowledges financial support from Singapore Management University’s Della Suantio Fellowship and Start-Up Research Grant. Email addresses: bizshen@nus.edu.sg (Michael Shen), samueltan@smu.edu.sg (Samuel T. Tan) April 5, 2023 1. INTRODUCTION In this paper, we examine how the revolving door at the Securities and Exchange Commission (hereafter SEC) affects its oversight on corporate disclosure. Specifically, we examine whether firms that involve former SEC lawyers in the comment letter process negotiate to a greater extent with SEC, and have more favourable outcomes as a result of the process. The flow of personnel and expertise from the SEC to the private sector has been the subject of scrutiny and debate by government officials, advocacy groups, the business press, and academics in recent years. A 2013 report by the Project On Government Oversight (POGO) found that over 400 former SEC employees filed statements between 2001 and 2010 that they intended to represent an external party before the SEC, and expressed the concern that “a rapidly spinning revolving door can weaken the agency’s protection of investors” (Smallberg, 2013, quotation at p. 37).1 In response to the POGO report, U.S. Senator Chuck Grassley wrote that “The SEC has to fix this problem once and for all” (Grassley, 2013). Furthermore, much of the recent media coverage on the SEC revolving door has skewed negatively, and the agency has also received numerous comments from the public about the issue.2 Critics of the revolving door cite the risk of regulatory capture—the risk that regulators “created to act in the public interest [...] end up acting directly or indirectly in the interests of those they regulate.” (Brown, 2016, p. 1). Based on data compiled by POGO, the majority of former SEC employees who filed post-employment disclosures revolved to private law firms, where they may represent clients before the SEC.3 1For two years after they have left the SEC, former SEC employees who intend to represent a client before the SEC or in communications with the SEC are required to file statements with the SEC’s Office of the Ethics Counsel (see 17 CFR §200.735-8b). 2Examples of negative media coverage include “SEC Lawyer One Day, Opponent the Next” (McGinty, 2010a, The Wall Street Journal); “SEC ’Revolving Door’ Under Review” (McGinty, 2010b, The Wall Street Journal); “SEC staff’s ’revolving door’ prompts concerns about agency’s independence” (Hilzenrath, 2011, The Washington Post); and “S.E.C.’s Revolving Door Hurts Its Effectiveness, Report Says” (Protess and Craig, 2013, The New York Times). A search for the term “revolving door” on the SEC’s website also reveals numerous and often strongly-worded comments from the public about the revolving door. 3Based on 2001–2010 post-employment statements filed by former SEC employees that disclose their intention to represent an external party before the SEC, obtained by POGO via Freedom of Information Act requests. Retrieved in August 2017 from https://protect2.fireeye.com/v1/url?k=31323334-50bba2bf-3132d782-4544474f5631-36c37628fcfc4bd8&q=1&e=0f8ec12d-e43b-4a7b-9ff0-a6793fd596e3&u=http%3A%2F%2Fwww.pogo.org%2Ftools-and-data%2Fsec-revolving-door-database%2F. 2 There are two mechanisms by which the revolving door might compromise regulatory oversight efforts. The first is that it could incentivize individuals at regulatory agencies to act in the interests of prospective employers before they join the private sectors. deHaan et al. (2015) provide empirical evidence against this form of regulatory capture, finding that SEC trial lawyers’ enforcement efforts are not compromised before they revolve to private law firms; instead, these lawyers are associated with more aggressive enforcement, consistent with the human capital theory of the revolving door.4 A second mechanism is that individuals formerly at a regulator may influence it after revolving to the private sector, on behalf of a private sector client. This mechanism is a primary concern of POGO’s 2013 report.5 Several studies have examined the impact of post-revolving SEC employees; for example, deHaan et al. (2015) and Shive and Forster (2017) find evidence that former employees of regulators, including the SEC, help their clients reduce damages from enforcement actions, although Shive and Forster (2017) cautions that they “do not find strong evidence” (p. 1480) that firms that hire former regulators face more lenient enforcement. The SEC itself has argued that ethics rules safeguard against conflicts of interest, and that the revolving door actually improves regulatory compliance.6 Our study extends this line of research by examining the impact of post-revolving SEC employees on the SEC comment letter process, which affects every issuer in the United States and could result in adverse consequences for firms. This is an ideal setting because each comment letter conversation involves a dialogue between representatives of the firm and SEC staff, and firms often involve lawyers in their response (see Bozanic et al., 2019), who may have formerly worked 4The human capital theory “focuses on incentives the revolving door creates for regulators to signal the type of human capital valued by industry employers” (Zheng, 2015, p. 1268). 5See Smallberg (See 2013, p. 2): “Former employees of the Securities and Exchange Commission (SEC) routinely help corporations try to influence SEC rulemaking, counter the agency’s investigations of suspected wrongdoing, soften the blow of SEC enforcement actions, block shareholder proposals, and win exemptions from federal law.” 6 In a 2005 speech, then-SEC Commissioner Paul Atkins said that the revolving door improves compliance in the private sector because “people with regulatory backgrounds become good proselytes for lawful behavior” (Atkins, 2005); and in 2007 he said that “people who leave government to return to industry can help to instill a proper sense of the importance of complying with legal obligations [...] strong ethics rules guard against potential conflicts of interest when people leave the government for the private sector.” (Atkins, 2007). Such ethics rules include former SEC employees being permanently barred from appearing before the SEC on a matter they worked on while at the SEC (see 18 USC §207). 3 at the SEC. The SEC issues comment letters when its staff identifies disclosure deficiencies, and may make requests of the firm, for example to amend prior SEC filings (see Bozanic et al., 2017). The firm may negotiate for more desirable outcomes, for example to revise future filings instead. Former SEC lawyers may be better positioned to help their clients secure more desirable outcomes than other lawyers due to their familiarity with former colleagues still at the SEC, for example prior studies have documented that negotiation outcomes may be affected by the extent of familiarity and trust between the parties (e.g. Lumineau and Henderson, 2012; Druckman and Olekalns, 2013). In Appendix A, we describe the comment letter process in more detail, and include examples of comment letter conversations involving lawyers formerly employed by the SEC. In order to identify the involvement of former SEC lawyers in SEC comment letters, we handcollect the characteristics of the individual lawyers involved in our sample of comment letter conversations from public sources, primarily LinkedIn profiles and biographies on law firm websites.7 Our dataset includes the former SEC employment status, educational background, and other characteristics of over 1,300 unique lawyers. After attrition our final sample comprises 4,524 comment letter conversations initiated between 2005 and 2016 that involve external counsel; about 7.2 percent of the sample involve lawyers who were formerly employed by the SEC. We first examine the determinants of hiring a former SEC lawyer over only non-SEC lawyers. We find that larger and older firms, firms that recently faced securities class action lawsuits, and firms that do not use top- or second-tier audit firms are more likely to retain a former SEC lawyer relative to only non-SEC lawyers. These results suggest that financial and reputational concerns, and the risk of litigation and misstatements, may contribute to the decision to hire a former SEC lawyer. As these incentives may confound our tests of the consequences of hiring such lawyers, in our main analyses we propensity-match conversations involving former SEC lawyers against conversations involving non-SEC lawyers along lawyer and law firm characteristics, the complexity 7Firms often involve external counsel in responding to comment letters: they were involved in about a third of our initial sample of comment letter conversations. We restrict the hand-collection to lawyer-law firms with at least two conversations in our initial sample. Please see Section 4 and Table 1 for details on our sample selection procedure. 4 of and the issues raised in the initial comment letter, and firm characteristics. We next examine whether resistance to the SEC is greater when the firm retains a former SEC lawyer relative to when it retains other lawyers, consistent with regulatory capture. We construct a composite measure of the extent of negotiation based on the length of the conversation in days, the number of letters exchanged, and whether the conversation took multiple rounds to resolve. We find that relative to matching firms that involve non-SEC lawyers, firms that involve former SEC lawyers negotiate to a greater extent with the SEC. We additionally find that they are more likely to request confidential treatment. Finally, we examine whether former SEC lawyer involvement is associated with fewer amendments. We find that relative to matching firms that involve non-SEC lawyers, firms that involve former SEC lawyers have about a 32 percent lower odds of amending their periodic filings after receiving a comment letter, and file 0.38 fewer amendments to periodic filings on average. We find that abnormal returns during the period are 3.9 percentage points greater for firms that involve former SEC lawyers. However, on aggregate we do not find that former SEC lawyer involvement is associated with the likelihood of adverse restatement or securities class action lawsuits, suggesting that they help firms reduce the cost of amending filings, but not to the extent of avoiding disclosures that would have resulted in adverse restatements or litigation. As our study documents a relationship between involvement of former SEC lawyers and outcomes of the comment letter process, an alternative interpretation of our findings is that firms predisposed to certain outcomes are more likely to hire former SEC lawyers. That is, our findings may be influenced to an extent by endogenous sorting between firms and lawyers. We acknowledge and caution readers that this possibility limits to some degree our ability to draw causal inferences from our empirical analyses. However, we mitigate selection bias with our research design in several ways. First, we identify the impact of former SEC lawyers by using firms that hire non-SEC lawyers in the control group, which precludes selection bias due to the choice of whether to hire a lawyer. Second, we propensity-match conversations involving former SEC lawyers against those that involve only non-SEC lawyers along a large set of observable lawyer, law firm, comment let5 ter, and firm characteristics, which mitigates selection bias due to the choice of whether to hire a former SEC lawyer relative to a non-SEC lawyer. Finally, we examine our proposed mechanisms for the impact of former SEC lawyers on outcomes of the comment letter process using of cross-sectional tests. As we detail in Section 2.2, former SEC lawyers are likely to be more familiar with current SEC staff, allowing them to more effectively influence the outcomes of the process on behalf of their clients. In cross-sectional tests, we find that the impact of former SEC lawyers on increasing the extent of negotiation, reducing the likelihood of amendments amendments, and reducing the likelihood of restatements or litigation, are driven to a larger extent by former SEC lawyers who had left the SEC more recently. To our knowledge, our study is the first to examine the impact of the revolving door between the public and private sectors on the SEC comment letter process. Our study therefore contributes to three streams of research. First, our study contributes to the literature on the revolving door itself. Extant literature on the revolving door examine how it affects regulatory agents’ actions before they join prospective employees (e.g. Lucca et al., 2014; deHaan et al., 2015; Cornaggia et al., 2016; Tabakovic and Wollmann, 2018), and the impact of former agents on their new employers’ outcomes (e.g. Luechinger and Moser, 2014; deHaan et al., 2015; Shive and Forster, 2017). We extend the latter literature by examining the impact of former SEC lawyers on the SEC comment letter process, a ubiquitous and consequential mechanism for disclosure oversight, in which negotiation between SEC staff and representatives of the firm plays a crucial role. Consistent with regulatory capture, we find evidence that former SEC lawyers impede the effectiveness of the comment letter process. Second, our study contributes to the growing literature on the SEC comment letter process. Recent studies (e.g. Bens et al., 2016; Kubick et al., 2016; Wang, 2016; Johnston and Petacchi, 2017; Calderon and Gao, 2022) have documented that SEC regulatory oversight over corporate disclosures improves firms’ information environments and accounting choices. For example, Kubick et al. (2016) find that firms decrease their tax avoidance behavior after receiving tax-related comment letters, and Johnston and Petacchi (2017) find that the comment letter process decreases 6 firms’ information asymmetry and increases their earnings response coefficients. However, few studies have examined factors that could impede effective oversight of the comment letter process. One factor that has been examined empirically (e.g. Kedia and Rajgopal, 2011; Ege et al., 2020) is resource constraints at the SEC. For example, Ege et al. (2020) find that waves of transactional filings (e.g., IPO) reduces the quality of comment letters for periodic firm filings. Our paper extends this line of research by documenting that the SEC revolving door also impedes effective oversight on corporate disclosure by the SEC. Last, our study contributes to the emerging literature on the role of firms’ external legal counsel in financial reporting (e.g. deHaan et al., 2015; Bozanic et al., 2019; Dechow and Tan, 2021). Bozanic et al. (2019) find that firms that involve law firms in SEC comment letter conversations are more likely to file confidential treatment requests and file fewer amendments, but have improved disclosures in the long run, consistent with law firms serving as client advocates while also improving their clients’ disclosures. In contrast, Dechow and Tan (2021) find that the spread of stock option backdating was facilitated by the network of shared external law firms. Our study examines former SEC lawyers within a sample of companies that retain law firms, so unlike Bozanic et al. (2019) we do not make inferences on the impact of hiring a law firm, relative to not hiring one. Instead, our study contributes to this line of research by documenting that the characteristics of individual lawyers—for example former SEC employment status—matter, and may be associated with adverse long-run accounting and shareholder outcomes. 2. RELATED LITERATURE AND HYPOTHESIS DEVELOPMENT 2.1. Related literature To our knowledge, our study is the first to examine the impact of the SEC revolving door on the SEC comment letter process, one of the SEC’s most important mechanisms for oversight over corporate disclosures. Our study therefore builds on literature that examines the impact of SEC lawyers on public companies after they have revolved from the SEC to the private sector. 7 In particular, deHaan et al. (2015) and Shive and Forster (2017) examine the impact of former regulators, including former SEC personnel, on their clients’ outcomes in enforcement actions.8 deHaan et al. (2015) focus on the regulatory actions of SEC lawyers while they are at the SEC, but also examine whether law firms with more former SEC lawyers with enforcement experience at the SEC help their clients secure more lenient enforcement outcomes. They find that law firms hiring more former SEC lawyers secure lower damages, and some evidence that they secure fewer criminal charges, for their clients. They find that the results are not driven by former SEC lawyers who had worked in the same team or office as a current SEC lawyer on the case, which is inconsistent with a social influence story, and more consistent with former SEC lawyers having gained “unique training and insider knowledge” from their tenure at the SEC (footnote 35). Shive and Forster (2017) find that financial firms that hire former employees of regulatory agencies including the SEC have lower firm risk and improved risk management activities, and “do not find strong evidence” (p. 1480) that these firms face more lenient enforcement, although they do not rule it out. Specifically, they find that no evidence that enforcement actions are less likely when a former regulator is hired, and some evidence that fines are smaller. Our study extends this line of research by examining the SEC comment letter process instead of enforcement actions. First, the SEC comment letter process is a setting over which former SEC employees are much more likely to be able to exert influence, because the process is based on negotiations between representatives of firms and SEC staff. Filing reviews are completed by staff accountants and staff examiners, and senior staff members are only involved upon the firm’s request.9 In contrast, the outcomes of enforcement actions are decided at the highest levels of the 8Several empirical studies have examined the revolving doors at other government offices or regulatory agencies, including state public utility commissions (Law and Long, 2012), the U.S. Department of Defense (Luechinger and Moser, 2014), legislators’ offices (Igan and Mishra, 2014), credit rating agencies (Cornaggia et al., 2016), and the U.S. Patent and Trademark Office (Tabakovic and Wollmann, 2018). 9We outline the comment letter process and provide examples of legal counsel involvement at Appendix A. Detailed information on how the SEC conducts the comment letter process is also publicly available on the SEC’s website, for example https://www.sec.gov/divisions/corpfin/cffilingreview.htm and https://www.sec.gov/oig/reportspubs/ aboutoigaudit259finhtm.html. 8 SEC, limiting the extent to which a representative of the firm may influence outcomes.10 Second, the comment letter process represents an early and preemptive stage of the SEC’s oversight on issuers’ disclosures, and therefore has important implications for firms’ future disclosure quality and shareholder value. For example, recent studies have found that the comment letter process improves firms’ information environments (e.g. Bens et al., 2016; Wang, 2016; Johnston and Petacchi, 2017; Calderon and Gao, 2022), and may also trigger restatements (e.g. see Kubic, 2020) and lead to enforcement actions (Karpoff et al., 2008). In contrast, enforcement actions are typically the end of a long investigation, occurring long after disclosure deficiencies are revealed and most of the damage done to shareholder value. Karpoff et al. (2008) document that abnormal negative returns are greatest around the trigger events (−25.24% per trigger on average), and smallest around the enforcement filings themselves (−6.56% per filing on average). Finally, outcomes of the SEC comment letter process affect thousands of firms every year, while few firms ever reach the stage where they would be attempting to influence the outcome of an enforcement action: Karpoff et al. (2008) document a total of only 585 enforcement actions initiated by the SEC and the DOJ between 1978 and 2002. In contrast, as discussed in more detail in Appendix A, the SEC reviews firms’ filings every three years; over 100,000 comment letter conversations were initiated between 2005 and 2016 (see Table 1). Another thread of related literature examines the impact of firms’ external legal counsel on financial and financial reporting outcomes (e.g. Krishnan and Masulis, 2013; deHaan et al., 2015; Bozanic et al., 2019; Dechow and Tan, 2021). In particular, Bozanic et al. (2019) compare firms that involve external legal counsel in SEC comment letter conversations against firms that do not, finding that the former are more likely to file confidential treatment requests and file fewer amendments as a result of the comment letters, but have improved disclosures and fewer restatements and 10For example, formal orders of investigation are authorized and issued by the director of the Division of Enforcement, and enforcement actions are voted upon by the SEC commissioners, who may question the recommendations of the staff. See the SEC’s Enforcement Manual, particularly sections 2.3.4 and 2.5, available at the SEC’s website at https://www.sec.gov/divisions/enforce/enforcementmanual.pdf, and see also Correia (2014) for a discussion of the SEC enforcement process. 9 comment letters in the two years following resolution of the comment letter. While our study also uses the SEC comment letters setting, we do not make inferences on the impact of hiring a law firm and instead examine a different research question: the impact of the SEC revolving door on the comment letter process. Our study therefore contributes to this line of research by documenting that the characteristics of individual lawyers—for example former SEC employment status— matter, and may be associated with adverse long-run accounting and shareholder outcomes.11 2.2. Hypothesis development We hypothesize that former SEC lawyers are better able to influence the outcomes of their clients’ comment letter conversations with SEC staff, relative to lawyers who had never been employed by the SEC. Former SEC lawyers are likely to be more familiar with current SEC staff, and prior literature has documented that the outcomes of negotiations are affected by the degree of trust between the parties (e.g. Lumineau and Henderson, 2012; Druckman and Olekalns, 2013). If former SEC lawyers exert their influence on the SEC comment letter process on behalf of their clients, we would expect to observe greater resistance by the clients to the process in the form of more protracted negotiation. The null hypothesis is therefore that there is no difference in resistance between firms that retain former SEC lawyers and other lawyers. We summarize this in the following formal hypothesis, in alternative form. Hypothesis 1. The involvement of a lawyer formerly employed by the SEC in the SEC comment letter process is associated with greater negotiation than the involvement only of lawyers who had not been employed by the SEC, ceteris paribus. If former SEC lawyers are more effective at influencing the SEC comment letter process, we would expect to observe more favourable outcomes of the comment letter, particularly a reduction 11Several studies (e.g. Kwak et al., 2012; Hopkins et al., 2015; Ham and Koharki, 2016) have also examined the role of general counsel on firms’ disclosure policies, financial reporting quality, and risk management. Our study focuses on external counsel rather than general counsel because based on a database of 2001–2010 post-employment disclosures compiled by POGO (retrieved in August 2017 from https://protect2.fireeye.com/v1/url?k=31323334-50bba2bf-3132d782-4544474f5631-4b533c6e4d38b6eb&q=1&e=0f8ec12d-e43b-4a7b-9ff0-a6793fd596e3&u=http%3A%2F%2Fwww.pogo.org%2Ftools-and-data%2Fsec-revolvingdoor-database%2F), former SEC employees are substantially more likely to revolve to law firms than public companies. 10 in the likelihood and number of amendment filings, and fewer adverse events including restatements and shareholder litigation. Hypothesis 2. The involvement of a lawyer formerly employed by the SEC in the SEC comment letter process is associated with more favourable firm outcomes than the involvement only of lawyers who had not been employed by the SEC, ceteris paribus. 3. RESEARCH DESIGN 3.1. Impact of involving former SEC lawyers We test Hypotheses 1 and 2 by regressing each outcome variable against a dummy variable (SEC) that takes the value of one if the firm referenced a former SEC lawyer in a comment letter in the conversation, and zero otherwise: Extent of Negotiation = α + β SEC + γ Lawyer, CL, and Firm Controls (1) Outcomes = α + β1 SEC + γ Lawyer, CL, and Firm Controls (2) We include controls for lawyer and law firm characteristics, comment letter complexity and issues, firm characteristics, and industry and year fixed effects. We estimate the models using ordinary least squares, logistic regressions, and poisson regressions for the models with continuous, dummy, and count dependent variables respectively. 3.2. Propensity score matching To further mitigate bias due to confounding by one or more control variables, we estimate each of the regression models after propensity-matching treatment (SEC = 1) and control (SEC = 0) conversations along the control variables (detailed at Section 3.5 below), including industry fixed 11 effects. Specifically, we estimate the following model using a logistic regression: SEC = α + γ Lawyer, CL, and Firm Controls (3) The logistic regressions allow us to obtain the predicted probability that a conversation with a given set of lawyer, comment letter, and firm characteristics involves a former SEC lawyer. We then apply matching procedures to minimize the difference in predicted probabilities, or the distance scores, between the treatment and control conversations. We first discard observations outside of the support of the distance score, which drops extreme observations from the sample and ensures that treatment and control conversations are comparable (see Stuart, 2010, Section 3.3). We next minimize the difference in distance scores using full matching (Rosenbaum, 1991; Hansen, 2004), which produces a set of weights that we apply to each observation remaining in the sample when estimating the regression models.12 3.3. The extent of negotiation We measure the extent of negotiation using a composite variable, Negotiation, which is based on the length of the comment letter conversation both in days and in the number of letter exchanged. Specifically, we use the first principal component of the log-transformed length of the comment letter conversation in days (ConvT ime), the number of comment letters in the conversation (NumLetters), and whether the conversation takes at least two rounds of exchanges (i.e. at least five letters) between the firm and the SEC to resolve (MultiRound). Please see Appendix B for detailed variable definitions. To further explore how the negotiation process is influenced by former SEC lawyers, we also examine whether the involvement of former SEC lawyers is associated with a greater likelihood of 12Full matching matches treatment and control units (or vice versa) one-to-many. To take into account potential changes in the underlying prediction model over time, we carry out the matching procedure within two-year windows. The matching procedures in our analyses are carried out using the MatchIt package (Ho et al., 2011) in R, and full matching uses functions from the optmatch package (Hansen and Klopfer, 2006). See also Armstrong et al. (2010) for one of the first studies in accounting that use propensity score matching. 12 confidential treatment requests. A lawyer who has greater influence on the SEC may use that influence to secure redaction of sensitive firm information in the comment letter disclosures. Bozanic et al. (2019), for example, documents that the involvement of lawyers in the comment letter process increases the likelihood of confidential treatment requests. 3.4. Outcomes of the comment letter conversation Of the disclosure corrections usually requested by the SEC, amendments of prior filings are the most costly, and managers would prefer to obscure any deficiencies in its prior filings by revising future filings instead. As the comment letter process aims to correct disclosure deficiencies, any undue delay or reduction in amendments associated with former SEC lawyer involvement would be disadvantageous to shareholder welfare. We therefore examine 10-K or 10-Q amendments filed as a result of the comment letter (Amendment, NumAmend). As in Cassell et al. (2013, see footnote 34), we use the start and disclosure dates of the conversation as the window in which a company would make disclosure corrections related to the comment letters.13 We also examine size-adjusted returns during the period (CLReturns): if firms involving former SEC lawyers file fewer amendments, we would expect them to have higher abnormal returns during the period.14 Finally, we also examine the relationship between former SEC lawyers involvement and whether the comment letter resulted in restatements or litigation. We use a composite variable, Adverse, that is equal to one if an adverse restatement or securities class action lawsuit was filed as a result of the comment letter.15 A significantly negative relationship would suggest that former SEC 13For the few conversations in our sample with multiple disclosure dates, i.e. the letters in the conversation were made public on EDGAR at different dates, we use the last available date. A small number of conversations in our final sample have disclosure dates more than two years after the end of the conversation; we truncate these at two years. 14We use CRSP size deciles based on the year before the start of the conversation, and include delisting returns, assuming that shareholders invest in the corresponding size portfolio after delisting. In our regression involving CLReturns, observations with missing CRSP size portfolio or missing portfolio returns are dropped. 15We examine shareholder lawsuits filed on or after the start of the conversation, for which the class period overlaps with the period referenced by the comment letter conversation, and we examine adverse restatements filed between the start of comment letter conversation and the disclosure of the conversation, for which the restatement period overlaps with the periods referenced by the comment letters. Assuming that Forms 10-K (10-Q) are filed 90 days (45 days) after the end of the fiscal period, we use the period spanned by start of the earliest fiscal period referenced and the last periodic filing date referenced as the period referenced by the comment letter conversation. 13 lawyers help firms avoid the disclosure of information that would reveal GAAP misstatements or fraud. 3.5. Control variables Following prior studies, we use three sets of control variables in this study: lawyer and law firm characteristics, SEC comment letter complexity and issues, and firm characteristics. Several of the lawyer and law firm characteristics are based on our hand-collected database of lawyers described in detail in Section 4. We outline the control variables in this section briefly for brevity; please see Appendix B for detailed variable definitions. Lawyer and law firm characteristics. We include proxies for the lawyers’ and law firms’ prior experience in advising clients on SEC comment letters, and the lawyers’ educational backgrounds. Lawyers often detail their prior experience and education in their online biographies, suggesting that this information is useful to current and prospective clients. Furthermore, in a different setting, Karsten et al. (2019) provide evidence that lawyers’ education and expertise could affect their clients’ outcomes. We include educational background proxies based on universities with topfourteen (“T14”) law schools, widely-regarded as the premier law schools in the United States.16 Finally, we control for the number of lawyers involved by the firm, which could be mechanically correlated with the probability that one of the lawyers was formerly employed by the SEC, and include proxies for gender and estimated age. The literature on the legal profession (e.g. see Bruck and Canter, 2008; Adams and Engel, 2015) has suggested that there are differences in education, career trajectories, or salary between male and female lawyers, even at the SEC (Choi et al., 2019), while the lawyer’s age provides an additional proxy for experience.17 16For example, see Adams and Engel (2015) (“Virtually all prospective law students who thoroughly research law schools across the nation are aware of a phenomenon referred to as the T-14, a list of 14 law schools that are annually ranked in the top 14 of the U.S. News & World Report Law School Rankings”, page 1216, internal quotation marks omitted). The T14 are the law schools of Columbia University, Cornell University, Duke University, Georgetown University, Harvard University, New York University, Northwestern University, Stanford University, the University of California, Berkeley, the University of Chicago, the University of Michigan, the University of Pennsylvania, the University of Virginia, and Yale University. We thank an anonymous referee for suggesting that we partition the T14 schools further into the top three schools (Harvard, Stanford, or Yale Law School) and the rest of the T14. 17Companies may involve more than one lawyer in the comment letter conversation; in our sample, about 14 percent 14 Comment letter complexity and issues. We control for the complexity of the initial comment letter from the SEC using the length of the letter and the number of SEC filings referenced in the initial letter. We control for the specific issues raised by the SEC at the initial comment letters, coded based on Audit Analytics’ issue variables and taxonomy keys. As in Bozanic et al. (2019), we use dummy variables that take the value of one if a specific issue was referenced. We include a dummy variable equal to one if the initial comment letter mentions the word “amend”, to proxy for whether the SEC suggested or requested that the firm file an amendment of prior filings in response. Finally, when estimating the relationship between the outcomes and former SEC lawyer involvement (Model 2), we also include a proxy for the extent of negotiation, NumLetters, in the matching and regressions to control for any indirect effect of former SEC lawyers via the length of the conversation. Firm characteristics We control for a large set of firm characteristics. These include proxies for the the size, growth, performance, volatility, and age of the firm, to proxy for the reputation and financial performance of the firm, which would affect incentives to retain higher-quality or more expensive external legal counsel. We also include recent evidence of disclosure issues including recent restatements and shareholder litigation, to control for underlying accounting and disclosure issues. Cassell et al. (2013), for example, find that companies that restated during the fiscal year under review by the comment letter, or the prior year, have longer comment letter conversations. Finally, we control for the tier of the firm’s auditor, which may be associated with the extent to which the firm requires external assistance in responding to the SEC. Please see Table B.2 for a more complete list of the variables and their definitions. Finally, we include industry fixed effects based on one-digit SIC codes to control for possible variation between offices at the SEC.18 We also include year fixed effects based on the calendar year in which the conversations began. of conversations do so. Since our unit of analysis is the comment letter conversation, we aggregate the lawyer and law firm controls variables when necessary. For example, F emale takes the value of one when any of the lawyers involved by the firm was female, and HSY School is equal to one if any of the lawyers involved in the conversation attended a university with a top-three law school (please see the notes to Table B.1 for further details). 18Review of filings at the SEC’s Division of Corporation Finance is organized by industry: filings are assigned to one of eleven offices, each of which covers a broad industry group. 15 4. SAMPLE AND DESCRIPTIVES 4.1. Sample selection and data sources We retrieve comment letter conversations initiated between 2005 and 2016 that reference a 10- K filing from Audit Analytics’ Comment Letter Conversations database. We begin our sample in 2005 because the comment letters data is sparse before then. We require the firms in the conversations to be covered by Compustat as of the most recent fiscal year, and by CRSP as of the starting date of the conversation, resulting in an sample of 16,734 conversations. Similar to Bozanic et al. (2019), we omit conversations with only one letter and conversations below the bottom percentile of the time to resolution. These steps result in an initial sample of 16,351 conversations involving 5,534 unique firms. Please see Panel A of Table 1 for details. Next, we retrieve and require legal advisor data (Panel B of Table 1). We retrieve data on law firm personnel referenced in firms’ replies to SEC comment letters (i.e., CORRESP files) using the People Search tool on auditanalytics.com, restricting the search to comment letters referencing periodic filings. Of the 16,351 conversations in our initial sample, 5,580 conversations have replies by firms in which lawyers were referenced.19 In other words, firms involved lawyers in about 34% of conversations. In these 5,580 conversations, firms referenced a total of 2,865 unique lawyers from 539 unique law firms. We next require the availability of hand-collected data on the characteristics of the individual lawyers referenced in firms’ replies. For each unique lawyer-law firm combination, we retrieve and read the lawyer’s biographies on public sources, primarily LinkedIn profiles and biographies on law firm websites. We collect each lawyer’s individual characteristics including the lawyer’s educational history, gender, and SEC affiliation before joining the law firm, if any. We restrict our hand-collection to the lawyer-law firms with at least two conversations in our initial sample.20 19We also omit 82 conversations in which the lawyers were not named. 20We avoid using lawyers who appear only once in the data, because they are more likely to have been involved incidentally, rather than as lawyers specialising in securities law and government relations. 16 We also identify and omit several individuals that Audit Analytics appeared to have miscoded as external counsel. This restricts the sample to 4,689 conversations involving 1,407 unique lawyers and 326 unique law firms. We were able to collect data on 1,384 of the lawyers. For several lawyers for whom we could not locate certain educational characteristics needed to construct the control variables (particularly LawyerAge and NumDegrees), we extrapolate the missing educational information from other available information.21 Finally, we require the availability of variables used in our analyses, resulting in a final sample of 4,524 conversations involving 2,140 unique firms. The number of unique firm-years in the sample, 4,507, is very close to the sample size, indicating that the conversations in our sample map almost uniquely to firm-years. For our analyses involving future lawsuits in the dependent variable (Adverse), we further omit comment letter conversations that follow a securities lawsuit filing (i.e. we omit conversations where P astLit = 1). We do so because securities lawsuits may take several years to resolve, and different lawsuit filings may be consolidated into a single case. Therefore if similar issues triggered both the comment letter conversation and the pre-conversation lawsuit filing, any future litigation related to the comment letter may be consolidated into the existing lawsuit. 4.2. Descriptive statistics 4.2.1. Lawyer characteristics Our hand-collected dataset includes 1,447 unique combinations of lawyers and law firms represented in our sample.22 There are 1,384 unique lawyers in the data, indicating that only a small proportion of the lawyers in our sample switched law firms, and 320 unique law firms in our sam21Specifically: (1) If the lawyer’s degree years were all unavailable, we use the year the lawyer was admitted to the bar as the year of his or her terminal degree. From our reading of the lawyers’ biographies, almost all were admitted to the bar at or around the year of their terminal degree. (2) If the year of the lawyer’s first degree was unavailable, we use the year of his or her next available degree minus three years, the median time between first and second degrees of the lawyers for whom the data was available. (3) Several lawyers only disclosed one of their degrees. We set the floor of NumDegrees to two because lawyers generally require a graduate degree (e.g. a Juris Doctor) to practice. 22We collect the data at the lawyer-law firm level to take moves between law firms into account. For example, to collect the SEC variable, we examine whether the lawyer was employed by the SEC before joining the focal law firm. 17 ple. Table 2 documents the descriptive statistics for our dataset of lawyers involved in the comment letter process.23 Panel A of Table 2 documents the characteristics of the 1,384 unique lawyers in our dataset. On average, a lawyer is referenced in about 4.1 comment letter conversations in our sample, and are almost always associated with only one unique law firm throughout the sample. 62 of the 1,384 lawyers had previously been employed by the SEC, about 4.5 percent of our sample. 21.1 percent of the lawyers had attended a university with a top-three law school, 45.2 percent had attended a university with other top-fourteen law schools, and most have up to two degrees. About 16.5 percent of the lawyers are female. Panel B of Table 2 compares the former SEC lawyers in our sample with the lawyers who were not formerly employed by the SEC. Former SEC lawyers are involved in more than one more conversation in our sample than non-SEC lawyers on average, although the statistical significance of the difference is slightly lower than conventional significance levels. There are statistically significant differences in educational background between former SEC and non-SEC lawyers: SEC lawyers are a little more likely to have earned more than two degrees, but are less likely to have attended a university with a top law school. The difference in the number of law firms the lawyers are associated with or the likelihood that the lawyers are female are not statistically significant at conventional significance levels. 4.2.2. Comment letter conversations Our sample comprises 4,524 comment letter conversations initiated between 2005 and 2016, after our sample selection procedure detailed at Table 1. Table 3 documents descriptive statistics for our sample at the conversation level, which is the unit of observation in our study.24 Panel A of Table 3 shows descriptive statistics of the outcome variables. As a principal compo23Unless otherwise stated, the descriptives in this subsection are based on raw values, without winsorization or other outlier removal method. 24At the conversation level, we reduce the impact of outliers by winsorizing annually count variables such as NumLawyers and NumF ilings at the top percentile, and other non-discrete variables such as ConvT ime and ROA at the top and bottom percentiles. We do not winsorize variables based on returns, specifically BHR and StdRet. Please see Appendix B for variable definitions. 18 nent, Negotiation is centered on zero, and about 19.6 percent of conversations discuss confidential treatment requests. About 17.1 percent of conversations result in a 10-K or 10-Q amendment, and the mean number of amendments is higher, at 0.243, because some conversations result in more than one amendment. On average, abnormal returns from the start to the disclosure of the conversation is close to zero, and about 7.6 percent of conversations result in adverse restatements or shareholder litigation. Panel B of Table 3 shows descriptive statistics for the lawyer and law firm characteristics at the conversation level. About 7.2 percent of conversations involve a former SEC lawyer. This is higher than the 4.5 percent at the lawyer level, likely driven by former SEC lawyers being involved in more conversations in our sample than non-SEC lawyers (see Table 2). On average, the most experienced lawyer (law firm) in the conversation was involved in less than one (more than eight) other conversation(s) in the year before the focal conversation, and a conversation involves about 1.2 lawyers. Most lawyers have at least two degrees, and about 26.0 and 49.5 percent of conversations involve a lawyer who attended a university with a top-three law school or other top-fourteen law schools respectively. About 17.3 percent of conversations involved a female lawyer, and the average log-transformed age of the oldest lawyer in the conversation is 3.298, corresponding to about 27 years between the start of the conversation and the lawyer’s first degree. Panel C of Table 3 shows descriptive statistics for proxies for comment letter complexity and issues. The average log-transformed length of the initial comment letters is 6.893, corresponding to 985 words, and the initial comment letters reference 1.8 SEC filings on average. The proportion of the initial comment letters that raise specific issues range between 10.8 percent (risk factor issues) and 77.1 percent (accounting issues). About 80.6 percent of the initial comment letters reference issues that do not fit standard issue categories as coded by Audit Analytics. About 58.2 percent of the initial comment letters mention the word “amend”. 19 Finally, Panel D of Table 3 shows descriptive statistics for the firm controls.25 The average log-transformed firm size and age are 6.675 and 2.513 respectively, corresponding to a market capitalization of $792 million, and over 11 years since the firm’s first appearance in CRSP. On average, the firms had raw buy-and-hold returns of 12.5 percent over the 12 months before the start of the conversation, the monthly returns had a standard deviation of 13.0 percent, and the book-to-market ratio was 0.580. The firms had return on assets close to zero on average, sales growth of about 20.2 percent, and R&D intensity of about 5.2 percent. Most firms (94.6 percent) are domestic, and 64.7 percent were incorporated in Delaware. About 78.3 percent were audited by a Big Four audit firm, and 11.2 percent were audited by a second-tier audit firm. Finally, 46.7, 5.2, and 16.9 percent of conversations had a prior comment letter conversation, a securities litigation filing, and a restatement within the two years leading up to the start of the focal conversation. 5. EMPIRICAL RESULTS 5.1. Determinants of retaining a former SEC lawyer We first examine the determinants of retaining a former SEC lawyer by estimating logistic regressions of SEC, the dummy variable indicating that the firm retained a former SEC lawyer, against the controls for comment letter complexity and issues, firm characteristics, and industry and year fixed effects. The results are documented at columns 1 to 3 of Table 4. Column 1 includes controls for comment letter complexity and issues, column 2 includes controls for firm characteristics, and column 3 include both sets of controls. We include industry and year fixed effects in all three specifications. Of the controls for comment letter complexity and issues, we find that the only two determinants are statistically significant. The length of the letter is positively associated with the likelihood of retaining a former SEC lawyer: the coefficient of 0.395 on LetterLength suggests that a ten 25Unless otherwise stated, the firm controls are based on the most recent fiscal year ended before the start of a conversation. 20 percent increase in the length of the initial letter increases the odds that the firm retains a former SEC lawyer by about 3.8 percent (= 1.1 0.395 − 1). However, the coefficient becomes insignificant in Column (3), suggesting that the impact of letter length is subsumed by firm characteristics. We also find that firms are significantly less likely to involve a former SEC lawyer when the initial comment letter mentions management discussion and analysis disclosure issues. Of the controls for firm characteristics, we find that larger and older firms are significantly more likely to retain a former SEC lawyer, as are firms that faced recent litigation or that do not have big four or tier two audit firms. We note that the pseudo-R-squared of our determinants model is only 4.7 percent, indicating that many other factors drive the decision to hire a former SEC lawyer. Some additional factors could include whether the law firm retained or usually used by the company has a former SEC lawyer, or simply the availability of former SEC lawyers when the firm receives a comment letter. We note that it is rare for a lawyer to have had SEC experience—only about seven percent of our sample of conversations involve former SEC lawyers—so we would not expect every firm to be able to hire a former SEC lawyer if they chose to do so. As these determinants may confound our main tests of the consequences of hiring a former SEC lawyer, we carry out a propensity matching of treatment (SEC = 1) and control (SEC = 0) observations along the control variables. 5.2. Impact of propensity score matching Due to the differences between former SEC and non-SEC lawyers, and the comment letter and firm determinants of retaining a former SEC lawyers, our analyses control for potential confounding. We include the control variables in our regressions to mitigate linear confounding, and also implement propensity score matching between treatment (SEC = 1) and control (SEC = 0) observations as described in our research design at Section 3.2. Table 5 shows the impact of matching treatment (SEC = 1) and control (SEC = 0) observations on the propensity scores and covariate balance. For brevity, we only tabulate descriptives for the first matching procedure in our analyses: 21 full matching for Model 1, along the controls for lawyer and law firms characteristics, comment letter complexity and issues, firm characteristics, and industry fixed effects. Table 5 reports the propensity score and covariate balance before matching (columns 2–5 on the left), and after matching (columns 6–9 on the right), in both cases after observations outside the support of the propensity score are discarded. We report the difference in means of each variable, and the absolute standardized difference between the variables for assessing covariate balance. We use a cutoff absolute standardized difference of 25 percent as an indicator for large differences in the covariate means (e.g. Rosenbaum and Rubin, 1985). After matching, the difference in weighted average propensity score reduces from 19 percentage points to -0.1 percentage points, with the absolute standardized difference decreasing from 120 percent to less than one percent after matching. This indicates that the matching procedure was successful in reducing the difference in the propensity of treatment. From columns 2 to 5, before matching, a number of covariates have absolute standardized differences above the cutoff even after discarding observations outside the support of the propensity score. From columns 6 to 9, after matching, all the individual covariates in the table have absolute standardized differences of less than the cutoff of 25 percent. 5.3. The extent of negotiation We test Hypothesis 1 at Table 6. Column (1) of Table 6 documents the results from estimating Model 1. The coefficient on SEC is 0.220 and statistically significant at 1%, indicating that the involvement of a former SEC lawyer is associated with a significant increase in the extent of negotiation, relative to matching control firms. To the extent that the extent of negotiation represents an undue delay in the correction of disclosure deficiencies, or efforts by the firm to avoid correcting disclosure deficiencies, this finding suggests that the involvement of former SEC lawyers is negatively related to shareholder welfare. Column (2) examines the relationship between confidential treatment requests and former SEC lawyer involvement. The coefficient estimate on SEC is 0.501 and is statistically significant at 22 1% level, which corresponds to about a 65 percentage points increase in the odds of a confidential treatment request, and suggests that former SEC lawyers are more likely to assist firms in securing redaction of sensitive firm information in the comment letter disclosures than lawyers who had never been employed by the SEC. Turning to the control variables, we find that the extent of negotiation is associated with comment letter that raise accounting or event disclosure issues, and interestingly with longer letter that reference fewer filings. 5.4. Outcomes of the comment letter process We test Hypothesis 2 at Table 7. The coefficients of SEC are significantly negative at conventional levels in Columns (1) and (2), suggesting that former SEC lawyer involvement is associated with both fewer amendments, and a decreased likelihood of filing any amendments. Specifically, former SEC lawyers are associated with a 31.9 percentage point decrease in the odds of an amendment and 0.380 fewer amendments on average, relative to non-SEC lawyers. This again suggests that the involvement of former SEC lawyers is negatively related to shareholder welfare, as it suggests that they are associated with a reduction in the correction of disclosure deficiencies. The coefficient on SEC in Column (3) is significantly positive at five percent level, indicating that firms that involve former SEC lawyers experience positive abnormal returns between the start and disclosure of the comment letter conversation, relative to matching firms. Finally, in Column (4), we do not find a statistically significant relationship between the involvement of former SEC lawyers and adverse restatements or shareholder litigation. This suggests that while former SEC lawyer involvement is associated with a reduction in amendments, there is not evidence that this is to the extent to avoiding disclosures that would have resulted in avoiding restatements or litigation. 6. CROSS-SECTIONAL ANALYSES The proposed mechanism for our findings is that former SEC lawyers are better able to influence the comment letter process than other lawyers due to greater familiarity with former colleagues 23 still at the SEC. We therefore perform cross-sectional analyses to validate this: our results should continue to hold when we restrict our treatment sample to comment letter conversations involving former SEC lawyers that had left the SEC more recently. For comparison, we examine two additional cross-sections more likely to capture the expertise and knowledge of the lawyer rather than their familiarity with SEC staff. Specifically, we examine whether the former SEC lawyer had experience with the Division of Corporation Finance (DCF), the SEC division that administers the comment letter process, and the extent to which the lawyer recently represented clients in response to SEC comment letters. We collect data on the year each lawyer left the SEC, and information on whether he or she had worked at the DCF, from publicly-available biographies. We proxy for experience representing clients using LawyerExp. For each of the cross-sectional analyses, we partition the treatment sample to observations above or below the median time since leaving the SEC (14 years), whether there was information that the lawyer had previously worked at the DCF, and whether the observation was above or below the annual median of LawyerExp. We then replicate our key regressions after the respective restrictions to the treatment sample, using the same propensity matching weights per observation to allow for comparability of the coefficients with our main results.26 Table 8 documents the results from replicating our main analyses with each cross-section of the treatment sample. Panels A, C, and E use treatment conversations involving former SEC lawyers who had left the SEC more recently, have prior experience at the DCF, and have greater experience in representing clients in responding to SEC comment letters, respectively. Panels B, D, and F use treatment conversations involving former SEC lawyers who had left the SEC further in the past, do not have prior experience at the DCF, and have less experience in representing clients in responding to SEC comment letters, respectively. 26For several lawyers, we had to infer the year of leaving the SEC from other information, for example based on information on how long the lawyers had spent working at law firms since leaving the SEC. For conversations involving more than one former SEC lawyer, we use the minimum available time since leaving the SEC, and whether there was information that any lawyer had worked at the DCF. The treatment observations for which we were unable to collect the year of leaving the SEC are retained in both cross-sections, and we omit LawyerExp from the regression controls for cross-sectional analyses that partition on LawyerExp. 24 The results in Panel A show that former SEC lawyers who had left the SEC recently are associated with greater negotiation and fewer amendments, as expected. The corresponding coefficients in Panel B, for former SEC lawyers who had left further in the past, are not statistically significant. Interestingly, we find that abnormal returns between the start and disclosure of the comment letter conversation is significantly positive only for former SEC lawyers who had left the SEC further in the past (Column 4). We also find that former SEC lawyers who had left the SEC more recently are associated with fewer adverse restatements or shareholder litigation. Comparing the corresponding coefficients between Panels A and B statistically, we find that difference in coefficients between Panels A and B are statistically significant only for Columns (1) and (5), that is, the results for the extent of negotiation, and adverse restatements or shareholder litigation. The cross-sectional analyses based on DCF experience (Panels C and D) and prior comment letter experience (Panels E and F) have mixed or null results, for example none of the coefficients on SEC is statistically significant in Column (2). We interpret these results as suggesting that the impact of former SEC lawyers is driven more by familiarity with colleagues at the SEC rather than experience with the comment letter process. 7. CONCLUSION Our study examines whether former SEC lawyers increase firms’ resistance to the SEC comment letter process, and affect the outcome of the process, relative to lawyers who had never been employed by the SEC. Using hand-collected data on the characteristics of individual lawyers involved in our sample of comment letters, we find that relative to matching firms that retain non-SEC lawyers, firms that retain former SEC lawyers negotiate to a greater extent with the SEC after receiving a comment letter, are less likely to issue amendments as a result of the comment letter, and issue fewer amendments on average as a results of the comment letter. They also experience higher abnormal returns between the start and disclosure of the comment letter conversation. Our results suggest that former SEC lawyers help companies secure more favourable outcomes for firms, but 25 that their involvement may be disadvantageous to shareholder welfare as it is associated with a lower likelihood of correcting disclosure deficiencies. In cross-sectional tests, we also find that our results on negotiation and amendments are driven more by former lawyers who left the SEC more recently, consistent with our main results being driven by former SEC lawyers having greater familiarity with their former colleagues at the SEC. Our study therefore sheds light on the impact of the SEC revolving door on the SEC comment letter process, and has important implications for firms, investors, and regulators. We find that the SEC revolving door impedes the effectiveness of one of the SEC’s most important mechanisms for disclosure oversight. Our study has implications for future research, which could examine whether former SEC employees play a role in influencing the SEC in other settings, and whether changes to existing SEC regulations and codes of ethics could be effective at mitigating the effects of the SEC revolving door. 26 REFERENCES Adams, E. S. and S. P. Engel. 2015. Gender diversity and disparity in the legal profession: An empirical analysis of the gender profile in national law firms and law schools. Buffalo Law Review 63: 1211–1261. Armstrong, C. S. , A. D. Jagolinzer, and D. F. Larcker. 2010. Chief executive officer equity incentives and accounting irregularities. Journal of Accounting Research 48 (5): 225–271. Atkins, P. S. 2005. Speech by SEC Commissioner: Charles Hamilton Houston Lecture. Available at: https://www.sec.gov/news/speech/spch040405psa.htm. Atkins, P. S. 2007. Speech by SEC Commissioner: American Society and the SEC’s Mission. Available at: https://www.sec.gov/news/speech/2007/spch101507psa.htm. Bens, D. A. , M. Cheng, and M. Neamtiu. 2016. The impact of SEC disclosure monitoring on the uncertainty of fair value estimates. The Accounting Review 91: 349–375. Bozanic, Z. , P. Choudhary, and K. J. Merkley. 2019. Securities law expertise and corporate disclosure. The Accounting Review 94: 141–172. Bozanic, Z. , J. R. Dietrich, and B. A. Johnson. 2017. SEC comment letters and firm disclosure. Journal of Accounting and Public Policy 36: 337–357. Brown, S. L. 2016. Mutual funds and the regulatory capture of the SEC. Journal of Business Law 19: 701. Bruck, A. and A. Canter. 2008. Supply, demand, and the changing economics of large law firms. Stanford Law Review 60: 2087–2130. Calderon, T. G. and L. Gao. 2022. Changes in corporate cybersecurity risk disclosures after SEC comment letters. Journal of Accounting and Public Policy 41. Cassell, C. A. , L. M. Dreher, and L. A. Myers. 2013. Reviewing the SEC’s review process: 10-K comment letters and the cost of remediation. The Accounting Review 88: 1875–1908. Choi, S. J. , M. Gulati, and A. C. Pritchard. 2019. Should I Stay or Should I Go? The Gender Gap for Securities and Exchange Commission Attorneys. Journal of Law and Economics 62: 427–456. Cornaggia, J. , K. J. Cornaggia, and H. Xia. 2016. Revolving doors on Wall Street. Journal of Financial Economics 120: 400–419. Correia, M. M. 2014. Political connections and SEC enforcement. Journal of Accounting and Economics 57: 241–262. Dechow, P. M. , A. Lawrence, and J. P. Ryans. 2016. SEC Comment Letters and Insider Sales. The Accounting Review 91: 401–439. 27 Dechow, P. M. and S. T. Tan. 2021. How Do Accounting Practices Spread? An Examination of Law Firm Networks and Stock Option Backdating. The Accounting Review 96: 431–464. deHaan, E. , S. Kedia, K. Koh, and S. Rajgopal. 2015. The revolving door and the SEC’s enforcement outcomes: Initial evidence from civil litigation. Journal of Accounting and Economics 60: 65–96. Druckman, D. and M. Olekalns. 2013. Motivational Primes, Trust, and Negotiators’ Reaction to Crisis. Journal of Conflict Resolution 57: 966–990. Ege, M. , J. L. Glenn, and J. R. Robinson. 2020. Unexpected SEC resource constraints and comment letter quality. Contemporary Accounting Research 37: 33–67. Grassley, C. E. 2013. SEC’s revolving door. Available at: https://www.grassley.senate.gov/news/ news-releases/sec%E2%80%99s-revolving-door. Ham, C. and K. Koharki. 2016. The association between corporate general counsel and firm credit risk. Journal of Accounting and Economics 61: 274–293. Hansen, B. B. 2004. Full Matching in an Observational Study of Coaching for the SAT. Journal of the American Statistical Association 99: 609–618. Hansen, B. B. and S. O. Klopfer. 2006. Optimal Full Matching and Related Designs via Network Flows. Journal of Computational and Graphical Statistics 15: 609–627. Hilzenrath, D. S. 2011. SEC staff’s ’revolving door’ prompts concerns about agency’s independence. Available at: https://www.washingtonpost.com/business/economy/sec-staffs-revolvingdoor-prompts-concerns-about-agencys-independence/2011/05/12/AF9F0f1G story.html. Ho, D. E. , K. Imai, G. King, and E. A. Stuart. 2011. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. Journal of Statistical Software 42: 1–28. Hopkins, J. J. , E. L. Maydew, and M. Venkatachalam. 2015. Corporate general counsel and financial reporting quality. Management Science 61: 129–145. Igan, D. and P. Mishra. 2014. Wall Street, Capitol Hill, and K Street: Political Influence and Financial Regulation. Journal of Law and Economics 57: 1063–1084. Johnston, R. and R. Petacchi. 2017. Regulatory oversight of financial reporting: Securities and Exchange Commission comment letters. Contemporary Accounting Research 34: 1128–1155. Karpoff, J. M. , D. S. Lee, and G. S. Martin. 2008. The cost to firms of cooking the books. Journal of Financial and Quantitative Analysis 43: 581–612. Karsten, C. , U. Malmendier, and Z. Sautner. 2019. Lawyer expertise and contract design – evidence from M&A negotiations. Working Paper. Kedia, S. and S. Rajgopal. 2011. Do the SEC’s enforcement preferences affect corporate misconduct? Journal of Accounting and Economics 51: 259–278. 28 Krishnan, C. N. V. and R. W. Masulis. 2013. Law Firm Expertise and Merger and Acquisition Outcomes. Journal of Law and Economics 56: 189–226. Kubic, M. 2020. Examining the examiners: SEC error detection rates and human capital allocation. The Accounting Review Forthcoming. Kubick, T. R. , D. P. Lynch, M. A. Mayberry, and T. C. Omer. 2016. The effects of regulatory scrutiny on tax avoidance: An examination of SEC comment letters. The Accounting Review 91: 1751–1780. Kwak, B. , B. T. Ro, and I. Suk. 2012. The composition of top management with general counsel and voluntary information disclosure. Journal of Accounting and Economics 54: 19–41. Law, M. T. and C. X. Long. 2012. What Do Revolving-Door Laws Do? Journal of Law and Economics 55: 421–436. Lucca, D. , A. Seru, and F. Trebbi. 2014. The revolving door and worker flows in banking regulation. Journal of Monetary Economics 65: 17–32. Luechinger, S. and C. Moser. 2014. The value of the revolving door: Political appointees and the stock market. Journal of Public Economics 119: 93–107. Lumineau, F. and J. E. Henderson. 2012. The influence of relational experience and contractual governance on the negotiation strategy in buyer-supplier disputes. Journal of Operations Management 30: 382–395. McGinty, T. 2010a. SEC Lawyer One Day, Opponent the Next. Available at: https://protect2.fireeye.com/v1/url?k=31323334-50bba2bf-3132d782-4544474f5631-3218b3eca19f0a31&q=1&e=0f8ec12d-e43b-4a7b-9ff0-a6793fd596e3&u=https%3A%2F%2Fwww.wsj%2F. com/articles/SB10001424052702303450704575160043010579272. McGinty, T. 2010b. SEC ‘Revolving Door’ Under Review. Available at: https://www.wsj.com/ articles/SB10001424052748703280004575309061471494980. Protess, B. and S. Craig. 2013. S.E.C.’s revolving door hurts its effectiveness, report says. Available at: https://dealbook.nytimes.com/2013/02/11/s-e-c-s-revolving-door-hurts-itseffectiveness-report-says. Rosenbaum, P. R. 1991. A characterization of optimal designs for observational studies. Journal of the Royal Statistical Society. Series B (Methodological) 53: 597–610. Rosenbaum, P. R. and D. B. Rubin. 1985. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. The American Statistician 39: 33–38. Shive, S. A. and M. M. Forster. 2017. The revolving door for financial regulators. Review of Finance 21: 1445–1484. Smallberg, M. 2013. Dangerous Liaisons: Revolving Door at SEC Creates Risk of Regulatory Capture. Available at: https://protect2.fireeye.com/v1/url?k=31323334-50bba2bf-3132d782-4544474f5631-0aef933f3de0fc0e&q=1&e=0f8ec12d-e43b-4a7b-9ff0-a6793fd596e3&u=http%3A%2F%2Fpogoarchives.org%2Febooks%2F20130211-dangerous-liaisons-sec-revolvingdoor.pdf. 29 Stuart, E. A. 2010. Matching methods for causal inference: A review and a look forward. Statistical Science 25: 1–21. Tabakovic, H. and T. Wollmann. 2018. From revolving doors to regulatory capture? evidence from patent examiners. NBER Working Paper. Available at: http://www.nber.org/papers/w24638. Wang, Q. 2016. Determinants of segment disclosure deficiencies and the effect of the SEC comment letter process. Journal of Accounting and Public Policy 35: 109–133. Zheng, W. 2015. The revolving door. Notre Dame Law Review 90: 1265–1308. 30 Table 1: Sample selection Unique Unique Unique conversations firm-years firms Panel A: Initial sample selection Conversations initiated between 2005 and 2016 that reference a 10-K filing, after requiring Compustat and CRSP coverage 16,734 16,531 5,584 Omit conversations with only one letter 16,470 16,376 5,550 Omit conversations in the bottom 1% of time to resolution 16,351 16,270 5,534 Panel B: Requiring legal advisor data and variables for analysis Restrict to conversations referencing legal counsel† 5,580 5,560 2,814 Require lawyer-law firm to have at least two conversations‡ 4,689 4,671 2,221 Require the availability of variables used in the analyses 4,524 4,507 2,140 † We also omit 82 conversations in which the lawyer was not named. The 5,580 conversations at this step involve 2,865 unique lawyers and 539 unique law firms. ‡ After requiring lawyer-law firms to be associated with at least two conversations, there are 1,407 unique lawyers and 326 unique law firms associated with the conversations in the sample. We were able to collect data for 1,384 of the unique lawyers from 320 unique law firms. 31 Table 2: Descriptives of the lawyer characteristics database Panel A: Characteristics per unique lawyer N Mean SD Q1 Q25 Q50 Q75 Q99 # Conversations 1, 384 4.081 3.446 2 2 3 5 18 # Law firms 1, 384 1.046 0.219 1 1 1 1 2 SEC 1, 384 0.045 0.207 0 0 0 0 1 HSY School 1, 384 0.211 0.408 0 0 0 0 1 OtherT opSch 1, 384 0.452 0.498 0 0 0 1 1 NumDegrees 1, 384 2.171 0.402 2 2 2 2 3 F emale 1, 384 0.165 0.371 0 0 0 0 1 Panel B: Former SEC lawyers (N = 62) versus other lawyers (N = 1,322) Overall SEC = 0 SEC = 1 Diff. p-valuet/χ2 mean mean mean in means # Conversations 4.08 4.01 5.53 1.52 (13.54%) # Law firms 1.05 1.04 1.08 0.04 (30.23%) HSY School 0.21 0.22 0.10 −0.12∗∗ (2.43%) OtherT opSch 0.45 0.46 0.32 −0.14∗∗ (4.87%) NumDegrees 2.17 2.17 2.27 0.11∗ (8.74%) F emale 0.16 0.16 0.16 −0.004 (100.00%) This table shows descriptive statistics of the characteristics of the 1,384 lawyers from 320 law firms in our sample. # Conversations is the number of comment letter conversations in our sample that reference the lawyer, and # Law firms is the number of unique law firms the lawyer is associated with in our sample, before requiring the availability of other variables; and the other variables are as defined in Appendix B. Panel A documents descriptive statistics for the full sample of lawyers. Panel B documents univariate differences in the variables between the lawyers who were formerly employed by the SEC (SEC = 1) and the lawyers who had not been employed by the SEC (SEC = 0). For the binary outcome variables (HSY School, OtherT opSch, and F emale) we examine the difference in proportions using chi-squared tests with p-values computed by Monte Carlo stimulation with 10,000 replicates. For the other outcome variables we examine the difference in means using t-tests. 32 Table 3: Descriptives of the comment letter conversations sample N Mean SD Q1 Q25 Q50 Q75 Q99 Panel A: Outcome variables Negotiation 4, 524 −0.000 1.549 −2.295 −1.405 −0.001 1.164 3.760 ConfT reat 4, 524 0.196 0.397 0 0 0 0 1 Amendment 4, 524 0.171 0.376 0 0 0 0 1 NumAmend 4, 524 0.243 0.609 0 0 0 0 3 CLReturns 4, 494 −0.003 0.335 −0.731 −0.139 −0.016 0.100 1.083 Adverse 4, 524 0.076 0.265 0 0 0 0 1 Panel B: Lawyer and law firm characteristics SEC 4, 524 0.072 0.259 0 0 0 0 1 LawyerExp 4, 524 0.725 1.131 0 0 0 1 5 LawF irmExp 4, 524 8.163 8.667 0 2 5 11 36 NumDegrees 4, 524 2.195 0.411 2 2 2 2 3 NumLawyers 4, 524 1.202 0.462 1 1 1 1 3 HSY School 4, 524 0.260 0.438 0 0 0 1 1 OtherT opSch 4, 524 0.495 0.500 0 0 0 1 1 F emale 4, 524 0.173 0.378 0 0 0 0 1 LawyerAge 4, 524 3.303 0.319 2.485 3.091 3.332 3.555 3.892 Panel C: Comment letter complexity and issues LetterLength 4, 524 6.893 0.478 5.999 6.553 6.847 7.197 8.135 NumF ilings 4, 524 1.821 0.943 1 1 2 2 5 AccIssue 4, 524 0.771 0.420 0 1 1 1 1 EventIssue 4, 524 0.185 0.389 0 0 0 0 1 MDAIssue 4, 524 0.664 0.472 0 0 1 1 1 NonF itIssue 4, 524 0.806 0.395 0 1 1 1 1 RegisIssue 4, 524 0.160 0.367 0 0 0 0 1 RegSKIssue 4, 524 0.562 0.496 0 0 1 1 1 RegSXIssue 4, 524 0.173 0.378 0 0 0 0 1 RiskIssue 4, 524 0.108 0.311 0 0 0 0 1 RefAmend 4, 524 0.582 0.493 0 0 1 1 1 Panel D: Firm characteristics BHR 4, 524 0.125 1.597 −0.840 −0.241 0.035 0.307 2.299 BigF our 4, 524 0.783 0.412 0 1 1 1 1 BTM 4, 524 0.580 0.644 −0.355 0.246 0.455 0.762 2.753 Delaware 4, 524 0.647 0.478 0 0 1 1 1 Domestic 4, 524 0.946 0.226 0 1 1 1 1 F irmAge 4, 524 2.513 0.843 0.740 1.924 2.617 3.087 4.369 Growth 4, 524 0.202 0.614 −0.644 −0.020 0.090 0.248 3.172 P astConv 4, 524 0.467 0.499 0 0 0 1 1 P astLit 4, 524 0.052 0.223 0 0 0 0 1 P astRes 4, 524 0.169 0.375 0 0 0 0 1 RD 4, 524 0.052 0.104 0 0 0 0.062 0.522 ROA 4, 524 −0.002 0.185 −0.752 −0.017 0.026 0.072 0..357 Size 4, 524 6.675 1.794 2.718 5.436 6.712 7.896 11.123 StdRet 4, 524 0.130 0.094 0.031 0.073 0.109 0.160 0.473 T ierT wo 4, 524 0.112 0.315 0 0 0 0 1 This table shows descriptive statistics for the variables used in the study, for the 4,524 comment letter conversations in our final sample. Please see B for variable definitions. 33 Table 4: Determinants of retaining a former SEC lawyer Model Dependent variable: SEC; Logistic Regressions Full Sample (1) (2) (3) LetterLength 0.395∗∗ 0.295 (0.192) (0.195) NumF ilings 0.025 0.061 (0.070) (0.071) AccIssue 0.061 −0.003 (0.166) (0.169) EventIssue 0.212 0.212 (0.150) (0.153) MDAIssue −0.234∗ −0.245∗ (0.142) (0.144) N onF itIssue −0.264 −0.230 (0.162) (0.164) RegisIssue −0.010 0.016 (0.174) (0.175) RegSKIssue −0.054 −0.023 (0.143) (0.144) RegSXIssue 0.059 0.091 (0.155) (0.157) RiskIssue −0.265 −0.295 (0.210) (0.213) RefAmend −0.015 0.050 (0.132) (0.134) BHR −0.186 −0.189 (0.131) (0.132) BigF our −0.640∗∗∗ −0.619∗∗∗ (0.203) (0.204) BTM 0.145 0.126 (0.096) (0.097) Delaware 0.004 0.004 (0.136) (0.136) Domestic −0.111 −0.088 (0.276) (0.277) F irmAge 0.349∗∗∗ 0.356∗∗∗ (0.081) (0.081) Growth −0.026 −0.033 (0.118) (0.117) P astConv −0.150 −0.137 (0.130) (0.130) P astLit 0.501∗∗ 0.477∗∗ (0.234) (0.235) P astRes 0.173 0.116 (0.154) (0.156) RD −1.479 −1.457 (0.957) (0.969) ROA −0.392 −0.368 (0.462) (0.465) Size 0.152∗∗∗ 0.151∗∗∗ (0.045) (0.045) StdRet −0.781 −0.861 (0.957) (0.964) T ierT wo −0.557∗∗ −0.557∗∗ (0.252) (0.253) Industry FEs Yes Yes Yes Year FEs Yes Yes Yes Observations 4,524 4,524 4,524 McFadden R2 0.019 0.041 0.047 The table documents the results from regressing SEC against the control variables for comment letter complexity and issues, firm characteristics, and industry and year fixed effects. Please see Appendix B for variable definitions. Year and industry fixed effects are based on the calendar year in which the conversation began, and one-digit SIC codes, respectively. 34 Table 5: Impact of propensity score matching on propensity scores and covariate balance After omitting observations outside the support of the propensity score (N = 3,429) Means before weighting Means after weighting SEC = 0 SEC = 1 Diff. |Std. Diff.| SEC = 0 SEC = 1 Diff. |Std. Diff.| Propensity score 0.07 0.26 0.19† 119.78% 0.26 0.26 −0.001 0..29% Covariate balance: lawyer and law firm characteristics LawyerExp 0.71 1.24 0.53† 42.03 1.24 1.24 0.01 0.36 LawF irmExp 7.87 6.84 −1.03 14.10 6.63 6.84 0.21 3.21 NumDegrees 2.21 2.24 0.03 6.82 2.22 2.24 0.02 4.46 NumLawyers 1.20 1.29 0.09 18.40 1.32 1.29 −0.03 4.59 HSY School 0.21 0.11 −0.10† 27.22 0.11 0.11 −0.0004 0.14 OtherT opSch 0.47 0.33 −0.13† 27.09 0.31 0.33 0.02 4.45 F emale 0.18 0.22 0.03 8.51 0.22 0.22 −0.002 0.42 LawyerAge 3.33 3.43 0.11† 37.76 3.45 3.43 −0.01 5.57 Covariate balance: comment letter complexity and issues LetterLength 6.90 6.95 0.04 9.25 6.93 6.95 0.01 2.71 NumF ilings 1.82 1.93 0.11 11.24 1.84 1.93 0.09 8.37 AccIssue 0.76 0.81 0.05 11.63 0.80 0.81 0.01 2.80 EventIssue 0.18 0.23 0.05 11.85 0.18 0.23 0.04 10.54 MDAIssue 0.69 0.65 −0.04 8.40 0.66 0.65 −0.003 0.72 NonF itIssue 0.81 0.80 −0.02 3.99 0.78 0.80 0.02 5.16 RegisIssue 0.17 0.18 0.01 2.22 0.18 0.18 −0.01 1.71 RegSKIssue 0.59 0.57 −0.02 3.48 0.55 0.57 0.02 3.99 RegSXIssue 0.17 0.21 0.03 8.95 0.17 0.21 0.03 7.94 RiskIssue 0.11 0.09 −0.02 6.19 0.09 0.09 −0.001 0.22 RefAmend 0.59 0.63 0.04 7.97 0.62 0.63 0.01 1.87 Covariate balance: firm characteristics BHR 0.12 0.06 −0.06 4.59 0.10 0.06 −0.03 4.48 BigF our 0.77 0.75 −0.02 5.45 0.74 0.75 0.01 3.06 BTM 0.60 0.64 0.03 4.67 0.69 0.64 −0.06 8.08 Delaware 0.63 0.60 −0.03 6.94 0.60 0.60 −0.01 1.51 Domestic 0.95 0.95 0.01 2.44 0.95 0.95 0.01 3.46 F irmAge 2.56 2.70 0.14 16.05 2.75 2.70 −0.05 5.85 Growth 0.18 0.16 −0.01 2.37 0.13 0.16 0.03 5.36 P astConv 0.50 0.47 −0.03 5.64 0.47 0.47 0.001 0.11 P astLit 0.05 0.06 0.01 2.42 0.06 0.06 −0.01 2.44 P astRes 0.16 0.17 0.01 2.81 0.17 0.17 0.01 1.93 RD 0.05 0.04 −0.01 11.16 0.04 0.04 0.001 1.21 ROA 0.001 0.01 0.01 4.95 0.002 0.01 0.01 4.30 Size 6.72 6.87 0.14 7.39 6.72 6.87 0.14 7.00 StdRet 0.13 0.12 −0.01 10.39 0.12 0.12 −0.004 4.71 T ierT wo 0.12 0.11 −0.005 1.56 0.13 0.11 −0.01 4.32 This table documents the impact of our full matching procedure on the propensity scores and covariate balance over all years. We report the treatment and control means and the difference in means, after omitting observations not in the support of the propensity score. We also report the absolute values of the standardized difference in means, defined as the difference between the treatment and control means scaled by the root mean squared standard deviations in the treatment and control subsamples. Columns 2 to 5 are based on the means and standard deviations before applying the weights constructed by the matching procedure; columns 6 to 9 are based on the weighted means and standard deviations. † indicates absolute standardized differences greater than 25 percent. 35 Table 6: Representation by a former SEC lawyer and the extent of negotiation Dep. Var. (model) Negotiation (OLS) ConfT reat (Logistic) (1) (2) SEC 0.220∗∗∗ 0.501∗∗∗ (0.079) (0.164) LawyerExp 0.019 −0.177∗∗∗ (0.019) (0.045) LawF irmExp 0.002 −0.009 (0.004) (0.008) NumDegrees 0.093∗ −0.106 (0.054) (0.123) NumLawyers −0.046 0.177 (0.048) (0.109) HSY School 0.069 0.307∗ (0.075) (0.161) OtherT opSch 0.159∗∗∗ 0.175 (0.054) (0.120) F emale 0.215∗∗∗ 0.242∗ (0.059) (0.125) LawyerAge −0.160∗ −0.808∗∗∗ (0.091) (0.196) LetterLength 1.610∗∗∗ 1.163∗∗∗ (0.076) (0.169) NumF ilings −0.201∗∗∗ −0.133∗∗ (0.029) (0.064) AccIssue 0.217∗∗∗ −0.420∗∗∗ (0.064) (0.141) EventIssue 0.127∗∗ −0.450∗∗∗ (0.062) (0.143) MDAIssue −0.091 −0.460∗∗∗ (0.057) (0.136) NonF itIssue −0.012 0.571∗∗∗ (0.063) (0.162) RegisIssue 0.084 0.112 (0.067) (0.140) RegSKIssue 0.019 0.124 (0.058) (0.135) RegSXIssue 0.022 0.069 (0.062) (0.137) RiskIssue −0.097 0.056 (0.083) (0.165) RefAmend −0.096∗ 0.311∗∗ (0.053) (0.123) Other Controls & FEs Yes Yes Observations 3,429 3,429 Adj. / McF. R2 0.322 0.191 This table documents the results from estimating Model 1. Negotiation is the first principal component of ConvT ime, NumLetter, and MultiRound, and we use full matching of the treatment and control subsamples. Please see Section 3 for details on our research design and matching, and Appendix B for other variable definitions. Year fixed effects are based on the calendar year in which the conversation began, and industry fixed effects are based on one-digit SIC codes. We report adjusted R2 s for the OLS model, and McFadden R2 s for the logistic models. 36 Table 7: Representation by a former SEC lawyer and outcomes of the comment letter Dep. Var. Amendment NumAmend CLReturns Adverse Regression Logistic Poisson OLS Logistic (1) (2) (3) (4) SEC −0.384∗ −0.380∗∗ 0.039∗∗ 0.067 (0.215) (0.163) (0.018) (0.295) LawyerExp −0.064 −0.020 −0.004 0.041 (0.050) (0.034) (0.004) (0.071) LawF irmExp −0.013 −0.0001 0.002∗∗ −0.012 (0.010) (0.007) (0.001) (0.016) NumDegrees 0.334∗∗∗ 0.121 0.027∗∗ −0.034 (0.124) (0.086) (0.013) (0.217) NumLawyers 0.046 0.051 −0.034∗∗∗ 0.643∗∗∗ (0.122) (0.081) (0.011) (0.176) HSY School 0.019 −0.061 0.003 0.419 (0.212) (0.154) (0.018) (0.290) OtherT opSch 0.205 0.126 0.030∗∗ −0.102 (0.142) (0.100) (0.013) (0.245) F emale 0.297∗∗ 0.104 −0.032∗∗ 0.055 (0.150) (0.106) (0.013) (0.233) LawyerAge −0.535∗∗ −0.216 −0.038∗ 0.126 (0.240) (0.157) (0.021) (0.376) LetterLength −0.357∗ −0.012 −0.079∗∗∗ 0.570∗ (0.207) (0.141) (0.019) (0.321) NumF ilings 0.206∗∗∗ 0.155∗∗∗ −0.002 0.182 (0.070) (0.045) (0.007) (0.121) AccIssue −0.084 −0.390∗∗∗ 0.062∗∗∗ 0.478 (0.169) (0.118) (0.015) (0.305) EventIssue −0.134 0.055 −0.017 0.172 (0.154) (0.098) (0.014) (0.243) MDAIssue 0.347∗∗ 0.169 0.005 −0.430 (0.168) (0.124) (0.013) (0.273) NonF itIssue 0.206 0.135 0.025∗ −0.348 (0.189) (0.136) (0.014) (0.282) RegisIssue 0.267∗ 0.021 0.009 0.552∗∗ (0.159) (0.107) (0.016) (0.243) RegSKIssue 0.704∗∗∗ 0.354∗∗∗ −0.001 −0.248 (0.163) (0.120) (0.013) (0.242) RegSXIssue −0.016 0.038 0.005 −0.219 (0.171) (0.108) (0.015) (0.266) RiskIssue −0.281 0.029 −0.014 0.440 (0.217) (0.136) (0.020) (0.295) RefAmend 1.389∗∗∗ 1.021∗∗∗ −0.044∗∗∗ −0.969∗∗∗ (0.160) (0.114) (0.012) (0.236) Other Controls & FEs Yes Yes Yes Yes Observations 3,309 3,309 3,284 2,783 Adj. / McF. R2 0.273 0.403 0.109 0.336 This table documents the results from estimating Model ?? using full matching of the treatment and control subsamples. Please see Section 3 for details on our research design and matching, and Appendix B for variable definitions. The control variables comprise the lawyer and law firm characteristics, comment letter complexity and issues, firm characteristics, and the number of comment letters in the conversation (NumLetters). Year fixed effects are based on the calendar year in which the conversation began, and industry fixed effects are based on one-digit SIC codes. We report adjusted R2 s for the OLS model, and McFadden R2 s for the poisson and logistic models. 37 Table 8: Cross-sectional analyses by SEC lawyer characteristics Dep. Var. Negotiation Amendment NumAmend CLReturns Adverse Regression OLS Logistic Poisson OLS Logistic (1) (2) (3) (4) (5) Panel A: Former SEC lawyers who left the SEC recently SEC 0.446∗∗∗ −0.609∗∗ −0.488∗∗ 0.028 −1.119∗∗ (0.104) (0.294) (0.222) (0.023) (0.530) Controls & FEs Yes Yes Yes Yes Yes Observations 3,308 3,188 3,188 3,163 2,668 Adj. / McF. R2 0.332 0.279 0.407 0.117 0.366 Panel B: Former SEC lawyers who left the SEC further in the past SEC 0.004 −0.202 −0.338 0.057∗∗ 1.007∗∗∗ (0.107) (0.282) (0.216) (0.025) (0.374) Controls & FEs Yes Yes Yes Yes Yes Observations 3,299 3,175 3,175 3,150 2,659 Adj. / McF. R2 0.279 0.410 0.238 0.112 0.355 Panel C: Former SEC lawyers with prior experience at the Division of Corporation Finance SEC 0.275∗∗∗ −0.350 −0.338∗ 0.024 −0.387 (0.090) (0.241) (0.188) (0.020) (0.385) Controls & FEs Yes Yes Yes Yes Yes Observations 3,366 3,246 3,246 3,221 2,722 Adj. / McF. R2 0.328 0.272 0.403 0.112 0.351 Panel D: Former SEC lawyers without prior experience at the Division of Corporation Finance SEC 0.189 −0.455 −0.580∗ 0.098∗∗ 1..215∗∗ (0.174) (0.484) (0.348) (0.040) (0.516) Controls & FEs Yes Yes Yes Yes Yes Observations 3,204 3,080 3,080 3,055 2,573 Adj. / McF. R2 0.355 0.288 0.417 0.122 0.371 Panel E: Former SEC lawyers with more experience with SEC comment letters SEC 0.215∗∗ −0.279 −0.425∗ 0.038 −0.644 (0.103) (0.285) (0.220) (0.023) (0.464) Controls & FEs Yes Yes Yes Yes Yes Observations 3,309 3,189 3,189 3,164 2,671 Adj. / McF. R2 0.333 0.281 0.411 0.118 0.361 Panel F: Former SEC lawyers with less experience with SEC comment letters SEC 0.226∗ −0.505 −0.311 0.038 0.791∗∗ (0.118) (0.319) (0.238) (0.027) (0.384) Controls & FEs Yes Yes Yes Yes Yes Observations 3,271 3,147 3,147 3,122 2,633 Adj. / McF. R2 0.334 0.278 0.408 0.113 0.355 This table documents the results from replicating the regressions at Panel A of Tables 6 and 7, under the restrictions to the treatment sample detailed at Section 6. Panels A, C, and E use treatment conversations involving former SEC lawyers who left the SEC more recently, previously worked at the Division of Corporation Finance, and have greater prior experience in representing clients in responding to SEC comment letters, respectively; Panels B, D, and F use the opposite cross-sections. The number of years since leaving the SEC was unavailable for several observations (28 of our final sample of 4,524); we retain these observations in both Panels A and B. The partitioning variable LawyerExp is omitted as a regression control in Panels E and F. Please see Section 6 for additional details. 38 A. THE SEC’S FILING REVIEW PROCESS AND OUTCOMES The Sarbanes-Oxley Act of 2002 requires the SEC to review firms’ filings “for the protection of investors”, at least once every three years.27 At the SEC, review of filings is carried out by the Division of Corporation Finance through eleven offices categorized by industry. The SEC sends comments to the firm when its staff “identifies instances where it believes a company can improve its disclosure or enhance its compliance with the applicable disclosure requirements”. Between 2005 and 2016, the SEC initiated over 100,000 comment letter conversations with firms (see Table 1). Initiation of a comment letter conversation leads to a dialogue between the firm and SEC staff. During the dialogue, the SEC may make requests of the firm, for example that the firm provide additional information, amend prior SEC filings, or revise future SEC filings. The firm, in turn, may propose alternative solutions. For example, in the following exchange between the SEC and the external counsel of Chase Issuance Trust, the SEC requested that the firm amend its Form 10- K or explain why an amendment is unnecessary. The firm declined amending through its external counsel and instead stated that they would clarify in future Form 10-Ks “if applicable”.28 SEC Staff comments to Chase Issuance Trust (September 13, 2012, excerpt): Please amend your Form 10-K to identify in the body of the 10-K each instance of noncompliance and the scope of the transactions to which each instance of noncompliance relates or tell us why such an amendment is not necessary. Chase Issuance Trust’s response, via external counsel (October 2, 2012, excerpt): In the future, beginning with our next Form 10-K, if applicable and to avoid confusion, 27See §408 of the Sarbanes-Oxley Act. Other information in this section on the SEC’s filing review process is obtained from publicly-available sources, specifically the SEC’s “Division of Corporation Finance Filing Review Process” webpage at https://www.sec.gov/divisions/corpfin/cffilingreview.htm, and comment letters on the SEC’s online EDGAR database. 28The letters are available on the SEC’s EDGAR database at https://www.sec.gov/Archives/edgar/data/869090/ 000000000012050064/filename1.pdf and https://www.sec.gov/Archives/edgar/data/1174821/000119312512412640/ filename1.htm. 39 we will add a sentence in the body of the Form 10-K to clarify whether or not a disclosed material instance of noncompliance is relevant to the Company’s transactions. Since the comment letter conversation often involves negotiations between the company and the SEC, persuasion and social influence may play important roles. A conversation between the SEC and Owens-Illinois, Inc. in 2006, for example, had carried on for multiple exchanges of letters, but was resolved after a conference call between the company and SEC staff. The SEC had requested that the firm either explain its basis of accounting related to a special purpose vehicle, or amend prior 10-K and 10-Q filings. This exchange followed multiple exchanges of letters between SEC staff and the firm in which the SEC questioned the firm’s accounting related to the special purpose vehicle. After a conference call with SEC staff, the SEC allowed the firm to resolve the issue by simply revising future filings.29 SEC Staff comments to Owens-Illinois, Inc. (August 21, 2006, excerpt): Please tell us your basis in GAAP for treating the collection of your accounts receivable by SPV1 as an operating cash inflow rather than an investing cash inflow. Otherwise, please amend your Forms 10-Q for the quarters ended March 31 and June 30, 2006 and, if material, your Form 10-K for the year ended December 31, 2005 to reflect subsequent collection of the receivables previously sold to SPV1 as an investing activity. Owens-Illinois, Inc.’s response, copied to former SEC counsel (October 10, 2006, excerpt): As agreed in the telephone conference with the SEC Staff on Friday, September 15, 2006, the Company will, in future filings, reclassify certain items in its statement of cash flows to reflect the cash inflows related to the subsequent collection of the accounts receivable held by SPV1 at the consolidation date of December 13, 2005 as cash flows from investing activities. Notably, the company copied an external counsel who was formerly the director of the SEC’s 29The letters are available on the SEC’s EDGAR database at https://www.sec.gov/Archives/edgar/data/812074/ 000000000006040461/filename1.pdf and https://www.sec.gov/Archives/edgar/data/812074/000110465906065678/ filename1.htm. 40 Division of Corporation Finance, the division of the SEC that reviews filings. When the firm has resolved all of the SEC’s comments, the SEC sends a final letter confirming completion of the review. A comment letter conversation with only one round of exchanges will therefore generally comprise at least three letters—the SEC’s initial comment letter, the firm’s response, and the SEC’s “no further comments” letter—and a conversation with two rounds of exchanges will generally comprise at least five letters, and so on. In the above example, the SEC replied with the following confirming review completion.30 SEC Staff’s response to Owens-Illinois, Inc. (November 21, 2006, truncated): We have completed our review of your Form 10-K and related filings and have no further comments at this time. The comment letter conversation is made public on EDGAR at least 20 business days after comments are resolved. Dechow et al. (2016) documents negative returns around and a negative drift after the date that comment letters related to revenue recognition are made public, and significant abnormal insider trading prior to that day. 30The letter is available on the SEC’s EDGAR database at https://www.sec.gov/Archives/edgar/data/812074/ 000000000006057629/filename1.pdf. 41 B. VARIABLE DEFINITIONS Tables B.1 and B.2 list definitions for variables used in this study. Table B.1 includes the definitions of the dependent variables, and variables related to lawyer and law firm characteristics. Because our unit of analysis is the comment letter conversation, in the case of conversations in which the firm referenced more than one lawyer or law firm, the lawyer and law firm characteristics are aggregated as explained in the table notes. Table B.2 includes definitions of variables capturing comment letter complexity and issues, and firm characteristics. Unless otherwise stated, the firm characteristics are based on the most recent fiscal year before the first comment letter in a conversation. In addition, the year fixed effects used in our study are based on the calendar year in which a conversation began, and the industry fixed effects are based on one-digit SIC codes. Please see Section 3.5 for a more detailed discussion of the control variables. 42 Table B.1: Outcome variables and lawyer characteristics Variable Definitions Outcome variables Negotiation The first principal component of ConvT ime, NumLetters, and MultiRound. ConvT ime The natural logarithm of one plus the length of the conversation in days. NumLetters The number of comment letters in the conversation. MultiRound One if NumLetters ≥ 5 and zero otherwise. ConfT reat One if the company requests confidential treatment, and zero otherwise Amendment One if NumAmend > 0 and zero otherwise. NumAmend The number of 10-K or 10-Q amendments filed between the start and the disclosure of the conversation. CLReturns Size- and delistings-adjusted buy-and-hold abnormal returns from the month after the start of the comment letter conversation to the month of disclosure Adverse One if Litigation or Restatement are one, and zero otherwise. Litigation One if a securities class action lawsuit was filed on or after the start of the conversation, and the class period of the first or reference lawsuit overlaps with the period referenced by the conversation, and zero otherwise.† Restatement One if an adverse restatement was filed between start of the conversation and the disclosure of the conversation, and the restatement period overlaps with the period referenced by the conversation, and zero otherwise.† Lawyer and law firm characteristics SEC One if the firm involved a lawyer formerly employed by the SEC in a comment letter in the conversation, and zero otherwise. LawyerExp The number of conversations in which the lawyer was involved over the year prior to the start of the focal conversation.‡ LawF irmExp The number of conversations in which the law firm was involved over the year prior to the start of the focal conversation.‡ NumDegrees The number of degrees earned by the lawyer, truncated on the left at two.‡ NumLawyers The number of lawyers involved in firms’ responses to the SEC. HSY School One if the firm referenced a lawyer who attended a university with a top-3 law school (Harvard, Stanford, or Yale Law School), and zero otherwise. OtherT opSch One if the firm referenced a lawyer who attended a university with a top-14 law school other than the top 3, and zero otherwise. F emale One if the firm involved a female lawyer in a comment letter, and zero otherwise. LawyerAge The natural logarithm of one plus the number of years between the start of a conversation and the year of the lawyer’s first degree.‡ This table lists the definitions for the dependent variables and variables related to lawyer or law firm characteristics. See Table B.2 for definitions of our comment letter and firm controls. † The period referenced by the conversation is defined as the period beginning with the start of the first fiscal period referenced, and ending with the filing date of the last 10-K or 10-Q referenced. ‡ We use the maximum values of LawyerExp, LawF irmExp, NumDegrees, and LawyerAge for conversations in which the firm referenced more than one lawyer or law firm. 43 Table B.2: Comment letter and firm control variables Variable Definitions Comment letter (CL) complexity LetterLength The natural logarithm of one plus the number of words in the initial CL. NumF ilings The number of filings referenced in the initial CL. Specific issues raised by the SEC AccIssue One if the initial CL mentions accounting rule issues, and zero otherwise. EventIssue One if the initial CL mentions event disclosure issues, and zero otherwise. MDAIssue One if the initial CL mentions MD&A disclosure issues, and zero otherwise. NonF itIssue One if the initial CL mentions issues not fitting standard categories, and zero otherwise. RegisIssue One if the initial CL mentions registration statement issues, and zero otherwise. RegSKIssue One if the initial CL mentions Regulation S-K issues, and zero otherwise. RegSXIssue One if the initial CL mentions Regulation S-X issues, and zero otherwise. RiskIssue One if the initial CL mentions risk factor disclosure issues, and zero otherwise. RefAmend One if the initial CL mentions “amend”, and zero otherwise. Firm characteristics BHR The firm’s raw buy-and-hold return over the 12 months before the start of the conversation. BigF our One if the firm is audited by a Big 4 accounting firm (Compustat: au is non-missing and strictly less than 9) and zero otherwise. BTM Common equity (ceq) divided by the market value of equity (prcc f × csho). Delaware One if the firm is incorporated in Delaware (incorp = “DE”), and zero otherwise. Domestic One if the firm is domestic (loc = “USA”), and zero otherwise. F irmAge Natural logarithm of one plus the public age of the firm, defined as the time between the start of the CL conversation and the firm’s first appearance on CRSP in years. Growth The change in annual total revenue (?revt) divided by the previous year’s total revenue. P astConv One if there was another CL conversation started in the two years prior to the focal conversation, and zero otherwise. P astLit One if a securities class action lawsuit was filed within the two years prior to the start of the conversation, and zero otherwise. P astRes One if the firm filed a restatement within the two years prior to the start of the conversation, and zero otherwise. RD R&D expense (rd) scaled by average assets (average at). ROA Income before extraordinary items (ib) scaled by average assets (average at). Size The natural logarithm of the firm’s market capitalization (prcc f × csho). StdRet The standard deviation of the firm’s raw monthly return over the 12 months before the first comment letter in a conversation. T ierT wo One if the firm is audited by a second-tier accounting firm (au is non-missing and equal to 11, 16, 17, or 21), and zero otherwise. This table lists the definitions for our comment letter and firm control variables. See Table B.1 for definitions of the dependent variables and variables related to lawyer or law firm characteristics. You guys have lots of work to do. It should be a crime the amount of time that you consume from the public with your shoddily written rules, that are better served by existing rules, and often, no rule at all. You are kill America through a death by a thousand cuts. Creating rules that long outlive those who wrote them, getting heavier and heavier. How many of your colleagues have ever tired to create something? Do you understand how hard, this destroys competition. Are you familiar with Stiglitz and the theory of regulatory capture? Regulatory capture is real, and you're doing it.