Federal officials are divided into three categories for purposes of the Appointments Clause. The President, with Senate advice and consent, must appoint principal officers, while the President, courts, or “Heads of Departments” may appoint “inferior Officers”;^1×1. U.S. Const. art. II, § 2, cl. 2. nonofficers are entirely exempt from either procedure.^2×2. See Buckley v. Valeo, 424 U.S. 1, 126 & n.162 (1976). Workable dividing lines between nonofficers and officers and between principal and inferior officers have eluded courts.^3×3. See Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 538 (2010) (Breyer, J., dissenting) (“Courts and scholars have struggled for more than a century to define the constitutional term ‘inferior officers,’ without much success.”). The Appointments Clause has lately been a flash point in litigation between agencies and regulated parties, putting these difficulties on display.^4×4. See id. at 544; PHH Corp. v. Consumer Fin. Prot. Bureau, 839 F.3d 1, 55 (D.C. Cir. 2016) (Randolph, J., concurring); Raymond J. Lucia Cos. v. SEC, 832 F.3d 277 (D.C. Cir. 2016); Gillian E. Metzger, Appointments, Innovation, and the Judicial-Political Divide, 64 Duke L.J. 1607, 1608–09 (2015) (“The federal appointments process is having its proverbial day in the sun.” Id. at 1608.). Recently, in Association of American Railroads v. Department of Transportation^5×5. 821 F.3d 19 (D.C. Cir.), reh’g and reh’g en banc denied, No. 12-5204 (D.C. Cir. Sept. 9, 2016). (Amtrak II), the D.C. Circuit held that an arbitrator who could be appointed to resolve a dispute between the National Railroad Passenger Corporation, also known as Amtrak, and the Federal Railroad Administration (FRA) was a principal officer ineligible for appointment by the Surface Transportation Board (STB). Amtrak II’s suggested dividing line between principal and inferior officers — the reviewability of the officer’s decision — avoided difficulties associated with another single-indicator test, but is ultimately overinclusive as an indicator of inferior-officer status.

Financial and operational difficulties have plagued Amtrak since its formation in 1970.^6×6. See Robert Puentes et al., Brookings, A New Alignment: Strengthening America’s Commitment to Passenger Rail 1–3 (2013). One factor in these performance troubles is Amtrak’s relationship with the freight railroads, which own most of the rail facilities and tracks Amtrak must use.^7×7. See U.S. Gov’t Accountability Office, GAO-07-15, Intercity Passenger Rail: National Policy and Strategies Needed to Maximize Public Benefits from Federal Expenditures 67, 73–76, app. V (2006). Under the legislation that created Amtrak, the railroads granted Amtrak access to infrastructure in exchange for taking the railroads’ common carrier obligations. See Rail Passenger Service Act of 1970, Pub. L. No. 91-518, 84 Stat. 1327 (codified as amended in scattered sections of 26 and 45 U.S.C.). Congress modified this relationship in the Passenger Rail Investment and Improvement Act of 2008^8×8. Pub. L. No. 110-432, div. B, 122 Stat. 4907 (codified in scattered sections of 49 U.S.C.). (PRIIA), which directs Amtrak and the FRA^9×9. The FRA, an agency within the Department of Transportation, primarily handles railroad safety, 49 U.S.C. § 103 (2012), and is further charged with facilitating “passenger-freight . . . integration,” id. § 103(j)(7). to set on-time performance and quality metrics and standards for Amtrak’s intercity trains.^10×10. PRIIA § 207, 49 U.S.C. § 24101 note (Metrics and Standards). Under section 207(d), if Amtrak and the FRA fail to agree, either entity can ask the STB^11×11. The STB, successor to the Interstate Commerce Commission, is an independent agency that regulates railroads. See 49 U.S.C.A. §§ 1301–1303 (West 2015). “to appoint an arbitrator to assist the parties in resolving their disputes through binding arbitration.”^12×12. PRIIA § 207(d), 49 U.S.C. § 24101 note.

The D.C. Circuit’s recent decision in Amtrak II is the latest installment in a drama that made its way to the Supreme Court in 2015.^13×13. Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225 (2015). In 2012, the Association of American Railroads (AAR), the freight railroad industry group,^14×14. AAR’s members had a stake in the metrics and standards. “To the extent practicable,” Amtrak and host railroads must adjust their agreements based on the metrics and standards. PRIIA § 207(d), 49 U.S.C. § 24101 note (Metrics and Standards); see also id. § 24308(f) (allowing the STB to fine railroads that contribute to Amtrak’s failure to meet metrics and standards). filed suit in the District Court for the District of Columbia seeking vacatur of metrics and standards set pursuant to PRIIA and asking the court to declare PRIIA unconstitutional on nondelegation and due process grounds.^15×15. See Complaint at 16–17, Ass’n of Am. R.Rs. v. Dep’t of Transp., 865 F. Supp. 2d 22 (D.D.C. 2012) (No. 11-1499). The district court granted the government’s motion for summary judgment.^16×16. Ass’n of Am. R.Rs., 865 F. Supp. 2d at 35. The D.C. Circuit reversed in Amtrak I,^17×17. Ass’n of Am. R.Rs. v. Dep’t of Transp. (Amtrak I), 721 F.3d 666 (D.C. Cir. 2013). holding that PRIIA unconstitutionally delegated legislative power to a private entity. The court reasoned that Amtrak was private, not governmental, because Congress designated it a “for-profit corporation.”^18×18. Id. at 675 (quoting 49 U.S.C. § 24301(a)(2)). The Supreme Court reversed, determining that Amtrak was a government entity for the purposes of the metrics and standards and remanding the case to the D.C. Circuit for consideration of three remaining issues: the due process challenge, the constitutionality of the appointments procedure for Amtrak’s president, and the section 207(d) nondelegation and Appointments Clause challenges.^19×19. Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225, 1233–34 (2015). Notably, Justice Alito, in concurrence, asserted that the arbitrator was a principal officer, remarking that the absence of supervision characterizes principal officers and that binding arbitration appeared unsupervised.^20×20. Id. at 1238–39 (Alito, J., concurring).

In Amtrak II, the D.C. Circuit invalidated PRIIA, giving two reasons: Amtrak’s involvement in setting metrics and standards violated the Due Process Clause, and PRIIA improperly vested power in an unconstitutionally appointed arbitrator.^21×21. 821 F.3d at 39. The court declined to reach AAR’s challenge to the appointment procedures for Amtrak’s president, citing possible preservation problems and noting that the outcome would not affect the case’s “ultimate disposition.”^22×22. Id. at 24. The due process claim was properly preserved and the arbitration challenge, although not properly preserved, was fit for adjudication. Id. at 24–26. AAR had argued that Amtrak’s president, who was appointed by the other Amtrak board members, was a principal officer eligible for appointment only by the President with Senate advice and consent.^23×23. Opening Brief for Appellant at 39–40, Amtrak II, 821 F.3d 19 (D.C. Cir. 2016) (No. 12-5204). After the Supreme Court’s decision in Amtrak I, Congress restructured Amtrak’s board. Compare 49 U.S.C.A. § 24302 (West 2015), with 49 U.S.C. § 24302 (2012). These changes were irrelevant in Amtrak II, as the challenged metrics and standards were from 2010, but the changes might foreclose future attacks on Amtrak’s structure.

Writing for the panel, Judge Brown^24×24. Judge Brown was joined by Senior Judges Williams and Sentelle. first concluded that permitting a self-interested entity to regulate competitors violates due process. Deriving from Carter v. Carter Coal Co.^25×25. 298 U.S. 238 (1936). a general principle that the opportunity “to co-opt the state’s coercive power to impose a disadvantageous regulatory regime on . . . market competitors would be problematic,”^26×26. Amtrak II, 821 F.3d at 31 (citing Alexander Volokh, The New Private-Regulation Skepticism: Due Process, Non-Delegation, and Antitrust Challenges, 37 Harv. J.L. & Pub. Pol’y 931 (2004)); see also Carter Coal, 298 U.S. at 311. the court held that PRIIA allows a self-interested entity (Amtrak) to regulate its resource competitors (the freight railroads).^27×27. Amtrak II, 821 F.3d at 31–34. Further, the court dismissed the argument that the scheme is best characterized as joint regulation by government and a self-interested group, as endorsed in Sunshine Anthracite Coal Co. v. Adkins^28×28. 310 U.S. 381 (1940). and similar cases, remarking that the FRA’s ability to counterbalance “overreaching by Amtrak is undermined by the power of the arbitrator . . . [who was] appointed unconstitutionally.”^29×29. Amtrak II, 821 F.3d at 34 n.4.

Judge Brown then introduced the appointments question by observing that the Framers were sensitive to questions of “who should be permitted to exercise the awesome and coercive power of the government.”^30×30. Id. at 36. Calling the Appointments Clause a “significant structural safeguard[],”^31×31. Id. (quoting Edmond v. United States, 520 U.S. 651, 659 (1997)). she explained that the procedure for principal officers promotes “accountability,” while the option for inferior officers is rooted in “convenience.”^32×32. Id. (quoting Edmond, 520 U.S. at 660). To start, the court rejected the government’s argument that considering the section 207(d) challenge was improper because arbitration had never occurred^33×33. Id. at 37 n.6 (first citing Amtrak I, 721 F.3d 666, 674 (D.C. Cir. 2013); then citing Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225, 1236 (2015) (Alito, J., concurring); and then citing Metro. Wash. Airports Auth. v. Citizens for Abatement of Aircraft Noise, Inc., 501 U.S. 252, 264–65 (1991)). and reaffirmed its conclusion from Amtrak I that section 207(d) was an unconstitutional delegation of lawmaking power to a private party because it permitted the appointment of a private arbitrator.^34×34. Id. at 37; see Amtrak I, 721 F.3d at 673–74.

Next, the court outlined why the appointment of a government arbitrator would be unconstitutional under the Appointments Clause. The analysis had two parts. First, the arbitrator would be an officer of the United States, not a nonofficer exempt from the Appointments Clause, because the authority to prescribe metrics and standards is “significant authority pursuant to the laws of the United States.”^35×35. Amtrak II, 821 F.3d at 37 (quoting Buckley v. Valeo, 424 U.S. 1, 126 (1976)). Second, the court turned to whether the arbitrator would be an inferior officer eligible for appointment by the STB.^36×36. The STB is a department for the purposes of the Appointments Clause. Id. at 38 (citing 49 U.S.C.A. § 1301(a)–(b) (West 2015); Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 511 (2010)). What distinguishes inferior from principal officers under Edmond v. United States,^37×37. 520 U.S. 651 (1997). Judge Brown explained, is that inferior officers are “directed and supervised at some level by” principal officers.^38×38. Amtrak II, 821 F.3d at 38 (quoting Edmond v. United States, 520 U.S. 651, 663 (1997)). Applying Edmond, the D.C. Circuit made two points about why the arbitrator was not an inferior officer. Generally, “[n]owhere does PRIIA suggest the arbitrator ‘is directed and supervised’” by a principal officer; specifically, the statute “doesn’t provide any procedure by which the arbitrator’s decision is reviewable by the STB.”^39×39. Id. at 39. After quoting an extended passage about train schedules from John Steinbeck’s East of Eden, the court summarized its holding that PRIIA was unconstitutional under the Due Process and Appointments Clauses, writing “there are limits to how far Congress may go to ensure Amtrak’s on-time performance.”^40×40. Id. The government moved for rehearing and rehearing en banc, which the court denied. See Amtrak II, No. 12-5204 (D.C. Cir. Sept. 9, 2016).

Courts have called the issues surrounding the section 207(d) arbitrator “banal”^41×41. Amtrak II, 821 F.3d at 39. and “mundane.”^42×42. Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225, 1234 (2015); see also id. at 1234–35. And commentaries on Amtrak have focused on nondelegation and due process, not appointments.^43×43. See, e.g., Alexander “Sasha” Volokh, The Shadow Debate over Private Nondelegation in DOT v. Association of American Railroads, 2014–2015 Cato Sup. Ct. Rev. 359; The Supreme Court, 2015 Term — Leading Cases, 129 Harv. L. Rev. 181, 341–50 (2015). But see Peter L. Strauss, Recent Developments in Administrative Law: The Tremors of Two March 9, 2015 Supreme Court Decisions, Part II: Association of American Railroads, Admin. & Reg. L. News, Summer 2015, at 12; Daniel Hemel, More on Amtrak and “Company A,” Yale J. on Reg.: Notice & Comment (May 4, 2016), http://yalejreg.com/nc/more-on-amtrak-and-company-a-by-daniel-hemel [https://perma.cc/PDV8-RH3A]. But Amtrak II’s Appointments Clause analysis deserves attention. In eschewing discussion of removability, a factor that cases since Edmond have emphasized as a key indicator of subordination, Amtrak II narrowed Edmond, and in an unexpected way. In focusing instead on the reviewability of the arbitration, the court avoided problems attendant to using removal to distinguish officers. Yet Amtrak II’s suggested dividing line is also imperfect: a reviewability test sweeps too many officials into the category of inferior officers.

Edmond held that Coast Guard Court of Criminal Appeals judges were inferior officers because they were subordinate to — that is, “directed and supervised” by — principal officers.^44×44. Edmond v. United States, 520 U.S. 651, 662 (1997); see also id. at 662–66. Specifically, the judges were subject to (1) oversight through court procedures, (2) removal at will, and (3) review by another executive branch court.^45×45. Id. at 664–66. Applying Edmond’s direction and supervision test to the arbitrator, the D.C. Circuit noted only the absence in the statute of any review of the arbitration, leaving off the other two factors.^46×46. Amtrak II, 821 F.3d at 39. The court’s essential consideration was therefore that the arbitrator’s decision appeared to be final and unreviewable.

Amtrak II’s review-centric reading of Edmond was anomalous. Relying on cases since Edmond, the government asserted in Amtrak II that the STB’s power to remove the arbitrator at will made the arbitrator an inferior officer, as the STB could “‘direct,’ ‘supervise,’ and exert some ‘control’” through the threat of dismissal.^47×47. Brief for Appellees at 47, Amtrak II, 821 F.3d 19 (D.C. Cir. 2016) (No. 12-5204) (quoting Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., 684 F.3d 1332, 1341 (D.C. Cir. 2012)). The recent emphasis on removal can be traced to Free Enterprise Fund v. Public Co. Accounting Oversight Board, 561 U.S. 477 (2010). In relevant part, Free Enterprise Fund held that certain independent agency board members were inferior officers based on removability at will and “other oversight.” Id. at 510. Section 207(d) is silent on ending the arbitrator’s tenure, but unrestricted removal power is typically incident to appointment power.^48×48. Free Enter. Fund, 561 U.S. at 509 (“Under the traditional default rule, removal is incident to the power of appointment.”); see also Keim v. United States, 177 U.S. 290, 293 (1900). Notably, in Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board,^49×49. 684 F.3d 1332. the D.C. Circuit concluded that the finality of Copyright Royalty Judges’ (CRJs’) decisions as well as restrictions on the Librarian of Congress’s ability to remove CRJs made the Judges principal officers ineligible for appointment by the Librarian.^50×50. Id. at 1339. As a remedy, the court severed and invalidated the removal restriction, suggesting that even an officer with the ability to make final decisions can be inferior if removable at will by a principal officer.^51×51. Id. at 1340–41 (“Without this restriction, we are confident that . . . the CRJs will be inferior rather than principal officers.”).

The Amtrak II court may have hesitated to explore removal for several reasons. Given the considerable doctrinal uncertainty that plagues boundary agencies like public-private Amtrak,^52×52. See Anne Joseph O’Connell, Bureaucracy at the Boundary, 162 U. Pa. L. Rev. 841 (2014); Strauss, supra note 43, at 13 (“The Court has had virtually nothing to say about structural limitations on intermediate institutions like AMTRAK, that Congress has created in such profusion.”). the panel may have been uncomfortable inferring the STB’s power to remove. In this vein, at oral argument, in response to the panel’s apparent discomfort with the arbitration scheme, AAR asserted that section 207(d) is “without precedent, as far as we can tell, in American law” and that constitutional problems arise from the provision’s silences.^53×53. Oral Argument at 15:05–15:08, Amtrak II, 821 F.3d 19 (No. 12-5204); id. at 16:51–17:06. Relatedly, in Amtrak I, the panel rejected the government’s argument that PRIIA’s silence on the type of arbitrator should be read to allow or even to require a public arbitrator. 721 F.3d 666, 673–74 (D.C. Cir. 2013). In addition, removal was inapt as a measure of subordination here because if the arbitrator were appointed dispute-by-dispute, the threat of dismissal may not translate into an instrument of control: by the time the STB realized that something was awry, the arbitration might be over.^54×54. But see Oral Argument, supra note 53, at 43:32–43:42 (the government suggested that the arbitrator might be appointed from “individuals within the STB”).

Far from being isolated to this case, the twin challenges of reading statutory silence and assessing abstract mechanisms in practice are endemic to considering removability. Other statutes are silent on removal.^55×55. See Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 546–48 (2010) (Breyer, J., dissenting). And scholars have documented that difficulties in translating the power to fire into control are fundamental, especially below the top echelons.^56×56. See, e.g., Aziz Z. Huq, Removal as a Political Question, 65 Stan. L. Rev. 1, 33–38 (2013); Gillian E. Metzger, The Constitutional Duty to Supervise, 124 Yale L.J. 1836, 1882 (2015) (“At middle and lower levels . . . indirect measures . . . may be more effective mechanisms [than removal] for controlling administrative behavior.”). At any level, removal provisions may produce counterintuitive consequences.^57×57. See, e.g., Daryl J. Levinson, The Supreme Court, 2015 Term — Foreword: Looking for Power in Public Law, 130 Harv. L. Rev. 31, 67–68 (2016). Yet courts are ill-equipped to ascertain from a statute’s text how power actually operates.^58×58. See Free Enter. Fund, 561 U.S. at 523 (Breyer, J., dissenting) (“Compared to Congress and the President, the Judiciary possesses an inferior understanding of the realities of administration, and the manner in which power . . . operates in context.”).

Although Amtrak II avoided these complications, its overemphasis on reviewability was doctrinally and practically problematic. Doctrinally, Amtrak II conflicts with the D.C. Circuit’s test for distinguishing nonofficers from officers. Officers exercise “significant authority pursuant to the laws of the United States” based on (1) the “significance of the matters” the official handles, (2) the official’s degree of discretion, and (3) the “finality of” decisions.^59×59. Raymond J. Lucia Cos. v. SEC, 832 F.3d 277, 284 (D.C. Cir. 2016) (emphasis added) (first quoting Buckley v. Valeo, 424 U.S. 1, 126 (1976); then quoting Tucker v. Comm’r, 676 F.3d 1129, 1133 (D.C. Cir. 2012)). Inability to make final decisions was the dispositive factor in two D.C. Circuit cases holding particular Administrative Law Judges to be nonofficers.^60×60. See id. at 287; Landry v. FDIC, 204 F.3d 1125, 1134 (D.C. Cir. 2000) (“As the ALJs . . . have no such powers [of final decision], we conclude that they are not inferior officers.”). Given this, using finality to separate inferior from principal officers seems less than coherent: if officials who cannot make final decisions are nonofficers and officials who can make final decisions are principal officers, then there appears to be little or no room for inferior officers. In Amtrak I, Justice Alito asserted that final decisions that “appear in the Federal Register” are reserved for principal officers.^61×61. Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225, 1239 (2015) (Alito, J., concurring) (“Inferior officers can do many things, but nothing final should appear in the Federal Register unless a Presidential appointee has at least signed off on it.”). Perhaps then inferior officers are permitted to make the minor or nonbinding decisions that bypass that journal.^62×62. See 44 U.S.C. § 1505 (2012) (describing documents that “shall be published in the Federal Register”). But this seems to relegate inferior officers to the status of employees. By definition, officers possess “significant authority,” and the D.C. Circuit has noted that principal and inferior officers may exercise “the power to bind third parties, or the government itself.”^63×63. Raymond J. Lucia Cos., 832 F.3d at 286 (citing Officers of the U.S. Within the Meaning of the Appointments Clause, 31 Op. O.L.C. 73, 87 (2007)) (defining officer as used in the Constitution based on evidence of the term’s meaning at ratification). Authority to make important, binding decisions is a less-than-helpful way to separate inferior and principal officers, as these features distinguish officers from nonofficers.

Amtrak II’s emphasis on reviewability also runs into practical problems. Judge Brown acknowledged that the arbitrator position felt too trivial to be appointed through presidential nomination with Senate advice and consent, writing, “[W]hile it may seem peculiar to demand ‘primary class’ treatment for a position as banal as the PRIIA arbitrator, it also seems inescapable.”^64×64. Amtrak II, 821 F.3d at 39. Agencies have always delegated authority internally, and the size and complexity of administrative government necessitate considerable delegation today.^65×65. See, e.g., Elizabeth Magill, Foreword, Agency Self-Regulation, 77 Geo. Wash. L. Rev. 859, 884–85 (2009); Metzger, supra note 56, at 1847–49 (“[T]he reality of power in modern government bureaucracies is much messier and more complex than the Weberian ideal, with lower-level staff and street-level employees often exercising substantial discretion . . . .” Id. at 1848.). Mid- and lower-level agency officials routinely make final decisions.^66×66. See, e.g., Magill, supra note 65, at 885 (“[F]rontline and midlevel decisionmakers make hundreds, if not thousands, of decisions each month . . . .”). Moreover, statutes often leave undetermined facets of agency structure, allowing agencies flexibility to design systems of oversight like review procedures.^67×67. See id. at 885–86; Jennifer Nou, Intra-Agency Coordination, 129 Harv. L. Rev. 421, 475–78 (2015). Because thousands of officials who might qualify as officers but who are not appointed through the principal-officer process might make decisions free from statutorily mandated review, Amtrak II’s test exposes agency decisionmakers — and their related statutes and agencies — to Appointments Clause challenges.

Finality is overinclusive as the touchstone of inferior-officer status, yet Amtrak II rejected potential limiting principles. One obvious way to help differentiate between principal and inferior officers is scope of duties.^68×68. See Morrison v. Olson, 487 U.S. 654, 671–72 (1988) (concluding that, among other factors, “limited duties,” id. at 671, made independent counsel an inferior officer). But Judge Brown expressly rejected using the narrowness of the arbitrator’s role as a sign of inferior-officer status.^69×69. Amtrak II, 821 F.3d at 38–39. Additionally, looking to multiple factors, as Edmond did, might have mitigated the inadequacies attendant to relying on removability or reviewability alone. Seen in this context, PRIIA’s arbitrator helps to illuminate the dangers of single-indicator tests for inferior-officer status.