Introduction

The seventy-fifth anniversary of the Administrative Procedure Act (“APA”) would be a fitting occasion to celebrate the present state of the agency adjudication process. Unfortunately, the process of agency adjudication is in worse condition today than it was when Congress enacted the APA in 1946. This contribution to a Symposium sponsored by the C. Boyden Gray Center for the Study of the Administrative State proceeds as follows: Part I recounts the history of the agency adjudication process beginning in the 1930s and extending to its unfortunate current state. Next, Part II describes the present and likely future state of the adjudication process. Finally, Part III explains potential changes to the agency adjudication process to allow it to function well in the future.

I.     The History of Agency Adjudication

Throughout the 1930s, participants in agency adjudications complained that the hearing examiners (later renamed administrative law judges or “ALJs”) who presided in agency adjudicative hearings were biased in favor of agencies and against private parties.¹See, e.g., Ramspeck v. Fed. Trial Exam’rs Conf., 345 U.S. 128, 131 (1953) (“Many complaints were voiced against the actions of hearing examiners, it being charged that they were mere tools of the agency concerned . . . .”). Participants also complained that they could not predict how agencies and ALJs would conduct a hearing. Agencies used a wide variety of procedures, and most agencies had few, if any, rules that governed the hearing process. Several studies supported those complaints.²See id. at 131–32.

After fifteen years of study and debate, Congress unanimously enacted the APA.³See George B. Shepherd, Fierce Compromise: The Administrative Procedure Act Emerges from New Deal Politics, 90 Nw. U. L. Rev. 1557, 1560 (1996) (“Congressional support for the bill was unanimous . . . .”). The APA responded to the complaints of bias and lack of predictability by describing in detail the process that an agency must use when it engages in adjudication. The procedures for adjudication are set forth in sections 554, 556, and 557 of the APA.⁴5 U.S.C. §§ 554, 556–57 (2018). Those sections require an agency to make available to participants in agency adjudications virtually all of the safeguards enjoyed by litigants in federal civil proceedings.

The APA also includes many provisions that are designed to assure that the ALJs, who have roles analogous to those of federal district judges, have almost as much decisional independence as federal judges.⁵See Butz v. Economou, 438 U.S. 478, 513 (1978); Ramspeck, 345 U.S. at 132, 134 n.4. Agencies have no control over an ALJ’s compensation.⁶See 5 U.S.C. § 5372. Agencies cannot evaluate an ALJ’s performance.⁷See id. § 7521; 5 C.F.R. § 930.206 (2011). Ex parte communications with ALJs, including communications from agency officials, are prohibited.⁸See 5 U.S.C. § 554(d). An agency must allocate cases among ALJs on a rotational basis except in unusual circumstances.⁹See id. § 3105. No one who has any role in investigation or enforcement can supervise an ALJ.¹⁰See id. § 554(d)(2). Agencies cannot assign ALJs duties that are inconsistent with their roles as independent adjudicators.¹¹See id. § 3105. Agencies have the power to appoint ALJs but only from a list of applicants who have been determined to be qualified by the Office of Personnel Management (“OPM”).¹²See id. § 5372(b)(2). That agency uses a variety of methods to determine whether an applicant has the experience, training, and judicial temperament required to be an ALJ. The most important safeguard against ALJ bias in favor of an agency is a provision that precludes an agency from removing or otherwise punishing an ALJ.¹³See id. § 7521. An agency can attempt to remove or otherwise discipline an ALJ only by bringing a case against the ALJ before the Merit Systems Protection Board (“MSPB”).¹⁴See 5 U.S.C. § 7521. That agency can remove or otherwise punish an ALJ only after it conducts an oral evidentiary hearing in which it finds that there is good cause to discipline the ALJ.¹⁵See id. § 7521(a).

Congress recognized that it was important for agencies to retain control over their policies.¹⁶See Paul R. Verkuil, Daniel J. Gifford, Charles H. Koch, Richard J. Pierce & Jeffrey S. Lubbers, The Federal Administrative Judiciary, in 2 Administrative Conference of the United States: Recommendations and Reports 771, 801–02 (1992). The APA empowers an agency to replace the initial decision of an ALJ with the agency’s own decision on appeal.¹⁷See 5 U.S.C. § 557(b). Accordingly, most agencies have procedures through which anyone who is dissatisfied with an ALJ’s initial decision can appeal that decision to the head of the agency. At that point, the agency has complete discretion to replace the ALJ’s initial decision with its own decision. Agencies use the appeals process to ensure that ALJs cannot make final decisions that are inconsistent with the agency’s policies. If a party is dissatisfied with the agency’s decision, it can obtain review of that decision in a federal court. In that review proceeding, the ALJ’s initial decision is merely part of the record that the court considers in deciding whether to uphold the agency decision.¹⁸See Universal Camera Corp. v. NLRB, 340 U.S. 474, 493–97 (1951).

During the 1950s, the Supreme Court unanimously and repeatedly praised Congress for the manner in which it responded to the complaints that agency hearing procedures were unpredictable and unfair, and that ALJs were biased in favor of agencies.¹⁹See, e.g., Ramspeck v. Fed. Trial Exam’rs Conf., 345 U.S. 128, 131–33 (1953); Wong Yang Sung v. McGrath, 339 U.S. 33, 41–45 (1950). In 1950, the Court commended Congress for creating an adjudication system that was carefully designed to assure that participants in agency adjudications were treated fairly.²⁰See Wong Yang Sung, 339 U.S. at 38–45. The Court expressed its belief that Congress had codified the minimum requirements of due process in the APA provisions that govern agency adjudication, with particular emphasis on the APA’s safeguards of the decisional independence of ALJs.²¹See id. at 46, 52–53. The Court then began to apply the APA to all agency adjudications whether or not Congress had explicitly made the APA applicable to the proceedings.²²See, e.g., Riss & Co. v. U.S., 341 U.S. 907, 907 (1951); see also Cates v. Haderlein, 342 U.S. 804, 804 (1952).

In an amendment to the Immigration and Naturalization Act (“INA”), Congress pointedly rejected the Court’s interpretation of the APA as codifying due process and made it clear that Immigration Judges (“IJs”) are not subject to the safeguards of the APA. In its 1955 opinion in Marcello v. Bonds,²³349 U.S. 302 (1955). the Court retreated from its view that Congress had codified the minimum requirements of due process when it enacted the APA.²⁴See id. at 308–11. The Court applied the amendment to the INA, rather than the APA, even though Congress had refused to confer on IJs any of the safeguards of decisional independence that apply to ALJs.²⁵See id. at 306–09. The Court held that the provisions of the APA that require agencies to use procedures that replicate those used by federal courts and to use ALJs who have statutory assurances of decisional independence apply only to agency adjudications that Congress has explicitly made subject to those provisions of the APA.²⁶See id. at 308–09.

In Marcello’s wake, adjudications in which the agency is required to provide the procedures described in sections 554, 556, and 557 of the APA are customarily referred to as formal adjudications. Adjudications that are not subject to those provisions of the APA are referred to as informal adjudications. Agencies that conduct informal adjudications have discretion to choose their own procedures. Most agencies that engage in informal adjudication have adopted rules of procedure that confer on participants procedural safeguards that satisfy the requirements of due process, but they do not use ALJs to preside in those proceedings.²⁷See Michael Asimow, Adjudication Outside the Administrative Procedure Act 13–15 (2016); Paul Verkuil, A Study of Informal Adjudication Procedures, 43 U. Chi. L. Rev. 739, 759–60 (1976) (“One goal of this empirical phase was to determine how much impact emerging notions of procedural due process were having upon the process of informal adjudication.”).

The agency officials who preside in informal adjudications have a variety of titles such as Administrative Patent Judge or Board of Contract Appeals Judge. They are often referred to collectively as Administrative Judges (“AJs”) to distinguish them from ALJs. Like the IJs that were the subject of the Court’s opinion in Marcello, AJs have few, if any, of the statutory safeguards of decisional independence that apply to ALJs.²⁸See Kent Barnett, Malia Reddick, Logan Cornett & Russell Wheeler, Non-ALJ Adjudicators in Federal Agencies: Status, Oversight, and Removal 1 (2018) (“[J]udicial doctrine over the past few decades has allowed agencies more discretion to use non-ALJs in place of ALJs. In contrast to ALJs, these non-ALJs almost never have statutory protections with respect to their independence.”). Thus, they are far more vulnerable than ALJs to the pro-agency bias that was common in all agency adjudications prior to the enactment of the APA.

Over the decades since the Court’s decision in Marcello, Congress has enacted scores of statutes that authorize agencies to conduct adjudications. Most of those statutes allow the agencies to conduct informal adjudications rather than formal adjudications and to use AJs rather than ALJs to preside in the hearings. As a result, there are now over five times as many AJs as ALJs.²⁹See id. at 3 (“In contrast to the 1,931 ALJs in the federal government, agencies reported at least 10,831 non-ALJs.”). Each of the thousands of AJs is highly susceptible to pro-agency bias, and there is abundant evidence that many AJs act in ways that reflect that bias.³⁰See, e.g., Catherine Y. Kim & Amy Semet, An Empirical Study of Political Control Over Immigration Adjudication, 108 Geo. L.J. 579, 587–88 (2020) (“Finally, using logistic regression and controlling for the same variables, we examined the influence that a sitting President might exercise over IJs’ decisions, regardless of which administration appointed the IJ. Here, we found clear differences across administrations—the identity of the administration in control at the time of decision (or presidential era) is a statistically significant predictor of removal rates, controlling for other variables.”).

The Administrative Conference of the United States (“ACUS”) published a lengthy study of this phenomenon in 1992.³¹See generally Verkuil et al., supra note 16. It expressed dismay that Congress and agencies had reverted to the methods of conducting agency adjudications that led to the well-supported and widespread complaints of pro-agency bias in the 1930s. ACUS urged Congress and agencies to abandon the practice of using potentially biased AJs to preside in adjudications and urged them to replace AJs with ALJs who are subject to the APA provisions that insulate them from sources of pro-agency bias. Unfortunately, Congress continued to decrease agency use of ALJs and to increase use of AJs. Thus, for instance, while there were twice as many AJs as ALJs when ACUS published its 1992 report, there were five times as many AJs as ALJs when ACUS published another study of agency adjudication in 2018.³²Compare id. at 788–89, with Barnett et al., supra note 28, at 3.

Scholars have also identified one unfortunate characteristic of a system of adjudication that relies on adjudicators who have decisional independence. Judges differ with respect to the combination of beliefs and values they bring to the task of adjudication. Those differences can create a pattern of inconsistent decisions. Thus, studies of Social Security disability decisions made by ALJs have found variations in patterns of decisions so great that the identity of the ALJ who was assigned to the case was the best predictor of the outcome of the case.³³See, e.g., Jerry L. Mashaw, Bureaucratic Justice: Managing Social Security Disability Claims 72–73 (1983); Jerry L. Mashaw, Charles J. Goetz, Frank I. Goodman, Warren F. Schwartz, Paul R. Verkuil & Milton M. Carrow, Social Security Hearings and Appeals: A Study of the Social Security Administration Hearing System 42 (1978).

Inconsistent patterns of decisions can exist in any context in which adjudicators have decisional independence, including decisions by federal judges. Eliminating or reducing the decisional independence of judges is not an appropriate response to that problem. There are many ways in which an agency can reduce the level of inconsistency in an adjudication system. They include: (1) issuing rules that reduce the discretion of judges; (2) establishing a system of binding precedents; (3) providing an opportunity for review by a multimember panel of judges; and (4) monitoring the patterns of decisions of individual judges to identify and counsel judges whose pattern of decisions deviate significantly from the norm.

The Supreme Court’s 2018 opinion in SEC v. Lucia³⁴138 S. Ct. 2044 (2018). created the potential for many more problems to develop in the agency adjudication process. In Lucia, the Court held that ALJs at the Securities and Exchange Commission (“SEC”) are inferior officers rather than employees.³⁵Id. at 2053. That holding undoubtedly applies to all ALJs who preside in adjudications at regulatory agencies. It almost certainly applies as well to AJs who perform functions analogous to ALJs at many agencies. The holding in Lucia may also apply to the thousands of ALJs and AJs who preside in adjudicative hearings at agencies that administer benefit programs. The functions of the ALJs who preside at agencies like the Social Security Administration differ so much from the functions of the AJs and ALJs who preside at hearings at regulatory agencies, however, that they may not qualify as inferior officers as the Court described and applied that term in Lucia. Hearings at agencies that administer benefit programs are not adversarial proceedings, as the only party that participates in the hearing is the applicant for benefits.

The Lucia holding creates serious doubts about the constitutionality of the statutory limits on the power of agencies to remove ALJs. In its 2010 opinion in Free Enterprise Fund v. PCAOB,³⁶561 U.S. 477 (2010). the Court held that the Take Care Clause prohibits Congress from providing two or more layers of insulation of inferior officers from presidential control in the form of good cause limits on the power to remove inferior officers.³⁷Id. at 484. Since ALJs can only be removed for cause by the MSPB, and MSPB members can only be removed for cause by the President, the Free Enterprise Fund holding suggests that the good cause limit on the power to remove an ALJ is unconstitutional. The Court included a footnote in its opinion stating that its holding did not necessarily apply to ALJs.³⁸See id. at 507 n.10.

Immediately after the Court issued its decision in Lucia, President Trump issued an executive order (“EO”) in which he removed ALJs from the category of civil servants who can only be appointed through use of a competitive meritocratic selection process.³⁹Exec. Order No. 13,843, 83 Fed. Reg. 32,755 (July 10, 2018) (“Excepting Administrative Law Judges from the Competitive Service”). The EO rescinded the rules that subjected applicants to be ALJs to an elaborate meritocratic selection process implemented by OPM.⁴⁰Id. The EO replaced those rules with a single criterion—to be eligible for appointment as an ALJ, an applicant must only be a member of a state bar association.⁴¹Id. In all other respects, the decision to appoint an ALJ is now completely within the discretion of the head of each agency. That raises the concern that some agency heads will appoint incompetent or biased ALJs. There is evidence that some agency heads have already used their discretion to appoint ALJs and AJs that are incompetent and biased in favor of the views of the head of the agency.⁴²See Catherine Rampell, Trump Didn’t Build His Border Wall With Steel. He Built It With Paper, Wash. Post (Oct. 29, 2020), https://perma.cc/5MX5-TSD4; Jennifer Yachnin, Appointments Signal Political Shift for In-House Judges, E&E News (May 1, 2020), https://perma.cc/4UEV-6WT7.

The combination of the Court’s opinion in Lucia and the Court’s opinions in other recent cases and the preexisting structure of the adjudication programs at many agencies raises other serious questions. In its opinions in Freytag v. Commissioner⁴³501 U.S. 868 (1991). and Edmond v. United States,⁴⁴520 U.S. 651 (1997). the Court seemed to say that officers are inferior officers, rather than principal officers, only if they can be removed by a principal officer and their decisions can be reviewed by a principal officer.⁴⁵See id. at 665–66; Freytag, 501 U.S. at 877–78, 881–82. The decisions of SEC ALJs are reviewable by the SEC⁴⁶15 U.S.C. § 78d-1 (2018).—an agency headed by principal officers—so SEC ALJs satisfy that criterion to be classified as inferior officers.

In many agencies, however, decisions of ALJs and AJs are not reviewable by principal officers. Thus, for instance, the decisions of the hundreds of Administrative Patent Judges (“APJs”) are not reviewable by any officer;⁴⁷See 35 U.S.C. § 314(d). the decisions of ALJs at the Social Security Administration are reviewable only by an Appeals Council that consists only of inferior officers or employees;⁴⁸See 20 C.F.R. §§ 404.967, 404.970 (2020). the decisions of ALJs at the Department of Agriculture are only reviewable by an inferior officer;⁴⁹See 7 C.F.R. § 2.27 (2020). and the decisions of IJs are only reviewable by the Board of Immigration Appeals that consists only of inferior officers or employees.⁵⁰See 8 C.F.R. § 1003.1 (2020).

In Arthrex, Inc. v. Smith & Nephew, Inc.,⁵¹941 F.3d 1320 (Fed. Cir. 2019). the Federal Circuit held that APJs are principal officers rather than inferior officers.⁵²Id. at 1335. That holding may be followed by holdings that thousands of other ALJs and AJs are principal officers rather than inferior officers. The Arthrex court held that the constitutional flaw in the system of adjudication at the Patent and Trademark Appeals Board (“PTAB”) could be remedied by making APJs removable at will. That remedy is controversial, however. The Supreme Court could easily affirm the holding in Arthrex but reject the remedy adopted in Arthrex.⁵³The Supreme Court granted the petition for writ of certiorari in Arthrex and will hear oral argument during the 2020 October Term. Id., cert. granted, No. 19-1434, 2020 WL 6037207, at *1 (Oct. 13, 2020). That would create a legal environment in which thousands of AJs and ALJs are principal officers rather than inferior officers.

II.     The Present State of Agency Adjudication

The present state of agency adjudication is troubling, and it may continue to deteriorate in the near future. Currently, eighty percent of the members of the administrative judiciary are highly vulnerable to pressure from the politicians that head their agencies. As AJs, they have no statutory safeguards against removal or other adverse actions that an agency might take if they fail to act in ways that please the head of the agency.⁵⁴See Barnett et al., supra note 28, 1, 4. In other words, they are highly vulnerable to the systematic pro-agency bias that plagued agency adjudications before Congress enacted the APA.

The proportion of agency adjudicators who are highly vulnerable to pro-agency bias will increase from eighty to one hundred percent if the Court extends the holding of Free Enterprise Fund to ALJs, thereby invalidating the statutory safeguards against removal that are designed to ensure that ALJs have decisional independence. Moreover, because of the combination of the Lucia holding and the EO that removed ALJs from the category of civil servants who are subject to a competitive meritocratic qualification process, agency heads now have complete discretion to hire AJs and ALJs who are incompetent and who are expected to act in ways that reflect the biases of the agency head who hired them.

Pending cases in both the Supreme Court and courts of appeal have the potential to destroy completely the viability of important agency adjudication programs. In the many contexts in which the decisions of AJs and ALJs are not subject to review by principal officers, the AJs and ALJs arguably are principal officers rather than inferior officers. As such they must be appointed through the process of nomination by the President and confirmation by the Senate. It is not at all clear that the President and the Senate are up to the task. Over the last couple of decades, the process of nomination and confirmation of principal officers has slowed and become increasingly politicized. That has produced a large and growing increase in the number of principal officer positions that are vacant for long periods of time.⁵⁵See Anne Joseph O’Connell, Actings, 120 Colum. L. Rev. 613, 651 (2020) (“Acting deputy administrators have cumulatively served 14.9% and 41.9% of the days under Presidents Bush and Obama, respectively, including the recess appointment under President Bush, but 66% under the Trump Administration when the position was occupied. Acting general counsels have racked up about one-third of the time when the job has been staffed under Presidents Bush and Trump (and 14.2% under President Obama).”). Requiring the President and the Senate to more than double the number of principal officers who must be nominated and confirmed could prove fatal to an appointment process that is already badly broken.

III.     Potential Ways of Repairing the Agency Adjudication System

To the extent that the modern Court’s opinions permit, the agency adjudication system could be repaired by returning to the system that Congress adopted unanimously in 1946 after fifteen years of debate and study. If the Court declines to apply the Free Enterprise Fund holding to ALJs, Congress could require all agencies to conduct agency adjudications as formal adjudications that are subject to APA sections 554, 556, and 557, and to the many statutory safeguards of the decisional independence of ALJs. If the Court applies the Free Enterprise Fund holding to ALJs or Congress is unwilling to require all agencies to use formal adjudication, agencies can bind themselves by rule to use formal adjudication and not to remove an ALJ except for good cause.⁵⁶See Kent Barnett, Regulating Impartiality in Agency Adjudication, 69 Duke L.J. 1695, 1744 (2020) (“First, the impartiality regulations limit the ramifications of applying Free Enterprise Fund to adjudicators because the regulations, by replicating the statutory status quo, serve as a backstop. In fact, if agencies acted quickly to promulgate them before current judicial proceedings end, the regulations may moot any challenge because the Court’s judgment would not lead to a remedy that changes anything. The adjudicators would have the same or more protection from at-will removal even absent the current statutory regime.”).

Congress cannot require agencies to use the competitive meritocratic system that OPM had long used before Lucia to determine whether applicants are qualified to become ALJs. The Court would hold that any such statute imposes unconstitutional conditions on the President’s power to appoint inferior officers. However, the President could issue an EO in which he authorizes OPM to implement a similar system, or agencies could adopt systems of that type by rule.

Finally, Congress can eliminate the risk that many AJs and ALJs will be held to be principal officers, rather than inferior officers, by amending every statute that now authorizes ALJs and AJs to make final decisions. If Congress adds a right to obtain intra-agency review of ALJ and AJ decisions by a principal officer, the ALJs and AJs become inferior officers rather than principal officers. Congress would have to amend at least the Social Security Act, the INA, the statutes that govern adjudication in the Department of Agriculture, and the statutes that apply to adjudications conducted by the PTAB to obtain that result. Of course, Congress would also have to increase the appropriations for each agency to the extent required to staff the new intra-agency appellate body, and the new appellate positions would have to be filled through the process of presidential nomination subject to Senate confirmation.

The disadvantages of this approach to the problem are obvious. Implementation is contingent on the willingness and ability of Congress, the President, and agencies to act in the face of powerful political forces. The resulting system of adjudication would also be fragile. Its persistence would depend on the willingness of future presidents to act in ways that assure that agency adjudications are conducted through the use of fair procedures and that agency adjudicators continue to be subject to safeguards on their decisional independence.

An alternative approach would be to create an entirely new system of agency adjudication that is modeled on the approach that some states have taken in recent years. A new Article I court could conduct all of the adjudications that are now conducted at agencies.⁵⁷See Richard E. Levy & Robert L. Glicksman, Restoring ALJ Independence, 105 Minn. L. Rev. 39, 96 (2020). To be consistent with the Take Care Clause, such a system would have to include a process through which the President, a department head, or a court appoints each judge; a process through which the head of the relevant agency or some other principal officer within the relevant agency can review ALJ decisions; and a process through which a principal officer who can be removed by the President at will can remove an ALJ for cause. If the principal officer with the power to remove an ALJ is the Chief Judge, rather than an agency official, this system could yield a return to the environment in which ALJs enjoy decisional independence and adjudications are conducted by unbiased adjudicators.

The disadvantages of this approach are also obvious. It would be a challenge to persuade Congress to make such a dramatic change in our administrative adjudication system. It would also be a challenge to design and to implement such a radical new system of adjudication.

A final approach is to build on, and to expand the scope of, successes in the field of administrative law. Those successes rely on written hearings to resolve many important types of disputes.

One such success is the use of the notice-and-comment process to issue rules that reflect and implement agency policies. When Congress enacted the APA, agencies often relied on adjudication as their primary means of making policy decisions. Over time, rulemaking has displaced adjudication as the primary means through which agencies make policy decisions. Policymaking through use of notice-and-comment rulemaking has enormous well-documented advantages over policymaking through adjudication.⁵⁸See 1 Kristin E. Hickman & Richard J. Pierce, Administrative Law Treatise § 4.8 (6th ed. 2019). A recent study of notice-and-comment rulemaking conducted by an outstanding team of scholars found that agencies have developed extraordinarily rich and effective means of using the notice-and-comment process to make policy decisions that are well supported by data and analysis.⁵⁹Wendy Wagner, William West, Thomas McGarity & Lisa Peters, Dynamic Rulemaking, 91 N.Y.U. L. Rev. 183, 183 (2017).

Rules can interact with adjudications in many ways. Sometimes they eliminate the need for an entire class of adjudications.⁶⁰See, e.g., Nat’l Petrol. Refiners Ass’n v. FTC, 482 F.2d 672, 674, 690, 697–98 (D.C. Cir. 1973), cert. denied, 415 U.S. 951 (1974) (upholding an agency rule that determines that all retail sales of gasoline without posting octane content are misleading, thereby eliminating the need to conduct adjudications with respect to each individual sale of gasoline). Sometimes they reduce the scope of a class of adjudications by resolving generically an issue that otherwise would be contested in every adjudication.⁶¹See, e.g., Vt. Yankee Nuclear Power Corp. v. Nat. Res. Def. Council, 435 U.S. 519, 525–27 (1978) (upholding an agency rule that had the effect of eliminating the need to litigate the issue of whether nuclear waste can be safely stored in every nuclear power plant licensing proceeding). Sometimes they transform a class of adjudications from disputes about how to interpret and apply a broad subjective standard to disputes about how to resolve one or more disputed but objectively verifiable issues.⁶²See, e.g., Am. Mining Cong. v. Mine Safety & Health Admin., 995 F.2d 1106, 1107, 1113 (D.C. Cir. 1993) (upholding an agency rule in which the agency interpreted the term “diagnosed” to refer to two objectively verifiable numerical diagnostic criteria).

A second major success is use of paper hearings to adjudicate many important types of disputes. Gradually, agencies and courts have come to realize that paper hearings are superior to oral hearings in the large and important classes of adjudications in which parties disagree only about issues of legislative fact.⁶³Cf. Hickman & Pierce, supra note 58, at § 6.3. Thus, for instance, courts have approved of the Environmental Protection Agency’s decision to switch from oral hearings to written hearings in cases in which a firm applies for a discharge permit under the Clean Water Act and opponents argue that that the discharges authorized by the permit will have unacceptable adverse effects on aquatic biota.⁶⁴Dominion Energy Brayton Point, LLC v. Johnson, 443 F.3d 12, 18–19 (1st Cir. 2006) (“Because we, like the Seacoastcourt, cannot discern a clear and unambiguous congressional intent behind the words ‘public hearing’ in the CWA and because the EPA’s interpretation of that term constitutes a reasonable construction of the statute, deference is due.” (relying on Seacoast Anti-Pollution League v. Costle, 572 F.2d 872 (1st Cir. 1978)). The written hearings require less time and resources and are more likely to yield accurate and consistent resolutions of contested issues in fields like biology and other hard sciences, as well as in economics and other social sciences.

Paper hearings also perform well in the context of contested issues of adjudicative fact if those issues are relatively objective. Thus, for instance, in Mathews v. Eldridge,⁶⁵424 U.S. 319 (1976). the Court held that written hearings are sufficient to satisfy due process in Social Security disability cases because doctors, whose opinions dominate in those cases, are good at expressing their opinions in writing and because the contested issues are relatively objective.⁶⁶Id. at 344–45 (“By contrast, the decision whether to discontinue disability benefits will turn, in most cases, upon routine, standard, and unbiased medical reports by physician specialists . . . . More important, the information critical to the entitlement decision usually is derived from medical sources, such as the treating physician. Such sources are likely to be able to communicate more effectively through written documents than are welfare recipients or the lay witnesses supporting their cause.” (internal quotation marks omitted)).

In Eldridge, the Court was aware that any applicant who was dissatisfied with the results of a paper hearing had the statutory right to contest the findings made in the paper hearing in an oral hearing before an ALJ, so it is plausible that the Court’s resolution of the due process issue was contingent on the subsequent availability of an oral hearing. That interpretation of Eldridge is dubious, however. The initial decision whether an applicant is eligible for disability benefits is made by a team that consists of a doctor and a disability expert. They base their findings on their consideration of all of the opinions of doctors that the applicant and the government provide to the team plus additional opinions that the team obtains when it finds gaps in the record that can be filled with additional opinions. If an applicant disagrees with the findings of the first team of decision makers, they can obtain a new paper hearing before a different doctor-disability expert team of decision makers. If the applicant disagrees with the findings of both of those decision-making teams based on paper hearings, the applicant can obtain an oral hearing before an ALJ. It seems unlikely that the availability of that third layer of decision-making improves the accuracy and consistency of the decision-making process. There is no logical reason to believe that an ALJ with no education or training in human health will make a more accurate decision than two doctor-disability expert teams of decision makers, particularly when the government does not even participate in the oral hearing before the ALJ.⁶⁷See Richard J. Pierce, What Should We Do About Social Security Disability Appeals?, 34 Regul. 34, 36 (2011).

Over time, more contexts will be discovered where the combination of simplifying rules and paper hearings can elevate the potential for enhanced accuracy and consistency in the decision-making process. Many scholars predict that advances in artificial intelligence will contribute to the replacement of the vast majority of oral hearings with paper hearings.⁶⁸See, e.g., David Freeman Engstrom, Daniel E. Ho, Catherine M. Sharkey & Mariano-Florentino Cuéllar, Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies 82 (2020) (“This [piece] makes three points about the future of hearing rights in the face of the AI revolution. First, while the most optimistic version of AI tools may improve accuracy and efficiency of adjudicatory decisions, such tools may also expose trade-offs in normative values underpinning hearing rights. Second, we articulate how procedural due process and statutory hearing rights may need to adapt if AI tools proliferate. A core challenge in the near-term will be crafting legal and institutional vehicles to detect and address systemic sources of error in light [of] the current structure of individualized decision-making. Third, the rise of AI tools in adjudication potentially raises longer-term, foundational questions: Do due process and statutory hearing rights imply a right to a human decision-maker? And what role is left for hearing rights in a world in which legal and regulatory mandates are crafted, adjudicated, and enforced with increasingly limited human involvement?”). There will always be an irreducible minimum of cases in which oral hearings are essential because the stakes are high and the disputed issues are who did what, when, where, and why. In those contexts, oral hearings are required by due process, but this is a small subset of the myriad classes of cases in which oral hearings are presently used.

Conclusion

The agency adjudication process is in poor condition today and recent Supreme Court opinions have the potential to make it far worse. Major systemic changes are needed to create the conditions in which adjudications produce accurate and unbiased results. This can be accomplished in one of three ways. First, Congress, the President, and agencies can take actions that recreate the fair and impartial system that Congress created when it enacted the APA in 1946. Second, Congress can create an entirely new system of agency adjudication centered on a new Article I court. Finally, the two most important successes in the field of administrative law—notice-and-comment rulemaking and the replacement of oral hearings with paper hearings—can be expanded. The third approach offers the best prospects for success, but aggressive adoption of one of the above approaches is required if agency adjudications are to perform well in the future.