From the founding to the Administrative Procedure Act
This piece was originally published in Artificial Weights on May 6, 2026.
In the first piece of this series that discussing ideas in my article, I described how administrative procedure gone wrong has both stymied abundance and facilitated autocracy. In the second piece of this series, I discussed the misapprehension that a more unconstrained executive is always a more powerful executive.
For those who are unfamiliar with the basics of administrative procedure, this third piece aims to provide some useful background information. I argue that the United States has had three systems of federal administrative procedure: a two-track system lasting from the Founding to the Gilded Age, a second system from the late 1800s to the 1960s that would be codified in the APA, and the third system that was constructed beginning in the 1960s and that we still live in today. This piece covers the first two of these systems: the development of federal administrative procedure from the Founding to the enactment of the Administrative Procedure Act of 1946.
Even for those who already know that history, my hope is that it helps to illustrate how previous regimes of administrative procedure grappled with the tension between inhibiting corrupt, autocratic, or unreasonable government action without inhibiting—and ideally abetting—effective governance. The failure of the first two eras of administrative procedure has lessons for those of us who wish to reform the third.
The two-track system: Administrative procedure from the founding to the Gilded Age
The failures caused by both autocratic and abusive government, as well as overly constrained and debilitated government, were not abstract for the Founders. Their generation had experienced both the tyranny of the British in the lead-up to the Revolutionary War, and the utter failure of excessively constrained governance under the too-weak Articles of Confederation. The Founders, “who had experience in colonial affairs, and who had suffered under the yoke of British administrative practices, were hardly naïve about the need to control government as well as empower it” (Mashaw at 38).
It is commonly asserted that, following the failures of British rule and the Articles of Confederation, Congress in the early years of the United States micromanaged the details of administration with little room left for executive or judicial judgment (id. at 44–46). But this is a myth. As is common today, statutes varied greatly in their level of generality and delegation. For an example of how sweeping early delegations could be, consider that in 1789, Congress passed a one-sentence statute providing that Revolutionary War pensions should be paid “to the invalids who were wounded and disabled during the late war . . . under such regulations as the President of the United States may direct” (An Act Providing for the Payment of the Invalid Pensioners of the United States, ch. 24, § 1, 1 Stat. 95, 95 (1789)).
And while it is often asserted that federal regulatory agencies did not exist prior to the Interstate Commerce Commission’s creation in 1887 (see, e.g., Fallon at 992), this is misleading. “From the earliest days of the Republic, Congress delegated broad authority to administrators, armed them with extrajudicial coercive powers, created systems of administrative adjudication, and specifically authorized administrative rulemaking” (Mashaw at 5). From early on, then, federal administrative law had to grapple with grants of authority to federal agencies that concerned important topics and required the exercise of substantial discretion (and gave rise to the temptation to abuse the powers granted).
To deal with this, early federal authorities continued to apply administrative law as it was known at English common law and in the colonies: the “law of officers” (id. at 63). This amounted to a two-track system (id. at 24).
The first track allowed for administrative action to be reviewed, albeit indirectly, by bringing a common law action—tort, contract, property, etc.—against a federal officer. The court would review whether the officer had acted legally and reasonably under the circumstances without any deference for the government actor (i.e., “de novo”). Accordingly, if an officer took any illegal action, they were personally liable for any harm done: for example, “without a valid law to justify it, tax collection . . . was simply a trespass” (Donahue at 1911). To us today, this is an extreme doctrine: “all officers, from the President of the United States, downwards . . . if they should overstep the limits of their official authority . . . would be liable to a civil or criminal prosecution, in the same manner as a private citizen” (id. at 1912). There were almost no immunities from suit: “the U.S. Supreme Court held a naval officer liable for implementing unlawful orders because ‘the instructions cannot . . . legalize an act’” (id. at 1920). And to ensure that lack of funds did not render an officer judgment-proof, Congress often layered requirements that officers post bonds or sureties into statutes (Mashaw at 62). Further, “qui tam” suits allowed anyone to bring suit and profit if a federal officer unlawfully withheld funds (id.).
The second track allowed for direct review of administrative actions through writs of mandamus seeking injunctive relief (that is, orders that the officer do something or refrain from doing something) (id. at 3). An important difference between the “law of officers” in England and federal administrative procedure in the United States was the scope of this mandamus relief. In England, as well as in many states, mandamus jurisdiction had been widened beyond cases where officers had no discretion whatsoever. “As early as 1758” in England, “Lord Mansfield had” ruled that mandamus was available “if it appeared to the Court . . . that [a Justice of the Peace’s] conduct was influenced by partial, oppressive, corrupt, or arbitrary views instead of exercising a fair and candid discretion” (id. at 211; see also Mechanick at 486, 490–94). But federal courts in the United States resisted this broader conception of mandamus as a vehicle to review the reasonableness of the discretionary actions of officers for some time, restricting mandamus to cases where an officer had absolutely no discretion (Mashaw at 212).
The result was that, from the Founding through at least the Jacksonian era, federal administration see-sawed between a timid and narrow form of mandamus review (if no common law action could be brought) and a juristocratic and aggressive second-guessing form of review (if common law actions were available).
How the two-track system shaped and misshaped government
Governments always shape their behavior in the shadow of the law; the Founding era was no exception. Supervisors, being on the hook financially for the actions of their subordinates, were often empowered by statutes to sue their own subordinates to collect a bond in the event that the subordinate’s action cost the supervisor (Mashaw at 62). Needless to say, this led to a form of governance that placed far more individual responsibility for assessing the legality of action in each officer, a notion foreign to our more hierarchical understanding today (Donahue at 1893).
In this era, there were no procedural restrictions on how agencies adjudicated cases or promulgated rules (Mashaw at 84). But congressional oversight pressure led to increasing requests (or statutory mandates) for agencies to provide reports on agency activity; that pressure, in turn, led agencies to forms of exercising power that reflected “formality, recordkeeping, and caution” (id. at 104). For example, in the middle of the nineteenth century the Board of Supervising Inspectors (an early steamboat safety agency) provided explanations of the basis and purpose of any new rule or amendment it made to a prior rule; it did so in preambles to its regulations as well as reports to the Secretary of the Treasury (id. at 203). Attentive agencies also sought expert input before regulating, even in this era. Returning to the Board of Supervising Inspectors, it “set[] aside some time at its annual meeting to hear orally from petitioners” and would “sometimes invite[] outsiders with special interests or competence to meet with the Board concerning particular issues” (id.). These were not just meetings with special interests—like steamboat transportation firms—but also meetings with professors, for example, to discuss new research shedding light on how to better protect safety valves from rust damage.
The problem was the way that the two-track system failed to function as the Jacksonian era shifted into the Gilded Age, and more and more government action fell outside of the common law track. Mostly, this was because the federal government increasingly engaged in “benefits provision, regulation, and licensing,” which common law actions did not easily apply to (Mashaw at 24).1 More generally, the fact that—depending on the nature of a legal claim—officers were free to act without judicial check when exercising discretion (if only mandamus was available) or were subject to “de novo second-guessing of administrative action” through common law actions led governments to shift authorities outside of the scope of the common law.
Consider the case Rees v. Watertown (86 U.S. 107 (1873)), where a plaintiff successfully won money damages against the city of Watertown, and secured a writ of mandamus compelling the members of its city council to levy a tax in order to pay those money damages. Before Rees could get the writ served on the city council, a majority of its members resigned. Rees then got another writ. And a majority of the council resigned again. The process played out once more. After the third attempt, Rees turned to the federal courts, but on appeal the Supreme Court continued to hold that a writ of mandamus could only run against the relevant officers (not the corporate body of the city itself), so Rees would just have to try for a fourth time. The narrowness of mandamus allowed for enormous mischief.
Yet at the same time that limits on mandamus relief were causing a great deal of trouble, “[j]udicial review in a de novo form could be enormously intrusive, indeed paralyzing” for federal officers who were at the mercy of how “nineteenth-century courts and juries” would later judge their actions (Mashaw at 308). Officers acting in good faith could be held liable for damages “because the laws they used to justify their actions were not passed according to proper procedure or were subsequently ruled unconstitutional,” or simply because they made a good faith and reasonable non-legal judgment that was nevertheless erroneous (Donahue at 1911–12).
The collapse of the two-track system and the rise of the second system
The pressure these seemingly absurd results put on this two-track federal administrative procedure regime caused it to begin to collapse across the end of the nineteenth and beginning of the twentieth centuries. Immunity regimes for officers acting in “quasi-judicial” capacities began to emerge in the middle of the nineteenth century (id. at 1958–73). Federal statutes began to replace direct liability with administrative appeals procedures in the back half of the nineteenth century (id.). Taking a different tack, federal statutes enacted during the Civil War protected “soldiers and other government officials from liability for following presidential orders” (id. at 1962). In the early twentieth century, the Supreme Court began to allow judicial challenges to agency actions that went far beyond the traditional scope of writs of mandamus, creating a general assumption that agency actions could be reviewed for legal error (that is, whether the agency exceeded its statutory or constitutional authority) (Mashaw at 248). Following English and state law, the Supreme Court blurred the formerly clear rule that discretionary actions are not reviewable (Mechanick at at 496–97). It eventually adopted the view that even if an agency action is “in form within the delegated power, nevertheless it must be treated as not embraced therein, because the exertion of authority which is questioned has been manifested in such an unreasonable manner as to cause it, in truth, to” exceed the proper meaning of the delegated power: what we have come to know as arbitrary or capricious review (id. at 498–501).
During the early twentieth century, federal courts began to require that federal agencies produce a record to review, in order to ensure that this arbitrary or capricious review could be meaningful (id. at 501). And by 1930, Congress had amended the Radio Act of 1927 to explicitly provide for arbitrary or capricious review, codifying into a federal statute the (by then, standard) common law of federal administrative procedure (id. at 503–04). Federal statutes would also increasingly authorize damages suits against the United States over this period, gradually rendering common law actions against officers obsolete until 1946, when that process culminated with the United States assuming liability generally for damages actions in the Federal Torts Claims Act (Donahue at 1973). 1946 is also when the Administrative Procedure Act (APA) was passed, which broadly codified the new regime of administrative procedure that had emerged from these changes to the old.
The Administrative Procedure Act of 1946
Administrative procedural reform was increasingly wrapped up in a fight over the New Deal as the Roosevelt administration wore on. As Roosevelt pivoted from early New Deal efforts that took an often corporatist, industrial cooperation approach to a more business-hostile effort to help “labor, tenant farmers, immigrants, and the urban poor,” conservative opposition to his administration grew (Schiller (2007) at 1564). Accordingly, “Republicans and Southern Democrats” pushed for stricter administrative procedure in order “to constrain liberal New Deal agencies” (id. at 1560). That was in part because “[e]ven after years of Roosevelt’s judicial appointments, the judiciary still contained many conservative judges,” including a majority of the Supreme Court (id. at 1613). As a result, how extensive and how strict judicial review of agency action would be (with the assumption that conservative judges would impede and delay liberal policies) became a particular sticking point in legislative debates between liberals and conservatives over administrative procedure reform (id. at 1613–14).
The legislative breakthrough that led to the Administrative Procedure Act was the increasing discomfort that liberals had with how some agencies were operating. This pressure led President Roosevelt to ask his Attorney General to form a committee advising on administrative procedure reform—composed of eight New Dealers, and four conservatives—and notably, its majority recommendations did not eschew substantial procedural constraints. The report noted that judicial review of agency action serves as a “check against excess of power and abusive exercise of power” on one hand, but also impedes and slows the “effective discharge [of] statutory obligations” (Committee at 76). So while some rulemaking should be issued promptly with limited process, other rulemaking should not be issued “until all those to be regulated have been given an opportunity to present facts and arguments to those in authority for the purpose of enlightening or persuading them towards this or that choice among many alternatives” (id. at 2, 103). And similarly, judicial review should be expected to ensure that agency action does not exceed “its lawfully delegated authority” and “check extremes of arbitrariness or incompetence”—but not to “[e]nsure ‘correct’ decisions,” as “[t]he correctness of a decision” is “a matter of judgment” (id. at 77–79). For that reason, the committee largely endorsed maintaining—and not weakening—the then-extant standard of reasonableness in judicial review of agency action, arbitrary or capricious review (id. at 78–79, 81–83, 90).
At the same time as liberal views were beginning to shift, and Roosevelt became more aligned with an explicitly liberal agenda, his court packing plan fueled voters’ fears that “the president sought the same absolute authority as the dictators who had recently achieved power in Europe” (Shepherd at 1581). This concern was serious; a 1936 poll—conducted before Roosevelt’s court packing plan, when he was at the peak of his popularity—already found that 45% of respondents were afraid that FDR’s policies could bring about a dictatorship in the United States (Kovacs (2021) at 578).
The defeat of the court packing plan and World War II agency failures led liberals even further towards compromise. Due to the war emergency, agencies had been given new powers and constraints on the exercise of their powers had been relaxed, but the result was more “abuses by the agencies” and “blunder[s]” (Shepherd at 1641–42). The public particularly loathed agency failures that caused chronic shortages and rationing of products like gasoline and clothing (id.). Both shortages and a spike in inflation was blamed on the new Office of Price Administration, which “rationed over ninety percent of consumer goods during the war” (id.). Frustration also flared when the agency issued regulations that seemed ludicrous, including one “prohibiting the sale of sliced bread” and a “40,000 word order mandating beef butchering techniques” (Schiller (2002) at 193). The War Production Board also “exercised unprecedented power” but was mismanaged, causing “endemic delay in the fulfillment of war contracts.” It also appeared “to be completely captured by the industries it was supposed to regulate,” funneling contracts exclusively to favored large corporations (id. at 194). And just as liberals grew more amenable to constraints on agencies due to wartime failures, they also feared the judiciary less, due to the rising share of Roosevelt appointed judges on the bench.
At the same time, the flip in Supreme Court jurisprudence that started in 1937 and the rising fraction of liberal judges on the bench made conservatives fear that “liberals might now use both broad judicial review and friendly courts to impede agency action that favored conservatives or business interests” during a future administration (Shepherd at 1645). Conservatives, accordingly, were increasingly open to imposing a less demanding set of administrative procedures. This allowed liberals and conservatives to compromise on a middle-ground form of judicial review that largely codified the status quo for most agencies, applying it across the board with only a few major changes. This was the APA: in 1946, after years of intense conflict, it passed by voice vote in both chambers with no dissenting votes recorded.
One genuine innovation of the APA was notice-and-comment rulemaking. In the New Deal era, rulemaking procedures were inconsistent, not only between agencies but “even within individual agencies” (Bremer at 98-99). Only a little had changed since the early days of regulating steamboats. The best agencies were praised for gathering expert input in rulemaking, yet “much of the [rulemaking] process was internal to the agency. But external consultation, through written comments and oral conferences, was increasingly common” in this period (id. at 101). The emphasis of external consultation was “targeted solicitation of views from representatives of organized industry or interest groups,” as well as “persons known to be knowledgeable about the relevant subject matter” (id. at 104). Yet because this process was ad hoc and discretionary, there was a concern in this era that “the consultative process was too closed” and captured, and that it did not do a good enough job “informing the agency’s expert judgment by giving the agency access to information it might not otherwise possess” (id. at 108, 116).
The APA’s solution was notice and comment. Agencies had to publish a “notice of proposed rule making . . . in the Federal Register” that included a “reference to the legal authority under which the rule is proposed” and “either the terms or substance of the proposed rule or a description of the subjects and issues involved” (5 U.S.C. § 553(b)). This notice had to be accompanied by “an opportunity” for members of the public to submit “written data, views, or arguments” about the proposed rule (id. § 553(c)). Following this public comment process, the agency would have to have the final rule accompanied by “a concise general statement of” the rule’s “basis and purpose” (id.) As the Senate committee report noted, this language “required agencies to ‘analyze and consider all relevant matter presented’ and ‘explain the actual basis and objectives of the rule’ ‘with reasonable fullness’” (Kovacs (2018) at 529).
Apart from notice and comment—which matched no agency’s practices, although it was inspired by many—the APA largely codified existing judicial doctrine and agency best practices at the time of its enactment. It divided rulemaking from adjudication, defining the latter as a residual category (5 U.S.C. § 551(4)–(7)). It provided different procedures for informal and formal versions of each type of action, with little process for informal adjudication and notice-and-comment process for informal rulemaking, along with more onerous trial-type procedures for formal versions of adjudication and rulemaking (id. § 553, 556, 557). And it provided that a category of rules would not be subject to the notice-and-comment process, which we today call guidance: “interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice” (§ 553(b)(A)). The APA also cemented the shift in mandamus by allowing “any person suffering legal wrong because of agency action” to bring suit (id. § 702), and directing courts to “set aside” any action that was sufficiently unreasonable or exceeded the agencies’ statutory or constitutional grant of authority (id. § 706). It also separated the functions of those in agencies who would investigate and prosecute lawbreaking from those who would adjudicate it (id. § 557).
The APA survives, but the second system ends
The APA was broadly successful, yet the second system would be abandoned in reaction to deep problems in the post-World War II period. If that seems odd, recall how serious those problems were. As Paul Sabin put it:
The government, often in partnership with industry and labor, was testing nuclear weapons in the atmosphere, spraying millions of tons of pesticides across the land, and plowing highways through urban neighborhoods. The government was allowing strip mines to ravage the Appalachian Mountains and leaving coal miners to suffer from black lung disease with little compensation. Government policies were permitting oil refineries to freely dump toxic emissions into low-income communities of color, and letting oil spills pollute the nation’s waterways and coasts.2
Distrust of government also grew as the civil rights struggle intensified and the American public became aware of lies in the official government narrative around the Vietnam War (Sabin at 48–53).
Were the APA’s shortcomings or inadequacies to blame for these problems? As I discuss in my article, evidence indicates that it was not.3 But as the next piece in this series will detail, a group of reformers eventually organized around the idea that solutions could be found in a new form of administrative procedure. One that layered on top of the APA a host of more exacting requirements that would ensure better government action.
Learning from history
The next post will dive into how this second era of administrative procedure, which began developing as the first era collapsed towards the end of the nineteenth century and fully realized with the enactment of the APA in 1946, was supplanted by a third era characterized by a hostile turn against state power paired with some narrow ad hoc exceptions to otherwise binding procedural constraints. But before turning to that subject, it should be emphasized that there is something to be learned from each of the first two eras of federal administrative procedure.
The breakdown of the first era illustrates that bad types of constraints will cause problems that eventually impair governance. In particular, constraints that vary with the type of suit brought, rather than the type of government action regulated, are liable to fail under stress.
The breakdown of the second era is a reminder that the connection between the fact of government failure and the causes of government failure are often unclear. As Hegel put it, “the owl of Minerva spreads its wings only with the falling of the dusk.” A system of government powers and constraints is only as strong as its weakest link, and depending on the political moment, a strong link may become the casualty of a weak one.
Sources cited
Emily S. Bremer, The Undemocratic Roots of Agency Rulemaking, 108 Cornell L. Rev. 69 (2022)
Nathaniel Donahue, Officers at Common Law, 135 Yale L. J. (forthcoming 2026)
Richard H. Fallon, Jr., Of Legislative Courts, Administrative Agencies, and Article III, 101 Harv. L. Rev. 915 (1988)
Kathryn E. Kovacs, Avoiding Authoritarianism in the Administrative Procedure Act, 28 Geo. Mason L. Rev. 573, 578 (2021)
Kathryn E. Kovacs, Rules About Rulemaking and the Rise of the Unitary Executive, 70 Admin. L. Rev. 515 (2018)
Jerry L. Mashaw, Creating the Administrative Constitution: The Lost One Hundred Years of Administrative Law (2012)
Alexander Mechanick, The Interpretive Foundations of Arbitrary or Capricious Review, 111 Ky. L.J. 477 (2023)
Paul Sabin, Public Citizens: The Attack on Big Government and the Remaking of American Liberalism (2021)
Reuel E. Schiller, Reining in the Administrative State: World War II and the Decline of Expert Administration, in Total War and the Law; The American Home Front in World War II (Victor Jew & Daniel R. Ernst eds., 2002)
Reuel E. Schiller, The Era of Deference: Courts, Expertise, and the Emergence of New Deal Administrative Law, 106 Mich. L. Rev. 399, 402 (2007)
George B. Shepherd, Fierce Compromise: The Administrative Procedure Act Emerges from New Deal Politics, 90 Nw. U. L. Rev. 1557, 1568 (1996)
United States Attorney General’s Committee on Administrative Procedure, Final Report, S. Doc. No. 8, 77th Cong., 1st Sess. (1941)
