Like police officers patrolling the streets for crime, the front lines for most large business regulators—Environmental Protection Agency engi­neers, Consumer Financial Protection Bureau examiners, and Nuclear Regulatory Commission inspectors, among others—decide when and how to enforce the law. These regulatory monitors guard against toxic air, financial ruin, and deadly explosions. Yet whereas scholars devote considerable attention to police officers in criminal law enforce­ment, they have paid limited attention to the structural role of regu­latory monitors in civil law enforcement. This Article is the first to chronicle the statutory rise of regulatory monitors and to situate them empir­ically at the core of modern administrative power. Since the Civil War, often in response to crises, the largest federal regulators have steadily accrued authority to collect documents remotely and enter pri­vate spaces without any suspicion of wrongdoing. Those exercising this monitoring authority within agencies administer the law at least as much as the groups that are the focus of legal scholarship: enforcement lawyers, administrative law judges, and rule writers. Regulatory moni­tors wield sanctions, influence rulemaking, and create quasi-common law. Moreover, they offer a better fit than lawyers for the modern era of “collaborative governance” and corporate compliance departments because their principal function—information collection—is less adver­sarial. Yet unlike litigation and rulemaking, monitoring-based deci­sions are largely unobservable by the public, often unreviewable by courts, and explicitly excluded by the Administrative Procedure Act. The regulatory-monitor function can thus be more easily ramped up or deconstructed by the President, interest groups, and agency directors. A better understanding of regulatory monitors—and their relationship with regulatory lawyers—is vital to designing democratic accountability not only during times of political transition but as long as they remain a central pillar of the administrative state.


    1. Regulatory Monitors as Distinct Actors
    2. Defining Large Regulators
    3. The Statutory Growth of Monitoring Authority
      1. Original Monitors: The Financial System, Transportation, and Utilities
      2. Gradual Monitors: Health, Safety, and the Environment
      3. Limited Monitors: Consumer Protection, Competition, and Labor
    4. Summary of the Statutory Rise
    1. Governance Changes Favoring Regulatory Monitors
      1. Collaborative Governance
      2. Compliance Departments and Self-Regulation
      3. Heightened Stakeholder Oversight
    2. Market Transformations Favoring Regulatory Monitors
      1. Increased Sophistication
      2. Faster Innovation
      3. Technological Tools
    1. Monitoring Firms
    2. Enforcing Law
      1. Citations, Recommendations, and Warnings
      2. Blocking Business Activity
      3. Public Shaming
      4. The Process as Punishment
      5. Investigations and Charges
    3. Making Law
      1. Creating Common Law
      2. Writing Rules
    1. External Accountability Mechanisms
      1. Public Disclosures
      2. Private Paper Trails
      3. Statutory Minimums
      4. Appointments
    2. Internal Accountability: Lawyers and Monitors as Rivals and Reviewers
      1. Resource Allocation
      2. Appeals
      3. Monitor–Lawyer Teams and Rivalries





Upton  Sinclair’s  1906  novel  The  Jungle  provoked  public  outcry  by  graphically  exposing  health  violations,   such  as  vermin  infestations,   in  the  American  meatpacking industry. 1 See Roger Roots, A Muckraker’s Aftermath: The Jungle of Meat-Packing Regulation After a Century, 27 Wm. Mitchell L. Rev. 2413, 2417–19 (2001). Lawmakers responded by charging the U.S. Department of Agriculture (USDA) with inspecting facilities nation­wide. 2 Meat Inspection Act, Pub. L. No. 59-242, 34 Stat. 1260 (1907) (codified at 21 U.S.C. §§ 601–695 (2012)). After the subprime mortgage crisis helped push the economy to the edge of a cliff in 2008, Congress created a new agency—the Consumer Financial Protection Bureau (CFPB)—with the first federal mandate to routinely examine mortgage servicers and payday lenders. 3 See Dodd–Frank Wall Street Reform and Consumer Protection Act, 12 U.S.C. §§ 5491–5492, 5493(c)(2)(A) (2012). When the Deepwater Horizon oil rig exploded and sank off the Gulf Coast in 2010, arguably the “worst environmental disaster in U.S. history,” 4 David M. Uhlmann, After the Spill Is Gone: The Gulf of Mexico, Environmental Crime, and the Criminal Law, 109 Mich. L. Rev. 1413, 1414 (2011). the Department of the Interior dissolved the responsible agency, created three in its place, and has since doubled the number of offshore energy inspectors. 5 See U.S. Dep’t of the Interior, Secretarial Order No. 3299, Amendment No. 2, Establishment of the Bureau of Ocean Energy Management, the Bureau of Safety and Environmental Enforcement, and the Office of Natural Resources Revenue (2011), [] (reassigning the Minerals Management Service into three new agencies).

These incidents expanded administrative agencies’ authority not only to litigate but also to monitor. 6 n policymakers’ broader responses to such major “regulatory crises,” see generally Policy Shock: Recalibrating Risk and Regulation after Oil Spills, Nuclear Accidents and Financial Crises (Edward J. Balleisen, Lori S. Bennear, Kimberly D. Krawiec & Jonathan B. Wiener eds., 2017). Monitoring authority enables agen­cies to regularly collect nonpublic information from firms without suspi­cion of wrongdoing. Under the Bush and Obama Administrations alone, in addition to the subprime mortgage crisis and Deepwater oil spill, pub­lic backlash prompted monitor-enhancing legislation to keep lead out of children’s toys; 7 See Consumer Product Safety Improvement Act of 2008, Pub. L. No. 110-314, 122 Stat. 3016 (codified in scattered sections of 15 U.S.C.). For a summary of how lead concerns in toys have influenced legislation, see Eileen Flaherty, Note, Safety First: The Consumer Product Safety Improvement Act of 2008, 21 Loy. Consumer L. Rev. 372, 380–84 (2009). prevent salmonella deaths from tainted peanut butter, ice cream, and other packaged foods; 8 See FDA Food Safety Modernization Act, Pub. L. No. 111-353, 124 Stat. 3885 (2011) (codified in scattered sections of 21 U.S.C.). For a discussion of the impact of salmonella deaths leading to the passage of the Act, see Debra M. Strauss, An Analysis of the FDA Food Safety Modernization Act: Protection for Consumers and Boon for Business, 66 Food & Drug L.J. 353, 353–54 (2011). and reduce prescription drug price manipulation. 9 See Medicare Prescription Drug, Improvement, and Modernization Act of 2003, Pub. L. No. 108-173, §§ 1111–1118, 117 Stat. 2066, 2461–64 (codified as amended at 21 U.S.C. § 355 (2012)). Whereas the literature has paid considerable atten­tion to administrative rulemaking and adjudication, it has left the story of the rise of regulatory monitoring largely untold. 10 The literature has provided broad accounts of administrative surveillance aimed at private individuals for other purposes. See, e.g., Daphna Renan, The Fourth Amendment as Administrative Governance, 68 Stan. L. Rev. 1039, 1043 (2016) (describing how the administrative state engages in “sweeping surveillance activity” that must be integrated with the “law and theory of the Fourth Amendment”). It has also covered the tangentially related function of court-ordered monitoring. See, e.g., Veronica Root, The Monitor-“Client” Relationship, 100 Va. L. Rev. 523, 524 (2014) (“Despite its name, a monitor is often not charged with ‘monitoring compliance.’”).

Some agencies describe monitoring as their “backbone” 11 Guy Hayes, A Day in the Life of an Inspector, BSEE,
newsroom/feature-stories/a-day-in-the-life-of-an-inspector [] (last visited Oct. 8, 2018).
or “core,” 12 USDA, One Team, One Purpose 15 (2013) [hereinafter USDA Inspection], []. and some administrative observers recognize that it is a meaningful part of what agencies do. 13 See, e.g., Gary Lawson, Federal Administrative Law 10 (6th ed. 2007) (acknowl­edging that most agency activity lies outside lawyerly roles); Julie E. Cohen, The Regulatory State in the Information Age, 17 Theoretical Inquiries L. 369, 396 (2016) (“[T]he two modalities [of rulemaking and adjudication] are not so much opposites as they are endpoints on a continuum, and . . . a great deal of agency activity occurs in the space between them.”); cf. Eric Biber & J.B. Ruhl, The Permit Power Revisited: The Theory and Practice of Regulatory Permits in the Administrative State, 64 Duke L.J. 133, 142 (2014) (“Topics such as . . . inspections and monitoring . . . deserve more attention than we can give here.”); William H. Simon, The Organizational Premises of Administrative Law, 78 Law & Contemp. Probs., nos. 1 & 2, 2015, at 61, 70 (describing both main adminis­trative law paradigms after World War II as relying on monitoring by agencies). Less obvious is why the responsi­ble bureau­crats—some of whom wear hard hats and goggles to inspect dangerous machin­ery, search for “[b]lack rots, yellow rots, white rots” in food manufac­turing plants, 14 FDA, U.S. Dep’t of Health & Human Servs., No. PB2013-110462, Food Code 410 (2013). or pore through accounting ledgers—merit the kind of sustained legal scholarly attention given to those writing rules and litigating cases.

This Article’s primary goal is to sketch regulatory monitors’ place in the federal regulatory architecture. It examines their statutory rise and workforce size at all nineteen “large” federal regulators. 15 See infra section I.B (defining large regulators and discussing the methodology used to identify them). By drawing on employee manuals, agency annual reports, congressional budget requests, job postings, and interviews, it also begins to piece together the enforcement role that regulatory monitors play and how that role relates to agency functions occupied by lawyers. 16 Publicly available documents were sufficient for understanding most of these agencies’ roles and responsibilities, but to fill in some gaps and to improve accuracy at least one interview was conducted with a current or former employee at each of the agencies or departments studied. Interviews were semistructured, with anonymous interviewees located through chain referral. For a similar interview methodology and review of the literature discussing limitations of such an approach, see, e.g., John Rappaport, How Private Insurers Regulate Public Police, 130 Harv. L. Rev. 1539, 1551 (2017). In short, it situates regulatory monitors at the center of administrative power.

Just as it would be incomplete to analyze criminal law enforcement without distinguishing police officers from prosecutors, this Article shows that a part of administrative law is missing without distinguishing regula­tory monitors from agency enforcement lawyers. To be clear, police offic­ers are unique in terms of state authority by having the discretion to use physical force and immediately take away life or liberty. Also, individuals are arguably more powerless in the face of police officers than businesses are in the face of bureaucrats.

While most regulatory monitors do not wield guns, 17 For a list of federal agencies with full-time staff that do bear arms, see Robert Longley, Firearms and Arrest Authority of U.S. Federal Agencies, ThoughtCo., [] (last updated Feb. 21, 2018) (listing the EPA as having 202 and the FDA as having 183 full-time personnel with firearms). they  stand  between  life  and  death  through  safety  inspections  of  airplanes,   nuclear  facilities,  highway  vehicles,   and  food.  Although  regulatory monitors can­not immediately arrest individuals, they may identify criminal wrongdo­ing, such as embezzlement, that could lead to imprisonment, 18 See, e.g., National Bank Act of 1864, 12 U.S.C. § 38 (2012) (charging bank examiners with identifying embezzlement and stating that deceiving a bank examiner is punishable by imprisonment); Nicholas R. Parrillo, Federal Agency Guidance: An Institutional Perspective 74 (2017), [] (describing how the EPA can often pursue either civil or criminal penalties). and  can  limit  a  business  owner’s  freedom  to  earn  a  livelihood  by  ordering  the  immediate  shutdown  of  oil-drilling  operations  or  food  manufacturing. 19 2015 BSEE Ann. Rep. 23–24 [hereinafter BSEE Annual Report], [] (discussing the Bureau’s enforcement approach, including using shutdowns). They also protect against devastating nonphysical threats by patrolling financial institutions for conduct that could cost families their homes or collapse the economy. Furthermore, regulatory monitors have a forceful informal sanction: the ability to ramp up inspection frequency and intensity, which itself inflicts pain and costs. 20 See infra section III.B.4. With monitoring, as with policing, some­times the process is the punishment. 21 On process punishment in criminal law, see, e.g., Malcolm M. Feeley, The Process Is the Punishment: Handling Cases in a Lower Criminal Court 3–5 (2d ed. 1979).

The analogy to police officers is illustrative because both groups have a patrol function at their core and make frontline law enforcement decisions. But the comparison structurally understates regulatory-moni­tor authority in three main ways. First, police have more constitutional constraints placed on them. Whereas police officers must generally have probable cause or a search warrant to enter a private space, the Supreme Court has held that the Fourth Amendment constrains regulatory searches far less. 22 See City of Los Angeles v. Patel, 135 S. Ct. 2443, 2452 (2015) (“Search regimes where no warrant is ever required may be reasonable where ‘special needs . . . make the warrant and probable-cause requirement impracticable’ and where the ‘primary purpose’ of the searches is ‘[d]istinguishable from the general interest in crime control.’” (first quoting Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 619 (1989); then quoting Indianapolis v. Edmond, 531 U.S. 32, 44 (2000); and then quoting id. at 44)); Marshall v. Barlow’s, Inc., 436 U.S. 307, 313–14, 321 (1978). Unlike police officers, for instance, Environmental Protection Agency (EPA) inspectors can enter private spaces without any suspicion of wrongdoing to make observations or collect samples so long as it is part of a “general neutral administrative plan.” 23 Nat’l-Standard Co. v. Adamkus, 881 F.2d 352, 361–63 (7th Cir. 1989) (holding that EPA inspectors can conduct searches based on administrative warrants, which require either that (1) there is “specific evidence of an existing violation,” necessitating a lesser degree of probable cause than criminal warrants; or that (2) the search is “part of a general neutral administrative plan”).

Second, the power of regulatory monitors in many agencies extends further along the spectrum of enforcement authority. According to one prominent account, “the most significant design flaw in the federal crimi­nal system” is prosecutors’ ability to enforce and adjudicate laws. 24 Rachel E. Barkow, Institutional Design and the Policing of Prosecutors: Lessons from Administrative Law, 61 Stan. L. Rev. 869, 871 (2009). In many agencies, regulatory monitors combine prosecutors’ enforcement and adjudication authority with the patrol function of police officers and investigatory function of detectives: They not only identify wrongdoers but also investigate, reach multimillion-dollar settlements, submit formal charges, and ultimately determine the fate of regulated entities. 25 See infra section III.B.

Third, regulatory monitors may have greater influence on policy­making. Police officers possess expansive authority to arrest people in light of the breadth of potential violations on the books. Those violations are, however, part of a detailed code. 26 To be clear, that code is expansive enough to give police officers tremendous power to arrest people. See William J. Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 577–78 (2001) (describing how, as the scope of criminal law expanded and became codified, “the legislative (and judicial) power have increasingly passed into the hands of law enforcers,” so that “[p]olice and prosecutors can choose whom to target from among the universe of potential offenders”). In the modern era of compliance, some regulatory monitors can go further by requesting internal business changes that advance the agency’s policy goals even if the original behavior was not clearly illegal—such as when a monitor believes a company’s internal pro­cess for reviewing legal complaints is likely to miss future violations. 27 On the pervasiveness of enforced compliance systems, see, e.g., Sean J. Griffith, Corporate Governance in an Era of Compliance, 57 Wm. & Mary L. Rev. 2075, 2124–25 (2016) (“The compliance function, in particular, is designed to inculcate norms of behavior that exceed narrow legal obligations.”); Kimberly D. Krawiec, Organizational Misconduct: Beyond the Principal-Agent Model, 32 Fla. St. U. L. Rev. 571, 572 (2005) (“Courts and agencies typically evaluate the level of care exercised by the organization by inquiring whether the organization had in place ‘internal compliance structures’ ostensibly designed to detect and discourage such conduct.”). In terms of rulemaking, regulatory monitors post their employee manu­als online, which businesses study intently to build com­pliance systems. Those manuals thereby shape industry behavior without any notice-and-comment process. 28 Parrillo, supra note 18, at 27 n.47. Courts have not, however, treated manuals as substantive rules having the force and effect of law in adjudications. See Disabled Am. Veterans v. Sec’y of Veterans Affairs, 859 F.3d 1072, 1078 (Fed. Cir. 2017) (holding that an employee manual was not binding on the agency in adjudications and therefore was not required to go through notice-and-comment procedures nor subject to judicial review); Nat’l Mining Ass’n v. McCarthy, 758 F.3d 243, 251–53 (D.C. Cir. 2014) (holding that a nonbinding guidance document cannot form “the basis for an enforcement action” or “a defense in a proceeding challenging the denial of a permit”). Additionally, post-visit exami­nation and inspection reports have become a meaningful body of com­mon law, used by busi­nesses to make their case in subsequent inspections. 29 See infra section III.C.1.

A key backstory to regulatory monitors’ current status is the advent in recent decades of “new governance” models emphasizing collabora­tive regulation. 30 See Ian Ayres & John Braithwaite, Responsive Regulation: Transcending the Deregulation Debate 4–7 (1992); Jody Freeman, Collaborative Governance in the Administrative State, 45 UCLA L. Rev. 1, 4 (1997) (calling for administrative law to follow a new normative direction in pursuit of “collaborative governance”); Orly Lobel, The Renew Deal: The Fall of Regulation and the Rise of Governance in Contemporary Legal Thought, 89 Minn. L. Rev. 342, 350–51, 371–76 (2004) (outlining a shift in the administrative state away from central control to a more partnership-driven model of governance focused on collaboration between agencies and various stakeholders). As this Article argues below, 31 See infra section II.A.1. the emphasis on collabo­ra­tive regulation syncs better with inspectors and examiners—who “work alongside, not against[ ] industry” 32 See Hayes, supra note 11. —than with litigators, whose main powers rest on adversarial court proceedings. Current governance models also emphasize “continuous” information flows so that rules respond rapidly to firms’ conduct, 33 See Freeman, supra note 30, at 22, 28–29 (“Monitoring and information exchange are crucial to an effective implementation and compliance regime . . . .”). inducing greater reliance on regulatory moni­tors’ real-time data. Moreover, as courts, Congress, and the President have increasingly constrained agency rule writing and litiga­tion, 34 See infra section II.A.3. agencies would be expected to rely more on less-constrained moni­toring activities to exercise authority.

By situating regulatory monitors at the center of administrative power, this Article places them at the intersection of leading administra­tive law conversations. One strand of scholarship has stressed the impor­tance of the structural design of public institutions in incentivizing optimal acqui­sition of information—the “lifeblood of effective governance.” 35 Matthew C. Stephenson, Information Acquisition and Institutional Design, 124 Harv. L. Rev. 1422, 1423 (2011). A major reason Congress created agencies was to undertake “specialized information-gathering” ill-suited for courts. 36 See Richard B. Stewart & Cass R. Sunstein, Public Programs and Private Rights, 95 Harv. L. Rev. 1193, 1273 n.338 (1982). This literature has also ana­lyzed agencies’ external strategies for acquiring information—but focus­ing on agencies as unitary entities rather than looking at internal groups. 37 See, e.g., Cary Coglianese, Richard Zeckhauser & Edward Parson, Seeking Truth for Power: Informational Strategy and Regulatory Policymaking, 89 Minn. L. Rev. 277, 279, 281–85 (2004) (“In this Article, we analyze regulators’ gathering of information from firms as a strategic game.”). Professors Coglianese, Zeckhauser, and Parson mention regulatory monitors in passing, but they examine a broader set of information-collection mechanisms (like phone conversations with industry experts) for a wider array of purposes (such as one-time rulemaking studies). See id. at 288–89, 305, 319–24.

Another related strand of scholarship argues that standard depic­tions of administrative law are incomplete because “agencies are typically treated as unitary entities.” 38 Elizabeth Magill & Adrian Vermeule, Allocating Power Within Agencies, 120 Yale L.J. 1032, 1035 (2011). Congress and agency leaders allocate clout among various subagency offices, divisions, and decisionmakers. 39 See id. at 1035–36 (offering a descriptive model of agencies that draws attention to how power is distributed between various offices and officials within an agency). Acknowl­edging these internal allocations improves understanding of “the most puzzling principles and doctrines of administrative law.” 40 Id. at 1035. Early studies provided rich insights into agency organizational design, includ­ing the role of inspectors, 41 See, e.g., Eugene Bardach & Robert A. Kagan, Going by the Book: The Problem of Regulatory Unreasonableness 73 (1982) (discussing how agencies and inspectors have configured their operations to meet legislative demands for rule enforcement); John Braithwaite et al., An Enforcement Taxonomy of Regulatory Agencies, 9 Law & Pol’y 323, 324 (1987) (“Deterrence or sanctioning strategies seek to identify and detect breaches of law through patrol and inspection; they then seek to develop a case for the courts through investigation.”); see also Colin S. Diver, A Theory of Regulatory Enforcement, 28 Pub. Pol’y 257, 258 (1980) (discussing inspectors from a theoretical perspective). This Article draws on those early studies. However, that literature focuses on (a) mostly inspectors; (b) a different set of agencies, including state and local agencies and typically excluding those that regulate trade or finance; and (c) agencies’ overall regulatory approach rather than on regulatory monitors. See, e.g., Bardach & Kagan, supra, at 7 (“The focus of this book is on the social dimension of unreasonableness: the experience of being subjected to inefficient regulatory requirements.”). The literature thus lacks any systematic study of regulatory monitors as a distinct group across the largest federal agencies, leaving open the question of regulatory monitors’ origins and power in the modern administrative state. “but the bulk of this work was done decades ago, largely in the context of administrative adjudication.” 42 Jennifer Nou, Intra-Agency Coordination, 129 Harv. L. Rev. 421, 428 (2015). Since then, agencies’ regulatory approaches have shifted significantly, and adjudica­tion has declined. 43 See, e.g., id. For an overview of the governance and market transformations behind this shift, see infra Part II. Consequently, scholars have recently revived the pro­ject of “crack[ing] open the black box of agencies to peer inside” 44 Magill & Vermeule, supra note 38, at 1035. the organizational structure of both rulemaking 45 See, e.g., Nou, supra note 42, at 422–25 (examining the internal divisions within agencies and how agency leaders deploy these divisions to advance the agency’s objectives). Professor Nou does not mention regulatory monitors and instead focuses on organizational mechanisms that give agency leaders control over information vital for decisionmaking, especially related to rulemaking. See id. at 429–31. and enforcement. 46 See, e.g., Rachel E. Barkow, Overseeing Agency Enforcement, 84 Geo. Wash. L. Rev. 1129, 1130–31 (2016) (acknowledging that “[d]espite the centrality of enforcement to agency practice, enforcement discretion receives relatively little attention,” and “begin[ning] to catalog approaches for overseeing [enforcement]”); Margaret H. Lemos, Democratic Enforcement? Accountability and Independence for the Litigation State, 102 Cornell L. Rev. 929, 942–43 (2017) (“[E]nforcement has inspired far less attention than rulemaking or adjudication. . . . This Article seeks to fill that gap.”). Others have looked more broadly at how to improve frontline decisionmakers, a category that includes inspectors and administrative law judges. 47 See, e.g., Daniel E. Ho, Does Peer Review Work? An Experiment of Experimentalism, 69 Stan. L. Rev. 1 (2017). Professor Ho underscores regulatory monitors’ importance by closely studying inspectors and emphasizing the “extensive discretion” of “frontline government officials carry[ing] out the law.” Id. at 5. His focus is on a broader function—frontline decisionmaking, which is exercised by other groups such as lawyers and judges—and a broader set of agencies, including local agencies that exercise adjudicatory power over individuals. See id. at 5–10. Nonetheless, his work produces significant empirical and policy insights into regulatory monitors. See id. at 11–13. For earlier valuable empirical studies of inspectors, see, for example, Bardach & Kagan, supra note 41; Braithwaite et al., supra note 41.

Despite the lack of sustained attention to regulatory monitors or artic­ulation of their distinct role in the modern administrative state, 48 When broad administrative law conversations mention monitoring, it is often of agencies, not firms. See Nou, supra note 42, at 423 (noting “administrative law’s overwhelming focus on the influence of agencies’ external monitors”). these strands of literature indirectly lay the foundation for understand­ing how regulatory monitors are crucial to administrative law. For most agencies, regulatory monitors are an organizationally distinct group at the heart of the policymaking and enforcement black boxes. 49 See infra section I.A. They are the gatekeepers for information, and thus for the “lifeblood” of agencies. 50 See infra section I.B, Part III.

As such, regulatory monitors are relevant to administrative law’s cen­tral preoccupations. The overriding purpose of administrative law is the accountability of delegated authority. The 1946 Administrative Procedure Act (APA) enables courts and the public to check agencies. 51 It does so by, for example, involving the public in notice-and-comment rulemaking. See 5 U.S.C. § 553 (2012). It also specifies judicial review of final agency action. See id. § 702. Yet regula­tory monitors operate in the “soft” administrative law 52 Adam J. Levitin, The Politics of Financial Regulation and the Regulation of Financial Politics: A Review Essay, 127 Harv. L. Rev. 1991, 2043–44 (2014) (noting that prudential regulators mostly operate using “soft law” rather than formal law such as notice-and-comment rulemaking). space largely exempted from the APA’s accountability mechanisms. 53 See 5 U.S.C. § 554(a)(3) (excepting “proceedings in which decisions rest solely on inspections” from the notice-and-comment process). Since regulatory monitors’ actions are less reviewable than those of more formal legal actors and the technical pro­cess of collecting information remains out of sight between crises, the rise of regulatory monitors potentially insulates agencies from public accountability.

Finally, scholars have debated how the law should address external stakeholders competing for influence over agencies. The literature identi­fies mechanisms, such as cost–benefit analysis, that alter the Presi­dent’s ability to control a defiant bureaucracy. 54 See, e.g., Martin S. Flaherty, The Most Dangerous Branch, 105 Yale L.J. 1725, 1819–21 (1996) (discussing the nondelegation doctrine); Abner S. Greene, Checks and Balances in an Era of Presidential Lawmaking, 61 U. Chi. L. Rev. 123, 176–79 (1994) (summarizing the checks and balances on presidential power over the administrative state); Elena Kagan, Presidential Administration, 114 Harv. L. Rev. 2245, 2253–72 (2001) (providing an overview of the ways agencies are constrained); Michael A. Livermore, Cost-Benefit Analysis and Agency Independence, 81 U. Chi. L. Rev. 609, 614–15 (2014) (describing the way cost–benefit analysis constrains agencies); Kevin Stack, The President’s Statutory Powers to Administer the Laws, 106 Colum. L. Rev. 263, 267 (2006) (arguing that the President does not have the authority to act directly under a statute or bind the discretion of lower-level officials unless Congress directly grants such authority, in contrast to the operating assumption). It also explores organiza­tional design features that insulate agencies from industry capture. 55 See Rachel E. Barkow, Insulating Agencies: Avoiding Capture Through Institutional Design, 89 Tex. L. Rev. 15, 17–18 (2010) (arguing that the analysis of an agency’s independence should shift from the traditional focus on insulation from the presidency to instead consider design features that prevent capture by interest groups). Regu­latory monitors add another dimension to these discussions. For instance, in 1961, about a month into a new job as a frontline Food and Drug Administration (FDA) examiner, Dr. Frances Kelsey received what her supervisors described as routine papers submitted for a new sleep aid used off-label for morning sickness. 56 Bara Fintel et al., The Thalidomide Tragedy: Lessons for Drug Safety and Regulation, Helix (July 28, 2009), []; see also Frances Oldham Kelsey, Autobiographical Reflections 44, 49–67 (unpublished manuscript), [] (last visited Nov. 3, 2018) (chronicling the start of Dr. Kelsey’s thalidomide assignment at the FDA through the drug company’s withdrawal of the FDA application, as recalled by Dr. Kelsey). Despite intense pressure from the drug’s manufacturer, she withheld approval by repeatedly demanding more rigorous clinical evidence than the FDA typically required. 57 See S. Rep. No. 87-1744, at 40–42 (1962) (detailing over forty-six contacts by the drug’s manufacturer attempting to “expedite clearance,” including one with Dr. Kelsey’s immediate supervisor calling her letter “somewhat libelous” and requesting that pressure be applied to her). It was ultimately discovered that in Germany alone the drug, thalidomide, had caused an estimated 10,000 incidents of deaths or shrunken or missing limbs in babies born to mothers who had taken the drug. 58 See Frederick Dove, What’s Happened to Thalidomide Babies?, BBC (Nov. 3, 2011), [] (“No-one knows how many miscarriages the drug caused, but it’s estimated that, in Germany alone, 10,000 babies were born affected by Thalidomide. Many were too damaged to survive for long.”). Mass harm was averted in the United States because a frontline examiner stood firm in exercising her agency’s statutory power. 59 See infra section I.C.2.

As powerful actors, regulatory monitors have in recent decades served as an important lever for any presidential ramp-up or drop-off in regulation. 60 See Holly Doremus, Data Gaps in Natural Resource Management: Sniffing for Leaks Along the Information Pipeline, 83 Ind. L.J. 407, 427–29 (2008) (identifying “information extraction” programs as early cuts during environmental deregulation); OMB Watch, The Obama Approach to Public Protection: Enforcement 4 (2010), [] (citing an increase in regulatory-monitor activity under President Obama). Most recently, as part of a planned “deconstruction of the administrative state,” 61 Gillian E. Metzger, The Supreme Court, 2016 Term—Foreword: 1930s Redux: The Administrative State Under Siege, 131 Harv. L. Rev. 1, 2 (2017) (quoting Steve Bannon’s statement as reported in Phillip Rucker & Robert Costa, Bannon Vows a Daily Fight for “Deconstruction of the Administrative State,” Wash. Post (Feb. 23, 2017), https:// []).
President Trump has taken steps to make the FDA drug-approval process “much faster,” 62 See, e.g., David Crow, Pharma Stocks Rally on Trump Pledge to Speed Drug Approvals, Fin. Times (Jan. 31, 2017), (on file with the Columbia Law Review) (internal quotation marks omitted) (quoting President Trump). and his appointees have moved to decrease federal inspections of polluting factories, examinations of banks, and monitoring of offshore oil platforms. 63 See Eric Lipton & Danielle Ivory, Under Trump, E.P.A. Has Slowed Actions Against Polluters, and Put Limits on Enforcement Officers, N.Y. Times (Dec. 10, 2017), (on file with the Columbia Law Review) (citing an EPA deputy administrator as stating that the agency “would back off some inspection” activity); Ted Mann, Regulators Propose Rollbacks to Offshore Drilling Safety Measures, Wall St. J. (Dec. 25, 2017), (on file with the Columbia Law Review). The ease with which such changes can be made varies by agency. At the FDA today, legal struc­tures constrain external influences far more than in the 1950s. Following the thalidomide incident, Congress codified the type of heightened report­ing requirements that Dr. Kelsey had sought. 64 Sam Peltzman, An Evaluation of Consumer Protection Legislation: The 1962 Drug Amendments, 81 J. Pol. Econ. 1049, 1049–52 (1973) (describing the statutory amendments passed in 1962 to strengthen FDA reporting requirements following congressional hearings related to the thalidomide incident). Streamlining the drug-approval process would now largely depend on changes to the law rather than convincing a frontline examiner. By contrast, in other agen­cies, legal rules and organizational structure leave regulatory moni­tors’ decisionmaking pro­cesses more susceptible to alteration without public knowledge. 65 See infra section IV.A.

The  analysis  below  maps  out  this  underappreciated  administrative  law  of  monitoring. 66 Administrative law here is meant in its broader sense, comprising not only judicial review but also “statutes, executive orders, and other legal instruments that structure the agencies and the procedures they use.” Magill & Vermeule, supra note 38, at 1056. It also adds to the toolbox of familiar accountability mechanisms by highlighting how the design of teams with both lawyers and monitors enables each group to check the other’s weaknesses. Given that monitoring occupies a central role in agency activity, an understand­ing of regulatory monitors and their surrounding legal framework is vital to improving the institutional design of agencies and to making adminis­trative law more administrative. 67 Cf. Edward Rubin, It’s Time to Make the Administrative Procedure Act Administrative, 89 Cornell L. Rev. 95, 97 (2003) (criticizing the APA for imposing an “essentially judicial concept of governance” that subjects agencies to “inappropriate procedural rigidities” instead of accommodating “new modes of governance” like priority setting and resource allocation).

The discussion is structured as follows. Part I provides an overview of regulatory monitors by defining their distinct place in agencies and surveying their statutory emergence. Part II articulates the changes in governance and markets that have organizationally favored regulatory monitors more than rule writers and litigators. Part III begins to map out major organizational design choices. It provides the first quantitative and qualitative evidence indicating regulatory monitors’ presence and influ­ence across the largest independent and cabinet-level regulators. Part IV considers how future agency architects might improve the regulatory-monitor framework for more optimal governance. Designers could improve many agencies through transparency, mandated minimum num­bers of inspections, appeals, appointments, and intra-agency coordina­tion among lawyers and regulatory monitors. Above all, whether the goal is to guard against abuse of agency authority or business capture of bureau­crats, administrative law could benefit from viewing regulatory monitors as what they have become: dominant state actors shaping the well-being of firms and citizens.

I. The Statutory Rise

Unlike other actors in the typical administrative narrative, such as the rule writer and enforcement lawyer, regulatory monitors have a less-well-documented core power. Accordingly, this Part begins by providing a definition and then offers a brief historical overview of the accumula­tion of statutory monitoring authority by large regulators.

A. Regulatory Monitors as Distinct Actors

This Article defines a regulatory monitor as an agency actor whose core power is to regularly obtain nonpublic information from businesses outside the legal investigatory process. Monitoring can be broken down into two main types: visitation and reporting. Visitation authority allows regulators to physically enter private business spaces to observe or collect infor­mation. Reporting requires firms to remotely transmit informa­tion—such as business records—that is then received by regulatory moni­tors within the agency. 68 These two categories are distinct from agencies monitoring publicly available data.

This seemingly straightforward authority does not easily fit into com­mon descriptions of the administrative state. Legal treatments of admin­istrative agencies typically break down their activities into rulemaking and enforcement, or sometimes into ex ante rulemaking and ex post enforcement. 69 See, e.g., James C. Cooper, The Costs of Regulatory Redundancy: Consumer Protection Oversight of Online Travel Agents and the Advantages of Sole FTC Jurisdiction, 17 N.C. J.L. & Tech. 179, 204 (2015) (describing ex ante rulemaking and ex post enforcement as “two tools in [agencies’] arsenals to enforce their statutory mandate”). Regulatory monitors arguably act ex ante because they aim to “secure compliance before violations occur.” 70 Heidi Mandanis Schooner, Consuming Debt: Structuring the Federal Response to Abuses in Consumer Credit, 18 Loy. Consumer L. Rev. 43, 49 (2005). But securing com­pli­ance from a particular regulated entity is very different from writing rules of general applicability, so categorizing monitoring as “ex ante” is a poor fit.

That leaves ex post enforcement as a more natural place for mon­itoring in the standard ex ante–ex post dichotomy. But as the Supreme Court explained, “Our cases have always understood ‘visitation’ as this right to oversee corporate affairs, quite separate from the power to enforce the law.” 71 Cuomo v. Clearing House Ass’n, L.L.C., 557 U.S. 519, 526 (2009). When the CFPB initially sent enforcement lawyers along on its regular on-site visits, called bank exams, the practice was met with “relentless opposition from bankers.” 72 Rachel Witkowski, CFPB Pulls Enforcement Attorneys from Its Exams, Am. Banker (Oct. 9, 2013), (on file with the Columbia Law Review). The agency ultimately ended the practice, with one former CFPB official explaining, “The bureau learned that the nature and logistics of the two jobs are very different . . . .” 73 See id. (quoting Ronald L. Rubin).

The U.S. Office of Personnel Management (OPM) also recognizes regu­latory monitors’ distinct role. It classifies attorneys in the “Legal and Kindred” category, but lists the most common titles used for regulatory monitors elsewhere: Inspectors, Auditors, and Examiners. 74 OPM, Employment Cubes, FedScope, (on file with the Columbia Law Review) [hereinafter FedScope] (last visited Apr. 14, 2017). Legal schol­ars’ frequent omission of regulatory monitors reflects the common view that this group is doing something apart from “Legal and Kindred” actors. 75 See infra section II.B.

Despite the confusion, it is important to recognize that internal agency groups can be distinguished by their core legal powers. Litigators hold the keys to the courts. Rule writers author text enacted as law. Regula­tory monitors peer inside firms.

B. Defining Large Regulators

While examples throughout the Article involve a variety of regula­tors, to manage the scope of the empirical analysis and investigation of statutory history this Article focuses on “large” regulators of business. The OPM defines an agency as “large” if it has more than 1,000 employ­ees. 76 OPM, Data Definitions, FedScope, [] (last visited Feb. 2, 2019). To identify the set of all large regulators within this group, I located every agency in the OPM’s database with over 1,000 employees and a mission focused on regulating businesses. 77 The one exception is the CFPB, which the OPM treats as a component of the Federal Reserve’s division of supervision and regulation, perhaps because the CFPB receives its funding from the Federal Reserve. But the Federal Reserve’s other functions are not listed. Thus, this Article treats the CFPB as an independent agency, and the Federal Reserve’s annual report was used to obtain personnel figures for its regulatory arm, which performs a similar bank-oversight function as the OCC and FDIC. See infra note 483 and accompanying text. This included both “Cabinet-Level” agencies and “Large Independent Agencies.” 78 OPM, supra note 76. The nineteen agen­­cies fitting this description were the CFPB, Federal Energy Regula­tory Commission (FERC), FDA, Food Safety and Inspection Service (FSIS), Mine Safety and Health Administration (MSHA), Occupational Safety and Health Administration (OSHA), Federal Aviation Administration (FAA), Federal Motor Carrier Safety Administration (FMCSA), Office of the Comptroller of the Currency (OCC), EPA, Equal Employ­ment Oppor­tunity Commission (EEOC), Federal Communications Com­mis­sion (FCC), Federal Deposit Insurance Corporation (FDIC), Federal Reserve, Federal Trade Commission (FTC), National Credit Union Admin­istration (NCUA), National Labor Relations Board (NLRB), Nuclear Regulatory Commission (NRC), and Securities and Exchange Commission (SEC).

Large regulators were chosen as the category, rather than medium or small regulators, under the assumption that any given large regulator is more likely to have a greater influence on the business world than any given small or medium regulator due to resource allocation. That focus, however, inevitably leaves out important regulators. Surely some medium and smaller agencies have considerable influence and, by some metrics, may be more influential than some large agencies. Also, significant monitoring of businesses happens at the state level. 79 See, e.g., Daniel E. Ho, Fudging the Nudge: Information Disclosure and Restaurant Grading, 122 Yale L.J. 574, 584–85 (2012) (listing states and localities that regulate and monitor restaurants through health inspections and cleanliness grading).

To differentiate business regulators from other agencies, a narrow def­inition was applied: The agency must focus on enforcing laws against businesses. Agencies focused on overseeing substantial personal activi­ties were eliminated. Thus, the Internal Revenue Service (IRS) was elimi­nated under this criterion because a substantial part of what it does is oversee individuals’ tax returns—even though the IRS also oversees revenue collection from businesses. 80 See SOI Tax Stats—Tax Stats at a Glance, IRS, [] (last updated Sept. 27, 2018). Much of this Article’s analysis would apply to agencies that collect information from individuals. But collection of information from individuals carries different implications for privacy, and it is less relevant to some of the discussions below about market transformations and compliance departments. 81 See infra Part II.

Agencies were also omitted if they did not enforce laws against busi­nesses but instead focused on some other activity. The U.S. Patent and Trademark Office (USPTO), for instance, is focused on “granting U.S. patents and registering trademarks.” 82 About Us, USPTO, [] (last visited Oct. 11, 2018). The USPTO leaves it to the patent and trademark holders, however, to enforce their intellectual prop­erty rights in court. 83 See, e.g., General Information Concerning Patents, USPTO, [
HGEM] (last visited Oct. 11, 2018) (“Once a patent is issued, the patentee must enforce the patent without aid of the USPTO.”).

There is no universally accepted definition of “business regulator,” and by other defensible definitions of the term, the USPTO and IRS could have been included. It is worth noting that the USPTO and IRS would, if included, presumably strengthen at least parts of this Article’s central thesis, since those agencies rely heavily on employees who regu­larly collect information. But it becomes less clear how to think about the role of lawyers in an agency that does not have a strong law enforcement role.

Large agencies may not be representative of agencies as a whole. It is possible that smaller agencies are inherently more likely to rely on enforcement lawyers than monitors, for instance, due to their limited resources. Further study would be needed to determine whether that is the case, although at least some excluded medium and small busi­ness regulators, such as offshore oil regulators, also rely heavily on monitor­ing. 84 See infra notes 169–171 and accompanying text (discussing monitoring outside the context of large agencies). Additionally, large independent agencies collectively com­prise 93% of all independent agency employees listed in the OPM data­base, meaning that they presumably reflect a substantial portion of the regulatory force. 85 See FedScope, supra note 74 (noting that large independent agencies have 160,524 total employees, medium independent agencies have 11,230, and small independent agencies have 1,440).

C. The Statutory Growth of Monitoring Authority

The modern monitoring framework is the product of numerous ad hoc statutes that give different agencies various levels of visitation and reporting powers. Today’s large business regulators can be historically classified into one of three categories: those that had strong monitoring authority from the outset, those that gradually accumulated monitoring authority, and those that have limited monitoring power today.

1. Original Monitors: The Financial System, Transportation, and Utilities. — Although historical treatments of the administrative state sometimes begin with federal control of the railroads in the 1880s, 86 See, e.g., Jed Handelsman Shugerman, The Dependent Origins of Independent Agencies: The Interstate Commerce Commission, the Tenure of Office Act, and the Rise of Modern Campaign Finance, 31 J.L. & Pol. 139, 143–45 (2015) (focusing on two events for their role in reshaping the executive branch, including the establishment of the Interstate Commerce Commission in the 1880s); Richard B. Stewart, The Reformation of American Administrative Law, 88 Harv. L. Rev. 1667, 1671 (1975) (beginning an “inquiry into the traditional model of American administrative law” by mentioning the regulation of railroads in the latter part of the nineteenth century). the first of today’s large business regulators was born during the Civil War, at a time when states implemented most inspection regimes. 87 See William J. Novak, The People’s Welfare: Law and Regulation in Nineteenth-Century America 88–89, 205–06 (1996) (describing state laws and regulations that implemented inspection regimes before the 1880s); Ross M. Robertson, The Comptroller and Bank Supervision 25–26 (1995) (describing state examination of banks prior to the Civil War). Monitoring has long been fundamental to federal administration. See Jennifer L. Mascott, Who Are “Officers of the United States”?, 70 Stan. L. Rev. 443, 522–23 (2014) (noting inspections of cargo ships since the nation’s founding); see also Robert L. Rabin, Federal Regulation in Historical Perspective, 38 Stan. L. Rev. 1189, 1197 (1986) (concluding that businesses began to be inspected regularly starting in the 1880s). In 1864, recognizing that a successful military campaign required a stable financial system, President Lincoln declared that a “national system . . . will create a reliable and permanent influence in support of national credit and protect the people against losses in the use of paper money.” 88 Lincoln and the Founding of the National Banking System, OCC, [] (last visited Oct. 11, 2018) (internal quotation marks omitted) (quoting President Lincoln’s 1864 State of the Union address). Later that year, he signed the National Bank Act, creating the OCC. 89 National Bank Act of 1864, ch. 106, § 1, 13 Stat. 99 (codified in scattered sections of 12 U.S.C.). The OCC’s mission included certifying compli­ance with federal banking laws, which sought to ensure a bank did not fail and thereby spark bank runs that could collapse the economy. 90 Eugene N. White, Lessons from the History of Bank Examination and Supervision in the United States, 1863–2008, in Financial Market Regulation in the Wake of Financial Crises 15, 21–22 (Alfredo Gigliobianco & Gianni Toniolo eds., 2009) (describing the creation and role of the OCC).

In pursuing these goals, the OCC’s main tool was monitoring. It could not litigate. Although the agency could write rules, 91 See National Bank Act of 1864 §§ 22, 24, 45, 47 (granting authority to the Secretary of the Treasury to prescribe certain regulations); see also 12 U.S.C. § 211 (providing the modern authority for the Comptroller to promulgate regulations). it rarely used that authority. 92 See White, supra note 90, at 21. Its chief sanction was revoking a bank’s national charter, 93 National Bank Act of 1864 § 53 (codified as amended at 12 U.S.C. § 93). Such decisions triggered formal procedures, such as appeals and hearings. See id. a seldom-used option given the OCC’s need to prevent bank closings. 94 See Eugene N. White, Lessons from American Bank Supervision from the Nineteenth Century to the Great Depression, in 17 Macroprudential Regulatory Policies: The New Road to Financial Stability? 41, 48 (Stijn Claessens et al. eds., 2012). OCC examiners still had the effect, when they appeared unan­nounced, of “terrorizing” lower-level bank cashiers. 95 See John D. Hawke, Jr., Comptroller of the Currency, Remarks Before a Conference on Credit Rating and Scoring Models 4 (May 17, 2004), https:// [] (“Sometimes it seemed as though terrorizing bankers was almost a requirement of the examiner’s job.”).
But as a statutory matter, the agency was built more to monitor than to litigate.

Initially, the OCC focused on reviewing quarterly bank reports and monthly statements. 96 See National Bank Act of 1864 § 34. It soon became clear that this enabled bankers to “window dress[ ]” reports. 97 See White, supra note 90, at 21. Congress responded by requiring a mini­mum of two surprise annual examinations of each national bank. 98 See id. The OCC already had the ability to conduct examinations in its originating statute. 99 National Bank Act of 1864 § 54. Former bank teller O. Henry depicted such an examination in one of his short stories, writing that an OCC examiner “[o]ne day . . . inserted an official-looking card between the bars of the cashier’s win­dow . . . [and] [f]ive minutes later the bank force was dancing at the beck and call of a national bank examiner.” 100 O. Henry, A Call Loan, in Heart of the West 240, 241 (1904); see also Hawke, supra note 95 (confirming O. Henry’s accounts of OCC bank examiners). Examiners had the author­ity to enter any room, open any drawer, and look at any document. 101 White, supra note 90, at 21.

Although the basic examination tool remained largely unchanged until recently, 102 See Peter Conti-Brown, The Power and Independence of the Federal Reserve 165 (2016). Minor changes were made, such as expanding the scope of what regulators could examine to include potential future earnings, management quality, and the local community’s needs. See Banking Act of 1935, Pub. L. No. 74-305, 49 Stat. 684 (codified as amended in scattered sections of 12 U.S.C.). the institutional and legal framework has swelled steadily. The 1907 financial panic led Congress to create the Federal Reserve, 103 See White, supra note 90, at 22. which — like  the  OCC — could  conduct  examinations  of  national  banks  and  of  state  banks  that  chose  to  become  “members.” 104 See Federal Reserve Act of 1913, Pub. L. No. 63-43, § 11(a), 38 Stat. 251, 261–62 (codified at 12 U.S.C. § 248 (2012)). After depositor panics sparked bank runs that nearly collapsed the banking system and the stock market crashed in the 1920s, more agencies were added, including the FDIC to insure bank deposits 105 Banking Act of 1933, Pub. L. No. 73-66, 48 Stat. 162 (codified in scattered sections of 12 U.S.C.). To become insured, banks had to accept federal examinations. Id. § 5. At first, the FDIC required approval from other banking regulators to conduct examinations, but in 1950 it received broader discretion to examine its member banks. White, supra note 90, at 26. While only some state banks had joined the Federal Reserve, “virtually all banks” signed up for FDIC oversight, thereby greatly expanding monitoring’s reach. Id. and the SEC “to protect . . . the national banking system” and investors. 106 Securities Act of 1934, Pub. L. No. 73-291, 48 Stat. 881, 881–82 (codified as amended at 15 U.S.C. §§ 77b–77s, 77ii–77jj, 78a–78qq (2012)). The SEC had visitation comparable to that of banking regulators, but over securities exchanges, credit rating organizations, and securities brokers and dealers. The SEC could require “reasonable periodic, special, or other examinations” of “accounts, correspondence, memoranda, papers, books, and other records . . . at any time.” Id. § 13(h)(4) (codified at 15 U.S.C. § 78m). Credit unions were also subject to federal examination. Federal Credit Union Act, Pub L. No. 86-354, 48 Stat. 1216 (1934) (codified in scattered sections of 12 U.S.C.). Authority was assumed in 1970 by the NCUA. See ​​​​A Brief History of Credit Unio​ns, NCUA, [] (last visited Oct. 11, 2018).

This early visitorial authority can also be seen in the infrastructure ser­vices industries of transportation, energy, and telecommunications agencies. The largest modern transportation agency, the FAA, built an early model for its contemporary safety program in 1932. 107 The FAA describes this program today as its “little-seen but still important . . . flight inspection program.” Scott Thompson, Flight Inspection History, FAA, https:// [] (last updated Aug. 6, 2014). Decades before airplanes even came into existence, Congress laid a foundation for the tradition of federal vehicle inspections when it authorized federal regulators to conduct inspections of steamboats. John G. Burke, Bursting Boilers and the Federal Power, 7 Tech. & Culture 1, 15 (1966) (discussing the law authorizing the appointment of steamboat-boiler inspectors in 1838).
The country was divided into six “[l]ighthouse dis­trict areas,” within which a single “patrol pilot[ ]” would fly around, able to enter any airplane, open any airport door, or review any flight-related document. 108 Scott A. Thompson, Flight Check!: The Story of FAA Flight Inspection 21 (1993) (describing the origins of modern flight inspection programs). The modern FAA grew out of the Aeronautics Branch of the Department of Commerce. Id. That predecessor’s authority originated in the Air Commerce Act of 1926. See Air Commerce Act of 1926, Pub. L. No. 69-254, 44 Stat. 568 (repealed 1938, 1958). That Act gave the FAA’s predecessor the power to conduct “periodic examination[s] of aircraft[,] . . . airmen serving in connection with aircraft of the United States as to their qualifications[,] . . . [and] facilities.” Id. § 3(b)–(d). The first airworthiness inspection of an American airplane occurred within the year. See FAA Historical Chronology, 1926-1996, FAA, [] (last visited Oct. 11, 2018). Like bank examiners, patrol pilots could sanction by recom­mending the “suspen­sion and revocation” of licenses. 109 See Air Commerce Act § (3)(f). Similarly extensive visitation can be found in the origins of today’s largest agencies overseeing energy and telecommunications: the FERC 110 The predecessor of today’s largest energy regulator, FERC, was established in 1920 and began overseeing hydroelectric facilities. See FERC Timeline, FERC, [] (last visited Oct. 10, 2018). The Commission’s originating statute listed, as the first of its general powers, the authority “to collect and record data concerning . . . the water-power industry.” Federal Power Act of 1920, Pub. L. No. 66-280, § 4, 41 Stat. 1063, 1065 (codified at 16 U.S.C. §§ 791–828c (2012)). When Congress expanded the Commission’s authority in 1935 to include electricity, it also more explicitly authorized inspections of energy facilities. See Richard A. Rosan, On the Fiftieth Anniversary of the Federal Energy Bar Association, 17 Energy L.J. 1, 25 (1996). and FCC. 111 The FCC’s 1934 originating statute grants authority to “inspect all transmitting apparatus.” Communications Act of 1934, Pub. L. No. 73-416, § 303(n), 48 Stat. 1064, 1083 (codified as amended in scattered sections of 47 U.S.C.). The FCC assumed the Federal Radio Commission’s responsibilities and personnel. See id. § 603(a). For common carriers, such as telephone companies, the Act provides that “[t]he Commission shall examine into transactions entered into by any common carrier” and “shall have access to and the right of inspection and examination of all accounts, records, and memoranda, including all documents, papers, and correspondence now or hereafter existing.” Id. § 215(a). This includes the submission of reports and inquiries into management. Id. § 218.

As these financial, transportation, telecommunications, and energy industries have evolved, monitoring statutes have mostly kept pace. Congress updated monitoring to reach new financial organizations (such as hedge funds), new products (such as credit cards), and even a shadow banking system that had by some measures become larger than the tradi­tional banking system. 112 For instance, some banks reorganized themselves by forming bank holding companies and thereby shielding new lines of business from examinations. White, supra note 90, at 27–28. Congress responded by extending Federal Reserve examinations to cover bank holding companies and subsidiaries. Id. (referring to the Bank Holding Company Act Amendments of 1970, Pub. L. No. 91-607, 84 Stat. 1760 (codified as amended at 12 U.S.C. § 1841 (2012)) and Bank Holding Company Act of 1956, Pub. L. No. 84-511, § 5(c), 70 Stat. 133, 137 (codified as amended at 12 U.S.C. §§ 1841–52)). Within the past few years, financial regulators also gained examination authority over hedge funds. See Dodd–Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111-203, § 404, 124 Stat. 1376 (2010) (codified as amended in scattered sections of 12 and 15 U.S.C.). As banks began to offer more products, such as credit cards, Congress enacted more laws, such as the 1968 Truth in Lending Act, thus widening the scope of examination. See Truth in Lending Act, 15 U.S.C. §§ 1601–1667(f) (2012). Banking crises between the 1980s and 2000s forced more comprehensive disclosures in regulatory reports. See White, supra note 90, at 34. Even third-party service providers that banks use—such as Amazon, IBM, Google, or other technology firms—have come under monitoring authority. See 12 U.S.C. §§ 1464(d)(7), 1867(c)(1). The CFPB has gained visitorial authority over most of the shadow banking system. Id. §§ 5321, 5322(a)(2); see also Steven L. Schwarcz, Regulating Shadow Banking, 31 Rev. Banking & Fin. L. 619, 620 (2012) (defining shadow banking and noting that it has grown larger than traditional banking). The FAA today has monitoring authority over drones. 113 National Defense Authorization Act for Fiscal Year 2014, Pub. L. No. 113-66, 127 Stat. 870 (2013) (codified at 49 U.S.C. § 40101 (Supp. II 2015)) (“The Secretary of Defense, the Secretary of Homeland Security, and the Administrator of the Federal Aviation Administration shall jointly develop and implement plans and procedures to review the potential or joint testing and evaluation of unmanned aircraft equipment and systems . . . .”). Regulators’ initial oversight of hydroelectric dams has extended to other energy sources, such as nuclear power. 114 See Atomic Energy Act of 1946, Pub. L. No. 79-585, § 10(c), 60 Stat. 755, 768 (codified in scattered sections of 42 U.S.C.) (nuclear energy); see also 15 U.S.C. § 717g (gas); 16 U.S.C. § 825(b) (electricity); 43 U.S.C. § 1348 (2012) (offshore oil and gas). The FCC, by classifying wireless phone companies as common carriers, broadened its visitation authority originally intended for landline telephone companies. 115 See Telecommunications Act of 1996, Pub. L. No. 104-104, 110 Stat. 56 (codified in scattered sections of 47 U.S.C.). Cable systems also came under FCC jurisdiction. See United States v. Sw. Cable Co., 392 U.S. 157, 167–73 (1968) (finding that the FCC had broad authority to regulate a mobile communication form using microwaves). Deregulation in these areas has not removed broad authority to extract information. See Joseph D. Kearney & Thomas W. Merrill, The Great Transformation of Regulated Industries Law, 98 Colum. L. Rev. 1323, 1325–26 (1998) (“The role of the agency has been transformed from one of protecting end-users to one of arbitrating disputes among rival providers and, in particular, overseeing access to and pricing of ‘bottleneck’ facilities that could be exploited by incumbent firms to stifle competition.”). Internet providers were also subject to FCC monitoring and had been classified as common carriers. See Open Internet Order, 80 Fed. Reg. 19737 (Apr. 13, 2015). That classification was removed in December 2017. See Restoring Internet Freedom, FCC, [] (last visited Oct. 11, 2018). Thus, regulators of the financial system, transportation, and utilities early on accumulated monitoring authority that has remained robust.

2. Gradual Monitors: Health, Safety, and the Environment. — Another set of agencies has gained monitoring authority more incrementally. This development pattern most closely fits those agencies, like environmental regulators, focused on protecting from physical harm. The earliest exam­ple arose in pharmaceuticals. After several children died from tainted vaccines in 1902, 116 Sharon B. Jacobs, Crises, Congress, and Cognitive Biases: A Critical Examination of Food and Drug Legislation in the United States, 64 Food & Drug L.J. 599, 601 (2009) (“[T]he deaths of children from contaminated vaccines provided the impetus for the passage of the Biologics Control Act of 1902.”). Congress authorized federal agents to “enter and inspect any establishment for the propagation and preparation of any virus, serum, toxin, [or] antitoxin.” 117 Biologics Control Act of 1902, Pub. L. No. 57-244, § 3, 32 Stat. 728, 729 (codified as amended at 42 U.S.C. § 262(c) (2012)). This function ultimately went to the FDA. See Bryan A. Liang, Regulating Follow-On Biologics, 44 Harv. J. on Legis. 363, 433 (2007). Related visitorial statutes soon fol­lowed for meat and therapeutic drugs. 118 See Meat Inspection Act, Pub. L. No. 59-242, 34 Stat. 1256, 1260–65 (1907) (codified as amended at 21 U.S.C. § 601 (2012)); Food and Drug Act of 1906, Pub. L. No. 59-384, 34 Stat. 768, repealed by Federal Food, Drug, and Cosmetic Act, Pub. L. No. 75-717, 52 Stat. 1040 (1938). These powers were more limited than those of banking and transportation regulators, 119 See supra section I.C.1. since inspectors could not examine documents. 120 See Winton B. Rankin, Inspection Authority, 18 Food Drug Cosm. L.J. 673, 673 (1963) (“[P]resent law and facilities only permit occasional spot checks through factory inspection . . . .”).

A shift began in 1938 when scores of people died after ingesting a new elixir used to treat sore throats. 121 David F. Cavers, The Food, Drug, and Cosmetic Act of 1938: Its Legislative History and Its Substantive Provisions, 6 Law & Contemp. Probs. 2, 20 (1939) (“At least 73, perhaps over 90, persons in various parts of the country . . . died as a result of taking a drug known as ‘Elixir Sulfanilamide’ . . . .”). Had the company run tests, the solution’s poisonous properties would have been evident. 122 See id. (“Tests on animals or even an investigation of the published literature would have revealed the lethal character of the solvent.”). This event prompted legislation requiring pharmaceutical companies to submit to the FDA information about drugs before any sale. 123 Federal Food, Drug, and Cosmetic Act (codified in scattered sections of 21 U.S.C.). The FDA had a sixty-day window after each submission during which it could intervene. 124 Id. § 505(c). Exam­iners could also postpone the effective date of an application, permit­ting consideration for an additional 120 days. 125 Id.; see also Kelsey, supra note 56, at 51, 55 (explaining what happened when the FDA found that the new drug application was incomplete). But the legislation did not set a minimum threshold for the rigor of test data, nor did it require a drug company to affirmatively gain approval, which happened automatically if the FDA examiner failed to respond in time. 126 Federal Food, Drug, and Cosmetic Act § 505(c). Also, the amount of time in which the FDA could consider an application was lim­ited. 127 Id. Thus, the laws allowed drug companies to engage in similar “win­dow dressing” that plagued banks’ early reports to the OCC. 128 See supra note 97 and accompanying text.

It was in this statutory context that Dr. Kelsey received, in her first few months on the job in 1961, the four-volume submission for thalido­mide. 129 See Kelsey, supra note 56, at 48–49. Her supervisor observed, “[T]his is a very easy one. There will be no problems with sleeping pills.” 130 Id. at 49. Even though Dr. Kelsey repeatedly requested more scientific evidence before each sixty-day window expired, the company did not have the data she sought, and the FDA lacked the authority to compel the production of that data. 131 See James L. Zelenay, Jr., The Prescription Drug User Fee Act: Is a Faster Food and Drug Administration Always a Better Food and Drug Administration?, 60 Food & Drug L.J. 261, 264–66 (2005) (noting that although examiners had the authority to reject a new drug application as unsafe, the FDA likely did not have the authority to delay an application on the basis of “insufficient information”). Consequently, the FDA was still negotiating with the pharmaceutical company over approval when reports of widespread birth defects emerged from Germany, which had approved the drug years earlier. 132 See Kelsey, supra note 56, at 65–67; see also Peltzman, supra note 64, at 1050–51 (discussing the thalidomide crisis as the catalyst for increased FDA monitoring of new drugs entering the market).

Fueled by public alarm that the United States had barely avoided trag­edy, 133 Jacobs, supra note 116, at 609–12 (discussing coverage of thalidomide that emphasized the episode as a potential “national tragedy [that] had been averted thanks only to the ‘skeptical FDA physician’” (quoting John M. Goshko, FDA Awaits Results on Thalidomide Check, Wash. Post, Aug. 3, 1962, at A4)). President Kennedy signed a law requiring pharmaceutical com­panies to submit heightened scientific evidence—a precursor to the FDA’s modern clinical trials. 134 See Kefauver Harris Amendment, Pub. L. No. 87-781, 76 Stat. 780 (1962) (codified as amended in scattered sections of 21 U.S.C.). Drug companies were also required to submit any reports of adverse effects, which they previously could have withheld. See Zelenay, supra note 131, at 266 (summarizing the increased reporting requirements included in the 1962 act). Starting in the 1960s, FDA officials could withhold drug approval 135 Compare Kefauver Harris Amendment § 102 (codified as amended at 21 U.S.C. § 355(d) (2012)) (listing grounds for “refusing to approve the application” that do not address safety concerns, including that there is “a lack of substantial evidence that the drug will have the effect it purports or is represented to have”), with 21 U.S.C. § 355(d) (1958) (listing only safety concerns as grounds for “refusing to permit the [drug] application to become effective”). See also Zelenay, supra note 131, at 265 & n.31 (noting that rejecting the thalidomide application in 1961 for “insufficient information” may not have been within the FDA’s statutory mandate). and  “inspect  records,   files,   papers,   processes,   controls  and  facilities”  of  pharmaceutical  companies 136 See Rankin, supra note 120, at 673. even without evi­dence that the drug would be unsafe. In 2011, after deaths and illnesses from tainted peanut butter, cookies, and ice cream products, 137 Recent Legislation, Food Safety Modernization Act Implements Private Regulatory Scheme, 125 Harv. L. Rev. 859, 859–60 (2012) (linking several high-profile deaths from salmonella to the Food Safety Modernization Act). Congress gave the FDA broad food-inspection powers, matching those the agency had received for drugs. 138 See FDA Food Safety Modernization Act, Pub. L. No. 111-353, 124 Stat. 3885 (2011) (codified in scattered sections of 7, 21, and 42 U.S.C.). Most notably, facilities now must maintain food safety plans among their business records. See 21 U.S.C. § 350c(a)(2) (2012). Federal on-site food and drug surveillance programs today reach manufacturers, distribution warehouses, grocery stores, and restaurants. See id.

The thalidomide incident marked the beginning of a period of rapid growth in health monitoring. Amidst worsening air quality and related health concerns, 139 Despite a broader mission, the EPA’s origins lie in health-related incidents. See William S. Eubanks II, The Clean Air Act’s New Source Review Program: Beneficial to Public Health or Merely a Smoke-and-Mirrors Scheme?, 29 J. Land Resources & Envtl. L. 361, 362 (2009) (discussing early air-pollution-control legislation, which resulted from thousands of sicknesses and deaths caused by smog). the federal government established the EPA in 1970. 140 See Reorganization Plan No. 3 of 1970, 3 C.F.R. 199 (1970), reprinted in 5 U.S.C. app. at 698 (2012). The Agency assumed duties from several preexisting agencies. See The Origins of EPA, EPA, [] (last visited Oct. 11, 2018). The agency has regularly received new visitation authority over private companies in a range of sectors. 141 See, e.g., Federal Insecticide, Fungicide, and Rodenticide Act, 7 U.S.C. § 136 (2012) (selling or distributing pesticides); Toxic Substances Control Act, 15 U.S.C. § 2601 (2012) (toxic substances); Federal Water Pollution Control Act, 33 U.S.C. § 1254 (2012) (transporting oil); Safe Drinking Water Act, 42 U.S.C. § 300f (2012) (drinking water suppliers); Resource Conservation and Recovery Act of 1976, 42 U.S.C. § 6927 (hazardous wastes); Comprehensive Environmental Response, Compensation, and Liability Act of 1980, 42 U.S.C. § 9604 (general pollutants). For a more detailed summary of these various inspection provisions, see James A. Holtkamp & Linda W. Magleby, The Scope of EPA’s Inspection Authority, Nat. Resources & Env’t, Fall 1990, at 16, 16–17. This authority covers organizational processes; remotely installed monitoring devices; and entrance onto private property to examine records, take samples, and inspect facilities. See id. (describing the monitoring authority granted to the EPA by these acts). Congress also requires firms to notify the EPA of the development of new chemicals. See 15 U.S.C. § 2604(a) (giving the EPA ninety days to write a rule following notice). In the same year as the EPA launched, Congress created OSHA, 142 Occupational Safety and Health Act of 1970, Pub. L. No. 91-596, 84 Stat. 1590 (codified as amended in scattered titles of the U.S.C.). whose originating statute empowered it to enter work­places to conduct inspections, examine documents, and question employees. 143 See 29 U.S.C. § 657(a)–(c) (2012).

Whereas prior federal visitorial powers targeted specific industries—drugs, food, banking, transportation, or mining 144 The federal government first gained inspection authority over mines in 1941, through the Department of the Interior. See Act of May 7, 1941, ch. 87, 55 Stat. 177 (repealed 1969). Inspections for noncoal mines came in the 1960s. See Act of Sept. 26, 1961, Pub. L. No. 87-300, 75 Stat. 649 (codified at 43 U.S.C. § 1457 (2012)). That authority was later transferred to the Department of Labor, see Federal Mine Safety and Health Act of 1977, § 301, Pub. L. No. 95-164, 91 Stat. 1290, 1317–19 (codified as amended at 30 U.S.C. § 961 (2012)), through the newly created Mine Safety and Health Administration in 1977, see id. § 302(a) (codified at 29 U.S.C. § 557a). —the EPA and OSHA obtained cross-industry reach, enabling the federal government to look inside almost every private business across the country. In 1978, in Marshall v. Barlow’s, Inc., the Supreme Court found a Fourth Amendment administrative search warrant requirement for industries without “a long tradition of close government supervision.” 145 436 U.S. 307, 313, 320–21 (1978). The EPA is held to similar standards. See Nat’l-Standard Co. v. Adamkus, 881 F.2d 352, 361 (7th Cir. 1989). In industries with a history of close regulatory oversight, an exception to the Fourth Amendment’s search warrant requirement is appropriate. See Marshall, 436 U.S. at 313–14. But this ruling left many domains subject to warrantless monitoring. 146 Marshall does not prevent warrantless administrative searches in various heavily regulated industries. See, e.g., Dow Chem. Co. v. United States, 476 U.S. 227, 239 (1986) (allowing the EPA to conduct warrantless aerial surveillance of private property); Donovan v. Dewey, 452 U.S. 594, 605–06 (1982) (allowing the Department of Labor to conduct warrantless searches to inspect worker health and safety in the mining industry); United States v. Chuang, 897 F.2d 646, 651 (2d Cir. 1990) (allowing the OCC to conduct warrant­less searches of bank documents). Moreover, inspectors in other industries regularly give a Miranda-style warning 147 See Miranda v. Arizona, 384 U.S. 436, 467–68 (1966) (establishing the duty of officers to inform those in custody of their right to remain silent). that the employer has the right to request a warrant, which businesses rarely exercise. 148 Interview with OSHA Deputy Regional Administrator and Regional Administrator (Apr. 7, 2017) [hereinafter OSHA Interview]. Despite the significance of a constitutional protection, Marshall’s practical impact is limited. The Court acknowledged that the Fourth Amendment was less relevant to OSHA than to criminal searches. See Marshall, 436 U.S. at 320. Unlike police officers, OSHA would not need “probable cause . . . based . . . on specific evidence of an existing violation.” Id. The agency could instead obtain a warrant if the search was part of a “general administrative plan.” See id. at 320–21. This ruling forced OSHA to develop national inspection plans. OSHA Interview, supra. If needed, OSHA inspectors can easily obtain a warrant without probable cause by showing the magistrate judge their plan. Id. Thus, despite some obstacles along the way, the largest federal health, safety, and environmental regulators incrementally over the past century obtained the type of visitorial tools that the OCC received for banks during the Civil War. 149 See supra section I.C.1.

3. Limited Monitors: Consumer Protection, Competition, and Labor. — Regulators focused on pro­tecting individuals from economic harms have more limited monitoring authority. 150 In contrast to the agencies discussed in this section, the SEC protects investors that are often institutional. Also, the agency was formed as part of a broader goal of protecting the financial system rather than individuals. See supra note 106 and accompanying text. Spurred by Ida Tarbell’s popular writings about the “auto­cratic powers in commerce” of John D. Rockefeller’s Standard Oil Company 151 2 Ida M. Tarbell, The History of the Standard Oil Company 229 (reprt. 1963) (Macmillan, two vols. in one 1933) (1904); see also 1 Tarbell, supra, at 158 (concluding that Standard Oil had “great power . . . resistless, silent, perfect in its might”). Tarbell’s writings would ultimately contribute to the breakup of Standard Oil. See Steve Weinberg, Taking on the Trust: The Epic Battle of Ida Tarbell and John D. Rockefeller 246–51 (2008). and the activism of President Theodore Roosevelt, 152 See F.M. Scherer, Sunlight and Sunset at the Federal Trade Commission, 42 Admin. L. Rev. 461, 462 (1990) (noting President Roosevelt’s role in providing the impetus for the founding of the Bureau of Corporations, the predecessor of the FTC). the FTC was founded in 1914. 153 Federal Trade Commission Act of 1914, Pub. L. No. 63-203, § 6, 38 Stat. 717, 721–22 (codified as amended in scattered sections of 15 U.S.C.). Its two main mis­sions are to protect consumers and to promote competition. 154 Most commentators agree that consumer welfare lies at the root of antitrust laws, although there is some debate. See Steven C. Salop, Question: What Is the Real and Proper Antitrust Welfare Standard? Answer: The True Consumer Welfare Standard, 22 Loy. Consumer L. Rev. 336, 336–38 (2010) (summarizing the debate). The FTC had from the outset the power “[t]o require . . . corporations engaged in commerce . . . to file with the commission . . . both annual and special[ ] reports or answers in writing to specific questions . . . as to the organization, business, conduct, practices, [and] management.” 155 Id. § 6(b). President Roosevelt had unsuccessfully advocated for a stronger moni­toring framework: man­datory notifications prior to mergers and acquisi­tions. 156 See Scherer, supra note 152, at 462–63 (discussing the monitoring framework that Roosevelt advocated for in a 1900 letter to the New York legislature). In 1976, Congress extended that authority. 157 Hart–Scott–Rodino Antitrust Improvements Act of 1976, Pub. L. No. 94-435, § 201, 90 Stat. 1383, 1390–94 (codified at 15 U.S.C. § 18a (2012)). In 2003, Congress added further mandatory notifications of contractual agreements between brand-name and generic drug companies. See Medicare Prescription Drug, Improvement, and Modernization Act of 2003, Pub. L. No. 108-173, § 1112, 117 Stat. 2066, 2461–63 (codified at 21 U.S.C. § 355 note (2012)). Despite its extensive report-collecting tools, the agency has never had explicit visitation authority for either competition or consumer protection.

The two leading regulators of employment have even more limited monitoring authority than the FTC. Amidst the labor unrest of the Great Depression, Congress tasked the NLRB with the “protection by law of the right of employees to organize and bargain collectively.” 158 National Labor Relations Act, Pub. L. No. 74-198, § 1, 49 Stat. 449, 449 (1935) (codified at 29 U.S.C. §§ 151–169 (2012)). The NLRB’s originating statute did not men­tion monitoring in the traditional sense. The agency perhaps comes clos­est to monitoring today through its on-site supervision of union elections. 159 See Representation Law and Procedures, ABA 17,
authcheckdam.pdf [] (last visited Oct. 11, 2018) (noting that elections are supervised by an NLRB agent on the employer’s premises). Since the NLRB’s main role is to conduct the elections, such as by overseeing the agreement as to time, place, and methods for voting, the main purpose is not as clearly to collect nonpublic information as to manage an event. See Conduct Elections, NLRB, [] (last visited Feb. 5, 2019).

In the face of nationwide protests and unrest, the 1964 Civil Rights Act established the EEOC and required companies to maintain employ­ment records. 160 42 U.S.C. §§ 2000e-4(a), 2000e-8(c) (2012). The original House bill for the agency had put forth an information-collection authority modeled after the FTC, but that lan­guage was removed in the face of intense Senate opposition. 161 See Michael Z. Green, Proposing a New Paradigm for EEOC Enforcement After 35 Years: Outsourcing Charge Processing by Mandatory Mediation, 105 Dick. L. Rev. 305, 320 (2001) (describing the much stronger authority for the EEOC envisioned in the committee version of the bills and the opposition that limited the agency’s authority). The final legislation specified that to collect records the EEOC must write rules. 162 42 U.S.C. § 2000e-8(c) (requiring employers to “make and keep such records” relevant to determining whether unlawful employment practices occurred but requiring employers to make reports only “as the Commission shall prescribe by regulation or order”). In both the EEOC and NLRB, “examination” occurs mostly after a firm is accused. 163 See Civil Rights Act of 1964, Pub. L. No. 88-352, §§ 709–710, 78 Stat. 241, 262–64 (codified as amended at 42 U.S.C. §§ 2000e-8 to 2000e-9); National Labor Relations Act § 11; EEOC v. Shell Oil Co., 466 U.S. 54, 64 (1984) (“[EEOC’s] power to conduct an investigation can be exercised only after a specific charge has been filed in writing.” (quoting 110 Cong. Rec. 7214 (1964))). But the EEOC has used its original statutory authority to write rules to require businesses to submit to the EEOC confidential employee data broken down by race, gender, and other categories. 164 See 29 C.F.R. § 1602.7 (2018) (requiring companies to file an EEO-1 report annually); EEO-1 Frequently Asked Questions and Answers, EEOC, [] (last visited Nov. 6, 2018) (noting that the survey “requires company employment data to be categorized by race/ethnicity, gender and job category”).

As yet, no crisis or national outcry has driven Congress to give explicit visitorial authority to these three agencies. But the creation of the CFPB in 2011 represented a break with the traditional absence of visitorial authority for regulators focused on protecting against economic harms to individuals. 165 anking regulators had a secondary mission of consumer protection, but this was rooted in stability concerns. See supra section I.C.1. The FTC had previously exercised consumer protection authority for many financial institutions implicated in the sub­prime mortgage crisis, such as nonbank mortgage servicers. 166 See Dodd–Frank Wall Street Reform and Consumer Protection Act, 12 U.S.C. §§ 5321, 5322(a)(2), 5491(a) (2012). Congress moved most of that authority to the CFPB after millions of families lost their homes to foreclosure, many due to unscrupulous lending. 167 Laura Kusisto, Many Who Lost Homes to Foreclosure in Last Decade Won’t Return—NAR, Wall St. J. (Apr. 20, 2015), (on file with the Columbia Law Review) (“More than 9.3 million homeowners went through a foreclosure, surrendered their home to a lender or sold their home via a distress sale between 2006 and 2014.”). Unlike the FTC, the CFPB was given broad visitorial authority to regularly appear on-site. 168 12 U.S.C. § 5514(b)(1) (noting that the “Bureau shall require reports and conduct examinations on a periodic basis”). Thus, despite having more limited authority than is pre­sent in other spheres, the largest regulators of individuals’ economic interests can monitor to some extent. Additionally, between the launch of the CFPB and the increase in FTC antitrust reporting, the overall trajectory of this sphere of regulation has been toward more statutory monitoring authority.

D. Summary of the Statutory Rise

Across diverse industries and under both Democratic and Republican party leadership, Congress has since the mid-1800s steadily expanded federal agencies’ ability to monitor private firms. This histori­cal accumulation of federal authority also spans industries that fall outside the scope of this Article because they are governed by  small  and  medium  regulators — areas  such  as  offshore  oil  drilling, 169 See Outer Continental Shelf Lands Act Amendments of 1978 § 208, 43 U.S.C. § 1348 (2012) (providing authority for the inspection and investigation of offshore oil-drilling platforms). liquor stores, 170 See Colonnade Catering Corp. v. United States, 397 U.S. 72, 76 (1970) (“Congress has broad power to design such powers of inspection under the liquor laws as it deems necessary to meet the evils at hand.”). and firearm manufacturers. 171 See United States v. Biswell, 406 U.S. 311, 316–17 (1972) (concluding that “inspections for compliance with the Gun Control Act pose only limited threats to . . . privacy” and when “regulatory inspections further urgent federal interest, and the possibilities of abuse and the threat to privacy are [minimal], the inspection may proceed without a warrant where specifically authorized by statute”). Over­all, among the nineteen large federal regulators, 172 See supra section I.B (listing the nineteen large regulators and describing the methodology for identifying them). only the NLRB is without substantial monitoring authority. Two others, the FTC and the EEOC, have the meaningful ability to collect records but not to conduct on-site inspections. Sixteen of the nineteen largest agencies have both strong visitorial monitoring and record-collection authority. 173 See infra Appendix A; supra section I.B. The laws are in place for a formidable regulatory-monitor state.

II. The Institutional Rise

Agency behavior is determined not just by its underlying statutes but also by stakeholders. Scholars have focused on the changing influence of external stakeholders such as Congress, the President, and special inter­est groups on the administrative state. 174 See, e.g., Kagan, supra note 54, at 2253 (arguing that President Clinton ushered in an era of “presidential administration,” but noting that “[a]t the dawn of the regulatory state, Congress controlled administrative action”). Internal agency groups also com­pete for control, but their history has been largely studied through the lens of policy instruments. 175 See, e.g., Daniel A. Farber & Anne Joseph O’Connell, Agencies as Adversaries, 105 Calif. L. Rev. 1375, 1407–08 (2017) (presenting a typology of inter- and intra-agency conflict, noting that agency conflicts “manifest in all forms of decision making: rulemaking, adjudication, and program-level policy,” and acknowledging that the scholarship focuses on rulemaking). A standard account holds that adjudica­tion dominated agency policymaking until the 1970s, when agencies entered “an age of rulemaking.” 176 See J. Skelly Wright, The Courts and the Rulemaking Process: The Limits of Judicial Review, 59 Cornell L. Rev. 375, 375–76 (1974); see also M. Elizabeth Magill, Agency Choice of Policymaking Form, 71 U. Chi. L. Rev. 1383, 1384–86 (2004) (noting the “detectable shift” toward rulemaking in the 1970s). The internal narrative then becomes vague, despite general recognition that in the 1990s and 2000s new gov­ernance models took hold. 177 See infra section II.A. Some observers believe that rulemaking still remains the dominant policy instrument, 178 See, e.g., Michael D. Sant’Ambrogio & Adam S. Zimmerman, The Agency Class Action, 112 Colum. L. Rev. 1992, 2017 (2012) (“[S]ince the 1970s, informal rulemaking has been the preferred means of implementing agency policy . . . .”). while others see a shift to either “policy through litigation, negotiated settlements, or the waiver of rules in individual contexts.” 179 See Magill, supra note 176, at 1398–99. Professors Magill and Vermeule identify various factors that reallocate power toward and away from lawyers, without distinguishing regulatory monitors or seeing an overall trend. See Magill & Vermeule, supra note 38, at 1077.

This  Part  adds  the  role  of  the  monitoring  group  to  that  internal  organiza­tion narrative. 180 At the core of existing internal narratives is a recognition that organizational dynamics of administrative agencies have shifted in response to new governance paradigms and market evolutions, but how those dynamics intersect with regulatory monitors has yet to be explored. It shows how prominent changes in governance and markets have plausibly moved regulators to rely more on monitors than on other groups. The governance changes include greater weight on collaborating with busi­nesses, the rise of compliance departments in corporations, and increased external stakeholder pressure. The market changes include the greater sophistication of modern businesses, the pace of innovation, and the ubiquity of information technologies. Although the focus is on recent historical shifts, the main goal is to lay the foundations for understanding the role of regulatory monitors today.

A. Governance Changes Favoring Regulatory Monitors

Over the past thirty years, agencies have adopted new approaches to governing firms. Prominent observers attribute these changes to a “crisis in confidence” 181 Ayres & Braithwaite, supra note 30, at 158. in regulation, or the perception that in “the administra­tive state . . . much is terribly wrong.” 182 See Freeman, supra note 30, at 8–9 (discussing widespread critiques of ossified regulation). Regulatory monitors are well situated to thrive in the resulting organizational landscape.

1. Collaborative Governance. — One major shift in the modern regu­latory approach is a greater emphasis on collaboration. 183 See id. at 4, 22 (identifying an emerging “model of collaborative governance”); see also Lobel, supra note 30, at 344. The U.S. House Budget Committee displayed this philosophy in OSHA’s 2017 budget hearing, encouraging the agency to minimize punishment and instead “partner with businesses to create safer workplaces.” 184 FY 2017 OSHA Cong. Budget Justification 14–16,
files/documents/general/budget/CBJ-2017-V2-12.pdf [].
The ex­tent to which any given agency has adopted this model varies, but one of its features is seeing rules as provisional, requiring the parties to flexibly “devise solutions to regulatory problems.” 185 Freeman, supra note 30, at 22. This depiction intersects with elements of Professors Ayres and Braithwaite’s “responsive regulation.” See Ayres & Braithwaite, supra note 30, at 35–36 (presenting a generic “enforcement pyramid” demonstrating that agen­cies seek regulatory compliance more frequently through efforts at “persuasion” than the use of civil or criminal penalties or license revocations); see also infra notes 297–301 and accompanying text.

The emphasis on partnership is important, in part, for the acquisi­tion of information. Agencies today generally believe rules should be “responsive to[ ] the particular contexts in which they are deployed” by relying on “feedback mechanisms” that are “continuous.” 186 Freeman, supra note 30, at 22, 28. Firms that are less afraid of punishment, it is thought, become more willing to share information. For instance, the EPA’s new cooperative model gave it “open access” to citrus-juice plants, whereas in the prior relationship “companies resist[ed] inspection and cooperate[d] with the EPA only grudgingly.” 187 Id. at 61. The cooperative model aims to free the parties to focus their energies on fixing mistakes and identifying causes instead of fighting over whether anything was wrong.

Litigation groups are seen as less well-suited to this model. Legal inves­tigations cause information exchange to become “bogged down as target firms resist[ ] compliance and pursue[ ] blocking actions in the courts.” 188 Scherer, supra note 152, at 471 (observing dynamics in the 1970s, from the perspective of having been an FTC economist). Consider, again, the example of how the CFPB in its early finan­cial examinations brought along enforcement lawyers. 189 See supra note 72 and accompanying text. Industry groups had criticized the practice, saying that “the presence of enforce­ment attorneys at routine examinations created a hostile regulatory envi­ronment.” 190 Alan Zibel, Consumer Regulator to Stop Bringing Lawyers to Firm Exams, Wall St. J. (Oct. 9, 2013), (on file with the Columbia Law Review). The CFPB’s Ombudsman had studied the matter and warned that the presence of attorneys would serve as “a barrier to a free exchange.” 191 CFPB Ombudsman’s Office, FY2012 Annual Report to the Director 13 (2012),
Office_Annual_Report.pdf [].
Asked to explain its subsequent termination of the policy, the CFPB said that it “wasn’t efficient.” 192 Witkowski, supra note 72.

A collaborative relationship with continuous information flow would naturally propel an agency to become more dependent on regulatory monitors. Although some regulatory monitors have been viewed as criti­cal and overbearing, 193 See Hawke, supra note 95, at 4. their information collection does not assume the regulated entity has misbehaved. Indeed, the scholarly depiction of the collaborative model of governance matches some historical descriptions of early bank examiners, who because of limited sanction authority “rec­ommended” rather than commanded 194 See White, supra note 90, at 21; see also White, supra note 94, at 48. and  relied  on  “cooperation”  to  achieve  compliance. 195 See Robertson, supra note 87, at 71. Banking  regulators  have  remained “famously  nonad­versarial,” 196 David Zaring, Administration by Treasury, 95 Minn. L. Rev. 187, 208 (2010). and  energy inspectors have retained a team-oriented approach. 197 See Hayes, supra note 11 (describing how energy inspectors “work alongside, not against, industry to ensure operators follow acceptable industry practices and federal safety standards”). An agency adopting collaborative governance might thus seek to shift more interactions from regulatory lawyers to regulatory monitors.

2. Compliance Departments and Self-Regulation. — Many regulators now emphasize “management-based regulation.” 198 See generally Cary Coglianese & David Lazer, Management-Based Regulation: Prescribing Private Management to Achieve Public Goals, 37 Law & Soc’y Rev. 691 (2003) (using case studies to illustrate when and how management-based regulation can be effective). Fiscal constraints simply make it impossible to monitor all private actions even for the most dangerous activities: For example, federal inspectors estimated that only 1–2% of all “safety-related” nuclear plant activities were subject to close, annual government monitoring. 199 Peter K. Manning, The Limits of Knowledge, in Making Regulatory Policy 49, 70 (Keith Hawkins & John Thomas eds., 1989). Self-regulation does not necessarily mean an absence of oversight but “that regulation should respond to . . . how effectively industry is making private regulation work.” 200 See Ayres & Braithwaite, supra note 30, at 4. This self-regulatory model encourages regulatory experimental­ism. 201 Cf. Michael C. Dorf & Charles F. Sabel, A Constitution of Democratic Experimentalism, 98 Colum. L. Rev. 267, 373–80 (1998) (describing “emergent experimentalism” in the environmental-regulation context). Instead  of  a  bottom-up  approach  of  examining  every  product,   doc­ument,   or  facility  for  strict  adherence  to  a  code,   the agency “inter­vene[s] at the planning stage, compelling regulated organizations to improve their internal management so as to increase the achievement of public goals.” 202 Coglianese & Lazer, supra note 198, at 694. In essence, the regulator engages in a top-down assess­ment of a firm’s self-monitoring.

The need for self-monitoring helps explain why “the compliance department has emerged, in many firms, as the co-equal of the legal department.” 203 Griffith, supra note 27, at 2077. When the legal department runs a company’s compli­ance, the concern is that the process may become “excessively legalis­tic.” 204 Robert C. Bird & Stephen Kim Park, The Domains of Corporate Counsel in an Era of Compliance, 53 Am. Bus. L.J. 203, 206 (2016). Compliance departments review employees’ practices or con­sumer complaints not only to ensure that the company is not breaking the letter of the law as determined by the legal department but in many cases to tell the company how to “comply with the spirit of the law.” 205 See Michele DeStefano, Creating a Culture of Compliance: Why Departmentalization May Not Be the Answer, 10 Hastings Bus. L.J. 71, 149 (2014) (quoting from the author’s interview with an anonymous chief compliance officer in the financial industry). The compliance department keeps internal records of violations and the firm’s responses 206 See generally id. at 91–97 (describing the function of the compliance department). —records that regulatory monitors can later examine.

EPA rules, for example, require companies producing hazardous chemicals to build a risk management plan 207 40 C.F.R. § 68.73(b)–(c) (2018) (requiring companies to develop and train employees concerning “procedures to maintain the on-going integrity of process equipment”). and perform inspections of their equipment. 208 See id. § 68.73(d). Companies must regularly submit the documentation to authorities, listing all incidents that have occurred. 209 See id. § 68.220(a)–(b). Environmental agencies then audit those internal reports, 210 Id. § 68.220(a). which may result in a “deter­mination of necessary revisions” to the company’s systems. 211 Id. § 68.220(e). Agen­cies also enlist a growing number of private third-party monitors to assess compliance. 212 See Jodi L. Short & Michael W. Toffel, The Integrity of Private Third-Party Compliance Monitoring, Admin. & Reg. L. News, Fall 2016, at 22, 22 (noting that third-party certification is used in “a wide array of domains, including food safety, pollution control, product safety, medical devices, and financial accounting”); see also Reinier H. Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 93–94 (1986) (giving examples of industries in which liability is imposed upon third-party monitors like the underwriters of securities to incentivize thorough and accu­rate gatekeeping in order to prevent fraudulent products from reaching the market). See generally Kraakman, supra, at 56–60 (outlining the benefits of relying on third-party monitors and noting that “[i]n general, third-party strategies can exploit private enforcement information ex ante . . . by disclosing it to enforcement officials or potential victims or by relying on private monitors themselves to take obstructive action short of direct disclosure”). The SEC uses a related model by overseeing a private regulator, the Financial Industry Regulatory Authority (FINRA), which performs examinations and has its own enforcement group. See FINRA, FINRA 2015 Year in Review and Annual Financial Report 12–13 (2016) [hereinafter FINRA Report], [].

Depending on how it is implemented, self-regulation can diminish the role of regulatory monitors relative to other agency groups because it privatizes core monitoring tasks. 213 See Ryan Beene, Is NHTSA Nominee Up to Task?, Tire Bus. (Dec. 1, 2014), [] (describing how the “[National Highway Traffic Safety Administration (NHTSA)] allocates just $10 million a year to its roughly 50 staffers,” while GM alone hired 35 safety investigators in a single year). This is particularly true when the agency delegates all monitoring to third parties. 214 Third-party private auditing has grown in recent years. See Lesley K. McAllister, Regulation by Third-Party Verification, 53 B.C. L. Rev. 1, 6 (2012). Private parties also often serve as monitors after courts determine wrongdoing. See Root, supra note 10, at 527. But replacement is not how most agencies have approached self-regulation. Many still conduct their own inspections, alongside industry self-monitoring. 215 See supra notes 207–211 and accompanying text for an example of how the EPA imposes self-monitoring obligations in addition to conducting its own inspections. Rather, the model transforms the agency into a manager of private monitors.

From an internal perspective, agencies’ regulatory monitors—not their litigators—normally assume this managerial role. 216 See, e.g., SEC, Agency Financial Report Fiscal Year 2016, at 9 (2016), [] (noting that the monitors in the Office of Compliance Inspections conduct examinations, not the litigators in the enforcement division or the Office of General Counsel). An agency group that is already the most knowledgeable about monitoring activities would be the natural home for such managing of private monitors. Thus, this managerial model moves regulatory monitors from examining the details of paperwork or safety valves to making sure others do those jobs. In some sense, this amounts to promoting regulatory monitors to a more senior supervisory role. As supervisors of large business departments rather than individual documents or equipment, regulatory monitors can col­lect more information in the same amount of time, since the company’s compliance employees create a data report that the regulatory monitors would have previously compiled.

Moreover, the compliance department is prominent inside large busi­nesses, with the Chief Compliance Officer typically reporting to the CEO and often the board. 217 See Griffith, supra note 27, at 2077. Consequently, any regulatory-monitor rec­ommendation for improving a firm’s compliance system can affect a broader portion of the business on a more enduring basis. Imagine, for instance, that a credit card company has been found to have illegally charged consumers fees. In a precompliance world, the regulator might rely on a legal settlement or court order requiring the company to stop charging that fee moving forward. In the era of compliance manage­ment, the regulator (today, the CFPB) can bypass the courts and simply ask the company to develop a system for internally reviewing customer complaints for legal violations. That internal change means that the com­pliance department moving forward will catch not only this particular illegal credit card fee but also other improper fees that might arise in the future. Furthermore, the CFPB examination group regularly checks to make sure financial institutions have such customer complaint monitor­ing systems in place, even without any evidence that the firm has done anything wrong. 218 Interview with Former CFPB Employee (Mar. 10, 2017) [hereinafter CFPB Interview].

In other words, the firm’s compliance team essentially serves as the regulatory monitors’ agents. Scholars have more broadly recognized that the compliance “revolution” in corporate governance means that “pros­ecutors can externalize a portion of their budget.” 219 See Griffith, supra note 27, at 2077, 2127. While that may be true, in terms of internal organizational dynamics, agencies would be expected to shift some of what was previously prosecutors’ domain—pro­moting compliance through litigation—to regulatory monitors.

The move to compliance management may also reallocate respon­sibilities between regulatory monitors and rulemakers. Compliance man­agement reflects how “[b]est practices are the new means through which Congress and federal agencies are making administrative law.” 220 David Zaring, Best Practices, 81 N.Y.U. L. Rev. 294, 296 (2006). In the Clean Water Act, Congress mandated that states and the EPA identify “best management practices” for tackling the biggest source of water pollution: runoff from cities and farms. 221 33 U.S.C. § 1329(a)(1)(C) (2012); see also Zaring, supra note 220, at 326, 329. The EPA then shares “success stories” that can be adopted elsewhere. 222 See Zaring, supra note 220, at 331. In a world of formal rules that must be strictly applied, the rulemaking group spells out the particular steps a firm must take to comply with the law. Conversely, in a world of best practices, there are often multiple ways to satisfy the mandate. A best practices regime thereby allows agency regulatory monitors not only to identify the best practices in the first place but also to assess whether a given firm’s practices come close enough to “best.”

3. Heightened Stakeholder Oversight. — Agencies have come under increasing scrutiny from Congress, 223 See, e.g., Jacob E. Gersen & Eric A. Posner, Soft Law: Lessons from Congressional Practice, 61 Stan. L. Rev. 573, 606–07 (2008) (“Congress uses a range of instruments to influence administrative agencies, including restrictions on the appointment and removal of personnel, specification of substantive or procedural restrictions, appropriations, oversight hearings, and deadlines.”). the President, 224 See, e.g., Kagan, supra note 54, at 2281–318 (discussing President Clinton’s role in shaping the regulatory activity of the executive branch agencies). and courts. 225 See, e.g., Jerry L. Mashaw & David L. Harfst, Inside the National Highway Traffic Safety Administration: Legal Determinants of Bureaucratic Organization and Performance, 57 U. Chi. L. Rev. 443, 444 (1990) (“Indeed, courts frustrated by the ineffectiveness of legal directives often try their own hand at reorienting agencies’ internal laws, cultures, and personnel.”). This oversight may drive agencies toward greater reliance on regulatory moni­tors for three main reasons. First, as a general matter, “[a]dministrative agencies, like trial judges facing appellate review, dislike having their deci­sions reversed.” 226 Jennifer Nou, Agency Self-Insulation Under Presidential Review, 126 Harv. L. Rev. 1755, 1756 (2013). To avoid wasted efforts and delays, agencies insu­late themselves from oversight. 227 See id. at 1782–813 (describing how agencies choose from various regulatory instruments to self-insulate from presidential review). They have substituted policy statements and interpretative guidelines for official rules to avoid having to go through notice and comment. 228 Thomas O. McGarity, Some Thoughts on “Deossifying” the Rulemaking Process, 41 Duke L.J. 1385, 1393 (1992) (observing the “increasing tendency of agencies to engage in ‘nonrule rulemaking’”); Zaring, supra note 220, at 297 fig.2 (showing a significant and steady increase from 1980 to 2004 in the annual number of regulations referencing “best practices” in the Federal Register). For enforcement, agencies have turned to extrajudicial strategies such as settlements and recommendations. 229 Sant’Ambrogio & Zimmerman, supra note 178, at 2034 (“Agencies . . . have, with modest success, adopted informal techniques in response to system-wide disputes that otherwise would overtax traditional, individualized adjudication.”). As the FDA explains of a regulatory-monitor tool it has used increasingly in recent years, a “Warning Letter is informal and advisory. . . . FDA does not consider Warning Letters to be final agency action on which it can be sued.” 230 See FDA, Regulatory Procedures Manual ch. 4, at 4 (2018) [hereinafter FDA Manual],
UCM074330.pdf [].
Courts have agreed. 231 See Holistic Candlers & Consumers Ass’n v. FDA, 664 F.3d 940, 944 (D.C. Cir. 2012) (“The letters plainly do not mark the consummation of FDA’s decisionmaking.”).

The same rulemaking and litigation groups could control informal activities. However, informal tools move further from the distinct func­tions and skillsets of legal actors, opening the door for other groups to assume related responsibilities. Moreover, court oversight has restricted even rulemakers’ informal alternatives. After industry complaints that the FDA was using “Good Guidance Practices” 232 The Food and Drug Administration’s Development, Issuance, and Use of Guidance Documents, 62 Fed. Reg. 8961, 8967–68 (Feb. 27, 1997) (codified at 21 C.F.R. § 10.115 (2018)). to write de facto rules, Congress required the agency to solicit public notice and comment prior to issuing major guidelines. 233 Food and Drug Administration Modernization Act, Pub. L. No. 105-115, § 405, 111 Stat. 2296, 2368 (1997) (codified at 21 U.S.C. § 371(h)(1)(A) (2012)) (“The Secretary shall develop guidance documents with public participation . . . .”); see also Lars Noah, Governance by the Backdoor: Administrative Law(lessness?) at the FDA, 93 Neb. L. Rev. 89, 98–99 (2014) (describing Congress’s requirement that the FDA “solicit comments before finalizing major guidance”). However, those constraints did not address regulatory monitors’ main textual outlets, such as their industry-wide in­spection manuals and case-by-case recommendations. 234 See infra section III.C.

Second, rulemaking has slowed considerably. Under the recent Bush and Clinton administrations, on average, over eight hundred days passed between a rule’s agenda publication and final adoption. 235 Stuart Shapiro, Presidents and Process: A Comparison of the Regulatory Process Under the Clinton and Bush (43) Administrations, 23 J.L. & Pol. 393, 416 (2007). When rules are not updated, frontline regulatory monitors or their supervisors must interpret old laws to apply them to new practices. If agencies are largely unable to write formal rules, and instead engage in soft rulemaking,   agen­cies  may  be  incentivized  to  write  vaguer  rules  that are  nonbinding. 236 See Zaring, supra note 196, at 208–09 (noting that financial regulators have adopted “principles-based regulation” that is largely unreviewable by courts and enforced informally, rather than by utilizing the rule of law). But see Daniel E. Walters, The Self-Delegation False Alarm: Analyzing Auer Deference’s Effects on Agency Rules, 119 Colum. L. Rev. 85, 157–60 (2019) (noting that even with the incentives for vague self-delegation created by the Auer decision, agencies have a “strong[ ] interest in promoting clarity in the regulatory text” to improve enforceability because “[i]n addressing the risk of hard look review, agencies will of necessity seek to reduce vagueness”). Imprecise rules may force agencies to rely more on frontline actors’ per­suasion and judgment. Instead of following a lawyer’s written instructions (the legal rule), regulatory monitors in such agencies can act more like clients, consulting lawyers only as needed with help in interpretation. 237 See, e.g., FY 2015 EEOC Performance and Accountability Rep. 23, https:// [] (mention­ing how the EEOC engaged in sixty “technical assistance” visits).

Third, one of the impulses behind greater external oversight is to “ensure[ ] that regulatory agencies exercise their policymaking discretion in a manner that is reasoned.” 238 Richard B. Stewart, Administrative Law in the Twenty-First Century, 78 N.Y.U. L. Rev. 437, 439 (2003). Most prominently, courts and the Presi­dent have imposed cost–benefit analyses, 239 See Richard L. Revesz & Michael A. Livermore, Retaking Rationality: How Cost-Benefit Analysis Can Better Protect the Environment and Our Health 10–12 (2008) (describing broad uses of cost–benefit analyses and concluding they are “here to stay.”). and “lawyers will have little to contribute to this quintessentially technocratic problem.” 240 Magill & Vermeule, supra note 38, at 1051. Additionally, the Paperwork Reduction Act (PRA) constrains rule writers’ ability to collect supportive information from firms. 241 See 44 U.S.C. §§ 3501–3521 (2012) (explaining the goal of “reduc[ing] information collection burdens on the public”).

In  contrast  to  these  legal  constraints  on  lawyers’  core  activities,  in  recent  years  Congress  has  imposed  widespread  monitoring  minimums,   such  as  annual  or  more  frequent  on-site  examinations  of credit rating organizations, 242 See Dodd–Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111-203, sec. 932(a)(8), 124 Stat. 1376, 1877 (2010) (codified at 15 U.S.C. § 78 (2012)). food manufacturers, 243 See FDA Food Safety Modernization Act, Pub. L. No. 111-353, § 201, 124 Stat. 3885, 3923 (2011) (codified as amended at 21 U.S.C. § 350j(a)(1) (2012)). High-risk facilities must be inspected at least every three years. Id. § 350j(a)(2)(B). and oil producers. 244 See 43 U.S.C. § 1348(c) (2012). To be sure, statutes in some contexts require regular actions by rule writers and litiga­tors if an agency chooses to act. For the EPA to ban a chemical, for instance, it must write a rule. 245 See, e.g., 15 U.S.C. § 2604(a)(5) (Supp. V 2018). But Congress does not mandate annual minimums for the number of chemicals banned, rules written, or trials litigated. Thus, whereas the external pressure for informed regulatory decisions slows down rule writers’ core activity—producing rules—it expands regulatory monitors’ basic function.

B. Market Transformations Favoring Regulatory Monitors

Whatever the inherent democratic accountability deficiencies of older governance models, new regulatory strategies were perhaps inevitable given the market transformations of recent decades. These changes have lessened or eliminated the sophistication gap between regulatory monitors and lawyers, expanded information asymmetries between regulatory monitors and legal groups, and provided regulatory monitors with technological tools that are more helpful to them than to rulemak­ers or litigators.

1. Increased Sophistication. — Modern businesses have reached unprec­edented size and complexity. All major industries have become more concentrated, creating bigger organizations with separate multimil­lion-dollar product lines. Oil companies have built ever larger floating cities drilling miles deeper under the ocean floor, 246 See BSEE Annual Report, supra note 19, at 15 (noting the increase in drill rigs).   manufacturers  release  thousands  of  new  chemicals  into  the  environment  annually, 247 Daniel C. Esty, Environmental Protection in the Information Age, 79 N.Y.U. L. Rev. 115, 163–64 (2004). and  large  businesses  deploy  big  data  computer  algorithms  for  key  decisions. 248 See Rory Van Loo, Helping Buyers Beware: The Need for Supervision of Big Retail, 163 U. Pa. L. Rev. 1311, 1331–32 (2015).

These transformations mean that an agency seeking to continue per­forming the same level of monitoring must now deploy additional regula­tory monitors. Until recently, an examiner could “storm[ ] into the bank, count[ ] the cash, add[ ] up the deposits, look[ ] at a sampling of the loans, and pronounce[ ] the work done.” 249 See Hawke, supra note 95, at 2. Today, “[t]he sheer depth of complexity that afflicts bank balance sheets prevents even experts from discerning what banks own and owe, what they sold and received, and whether they are compliant with . . . hundreds of banking statutes.” 250 Conti-Brown, supra note 102, at 165. At large banks, it takes a team of examiners many months to do what used to be wrapped up by one examiner in a half-day visit. 251 See Hawke, supra note 95, at 2–3.

More complex markets also require greater expertise, including advanced degrees, continuing education, and “leading experts in the most esoteric financial fields.” 252 See id. at 8. Regulatory monitors have varying back­grounds. In banking, examiners tend to have finance backgrounds. Oil inspectors often have engineering degrees. FDA drug reviewers are typi­cally scientists, doctors, or statisticians, 253 FDA’s Drug Review Process, FDA, [] (last updated Aug. 24, 2015). and many USDA facilities inspec­tors are veterinarians. 254 USDA Inspection, supra note 12, at 15. Agencies have raised salaries to accommo­date the additional educational requirements. 255 BSEE, United States Department of the Interior Budget Justifications and Performance Information Fiscal Year 2017, at 55, 64 (2017) [hereinafter BSEE Budget] (requesting more funding for inspectors due to “increased complexity in OCS oil and gas activities”).

As markets and businesses become more complex, monitors’ main object of analysis becomes more like lawyers’ main object of analysis—the law, which is also complex. Greater business sophistication may thus lessen the gap between monitors and lawyers, to the extent that both increasingly require great technical expertise.

2. Faster Innovation. — The rate of market changes has accelerated to unprecedented levels, meaning that many of today’s “routine” prod­ucts were until recently “exotic or nonexistent.” 256 See, e.g., Hawke, supra note 95, at 6. Therefore, new employees who join an agency will soon have large knowledge gaps with­out continual updates. They can obtain some of this through phone calls, conferences, and other voluntary mechanisms. 257 Coglianese et al., supra note 37, at 330. Yet much of the rele­vant information—the nature of Bank of America’s latest automated financial advisor or Ford’s self-driving car—is closely guarded as a trade secret and impenetrable from the outside. Complexity, secrecy, and inno­vation mean that inspectors “rely on industry representatives to explain the technology at a facility.” 258 Nat’l Comm’n on the BP Deepwater Horizon Oil Spill & Offshore Drilling, Deep Water: The Gulf Oil Disaster and the Future of Offshore Drilling 77 (2011) [hereinafter Deepwater Report], []; see also Conti-Brown, supra note 102, at 165.

Those  explanations  will  not  be  expressed  in  regulatory  monitors’  reports,   which  focus  on  violations. Nor  would  it  be  practical  or  even  legal  to  transmit  all  of  the  first-hand  data  observed  directly into a report. As a result, agencies’ other internal experts, such as scientists in the rulemak­ing division, will often lack understanding of the latest market develop­ments—an understanding that is indispensable for dynamic regulation. 259 Wendy Wagner et al., Dynamic Rulemaking, 92 N.Y.U. L. Rev. 183, 197 (2017) (positing that “some agencies operate in such rapidly changing technological environments that one would expect them to be adjusting their rules periodically to prevent entire programs from becoming obsolete”). Even if the raw monitoring data were somehow made available to agency actors other than monitors, processing that data would prove difficult for those who—unlike monitors—have not benefit­ted from industry repre­sentatives’ ongoing explanations.

Regulatory monitors may thus hold information monopolies com­pared not only to other legal actors but also to other technocrats in the agency, such as nonlawyer technical experts in the rule-writing depart­ment. Rapidly changing markets shift the locus of business expertise fur­ther inside the firm and, thereby, shift expertise within the agency more toward those who regularly operate inside the firm: regulatory monitors.

3. Technological Tools. — Every bureaucrat, including litigators, has more access to information than ever before. However, while information technologies can speed up legal research, they are less able to speed up court dockets or public notice-and-comment periods. To the contrary, information technologies enable more parties to participate in formal agency decisionmaking processes, even submitting tens of thousands of fake comments for proposed rules. 260 James V. Grimaldi & Paul Overberg, Many Comments Critical of ‘Fiduciary’ Rule Are Fake, Wall St. J. (Dec. 27, 2017), (on file with the Columbia Law Review). These advances slow down rule­making by increasing the information that must be processed and the stakeholders that must be managed.

In contrast, because regulatory monitors do not have the same exter­nal procedural constraints, their most substantial limit is the resources required to transmit and analyze information. When information submis­sion becomes too burdensome, businesses may object. Additionally, regu­latory monitors’ travel to business locations to look through paperwork has traditionally consumed considerable monitoring funds and time. Even if volumes of paperwork were obtained, human resources con­strained regulatory monitors’ ability to sift through that paperwork.

Technologies  have  reduced  these  barriers  by  providing  remote  monitoring  devices  that  continuously  transmit  data,  such  as  EPA  sensory  equipment  on  space  satellites  and  inside  factories  that  tracks businesses’ pollution. 261 See Esty, supra note 247, at 156. Billions of daily transactional data flow from energy compa­nies to FERC 262 2016 FERC Rep. on Enforcement 52,
2016/11-17-16-enforcement.pdf [] [hereinafter FERC Report].
and from securities firms to the SEC. 263 FINRA Report, supra note 212, at 1. Interagency pool­ing of these technologies multiplies the available data. 264 See, e.g., Report on NIH Collaborations with Other HHS Agencies for Fiscal Year 2017, NIH, [] (last updated June 30, 2018) (describing “interagency collaborations that enable agencies to combine their knowledge and diverse expertise to accomplish their collective mission”). Regulatory monitors then analyze these big data sets with advanced modeling and machine-learning algorithms. 265 See Cary Coglianese & David Lehr, Regulating by Robot: Administrative Decision Making in the Machine-Learning Era, 105 Geo. L.J. 1147, 1160–67 (2017). As a result, in various agencies, “on-site time as a percentage of overall examination hours dropped,” 266 See FINRA Report, supra note 212, at 5 (estimating a decrease from 32% to 19%). and “inspectors . . . conduct[ed] more thorough inspections.” 267 BSEE Budget, supra note 255, at 32. Today, hold­ing employees constant, regulatory monitors can process more nonpub­lic data more thoroughly, extending the reach of their core authority.

Thus, unlike in the mid-1800s, the appearance of national bank examiners today is less likely to get “the bank force . . . dancing at [their] beck and call.” 268 See Henry, supra note 100, at 241; see also Hawke, supra note 95. Instead, modern regulatory monitors more suitably meet with a senior executive or engineer running a large, self-regulating compliance system. Technologies convert what was previously a “one-time snapshot of performance taken on a particular inspection day” to a “‘movie’ of the plant’s processes.” 269 See Freeman, supra note 30, at 60 (quoting Interview with Bill Patton, Director of XL, EPA Region 4 (Mar. 14, 1997)) (describing EPA upgrades); see also Hawke, supra note 95, at 9 (describing the OCC’s “ongoing . . . on- and off-site monitoring”). Disruption is minimized because in some industries firms never stop working for—or collaborating with—regulatory monitors.

III. An Overview of Regulatory Monitors Today

The discussion so far has shown that changes over the past century in statutes, governance, and markets have formed the foundation for regulatory monitors’ ascendancy to a lead role within the administrative state. But authority on the books and authority demanded by exter­nal realities do not necessarily translate into authority used. Courts have held that an agency’s decisions about the extent to which it “‘monitors’ as well as ‘enforces’ compliance fall squarely within the agency’s exercise of discretion.” 270 Gillis v. U.S. Dep’t of Health & Human Servs., 759 F.2d 565, 576 (6th Cir. 1985); see also Madison-Hughes v. Shalala, 80 F.3d 1121, 1129–31 (6th Cir. 1996) (ruling that the Department of Health and Human Services’ decision not to collect data about racial disparities in health services was unreviewable). Inertia and internal politics influence organizational design. While the recent literature has helped lay the foundations for understanding why monitoring has become important, empirical evi­dence of actual regulatory monitors exercising that authority has been anecdotal or localized.

A fundamental empirical question thus remains unanswered: How big a role do regulatory monitors play in the regulatory state today? More specifically, how do regulatory monitors influence the administration of the law? While recognizing that “the sheer bewildering heterogeneity of the administrative state makes it impossible to generalize about the allo­cation effects of agency structure,” 271 Magill & Vermeule, supra note 38, at 1059. this Part provides the first systematic empirical evidence of regulatory monitors’ place in the federal govern­ment. That evidence begins to map out key agency organizational design choices shaping regulatory monitors’ influence.

A. Monitoring Firms

Resource allocation is one of many “modes of governance” 272 See Rubin, supra note 67, at 97 (noting that resource allocation is a “new mode[ ] of governance” not recognized by the Administrative Procedure Act). through which political leaders exercise power. 273 Eric Biber, The Importance of Resource Allocation in Administrative Law, 60 Admin. L. Rev. 1, 17 (2008) (discussing the “centrality of resource allocation to decisionmaking” and noting that Congress, the President, and other executive officers direct agency resources to prioritize “different problems, concerns, dreams, and goals”); see also Oil, Chem. & Atomic Workers Union v. OSHA, 145 F.3d 120, 123 (3d Cir. 1998) (denying a petition that would have the court “intrude into the quintessential discretion of the Secretary of Labor to allocate OSHA’s resources and set its priorities”). Statutes commonly pro­vide an “incomplete design,” leaving agency heads to finish the task of deciding how many regulatory monitors and lawyers to hire, as well as how to use them. 274 See, e.g., Communications Act of 1934, Pub. L. No. 73-416, § 4(f), 48 Stat. 1064, 1067 (codified at 47 U.S.C. § 154(f) (2012)); Mitchell Pearsall Reich, Incomplete Designs, 94 Tex. L. Rev. 807, 810 (2016) (explaining “the implicit delegation of institutional decisions to downstream actors”). This section provides the first data on how these decisions have allocated regulatory monitoring and legal resources across all large U.S. regulators. 275 For a description of how the agencies were chosen, see supra section I.B.

In many agencies—such as banking regulators, the Mine Safety and Health Administration, and the USDA’s Food Safety & Inspection Service—the federal personnel database or some public report provide a clear figure for the number of personnel devoted to monitoring. 276 See FedScope, supra note 74. They are supplemented by interviews, annual reports, and other sources as necessary. For instance, the Federal Reserve does not report its personnel, which necessitated relying on annual reports and interviews. In other agencies, such as the FCC, FDA, and EPA, monitors are officially listed in other categories such as scientists, veterinarians, and engineers. A category was counted toward an agency’s monitor total only when other sources suggested that it was mostly comprised of monitors. It is possible that some of these categories include personnel who do not directly monitor, which would cause my figures to overstate the number of monitors. It is also possible that other categories include monitors that I was unable to identify, thereby causing my figures to understate moni­tors’ presence in some agencies. Assumptions are noted in the appen­dices, and more focused study of those agencies’ subcategories would be needed to obtain more precise figures.

Data constraints also limit the figures for legal personnel. Although the main object of comparison here is between enforcement lawyers and monitors, for most regulators the legal figures available combine all legal positions—including those working in rule writing and the office of the general counsel. Consequently, the proportions below understate moni­tors’ presence relative to enforcement lawyers.

Among the nineteen agencies studied, only three—the FTC, NLRB, and EEOC—have relatively few regulatory-monitor personnel. These three are litigator-dominant, with law-related employees comprising over 85% of the total pool of regulatory-monitor and legal personnel. 277 See infra Appendix A. Those three are also the only agencies in the set that have no visitation authority. 278 See supra section I.A. Inter­views indicated that most of these agencies’ lawyers litigate. 279 Interview with FTC Bureau of Consumer Protection Employee (Apr. 12, 2017) [hereinafter FTC Interview]; Telephone Interview with EEOC Employee (Apr. 25, 2017); Telephone Interview with NLRB Employee (Apr. 4, 2017). This clas­sification as litigator-dominant differs from a prominent 1980s descriptor of some agency groups as “legalistic,” a term which could apply to regula­tory monitors. 280 The term “legalistic” is a broader concept that was used to describe, for example, some types of inspectors who operated in a more by-the-book manner. See Bardach & Kagan, supra note 41, at 93 (illustrating this concept).

The remaining sixteen agencies all have material numbers of regula­tory monitors, both in absolute terms and relative to legal personnel. The five hybrids have some balance between the groups: the CFPB, EPA, FCC, FERC, and SEC. 281 See infra Appendix A. It is worth noting that the FCC has a considerably lower percentage of monitors, and is the only one of these with fewer monitors than lawyers, suggesting that its commitment to monitoring could also have meaningful distinctions. In the remaining eleven agencies, regulatory monitors make up over 85% of the combined regulatory-monitor and legal workforce, making them monitor-dominant. 282 See infra Appendix A.

Figure 1: Monitors at Large Agencies


To what extent do personnel reflect monitoring activity? That ques­tion is one of the many in administrative law lacking empirical evidence showing the connection between agency design and agency behavior. 283 See Christopher R. Berry & Jacob E. Gersen, Agency Design and Political Control, 126 Yale L.J. 1002, 1007 (2017) (“[T]here has been very little quantitative scholarship that establishes a link between agency design and a similar agency output across agencies or over time.”). Activity data is less consistently available and comparable than human-resource data. 284 See infra section IV.A.1. Any given agency might decide to devote the same num­ber of workers to a small number of thorough inspections or a large number of light-touch inspections, meaning that one cannot infer that the agency with fewer inspections is monitoring less. Nor can this Article establish a definitive link between design and behavior. Nonetheless, as common sense would indicate, agencies with larger regulatory-monitor workforces (both hybrids and monitor-dominant agencies) tend to report more extensive monitoring activity. 285 See infra Appendix A.

Even litigator-dominant agencies exercise some amount of statutory monitoring authority, but their monitoring comprises a small part of their information collection. For example, the litigator-dominant EEOC uses its confidential data collected on gender and racial breakdowns to launch systemic discrimination investigations, but those account for less than 1% of its total investigations. 286 FY 2016 EEOC Performance & Accountability Rep. 12, 93, [] (identifying 245 systemic, agency-initiated Commissioner Charges and directed investigations in contrast to the 91,503 total charges investigated); see also EEOC, A Review of the Systemic Program of the U.S. Equal Employment Opportunity Commission 16 (2016), [] (explaining that “Commissioner Charges and directed investigations” are used “when the agency learns of a problem or there is reason to believe that discrimination may be more widespread or of a different nature than an individual charge alleges”). The EEOC receives cases mostly from employees. See 2016 EEOC Performance & Accountability Rep., supra, at 34. Although FTC competition lawyers regularly rely on a key monitoring program—premerger report submis­sions—for consumer protection, the agency depends on nonstatutorily acquired information sources such as industry conferences, online con­sumer complaints, or litigators watching television in search of deceptive ads. 287 See Lesley Fair, The Truth About False Advertising, Presentation at Boston University 16 (Apr. 14, 2017) (on file with the Columbia Law Review) (explaining the FTC’s “Ad Monitoring” and other sources of information in a presentation by an FTC attorney attended by the author).

The remaining sixteen agencies—84% of the group—conduct signi­ficant monitoring, albeit with great variation. 288 See infra Appendix A. Among hybrid agen­cies, for instance, the EPA completes over ten thousand on-site inspections annually. 289 See infra Appendix A. The FERC and the SEC analyze large volumes of business records and transactional data. 290 See infra Appendix A; see also FERC Report, supra note 262, at 34–35 (describing FERC’s extensive audit and accounting division); FY 2017 SEC Cong. Budget Justification 6–7 [hereinafter SEC Budget], [] (noting that “analysis of large datasets, including . . . trading data in equities, options, municipal bonds, and other securities” is important to detect misconduct and describing the SEC’s plan to “improve[ ] data analysis capabilities” by “invest[ing] in IT”). The CFPB has extensive on-site and remote records-examination programs, while the FCC inspects television and radio broadcasters nationwide and regularly collects business records. See infra Appendix A.

Monitor-dominant  agencies  tend  to  have  higher  monitoring  volumes  and  a  greater  likelihood  of  continuous  presence. In  2016,    the  FDA  con­ducted  164,696  surprise  tobacco  inspections  alone,   of retailers  ranging from  CVS  to  mom-and-pop  stores. 291 See Compliance Check Inspections of Tobacco Product Retailers, FDA, (on file with the Columbia Law Review) [hereinafter Compliance Check] (last visited Oct. 11, 2018). The NRC’s “resident inspectors” 292 Assessment of Efficiencies to Be Gained by Consolidating or Eliminating Regional Offices, NRC, [https://] (last visited Oct. 11, 2018).
and the Federal Reserve’s “examination teams” 293 See Levitin, supra note 52, at 2044. provide a year-round presence at nuclear plants and the largest banks.

Personnel numbers and activity figures provide only a partial per­spective on institutional design. Agencies with the same proportion of employees may distribute authority dissimilarly through divergent struc­tural decisions. Regulators may enforce only a small portion of the agency’s authority through on-site visits, as is the case with FCC television and radio station inspections, or a broader array of activities, as is the case with the CFPB examinations of financial institutions. 294 Interview with FCC Senior Attorney (Apr. 13, 2017) [hereinafter FCC Interview] (describing how engineers regularly inspect stations and both engineers and lawyers analyze mandatory reports submitted); Interview with Private Sector Attorney (Apr. 26, 2017) (stating that his clients, communication-sector companies, must regularly submit large volumes of information to the FCC); CFPB Interview, supra note 218. The follow­ing sections discuss those and other high-impact design choices. Nonethe­less, if the literature is correct that personnel numbers reflect power and priorities, 295 See supra note 273 and accompanying text. only 16% of the major regulators studied clearly favor lawyers, while more than half heavily prioritize regulatory monitors. 296 See infra Appendix A; see also supra Figure 1.

B. Enforcing Law

Regulatory monitors, like police officers, do more than patrol. To varying degrees across agencies, they also make enforcement decisions. Agencies have a “graduated enforcement continuum” 297 See, e.g., BSEE Annual Report, supra note 19, at 23. ranging from warning letters to prosecution. Figure 2 provides one illustration in which “the proportion of space at each layer represents the proportion of enforcement activity.” 298 See Ayres & Braithwaite, supra note 30, at 35. At the larger bottom layer of the pyramid are persuasion and warning letters, and above is smaller space for formal procedures such as civil penalties. 299 See id. The pyramid does not speak directly to groups within the agency, but it implies that those manag­ing the bottom layer of mostly unreviewable conduct control a large por­tion of enforcement. 300 Ayres and Braithwaite provide examples of regulatory monitors only in passing, and they do not explore the implications of responsive regulation for various internal agency groups. See id.

An agency’s designers can set up organizational processes that require regulatory monitors to hand over a case at the first sign of wrong­doing, reserving almost all major enforcement decisions in the pyramid for other groups, such as enforcement lawyers. Litigator-dominant agen­cies tend to adopt such a structure. Regulatory monitors at hybrid and monitor-dominant agencies, however, play a meaningful role in decisions far along the enforcement spectrum. Some regulatory monitors even act as something close to a prosecutor. An overview of that enforcement partic­ipation follows, broken down into (1) citations, recommendations, and warnings; (2) blocking business activities; (3) public shaming; (4) increased monitoring as punishment; and (5) control over investigations and charges.

Figure 2: Sample Enforcement Pyramid 301 This figure is based on Ayres & Braithwaite, supra note 30, at 35.



1. Citations, Recommendations, and Warnings. — Beginning at the base levels of the pyramid, there is evidence that regulatory monitors drive this enforcement activity at fifteen of the nineteen largest regula­tors. 302 This includes all agencies except the FCC, EEOC, NLRB, and FTC. See infra Appendix B. For example, FERC monitors possess the authority to issue public “noncompliance” notifications and direct nonpublic settlement agree­ments. 303 See, e.g., FERC Report, supra note 262, at 39. Although  not  all  agencies  release  such  figures,   those  that  are  available  in  agency  reports  reflect  the pyramid’s  space  allocation in  that  the  quantity  of  less formal activity is significantly greater than more for­mal proceedings. 304 See infra Appendix B. For instance, in fiscal year 2016, the FDA’s inspec­tions group issued 14,590 warning letters, while its legal division took only twenty-one enforcement actions. 305 FDA Enforcement Statistics Summary Fiscal Year 2016, FDA, [] [hereinafter FDA Enforcement] (last visited Oct. 11, 2018). Used here, the term “enforcement actions” encompasses injunctions and seizures. See id.

In  terms  of  behavioral  impact,   these  recommendations  can  be  far-reaching. Compliance  varies  across  time  and  agencies,   but  there   are  indica­tions  that  in  diverse  industries  companies  cooperate  when  infor­mally  advised  to  take  a  course  of  action. 306 See FERC Report, supra note 262, at 35 (reporting that in fiscal year 2016, energy companies implemented 98% of FERC’s “audit recommendations” within six months); Richard M. Cooper & John R. Fleder, Responding to a Form 483 or Warning Letter: A Practical Guide, 60 Food & Drug L.J. 479, 480 (2005) (noting that food companies typically comply with FDA inspectors’ requests); Interview with Former FDIC Employee (Mar. 10, 2017) [hereinafter FDIC Interview] (stating that financial institutions “almost always” comply with examiners’ requests). Even the recommendations of regulatory monitors at hybrid agencies can lead to substantial payouts, albeit less than those of litigators. In a recent six-month period, CFPB examinations prompted financial institutions to refund $44 million to consumers, while the enforcement group secured $82 million. 307 2016 CFPB Semi-Ann. Rep. 11 [hereinafter CFPB Report], [https://]. At FERC, auditors identified energy-company noncompliance that led to customer refunds and price reductions amounting to $5.3 million, less than a third of the $18 million for litigators. See FERC Report, supra note 262, at 12, 39.

Why would a firm comply with these expensive recommendations? 308 Cf. Parrillo, supra note 18, at 37 (discussing factors that incentivize regulated parties to follow guidance, including: “(A) pre-approval requirements, (B) investment in relationships to the agency, (C) intra-firm constituencies for compliance beyond legal requirements, and (D) the risks associated with one-off enforcement”). Despite being “advisory,” they carry the threat of harsher follow-up. As the FDA’s manual notes, the warning letter provides “an opportunity to take voluntary and prompt corrective action before [FDA] initiates an enforcement action.” 309 See FDA Manual, supra note 230, at 2. Moreover, regulatory monitors’ requests may not need backup from an agency’s litigation group, as the rest of this section explains.

2. Blocking Business Activity. — A more intrusive enforcement power comes in the form of preventing business operations ex ante or suspend­ing market access ex post. In at least eleven of the nineteen agencies, regulatory monitors exercise such authority. 310 The eleven agencies are the FDA, OCC, USDA (FSIS), FAA, FCC, FDIC, Federal Reserve, FMCSA, MSHA, SEC, and NRC. See infra Appendix B. Ex ante approval may be required only for new activities, such as launching new medical devices or opening a new bank branch. 311 See 12 C.F.R. § 303.40 (2018) (noting that banks must apply to the FDIC before establishing a branch); About FDA Product Approval, FDA, [] (last updated Dec. 29, 2017) (explaining which products are subject to ex ante review by the FDA). Other times agencies must approve daily activities, as is the case for every chicken carcass sold in the United States. 312 See USDA Inspection, supra note 12, at 15.

After a product enters the market, many regulatory monitors can order or request a halt in operations. Federal regulators can recall toys, automobiles, and food based on health or safety concerns. 313 See, e.g., Recalls, Market Withdrawals, & Safety Alerts, FDA, [] (last updated Sept. 27, 2018) (describing the scope of the FDA’s food recall powers and listing recent recalls); Safety Issues and Recalls, NHTSA, [] (last visited Dec. 1, 2018) (describing the NHTSA’s recall program); Toy Recall Statistics, Consumer Prod. Safety Comm’n, [] (last visited Dec. 1, 2018) (noting the number of toys recalled in each year from 2008–2018). Environmental inspectors can shut down companies that are discharging hazardous chemicals. 314 See 30 C.F.R. § 250.101 (2018) (providing an overview of BSEE’s authority); BSEE Annual Report, supra note 19, at 23–24 (describing BSEE’s enforcement approach and listing various incidents of noncompliance that the agency addressed in 2015); Telephone Interview with Former EPA Employee (Apr. 12, 2017) [hereinafter EPA Interview]. Restraints on business activity can significantly hurt a firm, both in terms of immediate lost revenues and longer-term loss of clients driven away by the disruption.

3. Public Shaming. — Whereas the other categories of sanctions rely on directly punishing the business, public shaming takes an indirect approach. Many agencies publicly post the name of the business along­side the violations identified by regulatory monitors. 315 In other industries, such as finance, examiners’ reports are private. The CFPB aggregate reports provide some detail about its examiners’ findings without identifying companies. See CFPB Report , supra note 307, at 75. One can learn, for example, that in 2014, oil inspectors shut down certain offshore Exxon operations thirteen times. 316 Incidents of Non-Compliance (INCs) Online Query, BSEE, https:// (on file with the Columbia Law Review) (last updated Feb. 3, 2019) (querying for INCs issued between January 1, 2014 and December 31, 2014).
A January 27, 2017, OSHA inspection of an Amazon warehouse uncovered a “serious” worker health violation leading to a $5,975 fine. 317 Inspection Detail, OSHA, [] (last visited Oct. 11, 2018). On  March 2,  2017,   FDA  inspectors  caught  Walmart  selling  tobacco  to  minors  in  cities  ranging  from  Memphis,  Tennessee,   to Scottsdale,   Arizona. 318 See FDA, No. 17AZ000611, Warning Letter Regarding Tobacco Retailer Inspection Violations, to Wal-Mart (Mar. 2, 2017), [] (notifying a Scottsdale, Arizona, Walmart that it violated federal tobacco laws and regulations by selling VUSE Menthol e-liquid to a minor); FDA, No. 17TN001357, Warning Letter Regarding Tobacco Retailer Inspection Violations, to Wal-Mart #1248, (Mar. 2, 2017), [] (notifying a Memphis, Tennessee, Walmart that it violated federal tobacco laws and regulations by selling electronic nicotine delivery system products to a minor).

The posting of such information can be seen as a form of transpar­ency—a means for the public to know what their government agents are doing—rather than as a sanction. But companies fear bad regulatory publicity, a risk that has grown in the internet era because sanction results can spread more easily. 319 See Nathan Cortez, Adverse Publicity by Administrative Agencies in the Internet Era, 2011 BYU L. Rev. 1371, 1373 (describing the use of negative publicity as an enforcement tactic employed by federal regulators). Given that a few thousand dollars in fines is insignificant to a large company, the public posting of monitoring violations enables some regulatory monitors to have greater enforcement power over businesses.

4. The Process as Punishment. — Another indirect enforcement mech­anism is agencies’ discretion to increase monitoring intensity. 320 rofessor Rubin has mentioned this as a possible use of monitoring. See Rubin, supra note 67, at 125 (“Agencies can use investigations themselves—repeated visits by inspectors or demands for documents—as sanctions.”). Regulators sometimes formally announce that good behavior will lessen oversight. 321 See, e.g., Parrillo, supra note 18, at 45 (“The relationship between an agency and a regulated party . . . may operate at an institutional and official level, if, say, the agency has an announced policy of reducing the frequency of inspections for parties who have a good track record.”). But they stop short of publicly describing monitoring as pun­ishment, which might provoke court challenges. 322 For example, that could imply that the inspection was a final determination of rights or not part of an “administrative plan.” See Marshall v. Barlow’s, Inc., 436 U.S. 307, 321 (1978) (holding, in part, that the Constitution requires agency searches of commercial facilities to be part of a “general administrative plan”).

Nonetheless, some agencies communicate that monitoring is both a consequence and a reward. OSHA, for instance, has a Voluntary Protec­tion Program in “recognition of the outstanding efforts of employers,” 323 All About VPP, OSHA, [] (last visited Oct. 11, 2018). which rewards firms by subjecting them to fewer inspections. 324 OMB Watch, supra note 60, at 6–7. OSHA’s “Severe Violator Enforcement Program” involves higher penalties and “increased OSHA inspections in these worksites, including mandatory OSHA follow-up inspections, and inspections of other worksites [owned by the violator].” 325 Press Release, OSHA, US Department of Labor’s OSHA Takes Action to Protect America’s Workers with Severe Violator Program and Increased Penalties (Apr. 22, 2010), []. The  agency  explains  this  policy  by  noting  that  “[h]igher  penalties  and  more  aggressive,   targeted  enforcement  will  provide  a  greater  deterrent.” 326 See id. The EPA’s audit policy program officially offers only reduced penalties for violations as a reward for good behavior, but a statistical study found that well-behaving firms were also subject to fewer inspections, even controlling for other factors. 327 See Parrillo, supra note 18, at 52.

Regulatory monitors’ scrutiny can be costly to firms, 328 See Freeman, supra note 30, at 14–17. and firms pre­dictably seek to avoid intense monitoring. 329 For instance, lawyers warn that a firm ignoring an FDA inspector’s request is “likely to be subject to extraordinarily intense and more frequent inspections.” Cooper & Fleder, supra note 306, at 480. In negotiated rulemaking with the EPA, industry representatives have pushed for rewarding exem­plary firms by giving them “tax credits” and “less frequent inspection audits.” 330 See Freeman, supra note 30, at 67. Thus, the threat of increased scrutiny provides one avenue for regulatory monitors to obtain compliance even without direct sanction authority.

5. Investigations and Charges. — For more significant sanctions, such as large fines and the revocation of licenses, an investiga­tory phase typi­cally follows the regulator’s identification of a violation. Regulators can allocate control over that investigatory process to differ­ent groups. At agencies with sizeable litigation divisions, such as at the SEC, enforce­ment lawyers control much of the investigatory function because they have their own investigation resources. Even at such agen­cies, regulatory monitors’ influence can extend beyond the handoff if the enforcement lawyer seeks regulatory monitors’ expertise or if regula­tory monitors originated the case. But regulatory monitors wield less influence overall in such agencies.

Agencies with smaller legal groups rely more on the inspector to investigate. FAA inspectors will investigate and recommend an airline’s civil penalty or a pilot’s suspension before attorneys take over the case. 331 See L. Ronald Jorgensen, The Defense of Aviation Mechanics and Repair Facilities from Enforcement Actions of the Federal Aviation Administration, 54 J. Air L. & Com. 349, 375 (1988); Peyton H. Robinson, An Overview of FAA Enforcement Actions, Utah B.J., Nov./Dec. 2012, at 29, 29–31 (describing the steps taken by FAA monitors before FAA attorneys become involved). The SEC and FAA models allow attorneys to decide the formal charges, but those models still reflect the relationships in federal criminal law enforcement, in which “iterated interactions between agents and prosecutors will affect investigative and adjudicative decisionmaking.” 332 See Daniel Richman, Prosecutors and Their Agents, Agents and Their Prosecutors, 103 Colum. L. Rev. 749, 751–52, 766–67 (2003).

Alternatively, regulatory monitors may lead cases through the formal charge phase. When an explosion or death occurs on an offshore oil platform, inspectors investigate and build the “case” for civil penalties. 333 See Telephone Interview with BSEE Employees (Mar. 31, 2017) [hereinafter BSEE Interview]; Civil Penalties Assessments and Appeals, BSEE, [] [hereinafter BSEE Civil Penalties] (last visited Oct. 12, 2018) (describing the process for investigating and building a case file in the event of a violation). Based on the inspector’s case and the company’s response, “the Review­ing Officer will issue a decision identifying the amount of any final civil penalty.” 334 See BSEE Civil Penalties, supra note 333 (emphasis added). That process led to over $6 million in civil penalties in 2015. 335 See BSEE Annual Report, supra note 19, at 23–24. OSHA inspectors in the vast majority of cases set fines and negotiate final settlements with businesses without ever involving litiga­tors. 336 See OSHA Interview, supra note 148. After OSHA inspectors and their supervisors decide on civil penalties, companies may then pay, negotiate, or file a legal appeal. See id. By one regional leadership’s estimate, firms rarely appeal, and about 80% of the time a negotiation ensues. See id. OSHA inspectors do not usually involve solicitors unless the negotiations falter. See id. Thus, regulatory monitors may serve as investigators, prosecutors, and de facto final decisionmakers.

In summary, the confluence of case-specific sanction control, as well as the degree of regulatory monitors’ information monopoly, 337 See supra section II.B.2. provides an overall sense of their influence over agency enforcement. Difficulties arise in comparing the external impact of regulatory monitors and litiga­tors. One legal case or rule can establish an industry standard. Tens of thousands of warning letters, incidences of noncompliance, and cita­tions do not attract as much attention as a $415 million SEC legal settle­ment with Merrill Lynch. 338 See, e.g., Suzanne Barlyn, Merrill Lynch to Pay $415 Million for Misusing Customer Cash: SEC, Reuters (June 23, 2016), []. But institutionalized through large firms’ compliance systems, and spread across millions of transactions, even non­quantifiable regulatory monitors’ interventions can have far-reaching impact.

Despite variation and comparison difficulties, regulatory monitors in at least fifteen of the nineteen large agencies have significant enforce­ment influence in several of the categories described above. 339 See infra Appendix B (detailing the techniques that monitors at the nineteen large agencies utilize to sanction firms). There was insufficient evidence to conclude that regulatory monitors at the FCC, FTC, EEOC, and NLRB had significant influence. See infra Appendix B. Further research into the inner workings of these agencies could produce such evidence, particularly at the FCC, which has a significant number of monitors and amount of monitoring activity. See infra notes 479–481, 517–519 and accompanying text. Multiple levers—including statutory authority, workforce size, internal infor­mation reliance, formal sanctions, and planning—can shift influence away from the legal division. As more of these levers align at a given agency and across the administrative state, regulatory monitors become the drivers of regulatory enforcement.

C. Making Law

Agencies make law through their determinations in individual cases and by issuing broader rules. Regulatory monitors contribute to each of these areas of policy development.

1. Creating Common Law. — Since the 1990s, FTC enforcement law­yers have created a common law of privacy with “hardly any judicial deci­sions to show for it.” 340 See Daniel J. Solove & Woodrow Hartzog, The FTC and the New Common Law of Privacy, 114 Colum. L. Rev. 583, 585 (2014). FTC lawyers have done so through settlement agreements, which set industry-wide practices. 341 See id. Individual regulatory-monitor determinations can have a similar effect. A plethora of reports, warnings, and other monitor decision results are available online. 342 See infra notes 370–372 and accompanying text. These documents offer great detail. For instance, one of the FDA’s 17,000 warning letters from 2015 reveals that during a Deerfield, Illinois, inspection of Walgreens’s over-the-counter drug preparation, the “[i]nvestigator observed what appeared to be hundreds of dead insects” throughout the facilities, and a follow-up laboratory analysis detected “spore-forming bacteria.” 343 FDA, 2017-DAL-WL-01, Warning Letter on Walgreens Infusion Services, to Paul Mastrapa, Chief Executive Officer, Option Care Enters., Inc. (Oct. 19, 2016), []. The FDA’s recommendations to Walgreens regarding behavioral changes are also specific. 344 See id. (requiring the laboratory management to assess operations, including “the prevention, destruction, repellence, or mitigation of the specific pests that were found in the warehouse” and in particular to “assess [the] aseptic processing operations” using a third-party consultant).

Like a lawyer to a judge, firms use these texts to plead their case. 345 See EPA Interview, supra note 314 (noting that companies use decisions from one site to negotiate with the EPA for different sites); OSHA Interview, supra note 148 (noting that attorneys routinely rely on OSHA citations to gather information about violations and develop the nuances of a case). The firm might argue that in a prior inspection at a different firm, simi­lar observations led to different recommendations. The EPA has warned its inspectors to follow national procedures because “[p]olicy decisions at one facility can have a precedential effect on all other facilities.” 346 EPA, Memorandum on Final National Policy: Role of the EPA Inspector in Providing Compliance Assistance During Inspections (2003),
production/files/2013-09/documents/inspectorrole.pdf [].
Firms  study  regulatory  monitors’  reports  to  learn  how  to  operate  in  the  future.  Since  the  reports  can  contain  specific  recommendations  not  required  by  law, 347 See supra section II.A.2. these regulatory monitors—and those who oversee them—wield the ability to not only interpret law but to create it.

2. Writing Rules. — Regulatory monitors’ most straightforward form of soft rulemaking is the writing of their employee manuals. Often running close to a thousand pages in length, these manu­als give instruc­tions as to what information the regulatory monitors should collect and how they should analyze the data they observe. 348 See EPA, EPA Pub. No. 305-K-17-001, NPDES Compliance Inspection Manual (2017), [] (totaling 918 pages); CFPB, CFPB Supervision and Examination Manual (2012),
and-examination-manual-v2.pdf [] (totaling 924 pages).
Firms meticulously study these texts to adjust behavior. 349 See McGarity, supra note 228, at 1393–96 (providing an example of a waste generator examining agency text for guidance). Manuals are most influential in indus­tries governed by best practices and principles-based rules, which are more subject to interpretation than in industries with detailed codes for every violation. 350 See supra section II.A.2. Manuals do not serve as the sole basis for court enforcement unless the agency treats them as substantive law and pro­cesses them through notice and comment. 351 See United States v. Bioclinical Sys., Inc., 666 F. Supp. 82, 83 (D. Md. 1987) (“Congress has mandated that a full and deliberate public process, including the making of recommendations by a broad-based advisory committee and the opportunity for public hearing, be followed before the FDA may establish a GMP.”). But a firm may still choose to follow the manual simply because it reflects the expectations of a powerful government actor. 352 See supra section III.B.1.

In a minority of industries, such as finance, regulatory monitors also lead formal rulemaking related to their expertise. 353 See FERC Report, supra note 262, at 58 (describing a FERC regulatory monitor’s recent writing of a rule for notice and comment); BSEE Interview, supra note 333 (stating that Department of the Interior regulatory monitors draft offshore-energy regulations). In those agencies, it would be standard for agency directors or the general counsel ultimately to scrutinize any rules written by regulatory monitors before subjecting them to notice and comment. 354 See Raymond P. Baldwin & Livingston Hall, Using Government Lawyers to Animate Bureaucracy, 63 Yale L.J 197, 198 (1953) (“The stated duties of an Office of General Counsel include: . . . preparing and reviewing administrative rules, regulations and reports, and drafting proposed legislation; and . . . participating in the policy-making process of the agency.”).

Regulatory monitors’ expertise enables them to influence both for­mal and soft rulemaking, but organizational configurations can lessen information asymmetries. Some agencies mandate the sharing of regula­tory monitors’ reports with a separate rulemaking group, which analyzes the reports for trends. 355 See supra section II.B; cf. Nou, supra note 42, at 425–31 (discussing broadly similar mechanisms). At many agencies, the regulatory monitors’ divi­sion leads authorship of manuals, subject to legal review. 356 See Bioclinical Sys., Inc., 666 F. Supp. at 83–84 (suggesting that the FDA’s Office of Compliance writes its “inspectional guidelines,” which are then published by the Center for Devices and Radiological Health); CFPB Interview, supra note 218; OSHA Interview, supra note 148. Others assign the manual writing to the rulemaking group, giving external groups more control over regulatory monitor–related policymaking. 357 See, e.g., USDA Inspection, supra note 12, at 18 (“[The Office of] Policy and Program Development develops regulations as well as instructions for inspectors to imple­ment these regulations . . . .”).

However, the location of the individuals managing the process does not give the full picture. The manuals are hundreds of pages long and often delve into esoteric considerations such as, in the case of FAA flight inspectors, the need to avoid “signals . . . that are greater than 48 µA in the 90 Hz direction from the glide slope crosspointer value.” 358 FAA, United States Standard Flight Inspection Manual 15-65 (2015), []. The rules themselves may be similarly detailed. Due to the technical density, even when the rulemaking group writes manuals or rules they may need help drafting the text unless they previously served as regulatory monitors. As a former EPA senior attorney described one major rulemaking process, a manual writer in Washington, D.C., without any field experience managed a working group of regional inspectors to draft the actual text. 359 EPA Interview, supra note 314.

IV. The Accountability Framework for Regulatory Monitors

The previous Part showed the breadth and structure of modern reg­ulatory monitors’ power. An individual regulatory monitor’s impact is rarely as salient as Dr. Kelsey’s was during the thalidomide period. 360 See supra notes 56–59 and accompanying text. Instead, such life-altering regulatory-monitor impact is broadly institu­tionalized. The FAA articulates the organizational trifecta by describing its inspectors as serving to “develop, administer, and enforce the regula­tions and standards relating to aviation safety.” 361 OPM, Position Classification Standard for Aviation Safety Series, GS-1825, at 2 (1973), []. These functions create a virtuous cycle. Regulatory monitors regularly write or advocate for rules and policies that give them more data. 362 See, e.g., Amendments to Form ADV and Investment Advisers Act Rules, 80 Fed. Reg. 33,718, 33,718–19 (June 12, 2015) (proposing significant new reporting require­ments for registered investment companies); FERC Report, supra note 262, at 52, 58 (proposing new energy-data submission requirements). Better data equips them to more forcefully advocate policy and enforcement priorities. As would be expected in an administrative state beset by rule ossification and intent on informed collaboration with industry, regulatory monitors have emerged in the compliance era wielding considerable administrative power.

The claim that regulatory monitors lie at the heart of the regulatory state implicates prominent administrative law and policy debates. With the administrative lens adjusted for their full status, they inevitably become targets in the tug-of-war among Congress, the Presi­dent, and interest groups for external control over agencies. 363 Currently, various stakeholders outside the agency can influence regulatory monitors. One study of President Obama’s first year cited mostly regulatory monitors’ activity in concluding that agencies “appear to be exercising their enforcement authority more strenuously than they had in recent years.” See OMB Watch, supra note 60, at 4. As President Trump has sought to reorganize the executive branch, regulatory monitors have provided options. See supra notes 62–63 and accompanying text. Regula­tory monitors also necessarily compete with other internal groups for influence over the agency’s actions. This Part takes up the questions of external and internal influence in turn, and identifies a set of legal and organizational design choices that determine how regulatory monitors can best serve their agencies’ missions.

A. External Accountability Mechanisms

An emerging web of legal and organizational constraints influences regulatory monitors’ accountability. 364 See Lemos, supra note 46, at 946; see also Jacob E. Gersen & Matthew C. Stephenson, Over-Accountability, 6 J. Legal Analysis 185, 186–87 (2014) (arguing that more accountability is not always necessarily in the public’s best interests). Both laws and organizational design alter the balance of accountability and independence. Some of these constraints guard against inactivity, while others guard against excess. This section describes the existing accountability framework and lays the groundwork for its expansion by showing how it is fragile, inconsistent, and incomplete.

1. Public Disclosures. — Visibility can bring accountability to une­lected officials, in the broader sense of improving the exercise of author­ity. Immediately after her 1981 appointment by President Reagan, EPA Administrator Ann Gorsuch suspended hazardous waste rules and reduced legal cases by 84%. 365 See Richard J. Lazarus, The Tragedy of Distrust in the Implementation of Federal Environmental Law, Law & Contemp. Probs., Autumn 1991, at 311, 344. An “awakened, angry and energized public,” 366 See William D. Ruckelshaus, Opinion, A Lesson Trump and the E.P.A. Should Heed, N.Y. Times (Mar. 7, 2017), (on file with the Columbia Law Review). sensing that businesses had captured the agency, paved the way for Gorsuch’s resignation in less than two years. 367 See Lazarus, supra note 365, at 344–46. Visibility can also curtail excesses, as demonstrated by the increased oversight that viral videos of police officer abuses prompted. 368 Scott Calvert & Valerie Bauerlein, Viral Videos Shape Views of Conduct, Wall St. J. (Dec. 30, 2015), (on file with the Columbia Law Review).

Changes to regulatory monitors are less salient. Whereas agency rules and litigation are by default public, regulatory monitors’ reports need not be. Bank examiners and occupational inspectors—unlike police officers and enforcement lawyers—operate mostly in private spaces, mak­ing it difficult for third parties to document excesses. 369 See, e.g., OCC, PPM 5310-3, Policies and Procedures Manual: Bank Supervision 14 (2017) [hereinafter OCC, 2017 Manual], []; OCC, PPM 5000-7, Policies and Procedures Manual: Bank Supervision 4–7 (2016), [].

Elected officials have begun to chip away at regulatory-monitor secrecy. In 2011, President Obama ordered agencies to “make . . . infor­mation concerning their regulatory compliance and enforcement activi­ties” such as “administrative inspections, examinations, reviews, war­nings, [and] citations” available for online search. 370 See Memorandum on Regulatory Compliance, 3 C.F.R. 326, 327 (2012). Executive agen­cies have accommodated. For instance, for each inspection, the FDA posts any noncompliance identified, “voluntary” recommendation made, 371 Data Dashboard: Inspections, FDA,
inspections.htm [] (last visited Feb. 2, 2019).
and overturned findings. 372 See Inspection Classification Definitions, FDA,
Inspections/ucm223231.htm [] (last updated Nov. 28, 2017) (noting that findings from FDA inspections may be overturned during Agency review and that such reversals will be reflected in a public database); see also BSEE Data Center, BSEE, [] (providing similar information for oil regulation).

Congress has also contributed to the transparency framework. In 2010 it required agencies to publicize “the tabulation, calculation, or recording of activity or effort that can be expressed in a quantitative or qualitative manner.” 373 GPRA Modernization Act, Pub. L. No. 111-352, § 3, 124 Stat. 3866, 3867–71 (2011) (codified at 31 U.S.C. § 1115 (2012)). Although this law does not mention regulatory monitors, major regulators release statistics such as the number of exami­nations. 374 See infra Appendix A. Consequently, aggregate changes, like cuts in examination numbers, are now more visible in many agencies.

In some agency-specific statutes, Congress has gone further. The Clean Air Act, for example, requires publication of any auditor’s “preliminary determination” that an internal system should be revised. 375 See Clean Air Act Amendments of 1990, Pub. L. No. 101-549, 104 Stat. 2399, 2571 (codified at 42 U.S.C. § 7412(r)(7)(B)(iii) (2012)) (requiring the EPA to promulgate regulations providing for agency audits of risk management plans and requiring such plans to be available to the public); 40 C.F.R. § 68.220(i) (2018) (implementing the directive of § 7412(r) by providing for audits and requiring the public to have access to “the preliminary determinations, responses, and final determinations under this section”). Dodd–Frank mandated that the SEC release reports summarizing examination findings, 376 See Dodd–Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111-203, § 932(a)(8), 124 Stat. 1376, 1878 (2010) (codified at 15 U.S.C. § 78o-7(p)(3)(C) (2012)). a break with the financial regulation tradition of “on-site examiners who enforce quite informally and often on a face-to-face and confidential, instead of a written and public, basis.” 377 See Zaring, supra note 196, at 209.

This transparency framework, despite some value, is variant and unstable. Independent agencies, except when required by statute, 378 See supra note 376 and accompanying text. have complied less thoroughly with President Obama’s directive than have executive agencies, 379 They do not, for instance, post company-specific or inspection-specific informa­tion. See, e.g., Compliance, FERC, [] (last updated Nov. 15, 2018). and a new president could easily issue a contrary order. Additionally, in many agency-specific statutes, Congress over­looked monitoring. The main regulator of offshore oil platforms, for instance, must publish information about its postaccident investigations, but not its regular inspections. 380 Outer Continental Shelf Lands Act, 43 U.S.C. §§ 1331–1356b (2012 & Supp. I 2014) (detailing the Department of the Interior’s responsibilities).

Moreover, many transparency mandates focus on aggregate disclo­sures, which provide limited insight. An agency that conducts fewer examinations over time may be doing so because industry has captured it or because it is conducting more thorough examinations. An agency meting out fewer regulatory-monitor sanctions for violations could mean less vigilant agencies or more compliant firms.

The design of many monitoring-transparency statutes also leaves open a window for obfuscation. For example, although the Clean Air Act mandates the publication of any preliminary audit determinations, it does not require a decision or report upon inspection, stating only that regulators “may issue the owner or operator of a stationary source a writ­ten preliminary determination.” 381 40 C.F.R. § 68.220(e) (2018) (emphasis added). That leaves the sequence of deci­sionmaking unclear as to what the frontline inspector’s determinations were, rather than the managerial pressures that followed. In contrast, in the Food, Drug, and Cosmetic Act, for instance, Congress mandated that “prior to leaving the premises, the officer or employee making the inspection shall give to the owner, operator, or agent in charge a report in writing . . . . A copy of such report shall be sent promptly to the [Health and Human Services] Secretary.” 382 21 U.S.C. § 374(b) (2012) (emphasis added).

One policy response would be to require more comprehensive trans­parency. Default requirements might include those adopted by the FDA, such as (1) visibility into the entire regulatory-monitor chain of com­mand; and (2) identification of the company. Transparency has well-known drawbacks that would need to be considered before expanding it. In particular, transparency could prompt firms to constrict the exchange of regulatory information to avoid more stringent regulation. 383 See Coglianese et al., supra note 37, at 290–92. And chain-of-command disclosures may also leave much unclear, as “the inner work­ings of complex bureaucracies [cannot] be captured neatly in charts or guidelines.” 384 See Nou, supra note 42, at 482. Some activities might need to remain private due to the necessity of protecting companies’ trade secrets. Transparency has also been used as a political tool for deregulatory goals. 385 See generally David E. Pozen, Transparency’s Ideological Drift, 128 Yale L.J. 100, 102 (2018) (arguing that the dominant policy rationale for increased government transparency in the twenty-first century emphasizes the capacity of transparency mechanisms “to make government leaner and less intrusive”).

But even without identifying the company, chain-of-command reports can have value. If the number of overturned frontline regulatory-monitor decisions changes significantly over time, the reports could suggest that leaders are captured by industry or that they are inadequately supervising frontline monitors. The data could also enable third parties to identify regulatory-monitor best practices or abuses of power. A recent study of publicly available health inspection microdata found that incon­sistent application of the law subjected restaurants to an “inspector lot­tery.” 386 See Ho, supra note 79, at 635–38 (analyzing data from a restaurant-sanitation grading system in New York and concluding that grade distributions are “essentially random” and that current grades have little correlation with grades in future inspection cycles). At least one agency subsequently adopted institutional improvements indicated by those findings. 387 Ho, supra note 47, at 12–13. This field experiment tested interventions indicated by the original database study. See id. at 50. For such advancements to be made, external parties need access to data. Despite limits, transpar­ency mechanisms can improve public oversight of both regulatory monitors and those who seek to coopt them.

2. Private Paper Trails. — Given the limits of public disclosures, Congress has sometimes turned to private disclosures. Even when kept private, an agency paper trail could deter problematic managerial behav­ior because it leaves open the possibility of subsequent investigation. For example, OCC examiner Victor Del Tredici caught a bank president ille­gally diverting loan fees into his personal account, 388 Quiet Hero: Victor Del Tredici and the Fall of the San Francisco National Bank, OCC, [https://] (last visited Oct. 12, 2018).
but Del Tredici’s superiors ignored his report for nine months. 389 See id. After  the  bank  failed  and  its  president  went  to  jail,  congressional  inquiries  into  the  agency’s  inaction  on  the  report  publicly  embarrassed  OCC  leadership, even though the report itself had been private. 390 Eugene N. White, The Comptroller and the Transformation of American Banking, 1960-1990, at 7 (1992). The paper trail also helped restore Del Tredici’s standing after OCC leadership had stripped him of his authorities over the incident. 391 See id. A manager who is made aware of the possibility of subsequent legal investigations or public criticism is more likely to internalize diverse constituents’ views—an “observer effect.” 392 Ashley S. Deeks, The Observer Effect: National Security Litigation, Executive Policy Changes, and Judicial Deference, 82 Fordham L. Rev. 827, 862 (2013) (“The premise of the observer effect is that the executive responds to certain or probable judicial [scrutiny] . . . . [T]he executive is more likely to perceive that a court may intervene . . . when the courts sense a shift in [public opinion].”).

Mandated paper trails for manager reviews have other accountability benefits. A paper trail makes reviews more likely to happen in the first place, which is important because reviews can improve the accuracy of frontline decisions. 393 See, e.g., Ho, supra note 47, at 96 (noting that a paper trail makes direct oversight easier, which in turn enables supervisors to moderate inconsistencies between decisions made by frontline monitoring staff). Also, managerial reviews of regulatory monitors help fulfill what is arguably a “constitutional duty to supervise” agency employees. 394 See Gillian E. Metzger, The Constitutional Duty to Supervise, 124 Yale L.J. 1836, 1874–904 (2015) (defining the “duty to supervise,” describing its constitutional basis, and delineating its scope).

3. Statutory Minimums. — Whereas both public disclosures and pri­vate paper trails rely on informational mechanisms, Congress can impose direct constraints through statutory “timing rules.” 395 Jacob E. Gersen & Eric A. Posner, Timing Rules and Legal Institutions, 121 Harv. L. Rev. 543, 545 (2007) (“A timing rule, as we define it, is a rule that substantially affects the timing of a government action, including legislation and executive action.”). Lawmakers some­times imposed a minimum frequency of inspections along with the origi­nal authorization of monitoring authority. 396 See, e.g., Burke, supra note 107, at 15 (noting semiannual inspections of steamboats). More often, however, minimums were mandated or increased in response to an often-observed regulatory pattern in which “[h]istory keeps repeating itself.” 397 George M. Burditt, The History of Food Law, 50 Food & Drug L.J. (Special Issue) 197, 200 (1995). After monitoring authority already existed in an industry, subsequent oil spills, 398 Deepwater Report, supra note 258, at 28–30 (describing government reaction to a series of offshore disasters); see also 43 U.S.C. § 1348(c) (2012) (providing for “scheduled on-site inspection” and “periodic on-site inspection without advance notice” of offshore facilities subject to environmental regulation). economic crises, 399 White, supra note 90, at 31; see also Securities Exchange Act of 1934, 15 U.S.C. §§ 78a–78qq (2012). mining deaths, 400 Federal Mine Safety and Health Amendments Act of 1977, Pub. L. No. 95-164, § 103(a), 91 Stat. 1290, 1297 (codified as amended in scattered sections of 30 U.S.C.) (requiring at least four annual inspections for all underground mines and at least two annual inspections for all surface mines); Federal Coal Mine Health and Safety Act of 1969, Pub. L. No. 91-173, § 103(a), 83 Stat. 742, 749 (codified as amended in scattered sections of 30 U.S.C.) (mandating four annual inspections at each underground coal mine); Federal Coal Mine Safety Act, Pub. L. No. 82-552, § 202(a), 66 Stat. 692, 693 (1952) (repealed 1969) (requiring annual inspections in some coal mines); Anne Marie Lofaso, What We Owe Our Coal Miners, 5 Harv. L. & Pol’y Rev. 87, 98 (2011) (“[T]he Federal Coal Mine Health and Safety Act of 1969 . . . came after the Farmington No. 9 mine explosion in West Virginia . . . . In response to the 1976 Scotia mine disaster in Kentucky, . . . Congress passed the 1977 Federal Mine Safety and Health Act . . . .”). and food poisoning outbreaks 401 See FDA Food Safety Modernization Act, Pub. L. No. 111-353, sec. 421(a), 124 Stat. 3885, 3923 (2011) (codified as amended at 21 U.S.C. § 350j(a)(1) (2012)) (providing that the “Secretary shall identify high-risk [food manufacturing] facilities and shall allocate resources to inspect facilities according to the known safety risks of the facilities”); Jacobs, supra note 116, at 600–01 (positing that, although crises are not the only factor motivating the passage of new legislation, many “key food and drug laws” can be “trac[ed] . . . to calamities in the last century”). have led Congress to impose activity floors, such as annual inspections. These minimums guard against the “problem of public underinvestment in information.” 402 See Stephenson, supra note 35, at 1427–37 (suggesting solutions for the problem of “misalignment” between the “marginal social costs . . . [and] the relevant government agent’s private marginal costs,” which “leads to socially suboptimal investment in information”).

Minimums alone, like transparency or paper trails, have limits. Regulatory monitors may not comply with legislative agendas, particularly following budget cuts. 403 See, e.g., U.S. Dep’t of Labor, No. 05-08-001-06-001, Underground Coal Mine Inspection Mandate Not Fulfilled Due to Resource Limitations and Lack of Management Emphasis 1 (2007), [] (reporting that the MSHA “did not complete one or more statutorily-required inspections at 107 . . . of the Nation’s 731 underground coal mines” in part due to the Administration’s “decreasing inspection resources”). Indeed, agencies such as the EPA usually face more than ten deadlines in a given year across all of their activities, and sometimes over fifty deadlines. Jacob E. Gersen & Anne Joseph O’Connell, Deadlines in Administrative Law, 156 U. Pa. L. Rev. 923, 982 fig.2 (2008). Courts have shown a willingness to compel agencies to take action after missing deadlines. 404 See id. at 952–54 (noting that despite limits on judicial review of agency inaction, missed statutory deadlines “may spur a court to order the agency to act, but will almost never allow the court to specify the content of that action”). But the “end-game” in such situations is unclear because higher courts have “exhib­ited a virtually complete unwillingness” to imprison agency leaders. 405 Nicholas R. Parrillo, The Endgame of Administrative Law: Governmental Disobedience and the Judicial Contempt Power, 131 Harv. L. Rev. 685, 697 (2018); see also Gersen & O’Connell, supra note 403, at 964 (“Most statutes that impose deadlines are silent about what should happen if the agency misses the deadline.”). Moreover, agencies can satisfy minimums perfunctorily, as many believe bank regulators and examiners did leading up to the financial crisis. 406 See, e.g., Levitin, supra note 52, at 2041–45 (explaining various ways in which financial regulators may be captured by industry). Minimums may also hinder agencies’ ability to adjust to fast-changing markets if, for example, effective remote monitoring becomes achievable.

Still, legislative strictures generally, and deadlines in particular, likely influence agencies. 407 See Gersen & O’Connell, supra note 403, at 977 (“Deadlines likely force agencies to reallocate resources away from programs without deadlines and toward programs with deadlines.”); Daryl J. Levinson, Making Government Pay: Markets, Politics, and the Allocation of Constitutional Costs, 67 U. Chi. L. Rev. 345, 383 (2000) (noting that legislatures “exercise control over agencies by drafting and revising statutes governing agency authority, authorizing appropriations, and monitoring agencies’ activities”). Even independent regulators, over which Congress has less influence, report compliance with statutory floors. 408 See, e.g., 2016 FDIC Ann. Rep. 25 [hereinafter FDIC Report], https:// [
DZ7M-82DH] (stating in its annual report that “the FDIC conducted all required . . . examinations”).
Regulatory monitors are highly skilled and likely could have earned more working elsewhere, which means some are presumably driven by a sense of public service. Allowing these employees to evaluate questionable business con­duct could provide avenues for prompting enforcement, even in a cap­tured agency. For example, the regulatory monitors might convince reluctant superiors to take action.

Statutory minimums also undermine industry capture of agencies because of leaks. In 2013, Federal Reserve compliance examiner Carmen Segarra unsuccessfully asked her superiors to take action against Goldman Sachs. 409 Jake Bernstein, The Carmen Segarra Tapes, ProPublica (Nov. 17, 2014), https:// [].
She later released forty-six taped hours of “cozy” conversations between examiners and bankers, and nonaction despite “window dressing” of reports and “shady” behavior. 410 See id. (internal quotation marks omitted) (first quoting Sen. Sherrod Brown; then quoting former Federal Reserve Senior Supervisory Bank Examiner for Goldman Sachs Michael Silva). The incident prompted congressional scrutiny and foreshadowed later criminal charges resulting from blurred lines between the regulator and bank. 411 See Ben Protess & Peter Eavis, Ex-Goldman Banker and Fed Employee Will Plead Guilty in Document Leak, N.Y. Times (Oct. 26, 2015),
2015/10/27/business/dealbook/criminal-charges-and-50-million-fine-expected-in-goldman-new-york-fed-case.html (on file with the Columbia Law Review).
Other bureaucrats have used Wikileaks to reveal documents. 412 David E. Pozen, The Leaky Leviathan: Why the Government Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512, 514 (2013). Whether these avenues improve governance is beyond the scope of the current discussion. Nonetheless, minimums can stifle complacency and capture by forcing agencies to deploy resolute regulatory monitors.

4. Appointments. — Another mechanism for involving heightened oversight is through the appointments process. Many agencies’ legal divi­sion heads are considered “inferior officers,” which triggers an appoint­ment process mandated by the Constitution. 413 The Supreme Court has recently resolved a circuit split about the meaning of “officer,” finding that administrative law judges are officers subject to the Appointments Clause. See Lucia v. SEC, 138 S. Ct. 2044, 2055–56 (2018). That process can enable external stakeholders to have a say in whether  the  appointee  is  fit  for  a  post  that  could  have  a  major  effect  on  people’s  rights. The  heads  of  large  regulatory  monitoring groups are not given the same status. 414 Cf. U.S. Gov’t Printing Office, 76-304, United States Government Policy and Supporting Positions, at v (2012), [] (listing the types of appointments required for various government positions).

This appointments asymmetry may in some cases be inconsistent with the actual influence that monitors have on the administration of the law. Directors of regulatory monitors in some agencies have similar or greater ability to oversee the final legal rights of regulated entities as do those leading attorney divisions. 415 See supra section III.B. Congress has in the past recognized the appropriateness of overseeing the appointment of regulatory moni­tors. In 1852, lawmakers required the President to appoint the bureau­crats who managed steamboat inspectors. 416 See Burke, supra note 107, at 20.

Given the size of the federal bureaucracy today, it may not be practi­cal to require an appointments process for all federal employees who have a significant effect on rights. But the appointments process offers a potential additional mechanism for ensuring that the individuals entrusted with monitoring are fit for their immense power. At the very least, it is worth reexamining the statutory designation of monitor leaders for appointments processes to remove any inconsistencies with compara­ble attorney counterparts.

B. Internal Accountability: Lawyers and Monitors as Rivals and Reviewers

Scholars have in recent years shown how internal “administrative rivals—perhaps as much as Congress, the President, and the courts—shape agency behavior.” 417 Jon D. Michaels, Of Constitutional Custodians and Regulatory Rivals: An Account of the Old and New Separation of Powers, 91 N.Y.U. L. Rev. 227, 229 (2016) (describing the dynamic among three categories of “rivalrous actors” internal to the administrative state: political appointees, career civil servants, and a “large and diverse civil society” that participates in administrative policymaking); see also Neal Kumar Katyal, Internal Separation of Powers: Checking Today’s Most Dangerous Branch from Within, 115 Yale L.J. 2314, 2317 (2006) (arguing that “bureaucracy creates a civil service not beholden to any particular administration,” which “promote[s] internal separation of powers”); Gillian E. Metzger, The Interdependent Relationship Between Internal and External Separation of Powers, 59 Emory L.J. 423, 425–26 (2009) (describing the reciprocal relationship between “internal and external checks on the Executive Branch”). That literature has focused on other groups or functions: how civil servants can check agency leaders, 418 See, e.g., Michaels, supra note 419, at 236–38. how separation of enforcers and adjudicators advances due process, 419 See, e.g., Barkow, supra note 24, at 890, 896. and how little-noticed inspectors general provide agency oversight from within. 420 See generally Shirin Sinnar, Protecting Rights from Within? Inspectors General and National Security Oversight, 65 Stan. L. Rev. 1027 (2013) (describing the features and conduct of inspectors general that allow them to provide agency oversight). Inspectors general are different from inspectors, with the former inspecting government actors and the latter inspecting private (external-to-the-agency) entities. This Article underscores how regulatory monitors—including those who lead them—are also potentially influential internal actors who can help con­tribute to a healthy balance of internal agency power. 421 This issue touches on two larger debates that scholars have covered. The first is the tradeoffs between lawyers and technocrats. See generally Frederick Schauer, Thinking Like a Lawyer: A New Introduction to Legal Reasoning (2009) (explaining how lawyers approach problems differently from others). Second, scholars have explored how to design agencies for the optimal collection of information. See generally Stephenson, supra note 35 (offering a framework for designing public institutions with adequate incentives for acquiring policy-relevant information). Three fundamental design decisions influence the extent to which regulatory moni­tors operate as agency rivals: resource allocation, formal appeals processes, and cross-functional independence.

1. Resource Allocation. — Agency architects have settled on greatly differing allocation of resources to regulatory monitors—from compris­ing almost all of the enforcement workforce to almost none. 422 See supra section III.A. A crucial agency-specific question is what regulatory-monitor allocations are opti­mal, weighing the costs of different regulatory configurations and the benefits in terms of deterrence and, ultimately, general welfare. Definitive answers to such complex questions must await empirical studies com­paring different monitoring models in similar contexts. One hypothesis to test is whether a balance of powers among monitors and attorneys provides benefits over the alternatives.

There are reasons to posit that hybrid agencies might function best. At one extreme, agencies with limited regulatory-monitor power presumably risk being too blind to regulate effectively. The many histori­cal examples of crises associated with insufficient monitoring lend sup­port to this hypothesis. 423 See supra section I.C. Additionally, observers in different regulatory spheres have recently identified many legal problems in need of greater agency monitoring, particularly in areas governed by litigator-dominant agencies. 424 For example, Professor Hemphill and I have, for different reasons, called for the FTC to use monitoring authority more for antitrust and consumer protection. See C. Scott Hemphill, An Aggregate Approach to Antitrust: Using New Data and Rulemaking to Preserve Drug Competition, 109 Colum. L. Rev. 629, 643 (2009); Van Loo, supra note 248, at 1311. For instance, a government task force concluded that the EEOC should collect more data to identify systemic discrimination. 425 See Leslie E. Silverman et al., EEOC, Systemic Task Force Report 11–12 (2006), []. Additionally, Professor Pasquale has argued that more monitoring of medi­cal devices could save lives. See Frank Pasquale, Grand Bargains for Big Data: The Emerging Law of Health Information, 72 Md. L. Rev. 682, 683 (2013) (arguing that “[p]roviders have kept vital information about price, quality, and access secret to maintain a competitive advantage or hide shortcomings” and have thus “impeded the types of large-scale analysis common in other industries”).

At the other extreme, it is important to study the potential pitfalls of overreliance on regulatory monitors. This inquiry takes on particular importance in light of new governance models that might drive the administrative state toward greater reliance on administrative moni­tors. 426 See supra section II.A. Policymakers have repeatedly turned to litigators following monitor-dominant regulators’ failures. After the Exxon Valdez oil tanker crashed into an Alaskan reef in 1989, releasing eleven million barrels of oil, 427 Alan Taylor, The Exxon Valdez Oil Spill: 25 Years Ago Today, Atlantic (Mar. 24, 2014), (on file with the Columbia Law Review). Congress passed the Oil Pollution Act to strengthen oil regulators’ civil penalties. 428 See Oil Pollution Act of 1990, Pub. L. No. 101-380, 104 Stat. 484 (1990) (codified as amended in scattered sections of 33, 43, and 46 U.S.C.). The 2002 Enron scandal “converted FERC from an eco­nomic regulator to an enforcement agency” by prompting an expansion of FERC’s ability to prosecute “market manipulation.” 429 Principal, SJC Energy Consultants, LLC, (on file with the Columbia Law Review) (last visited Oct. 12, 2018) (describing the effect of the Energy Policy Act of 2005 from the perspective of the then-Director of Enforcement); see also Energy Policy Act of 2005, Pub. L. No. 109-58, §§ 315, 1284, 119 Stat. 594, 691, 980 (2005) (codified at 16 U.S.C. §§ 825o-1, 824v (2012)). Following the 2008 financial crisis, lawyers began to play a larger role at the agencies responsible for regulating banks. 430 See Conti-Brown, supra note 102, at 93. Each of these agencies, prior to the scandal, was monitor-dominant. 431 The Enron scandal shifted FERC from a regulatory, monitor-driven agency into a litigator-driven one. See Impacts of H.R. 3795, the Over-the-Counter Derivatives Markets Act of 2009, on Energy Markets: Hearing Before the Subcomm. on Energy & the Env’t of the H. Comm. on Energy & Commerce, 111th Cong. 14–15 (2009) (testimony of Jon Wellinghoff, Chairman, FERC) (noting that the FERC’s “oversight and enforcement ha[d] increased greatly” since 2005 and that, by 2009, FERC had grown its investigatory staff from 14 attorneys to 180); Telephone Interview with FERC Employee (Apr. 5, 2017) [hereinafter FERC Interview]. Banking and oil regulators remain regulatory-monitor dominant. See infra Appendix A.

Capture by industry is a common explanation for such failures. 432 See Deepwater Report, supra note 258, at 77–78 (describing a culture in some offices of the federal Minerals Management Service of “accepting gifts from oil and gas companies,” which “cast[s] a shadow on an entire bureau” (internal quotation marks omitted) (quoting Letter from Earl E. Devaney, Inspector Gen., to Dirk Kempthorne, Sec’y, Dep’t of the Interior 3 (Sept. 9, 2008) (on file with the Columbia Law Review))); Dep’t of the Interior, Office of the Inspector Gen., Investigative Report: Island Operating Company et al 1 (2010), [
59FV-MD57]; Levitin, supra note 52, at 2041–49.
Regulatory monitors’ regular and frequent contact with businesses may make them particularly susceptible to leniency, giving them “empathy bred by personal contact.” 433 Cf. Diver, supra note 41, at 286 (describing a “sense of empathy or allegiance bred by personal contact or professional kinship” that can cause inspectors to “become reluctant to report violations”). Lawyers are not immune to capture or what is sometimes given as its principal explanation: the revolving door of employees working for regulators one day and regulated entities the next. 434 See, e.g., David Zaring, Against Being Against the Revolving Door, 2013 U. Ill. L. Rev. 507, 511–12 (describing and critiquing common concerns about the revolving door). But enforcement lawyers’ more arms-length removal from industry—and perhaps their unique professional thought process 435 See generally Schauer, supra note 423 (explaining that “certain techniques of reasoning are thought to be characteristic of legal decisionmaking”). —could make resource allocation to them an internal agency check on captured monitors. Resource allocation to monitors, on the other hand, helps ensure an agency does not operate in the dark.

2. Appeals. — Formal appeals provide a potential check on some regulatory-monitor actions. Some regulatory-monitor enforcement deci­sions, such as those suspending access to markets, constitute final agency actions, trigger formal administrative processes, and will likely get trans­ferred to legal groups and ultimately public courts if appealed. 436 See, e.g., Biber & Ruhl, supra note 13, at 145–48. How­ever, Congress has typically imposed less procedural oversight of regula­tory monitors. A Department of the Interior authorizing statute requires formal adjudicative processes including, for example, subpoena power mirroring that in “the district courts of the United States” for offshore oil platform investigations, but not for inspections. 437 See 43 U.S.C. § 1348(c)–(d), (f) (2012). The CFPB’s founding statute requires administrative law appeals for CFPB enforcement actions, but not for examination findings. 438 See 12 U.S.C. §§ 5515(e)(1), 5516(c), 5563 (2012). Such agency-specific statutes mirror the APA’s exemption of “proceedings in which decisions rest solely on inspections.” 439 5 U.S.C. § 554(a) (2012).

Despite statutory lenience regarding regulatory-monitor appeals, some agencies have built formal processes enabling firms to appeal regulatory monitors’ decisions, even when not required by statute. One model leaves appeals within the regulatory-monitor chain of com­mand. 440 See, e.g., CFPB, Appeals of Supervisory Matters 1 (2015), https://files. [
PUY2-W3CR] (CFPB appeals); Cooper & Fleder, supra note 306, at 492 (FDA appeals).
That procedural design would lessen the influence of the front­line monitor but overall still retain enforcement influence within the larger monitoring group. Other agencies have routed regulatory moni­tors’ appeals outside the monitor group, such as through administrative law judges. 441 See, e.g., 30 C.F.R. § 290.2 (2018) (permitting those adversely affected by a final decision of an official from the Department of the Interior’s BSEE to appeal the decision to the Department’s Interior Board of Land Appeals).

These design choices have limits. Even when agencies set up an appeals process outside the regulatory-monitor group, the fear of infor­mal repercussions, such as a damaged relationship and stricter inspec­tions, may deter appeals. Additionally, for many decisions, such as a temporary halting of activities or blocking of a chicken entering the stream of commerce, the appeals process may be impractical given the magnitude or timing of the decision.

3. Monitor–Lawyer Teams and Rivalries. — Once an agency’s leaders have decided to deploy both regulatory monitors and regulatory lawyers, a number of questions remain about how these groups should interact on an ongoing basis. Numerous models exist. At some agencies, lawyers and monitors function as teammates. At others, enforcement lawyers “become prisoners of the work done by inspectors.” 442 Cf. Diver, supra note 41, at 280 (characterizing inspectors’ role in the enforc­ement process).

As discussed above, various organizational design choices influence the extent to which agency lawyers and monitors are interdependent. When lawyers are required to have visibility into monitors’ activities, such as through the mandatory sharing of inspection reports, lawyers become more independent in taking action. When monitors receive sanction authority, they become more independent in securing compliance. 443 See supra section III.B.

Even hybrid agencies have deployed greatly divergent models for how their powerful groups of monitors and lawyers should interact. The CFPB organizationally imposes more separation between the two groups. CFPB examiners and lawyers coordinate some actions. 444 Cf. Witkowski, supra note 72 (“[E]nforcement attorneys will continue to coordinate with examiners offsite.”). But they organizationally occupy separate offices and ultimately can pursue sepa­rate tracks for resolving even multimillion-dollar wrongdoing. 445 See Bureau Structure, CFPB, [] (last visited Oct. 12, 2018) (showing a separate office for supervision examinations and enforcement); supra notes 353–355 and accompanying text (discussing the separate tracks).

In  contrast,   the  EPA  does  not  organizationally  separate  out  the  inspection  function. 446 See EPA Organization Chart, EPA, [] (last visited Oct. 12, 2018). Once inspectors identify anything beyond a minor violation, they work side by side with lawyers. EPA collaboration means that both engineers and lawyers are often involved in deciding on sanc­tions, negotiating with firms, and even coauthoring legal briefs. 447 See EPA Interview, supra note 314; see also Joel A. Mintz, Enforcement at the EPA 113 (rev. ed. 2012). Conse­quently, each meaningful regulatory-monitor decision is peer-reviewed both by someone trained within a professional code of ethics for the administration of justice and by someone familiar with the science and industrial organization. 448 See EPA Interview, supra note 314. See generally Schauer, supra note 423 (discussing lawyers’ approach to reasoning). Peer review alone can improve regulatory-monitor performance. See Ho, supra note 47, at 79–82 (discussing the evidence that shows how peer review can improve the accuracy and consistency in administering the law).

The institutional relationships between lawyers and regulatory monitors presumably can influence enforcement and policy outcomes. Some agencies’ enforcement orders make it clear that they believe lawyer–monitor organizational design matters—albeit for private entities. The SEC and the Department of Health and Human Services (HHS) have mandated that malfeasant companies separate their compliance and legal departments. 449 For a critique of these mandates, see DeStefano, supra note 205, at 122–55. In other words, the SEC and HHS have man­dated for businesses a level of separation that the EPA does not have for its own lawyers and compliance-related personnel. To the extent the com­pany’s compliance and legal departments serve as internal regulators, similar organizational principles may be appropriate for both public and private monitors. 450 See supra section II.A.2 (discussing self-regulation).

Since these organizational questions about regulatory monitor–lawyer peer review and independence have yet to be studied, it is difficult to assess the merits of these approaches. 451 Peer review of inspectors has been studied in great depth, but peer review across these two groups has not been. See supra notes 393–395 and accompanying text. Nor have scholars turned their attention to the ideal level of organizational dependence among regulatory monitors and regulatory lawyers. But regulatory lawyers and regulatory monitors have different expertise, worldviews, and legal authority. It is plausible that a set of agency-mandated processes for cross-functional peer review and information sharing could better organi­zationally set regulators up for success in overseeing complex markets.


Legal scholars commonly describe agencies as engaging in ex ante rulemaking and ex post enforcement. Ongoing monitoring should be added to that standard account of agency activity and studied more closely. Those who regularly extract information from firms influence much of the administrative state’s law-related activity. Any regulatory analysis that ignores regulatory monitors or groups them together with enforcement actors risks obscuring agencies’ vital “internal laws.” 452 Mashaw & Harfst, supra note 225, at 443 (“Bureaucratic institutions have their own internal laws, expressed both in regulation and in routine.”).

This administrative-monitoring ecosystem is ripe for systematic study to identify best practices for weeding out extremes of overbearing, blind, or captured agencies. Congress, the President, and agency directors have begun to construct a framework for promoting transparency and discouraging complacency. A key question is how much of the nascent regulatory-monitor oversight structure should be ingrained in the law rather than left to bureaucratic discretion.

Perhaps most importantly, agency designers should add regulatory-monitor resource allocation and intergroup processes to the toolbox for improving effectiveness, independence, and accountability. 453 For an overview of anticapture organizational-design mechanisms, see generally Barkow, supra note 55. Regulatory moni­tors are vital to the front line of business compliance. But lawyers—as judges, drafters of laws, and intra-agency rivals—are the “foot soldiers of our Constitution.” 454 Lee R. West, Judicial Independence: Our Fragile Fortress Against Elective Tyranny, 34 Okla. City U. L. Rev. 59, 73 (2009) (internal quotation marks omitted) (quoting Rennard Strickland & Frank T. Read, The Lawyer Myth: A Defense of the American Legal Profession 13 (2008)). The organizational design of these two groups’ intersection is crucial to a healthy system of checks and balances with regulatory monitors as a powerful internal branch of administration.



Appendix A: Employees and Monitoring 455 Unless otherwise specified, figures are all examiner, inspection, or compliance positions for regulatory monitors and all “Legal and Kindred” employees from the U.S. Office of Personnel Management. See FedScope, supra note 74. The Monitor Percent is calculated as Monitor Personnel / (Monitor Personnel + Legal Personnel). Figures reflect those reported through the end of 2016, although some figures have been updated since then.

The nineteen large regulators are the Consumer Financial Protec­tion Bureau (CFPB), Federal Energy Regulatory Commission (FERC), Food and Drug Administration (FDA), Food Safety and Inspection Service (FSIS), Mine Safety and Health Administration (MSHA), Occu­pational Safety and Health Administration (OSHA), Federal Aviation Administration (FAA), Federal Motor Carrier Safety Administration (FMCSA), Office of the Comptroller of the Cur­rency (OCC), Environ­mental Protection Agency (EPA), Equal Employ­ment Opportunity Commission (EEOC), Federal Communications Commission (FCC), Federal Deposit Insurance Corporation (FDIC), the Federal Reserve, the Federal Trade Commission (FTC), National Credit Union Adminis­tration (NCUA), National Labor Relations Board (NLRB), Nuclear Regulatory Commission (NRC), and the Securities and Exchange Commission (SEC). Data in the appendices aim to provide a survey of the level of activity across large regulators, but the data should not be viewed as comprehensive. Additionally, the data provide a snap­shot based on the most recent year readily available, and activity may vary over time. Drawing firm conclusions about the level of monitoring and the number of monitor employees would for many agencies require a more in-depth study focused on the full array of an agency’s activities and employees over a longer timeframe.



Monitor Personnel

Legal Personnel

Monitor Percent

Annual Monitor Activity

CFPB 416 349 54% 177 examinations and related 456 CFPB, CFPB Strategic Plan, Budget, and Performance Plan and Report 38–40 (2016), [] (listing “supervisory activities”). For a review of the CFPB’s early examination activities, see generally Jean Braucher & Angela Littwin, Examination as a Method of Consumer Protection, 87 Temp. L. Rev. 807 (2015).
FSIS 8,107 440 95% 1.7 million products inspected 457 Quarterly Enforcement Report: October Through December 31, 2016, USDA 3, [] (last visited Oct. 10, 2018).
FERC 509 458 This figure includes accounting, auditing, engineering, and general business. FERC Interview, supra note 434 (clarifying classifications). 308 62% 398 account reviews, 423 reports, 2,330 inspections 459 See FERC, Fiscal Year 2017 Congressional Performance Budget Request 48–51 (2017), [https://].
FDA 11,493 460 This figure includes scientists, engineers, consumer protection, and medical officers. Telephone Interview with FDA Employee (Mar. 24, 2017) (describing job responsibilities). 203 98% >160,000 inspections 461 See Compliance Check, supra note 291.
MSHA 1,521 462 Of these, about 1,145 actually conduct inspections, whereas the rest engage in related monitoring support and oversight activities. U.S. Dep’t of Labor, Report No. 05-10-001-06-001, Journeyman Mine Inspectors Do Not Receive Required Periodic Retraining 1–2 (2010), [https://].
141 463 This figure was determined using the same methodology (for the same reasons) that was used to determine the legal personnel figure for OSHA. See infra note 469. 91% 19,642 inspections 464 FY 2016 U.S. Dep’t of Labor Agency Financial Rep. 19,
sites/default/files/media_0/_Sec/2016annualreport.pdf [] (putting the figure at 3,095 for coal mines and 16,547 for metal and other noncoal mines).
OSHA 1,827 465 OSHA, supra note 184, at 28–29. 277 466 Legal employees are listed as zero for OSHA in the database because legal is centralized in the Department of Labor (DOL). This figure is calculated as “Legal and Kindred” (except Worker’s Compensation Claims examiners) from DOL proportioned out to OSHA’s percent of DOL employees. See FedScope, supra note 74; OSHA Interview, supra note 148 (explaining how DOL solicitors serve the department’s various agencies). 93% 35,822 inspections 467 OSHA, supra note 184, at 45. This figure corresponds to the number of inspections performed in fiscal year 2015, not including inspections of federal agencies.
FAA 4,388 468 This figure excludes 418 employees categorized as “General Inspection, Investigation, Enforcement, and Compliance,” due to the inability to obtain information differentiating the responsibilities within this category. 342 93% Inspect 227,900 aircraft 469 FAA, FY 2009 Citizens’ Report: Summary of Performance and Financial Results 4 (2009) [hereinafter FAA Citizens’ Report],
media/2009_Citizens_Report.pdf []. This statistic is from fiscal year 2009 because the FAA has not published updated figures; however, the agency’s more recent reports indicate no lessening of inspection responsibilities. See, e.g., FY 2017 FAA Performance & Accountability Rep. 50 [hereinafter FAA Accountability], https:// [] (“Since 2010, the FAA has seen an increase of approximately . . . 800 percent . . . in the number of inspections FAA performs to ensure safety compliance.”).
FMCSA 644 470 See FMCSA, 2017 Pocket Guide to Large Truck and Bus Statistics 18 (2017),
2017-pocket-guide-large-truck-and-bus-statistics-final-508c-0001.pdf [
WKJ6]. This figure counts only FMCSA Employees engaged in safety inspections, rather than the larger group of monitors, which would include managerial, support, and oversight positions, since they are not differentiated in the OPM database. Note that federal inspectors represent 5% of the total inspector force, most of whom are state employed. See id.
46 93% 118,494 inspections 471 See id. at 18. This total refers to the number of federal inspections conducted in 2016.
OCC 2,715 209 93% 768 applications 472 2016 OCC Ann. Rep. 30 [hereinafter OCC Report], [].
EPA 1,682 473 This figure corresponds to employees categorized as “Environmental Engineers” in the OPM database. See FedScope, supra note 74; see also Mintz, supra note 450, at 11 (confirming that the number of personnel conducting inspections for the EPA is approximately 1,600). 1,102 60% 13,500 inspections 474 Enforcement Annual Results Numbers at a Glance for Fiscal Year 2016, EPA, [] (last visited Oct. 11, 2018) (listing an overview of the enforcement numbers in the “Numbers at a Glance” tab).
EEOC N/A 522 0% Analyses of 67,146 employer reports 475 Agency Information Collection Activities: Revision of the Employer Information Report (EEO–1) and Comment Request, 81 Fed. Reg. 51,113, 51,115 (Feb. 1, 2016) (stating that there were 67,146 employer-submitted EEO-1 reports for 2014).
FCC 308 476 This figure reflects engineers and analysts from FedScope, supra note 74; see also FCC Interview, supra note 294 (explaining employee breakdowns). 602 477 This figure is roughly evenly divided between enforcement and other legal functions, such as central legal staff and rule writers. See FCC, Fiscal Year 2017 Budget Estimates to Congress 12 (2016) (on file with the Columbia Law Review) (stating that the enforcement division had 240 total employees in fiscal year 2016). 34% Undisclosed number of radio inspections and transaction reviews 478 See Inspection Fact Sheet, FCC,
inspection-fact-sheet [] (last visited Nov. 8, 2018) (describ­ing why and how FCC inspections of radio installations occur); Mergers and Acquisitions, FCC, [
THJ2-KFCG] (last visited Nov. 8, 2018) (describing the FCC’s respon­sibility for reviewing business transactions in which an FCC license will be transferred). The FCC does not provide readily accessible data about its monitoring activities, making it difficult to assess how extensively it uses its monitoring authority. Interviews indicated, however, that the agency engages in regular inspections of radio stations and processing of information submitted by businesses. See FCC Interview, supra note 294.   
FDIC 2,719 454 86% 6,892 examinations 479 FDIC Report, supra note 410, at 25.
Federal Reserve 1,382 480 See 2015 Fed. Reserve 102nd Ann. Rep. 308 [hereinafter Federal Reserve Report], [] (noting that full-time employees in the Boston branch of the Federal Reserve account for approximately 5.79% of 16,686 total employees); Interview with Federal Reserve Employee in Bos., Mass. (Mar. 22, 2017) [hereinafter Federal Reserve Interview] (estimating that the Boston office has eighty examiners and four lawyers). The figures in this table assume that Boston reflects national Federal Reserve breakdown. The Federal Reserve is not included in the OPM data and does not release examiner breakdowns. 69 481 See Federal Reserve Report, supra note 483, at 308; Federal Reserve Interview, supra note 483. 95% 4,190 482 See Federal Reserve Report, supra note 483, at 308; Federal Reserve Interview, supra note 483.
FTC 20 483 This figure is an estimate of the number of employees who work on the Consumer Sentinel Network. See FTC Interview, supra note 279 (estimating the size of the Consumer Sentinel group); Consumer Sentinel Network Data Book 2017, FTC (Mar. 2018), [] (explaining that the Consumer Sentinel Network stores consumer complaints from various data contributors and makes them available to law enforcement). 711 3% ~1,200 merger transactions 484 This figure is limited to Hart–Scott–Rodino Act (HSR) transactions. Since the annual aggregate figures released combine HSR transactions for the FTC and DOJ, in order to estimate the HSR transactions reviewed by FTC monitors, this figure assumes that the total number of HSR transactions reviewed by each entity is proportional to the figures for acquisition clearance granted to each agency. See 2015 FTC & DOJ Antitrust Div. Hart-Scott-Rodino Ann. Rep., at exh. A tbl.I,
reports/federal-trade-commission-bureau-competition-department-justice-antitrust-division-hart-scott-rodino/160801hsrreport.pdf [] (noting that there were 1,794 total HSR transactions reviewed by both agencies, there were 179 clearances granted to the FTC, and there were 79 clearances granted to the DOJ). Taking the data from the FTC and DOJ’s HSR annual report, the approximate number of HSR transaction reviews completed by FTC monitors was calculated as follows: 1,794 x   = 1,217.
NCUA 886 31 97% 9,465 contacts 485 2016 NCUA Ann. Rep. 13,
annual-report-2016.pdf [].
NLRB 0 797 0% Minimal clear monitoring 486 The closest activity to monitoring is the NLRB’s conducting of union elections. See supra note 159 and accompanying text. NLRB agents conducted 1,496 labor elections between October 1, 2015, and September 30, 2016. See NLRB, Election Report for Cases Closed Between 10/1/2015 and 9/30/2016, at 1 (2016),
files/attachments/basic-page/node-4626/Total%20Elections%202016.pdf [
H5QE-XG8C]; see also ABA, supra note 159 (explaining that the NLRB observes all union elections).
NRC 1,641 115 93% Continual presence, 99 plants 487 See NRC, A Day in the Life of an NRC Resident Inspector (Aug. 10, 2017), (on file with the Columbia Law Review); Power Reactors, NRC, [
CJ8P-XE4N] (last updated Oct. 31, 2018).
SEC 1,631 488 See SEC Budget, supra note 290, at 14 (providing figures for full-time equivalent employees in fiscal year 2015). This figure reflects the number of full-time equivalents in fiscal year 2015 for employees labeled “Compliance, Inspections, and Examinations,” “Corporation Finance,” and “Trading and Markets,” since the database left the employee breakdown unclear for monitor-like activities conducted by groups like the “Economic and Risk Analysis” and “Investment Management” employees. See id. 1,466 489 See id. This figure reflects the number of full-time equivalents in fiscal year 2015 for employees labeled “Enforcement” and “General Counsel.” Id. 53% 2,400 examinations 490 SEC, supra note 216, at ii.


Appendix B: Sanction Control


Monitor Citations, Voluntary Actions

Monitor Blocking Access

Monitor Formal Charges

CFPB $44 million in redress 491 CFPB Report, supra note 307, at 11. This figure represents the total amount of redress paid from October 1, 2015, to March 31, 2016. See id. at 8.
FSIS 25,516 noncompliances documented 492 See USDA, supra note 460, at 1 tbl.1. Pre-approve each meat and poultry product 493 Carmen Rottenberg, Food Safety Professionals Ensure that “What’s in Your Meat” Is Safe and Wholesome, USDA Food Safety & Inspection Serv. (Aug. 29, 2018), https:// [].
FERC 214 recommendations, $5.3 million in refunds 494 See FERC Report, supra note 262, at 5. Charge: license revocation 495 See FERC Interview, supra note 434 (noting that monitors have the authority to influence license revocations but that, in practice, licenses are almost never revoked).
FDA 14,590 warning letters 496 See FDA Enforcement, supra note 305, at 1. 2,847 recalls 497 See id. For additional context on the FDA’s recall procedure, see Dep’t of Health & Human Servs., Office of Inspector Gen., A-01-15-01500, Early Alert: The Food and Drug Administration Does Not Have an Efficient and Effective Food Recall Initiation Process 1 (2016), [
VGQ8] (finding that the FDA does not have “an efficient and effective food recall initiation process that helps ensure the safety of the Nation’s food supply”).
Investigate: penalties & recommend charges 498 See FDA, Regulatory Procedures Manual ch. 5, at 16–17, 88 (2018),
ucm176972.pdf [].
MSHA 97,255 citations and orders 499 Mine Safety and Health at a Glance, MSHA (July 7, 2017),
sites/default/files/Data_Reports/msha-at-a-glance-7-7-2017.pdf [
FEPG] (providing the total number of citations and orders issued for calendar year 2016).
Inspectors order mine evacuations 500 Laura E. Beverage, Litigation Under the Federal Mine Safety and Health Act Today: A Practical Guide, 16 Am. J. Trial Advoc. 305, 310–12 (1992) (“The inspector may issue a withdrawal order for the affected area . . . .”). Charge: $48 million in civil penalties 501 MSHA at a Glance (FY 1978-2016), MSHA,
factsheets/fy/at-a-glance-fy1984-2016.pdf [] (last visited Oct. 12, 2018) (providing 2016 figures); Mine Inspections, MSHA,
compliance-enforcement/mine-inspections [­3ZZF] (last visited Oct. 12, 2018) (describing the requirements of the MSHA, including inspections of underground mines four times a year and of surface mines twice a year).
OSHA 65,044 violations 502 Occupational Safety and Health Administration Enforcement, OSHA, https:// [] (last visited Oct. 12, 2018).
Charge: civil fines 503 See supra note 336.
FAA Warning letters, pilot retraining 504 See, e.g., Robinson, supra note 331, at 29–30. Pre-approve aircraft design 505 See FAA Accountability, supra note 472, at 12 (“The old standards ensured adequate levels of safety, but lacked flexibility to accommodate rapidly developing technological innovations. Today, instead of telling manufacturers how to build airplanes, the FAA’s regulations set performance standards and allow general aviation manufacturers to develop the designs and innovations to meet those standards.”); see also FAA Citizens’ Report, supra note 472, at 6. Prior to issuing a voluntary automobile recall, the DOT requires monitoring groups to obtain consent from the legal department. See Interview with DOT Employee (Mar. 26, 2017). Investigate: civil penalties, license 506 See Robinson, supra note 331, at 31.
FMCSA 35,756 Warning Letters 507 FMCSA, supra note 473, at 28. Registers and audits new vehicle entrants 508 49 C.F.R. § 385.319 (2017). The agency conducted 36,756 new entrant safety audits in 2016. See FMCSA, supra note 473, at 30.  
OCC Non-public MOUs and Commitment Letters 509 OCC, PPM 5310-3, Policies and Procedures Manual: Bank Supervision Operations 15, 18 (2011),
2017/11/PPM-5310-3-Old-2011.pdf [].
Pre-approve branches, notified of mergers 510 See OCC Report, supra note 475, at 31. Charge: civil penalties, $226 million 511 See id. at 32; OCC, 2017 Manual, supra note 369, at 4–7.
EPA Minor citations 512 See EPA Interview, supra note 314 (stating that notices of minor violations found in inspection can be sent to the company without legal review or enforcement action if corrected within thirty days). Joint charge: $6 billion in civil penalties 513 See EPA Interview, supra note 314; Enforcement Annual Results for Fiscal Year 2016, EPA, [] (last visited Oct. 12, 2018).
FCC Joint 514 See FCC Interview, supra note 294. Changes by licensees 515 See id. Joint charge: license revocation 516 See id.
FDIC Noncompliance notifications 517 FDIC Report, supra note 410, at 25–27. Pre-approve new branches Charge: civil money penalties 518 FDIC Interview, supra note 306.
Federal Reserve Noncompliance notifications Pre-approve branches, notified of mergers Charge: $2.2 billion in civil penalties 519 See Federal Reserve Report, supra note 483, at 57.
NCUA 303 actions 520 This figure is from 2016. See NCUA, supra note 488, at 16.      Charge: civil penalties 521 Telephone Interview with NCUA Employee (Apr. 11, 2017).
NRC 715 non-cited violations; 61 cited violations 522 See 2015 NRC Enforcement Program Ann. Rep. 4, 18,
docs/ML1606/ML16069A146.pdf [] [hereinafter NRC 2015 Enforcement Report]; see also generally NRC, Enforcement Manual (2017), https:// [] (explain­ing how inspections document violations).
Pre-approve equipment changes and construction 523 See NRC 2015 Enforcement Report, supra note 525, at 26. Investigate: civil money penalties & recommend charge 524 NRC, NRC Enforcement Policy 16–25 (Nov. 2016),
docs/ML1627/ML16271A446.pdf []; Interview with NRC Employee (Apr. 11, 2017).
SEC $60 million returned to investors in 2016 525 See SEC, supra note 216, at 21. Firm licenses and suspension of trading 526 See, e.g., Securities Exchange Act of 1934, Pub. L. No. 73-291, § 15, 48 Stat. 881 (codified at 15 U.S.C. § 78o (2012)) (describing the SEC’s registration requirements); SEC, supra note 216, at 5 (mentioning registration); Statement on Order of Suspension of Trading of Certain Bitcoin/Ether Tracking Certificates, SEC,
public-statement/suspension-trading-certain-bitcoinether-tracking-certificates [https://] (last visited Feb. 2, 2019) (providing an example of the Division of Trading and Markets and Division of Corporate Finance suspending trading).
Charge: license 527 17 C.F.R. § 240.15c3-1(c)(2)(vi)–(vii) (2018). Manage: $94 million in SRO fines 528 FINRA Report, supra note 212, at 3.