In Leaders of a Beautiful Struggle v. Baltimore Police Department, the Fourth Circuit Court of Appeals held that Aerial Investigation Research (AIR), Baltimore’s aerial surveillance program, violated the Fourth Amendment because it was not authorized by a warrant. AIR was constitutionally problematic, but not for the reason given by the Fourth Circuit. AIR, like many other technologically-enhanced policing programs that rely on closed-circuit television (CCTV), automated license plate readers and the like, involves the collection and retention of information about huge numbers of people. Because individualized suspicion does not exist with respect to any of these people’s information, an individual-specific warrant requirement can never be met by such a program. When police engage in suspicionless searches and seizures of the type exemplified by AIR, a different regulatory approach is needed, one that provides the protection against arbitrariness that the warrant process affords but does not require findings that specific people have violated the law. This Article argues that this regulatory alternative can be derived from administrative law principles. The logic of administrative law dictates that legislatures and agency rulemaking must be involved any time a policing agency wants to establish a program that will intentionally affect sizeable numbers of concededly innocent people. If administrative law principles applied, programs like AIR would not be permitted unless a legislature has delegated appropriate authority to the relevant police agency, implemented regulations have survived notice-and-comment and hard look judicial review, and the agency carried out the program in an even-handed fashion that minimizes discretion. At the same time, contrary to the holding in Leaders of a Beautiful Stuggle, if these requirements are met, the Fourth Amendment—at least the part of it requiring warrants and probable cause—would be irrelevant.
In Leaders of a Beautiful Struggle v. Baltimore Police Department, the Fourth Circuit Court of Appeals held that Aerial Investigation Research (AIR), Baltimore’s aerial surveillance program, violated the Fourth Amendment. Relying on the U.S. Supreme Court’s declarations in Carpenter v. United States and United States v. Jones that the Fourth Amendment governs both real-time and digital-record tracking, the Fourth Circuit correctly held that AIR’s constant recording of Baltimore’s pedestrian and car traffic constituted a “search” for Fourth Amendment purposes. But it erred in holding that such a program requires a warrant—a conclusion that, in effect, made AIR impossible as a legal matter. Because jurisdiction-wide systems like AIR do not purport to investigate a particular suspect, but rather aim only at gathering information that can later be used to carry out such investigations, they can never satisfy the Fourth Amendment warrant requirement, which regulates searches aimed at specific individuals.
There was another significant legal problem with AIR, however. In setting up the program, Baltimore and its police department failed to follow proper democratic and administrative processes. This Article argues that this oversight should have been the focus of the Fourth Circuit’s decision and should have been the death knell for the program. More generally, it argues that administrative law principles can and should fulfill the Fourth Amendment’s reasonableness requirement when the government wants to conduct suspicionless searches and seizures. While Carpenter, Jones, and other Fourth Amendment cases that address the constitutionality of searches aimed at specific individuals are inapposite in such situations, administrative law’s programmatic focus is ideally suited to govern data collection programs.
The aptly named Leaders of a Beautiful Struggle is a harbinger of cases to come as law enforcement continues to experiment with technologically-aided investigative techniques. In recent years, police departments and municipalities have avidly adopted surveillance systems using Automated License Plate Readers (ALPRs), closed-circuit television (CCTV), and facial recognition technology (FRT), and have also moved toward database analytics using “fusion centers” and private companies. Each of these programs involves the collection and retention of information about thousands, tens of thousands, or even hundreds of thousands of people. Because individualized suspicion does not exist at the time these law enforcement initiatives acquire information about people, imposition of an individual-specific warrant requirement makes little sense and, rigidly applied, would spell doom for all such programs, no matter how effective they are. A different regulatory approach is needed, one that provides the protection against arbitrariness that the warrant process affords but does not require findings that specific people have violated the law.
Because programmatic policing usually affects entire jurisdictions, this Article follows earlier work of mine in arguing that the regulatory regime in this setting should involve not just the courts, but significant engagement by the legislative and executive branches as well, as mediated through administrative law principles. In fact, the logic of administrative law requires agency rulemaking any time a policing agency proposes an investigative technique that will intentionally affect sizeable numbers of innocent people—a phenomenon that occurs not only with surveillance programs like AIR, but in connection with a wide array of physical searches and seizures such as health and safety inspections, checkpoints, and drug testing programs. These “panvasive” actions should not be permitted unless the requirements that typically govern informal rulemaking by all other agencies are met: a legislature must delegate appropriate authority to the relevant police agency, regulations must survive notice-and-comment and hard look judicial review, and the agency must carry out the program in an even-handed fashion that minimizes discretion. At the same time, contrary to the holding in Leaders of a Beautiful Struggle, if these administrative law requirements are followed, the Fourth Amendment—at least that part of it requiring warrants based on probable cause—should be irrelevant.
On May 1, 2020, the City of Baltimore, with the help of a company called Persistent Surveillance Systems (PSS), initiated a six-month trial of an aerial surveillance program called AIR. AIR relied on cameras positioned on high-flying planes to monitor the city during the daytime. If a crime was caught on camera or it otherwise came to the attention of the police, the aerial recordings were used to trace the people and cars near the crime scene when it occurred both forward and backward in time to help identify who they were. Because any individuals picked up on the cameras appeared merely as blurry dots, facial features were not observable. But people could be identified by connecting them to certain residences and in various other ways. If identification occurred, subsequent interviews, interrogations, stops, and arrests were presumably governed by traditional Fourth and Fifth Amendment law.
Perhaps because of concerns about disparate racial impact, but also based on straightforward cost-benefit calculations, Baltimore ended its program in 2021. That response is, of course, the government’s prerogative. But in Leaders of a Beautiful Struggle, the Fourth Circuit Court of Appeals took it upon itself to hold that, had the police department not ended AIR, it would have found the program unconstitutional under the Fourth Amendment. In an eight to seven en banc opinion, the court analogized AIR’s surveillance of Baltimore’s populace both to Carpenter, which held that police investigating a bank robbery engaged in a search when they accessed multiple days of Carpenter’s cell-site location information from his common carrier, and Jones, which held that police investigating Jones for drug dealing engaged in a search when they tracked him for twenty-eight days using GPS signals. In light of these opinions, the majority reasoned, the day-to-day aerial surveillance that took place under AIR was a Fourth Amendment search that required a warrant. Because under well-established case law and the language of the Fourth Amendment itself, a warrant is impossible to obtain until a crime has occurred or a suspect has been identified, the Leaders of a Beautiful Struggle holding prohibited the pre-crime and citywide recordings on which AIR depended.
The major flaw in Leaders of a Beautiful Struggle is its conceptualization of the problem. The majority opinion described AIR as a program that was aimed at “capturing everyone’s movements outside during the daytime . . . .” Citing Carpenter and Jones, it then declared that “prolonged tracking that can reveal intimate details through habits and patterns . . . invades the reasonable expectation of privacy that individuals have in the whole of their movements and therefore requires a warrant.” But that type of reasoning applies only to long-term tracking of identified suspects, as occurred in Carpenter and Jones. The issue before the court was not the constitutionality of such suspect-driven or crime-driven searches but rather the constitutionality of a program that collects, in the absence of suspicion about any given individual, the information used for such searches.
This confusion between the legitimacy of the data collection and the legitimacy of its use permeated the majority’s opinion. The Fourth Circuit appeared to be particularly concerned that the program retained recordings of the movements of everyone caught on camera for forty-five days. But these recordings were not accessed unless a violent crime was caught on camera. If at that point, the police wanted to view recordings surrounding the time and place of the crime, judicial authorization—a warrant or court order—might well be required. But a warrant process aimed at determining whether there is cause to search for evidence against a specific person is ill-fitted to decisions about whether to create those records in the first instance and, if so, the length of time recordings are kept, the types of crimes they can be used to investigate, and other programmatic matters.
Another example of the majority’s confusion was its reaction to the district court’s finding that AIR images show only “a series of anonymous dots traversing a map of Baltimore.” In rebuking the lower court for relying on this fact, the Fourth Circuit correctly observed that the habitual behavior of those “dots” (such as starting and ending the day at home), when “analyzed with other available information, will often be enough for law enforcement to deduce the people behind the pixels.” But, again, that use of AIR was not before the court; had it been, the court could rightly have demanded a Fourth Amendment justification based on individualized suspicion.
In another passage, the majority—still conflating collection with use—asserted that “[t]he AIR program is like a [twenty-first] century general search, enabling the police to collect all movements, both innocent and suspected, without any burden to ‘articulate an adequate reason to search for specific items related to specific crimes.’” But in real time, AIR’s day-to-day collection of pixels was not seeking information about any specific items, crimes, or people. And its “general” nature made it a perfect candidate for jurisdiction-wide legislative regulation rather than case-specific judicial determinations.
That was the gist of Judge Wilkinson’s dissenting opinion. As he put it, “I have no problem if the AIR program is discontinued. I have a big problem, however, if this court and not the citizens of Baltimore are the ones to terminate it.” Judge Wilkinson noted that AIR had been established on an experimental basis, after obtaining endorsements from the governor of Maryland, the mayor of Baltimore, the Baltimore City Chamber of Commerce, and a number of other high profile groups, including community leaders in East and West Baltimore, the Greater Baltimore Committee (“the region’s premier organization of business and civic leaders”), the presidents of local universities, and religious leaders from the United Baptist Missionary Convention (representing 100 churches across the state). The reason for this wide-ranging support was summarized by the local head of Neighborhoods United: “We have to do something. The murders are doing a lot of disruption to our city, especially in the black population.” Pointing to this community support, Judge Wilkinson concluded: “The people most affected by a problem are denied by this court a say in ameliorating it.”
In short, the Fourth Circuit had no business ending the program on the grounds it did. At the same time, the Judiciary was not the only branch of government that overreached in connection with the AIR program. The Baltimore Police Department was also too dominant in the decisionmaking process. Despite the widespread official support for AIR, the Baltimore City Council never formally authorized it, even on a test basis, and the police department’s guidelines under which it operated were not subject to community input. The type of democratic process normally associated with the initiation of civilian, citywide programs was—at most—tangentially involved in vetting law enforcement’s use of AIR.
In critiquing the majority’s ruling, Judge Wilkinson quoted a statement from the concurring opinion of Justice Alito in Jones (joined by Justices Ginsburg, Breyer, and Kagan): “A legislative body is well situated to gauge changing public attitudes, to draw detailed lines, and to balance privacy and public safety in a comprehensive way.” That call for legislative involvement is particularly apt when it comes to programmatic searches and seizures like AIR, given the inapplicability of the traditional, judicially-oriented warrant process. Assuming so, there remains the crucial tasks of drawing the “detailed lines” to which Justice Alito refers and figuring out how to cajole legislatures and police agencies into promulgating them.
There are two broad issues raised by programmatic searches and seizures. The first is whether a particular program should be authorized. For what purposes may a municipality, state, or federal government establish panvasive dragnets like AIR, fusion centers, or random inspection regimes, and what types of information may these policing efforts collect? Assuming the program is authorized, the second issue is how long the information it obtains may be retained and under what conditions. May the data collected be maintained indefinitely or should it be destroyed after a finite period, and how can the accuracy and security of the data be assured?
A helpful conceptualization of how data collection and retention might be regulated comes from the American Law Institute (ALI), which recently completed its Principles of Policing Project after six years of deliberation. The fourteen chapters of the Principles—which cover every aspect of the policing endeavor—were officially adopted by the full membership of the ALI in May 2022, after vetting by both an advisory committee (composed of judges and lawyers who work in the criminal justice system, police and advocacy organizations from both the left and right) and the ALI Council (composed of a select group of the full ALI membership). As an associate reporter for the project, I was principally responsible for two of the Principles’ chapters: “Policing in the Absence of Individualized Suspicion” and “Policing Databases.” Respectively, these two chapters provide principles that could govern the collection and retention issues raised by programmatic searches.
“Policing in the Absence of Individualized Suspicion” is defined by the ALI Principles as policing “conducted in the absence of cause to believe that the particular individual, place, or item subject to [the policing action] is involved in unlawful conduct . . . .” It should be clear from previous discussion that surveillance programs like AIR fit within this definition; because they are panvasive, they are designed to obtain information before any individualized suspicion develops. In this way, they are like various types of physical search programs—such as checkpoints, health and safety inspections, and drug-testing programs—that operate in the absence of individual suspicion.
The ALI chapter on this type of policing provides that “[l]egislatures and agencies should authorize suspicionless policing activities only when there is a sound basis for believing that they will accomplish an important law-enforcement or regulatory objective, and when achieving that objective outweighs their infringement on individual interests such as privacy, dignity, property, and liberty.” If a suspicionless program is authorized, written policies should identify, among other things, “(a) the specific harm sought to be detected or prevented; (b) the permissible scope of the suspicionless policing activity; [and] (c) the persons, entities, or activities subject to the policing activity.” Additionally, the Principles state that any suspicionless policing activity so approved “should be conducted in a manner that ensures agency discretion is guided by neutral criteria that are applied evenhandedly and developed in advance,” which must be accomplished by applying the procedure to every person within the target group, “a subset of that group that is selected on a random or neutral basis,” or “a subset of that group that there is a sound basis for believing is more likely to be engaged in unlawful conduct or pose a greater risk of harm than the rest of the target group.” There are several more detailed principles in the chapter, but these three capture its gist: suspicionless searches and seizures should only occur when (1) there is a strong rational basis for the program after considering its impact on collective and individual interests; (2) policies explicitly identify its purpose and scope; and (3) the program is applied in a neutral, even-handed fashion.
The primary rationale for these principles is straightforward: “In the absence of warrants and individualized suspicion, it is essential that there be alternative mechanisms in place to ensure that searches and seizures and other policing activities are justified, are not directed at individuals or groups in an arbitrary or discriminatory fashion, and are limited in scope consistent with their justification.” The “sound basis” requirement for surveillance legislation is admittedly vague, but necessarily so. As Justice Alito’s comments in Jones suggest, the initiation of panvasive programs calls for the kind of multifactor judgment that is best made initially by legislatures. The requirement that the purpose and scope of the program be explicitly identified ensures that these matters receive due deliberation by the appropriate decisionmaking bodies and that the policing agency has sufficient direction. The neutral criteria and even-handed application requirements minimize discretion and increase the likelihood that the program will be viewed by the public as both more legitimate and less intrusive (think, for instance, of TSA checkpoints at airports). These latter two requirements also make it likely that the program will affect those with political power, which acts as a brake on overly aggressive programs. As Justice Jackson stated in another context: “[T]here is no more effective practical guaranty against arbitrary and unreasonable government than to require that the principles of law which officials would impose upon a minority must be imposed generally.”
The gist of the ALI chapter on policing databases can also be succinctly stated. It provides that “[a] policing database [defined to include databases that might be generated by AIR, CCTV, fusion centers, and the like] should be created only if necessary to facilitate a legitimate policing objective” and continues:
Any policing database that contains information about identified or identifiable individuals should be governed by written policy or policies that specify: (1) the purpose of the data collection, including the criteria for inclusion in the database; (2) the scope of data to be collected, including the types of individuals, locations, or records that will be the focus of the database; and (3) the limits on data retention, the procedures for ensuring the accuracy and security of the data, the circumstances under which the data can be accessed, and mechanisms for ensuring compliance with these rules.
The first two principles overlap with the provisions in the chapter on suspicionless policing requiring delineation of the purpose and scope of such search and seizure programs. The third principle is given more detailed treatment in the database chapter’s subsequent provisions, which focus on:
These two sets of ALI principles, which are consistent with suggestions made by other entities and scholars, provide guardrails for thinking about data collection and retention. But they are aspirational and expressed at a high level of generality. How can legislatures and policing agencies be pushed toward adopting something like them and then fleshing out the details in connection with specific policing programs? Supreme Court case law has provided very little impetus in this direction, and legislative inertia or resistance has led, at best, to piecemeal statutory regulation. After documenting those assertions, the remainder of this Article explains why another source of rules—administrative law—must play a central role in this regulatory framework. If the well-established administrative law principles that govern virtually all other government agencies are made applicable to policing agencies as well, they can force the legislative and executive branches to produce reasonable regulations of search and seizure programs.
Fourth Amendment jurisprudence to date has largely ignored both the data collection and data retention issues. While legislatures have been more active on these issues, some types of search and seizure programs today are not subject to any statutory constraints and those that are often only loosely so.
Contrary to the Fourth Circuit’s assertion in Leaders of a Beautiful Struggle, the Supreme Court’s decisions Carpenter and Jones provide no help on the issue of when a surveillance program or other search and seizure programs may be authorized; as explained earlier, these cases involved suspect-driven searches, not panvasive ones. Of course, Carpenter and Jones do not exhaust the Fourth Amendment’s potential. As David Gray has argued, for instance, the Fourth Amendment’s language guaranteeing “the right of the people” to be secure from unreasonable searches and seizures could form the basis for protecting the collective interests of citizens against indiscriminate, arbitrary surveillance, and voracious and insecure databases. Neil Richards has contended that the First Amendment’s protection against chilling speech and assembly could fulfil much the same function. While no Supreme Court case has endorsed these precise themes, scattered dicta hint that there may be constitutional limits on “data greed.” In United States v. Knotts, after holding that short-term tracking of an individual was not a Fourth Amendment search, the Court suggested that “different constitutional principles” might apply to “twenty-four hour surveillance of any citizen of this country . . . without judicial knowledge or supervision.” And in Whalen v. Roe, it recognized “the threat to privacy implicit in the accumulation of vast amounts of personal information in computerized data banks or other massive government files.” It remains the case, however, that the Court has yet to put any meat on these bones or even identify the specific constitutional provision that might do so.
The one existing Fourth Amendment doctrine that could potentially lead to something more comes from a convoluted series of Supreme Court decisions that govern situations involving “exceptional circumstances in which special needs, beyond the normal need for law enforcement, make the warrant and probable-cause requirements impracticable.” In some of these so-called “special needs” cases, particularly those involving inspections and checkpoints conducted in the absence of individualized suspicion, the Court has appeared to endorse something akin to the ALI Principles’ requirements that search and seizure programs have a rational basis and defined scope and be governed by a neutral plan that is administered even-handedly.
For instance, in Donovan v. Dewey, after noting the many dangers associated with operating coal mines, the Court upheld a warrantless mine inspection program because the statute governing the program:
[R]equires inspection of all mines and specifically defines the frequency of inspection. . . . [T]he standards with which a mine operator is required to comply are all specifically set forth in the [Mine Safety] Act or in . . . the Code of Federal Regulations. . . . [R]ather than leaving the frequency and purpose of inspections to the unchecked discretion of Government officers, the Act establishes a predictable and guided federal regulatory presence.
The Court made the same type of point in a case involving a checkpoint designed to catch undocumented immigrants near the United States-Mexico border. In upholding the checkpoint, the Court emphasized that it was set up by higher-level authorities and that everyone who came to the checkpoint was subject to initial seizure. In another case, while finding random license checks made on the whim of individual officers unconstitutional, the Court noted that statutorily-authorized license checkpoints that stop everyone—or every third or fifth person—would pass constitutional muster.
These cases resonate with the ALI’s principles governing suspicionless searches and seizures. As I have detailed elsewhere, however, many of the Court’s other special needs cases ignore these types of strictures. As a result, at best, the Court’s message in this area is muddled.
The Court has been similarly circumspect about what the Constitution has to say concerning data retention, accuracy, and security rules. For instance, in Herring v. United States, it suggested, without holding, that evidence found during an arrest based on an expired warrant would require exclusion if the defendant could demonstrate “routine” or “widespread” systemic errors in the arrest warrant database. In Whalen, it noted with favor the fact that a state statute governing collection of prescription drug information prohibited unwarranted disclosures to the public, a protection it said “arguably has its roots in the Constitution” in some circumstances, and in later cases it “assumed, without deciding, that the Constitution protects a privacy right of the sort mentioned in Whalen . . . .” Similarly, in Maryland v. King, which upheld a statute permitting collection of DNA samples from arrestees, the Court emphasized that the state was prohibited from using the samples for any purpose other than identifying arrestees. But none of these cases explicitly stated that there would be a constitutional cause of action if personal information in government-run databases is inaccurate, retained indefinitely for no reason, or gratuitously disclosed to the public.
Legislatures have been more engaged in regulating the content and use of databases. Most prominently, the federal Privacy Act and similar state statutes permit agencies to obtain and keep information only if it is “relevant and necessary to accomplish a purpose [the agency is authorized to accomplish].” These statutes also prohibit public disclosure of personally identifiable information that has been collected without consent. Various other federal and state statutes require the destruction of information after a limited period—or at least the creation of policies that dictate how long that period should be—and call for procedures for assuring accuracy and security.
But these statutes can be remarkably lax when it comes to law enforcement. For example, the federal Privacy Act exempts from its strictures any agency “which performs as its principal function any activity pertaining to the enforcement of criminal laws” and specifically permits police agencies to retain any “investigatory material compiled for law enforcement purposes . . . .” Neither the federal government nor most states have developed statutes that specify nor define what “investigatory material” is, or whether there are limits on the types of investigatory material that may be collected.
Of course, there are specialized statutes that deal with specific types of programmatic searches. For instance, at both the federal and state levels, there are statutes regulating electronic surveillance, drone surveillance, ALPRs, and various types of records searches, including communications, tax, and video records. But with the exception of electronic surveillance laws, these statutes are usually extremely vague with respect to the types of information that may be collected, the purposes for which the information may be used, and how long the information collected may be retained. More significantly, for some types of programmatic searches, there is no authorizing statute at all; any rules that exist come solely from individual police departments.
This is where administrative law principles can change the game.
Even if Fourth Amendment jurisprudence in this area remains moribund, well-worn principles of administrative law applied to program-driven searches and seizures could require both legislatures and police departments to provide more substantive regulation about the type of data collection and retention that occurs in connection with programs like AIR. Most government agencies—at the federal level, ranging from Food and Drug Administration to the Commerce Department, and at the state level, from environmental agencies to health services—are governed by administrative procedure acts (APAs). Even some municipalities, such as New York City, have enacted such statutes meant to govern agencies. These acts control how government agencies make decisions and promulgate rules. Of most relevance here is the stipulation found in the typical APA, including the federal APA upon which most other APAs are modeled, that whenever an agency develops a policy that is a “statement of general or particular applicability and future effect” which affects “the rights and obligations” of the citizenry, it must follow certain procedures.
Those procedures have four important components. First, whenever an agency plans to engage in programmatic actions, it must engage in a rulemaking process. Second, that process calls for and encourages public participation in the fashioning of the rules, through the well-known notice-and-comment mechanism. Third, under commonly accepted administrative law jurisprudence, agencies can be required to justify to a court the rules they create through what is known as “hard look” review to ensure both that the rules have a rational basis and that they are implemented even-handedly; as the Supreme Court has said in the federal context, “The APA ‘sets forth the procedures by which federal agencies are accountable to the public and their actions subject to review by the courts. . . . [i]t requires agencies to engage in ‘reasoned decisionmaking.’” Finally, before any of this can happen, legislation must set out an intelligible principle for the agency to follow. This requirement forces democratic consideration about whether the agency should be engaging in the practice at all and, if so, how it should accomplish its aims.
All four of these procedural components could have a significant, concrete impact on policing programs. The rulemaking requirement would force police departments to think through how a given program will work and, combined with the rational basis requirement, should nudge them toward rules similar to those recommended by the ALI. The notice-and-comment process would facilitate local input about the program, something that today seldom occurs in the policing setting. The even-handed application requirement could also have a potent impact on police departments, which have been known to arbitrarily focus on particular neighborhoods or groups. Likewise, the requirement that programs be legislatively authorized is crucial in the police setting: programs like AIR are rarely the focus of legislation, meaning that the source of their authority often consists solely of the omnibus direction, found in virtually every state code, that police departments detect, prevent, and deter crime. That vague, capacious language is unlikely to satisfy even the lax mandates of the nondelegation doctrine.
In other work, I have elaborated further how these four requirements might help implement access and retention rules of the type described in the ALI’s principles. Here, I will only reiterate why administrative law must have a role in regulating policing. That discussion is important because, even though police departments are administrative agencies and even though most APAs do not specifically exempt them from their provisions, the accepted wisdom to date has been that police do not have to follow administrative rulemaking procedures. A typical pronouncement one finds in administrative law treatises is that “[a]dministrative law includes the entire range of action by government with respect to the citizen or by the citizen with respect to the government, except for those matters dealt with by the criminal law.”
On its face, this is a puzzling statement. As I have pointed out:
[T]he APA requires that agencies abide by its rulemaking dictates when dealing with such matters as workplace ergonomics, the height of a fence around animals, and the precise manner in which farm yields are reported. A regime that requires the relevant agency to submit to administrative law constraints in these situations, but not when police want to require citizens to submit to drug testing, checkpoints, and surveillance, is seriously askew.
The likely reason for this anomaly is the notion that police search and seizure rules are already governed by the criminal law or the Fourth Amendment; thus, they do not need to be subject to the process agencies must follow when they promulgate rules that apply to and that regulate the law-abiding public. While this rationale may exempt law enforcement from administrative law oversight, the rules that tell police when to conduct a stop, arrest, or search based on reasonable suspicion or probable cause are altogether different from rules governing data collection and retention programs. Stops, arrests, and traditional searches are suspicion-based; police have, or should have, reasonable suspicion or probable cause to believe the target has violated the criminal law. In contrast, the policies that established AIR, CCTV systems, or fusions centers are not interpreting a criminal statute or a judicial search and seizure decision about who may be subject to police action, but rather authorize suspicionless actions against the entire populace or large segments of it. As is true with the rules of other agencies that must abide by rulemaking procedures, these policies have “general and future effect” on the “rights” of citizens because their panvasive nature has a direct impact on thousands of concededly innocent people, who must either submit to surveillance or modify their legitimate behavior if they want to avoid police intrusion.
A second possible reason for the de facto exemption of police agencies from the ambit of administrative law is that much policing is local, carried out by municipal police, not federal or state agents, and so is thought to be outside the purview of federal and state APAs. But that would not explain why federal and state police agencies are not governed by their respective APAs, nor why the police departments in the nine states that consider municipal departments to be agencies of the state or the departments in those cities that have their own APA—such as New York—are exempt. Furthermore, every municipality, not just those directly governed by an APA, enforces federal or state criminal laws; at least to that extent, they should be covered by federal or state APAs. Additionally, many data collection programs, such as those undertaken by fusion centers, have a much broader scope than a purely municipal program and are often funded by federal and state governments. If none of these arguments win out, however, municipalities could still be brought into the fold through authorizing legislation that either is specific enough that little discretion is left to the department or mandates that departments follow administrative principles such as notice-and-comment review and even-handed application.
A third argument for exempting police agencies from the rulemaking process is that they have neither the wherewithal nor the expertise to develop rules. But, in fact, big police departments create detailed rules governing things such as use of force, stop and frisk, and surveillance techniques all the time. Smaller departments can piggyback on this work and on the recommendations made by various organizations, ranging from the International Association of Police Chiefs to the ACLU, for specific guidance.
Properly construed, Supreme Court case law provides a basis for this regulatory restructuring of suspicionless programs around administrative law principles. While the Supreme Court’s special needs jurisprudence does not explicitly reference APAs, its decisions in the business inspection setting suggest that the Fourth Amendment requires some sort of rulemaking in situations where the warrant requirement is inapposite. In addition to Dewey’s mention of the Code of Federal Regulations when it upheld warrantless inspections of coal mines, there is the Court’s pronouncement in Marshall v. Barlow’s, Inc. that, to protect business owners from the “unbridled discretion [of] executive and administrative officers,” the Judiciary must ensure that there are “reasonable legislative or administrative standards for conducting an . . . inspection . . . with respect to a particular [establishment].” And in Colonnade Catering Corp. v. United States, the Court said, in the course of authorizing warrantless inspections of liquor stores: “Where Congress has authorized inspection but made no rules governing the procedure that inspectors must follow, the Fourth Amendment and its various restrictive rules apply.” Conversely, this passage suggests, if the Fourth Amendment’s warrant requirement does not apply or cannot work, Congress or its agency delegee must act.
These various statements from the Supreme Court indicate that administrative law can perform the Constitution’s regulatory function if the legislature or agency promulgates constraining rules. They also suggest that, if such rules are developed, they establish a safe harbor from aggressive judicial intervention. Most importantly, they can be read to hold that the Fourth Amendment requires formal rulemaking by the police before a programmatic search or seizure may take place.
In Leaders of a Beautiful Struggle, the Fourth Circuit had an opportunity to reconceptualize the way the Fourth Amendment applies to programmatic police action. It could have held that, when analyzing the constitutionality of policing programs aimed at large swaths of the population, administrative law principles provide a meaningful substitute for traditional Fourth Amendment strictures, even when no suspicion with respect to any given individual in that population exists. But instead, the court applied Fourth Amendment doctrine meant to govern individual cases to the panvasive setting; a move that, followed to its logical conclusion, renders information-gathering programs such as AIR unconstitutional. Administrative law principles, in contrast, allow governments to experiment with such programs, assuming sufficient legislative delegation, the development of rational, community-vetted regulations, and non-arbitrary implementation. This type of regime fits much better with the Fourth Amendment’s overall admonition that searches and seizures be “reasonable.” A key advantage of subjecting police agencies to the administrative law principles that all other agencies must follow, however, is that they apply regardless of whether the Fourth Amendment does. APAs require a rulemaking process whenever agencies take actions, like Baltimore’s police department did with AIR, that have general and future effect on the rights of citizens.
Portions of this article are taken from chapter seven of my book Virtual Searches: Regulating the Covert World of Technological Policing (NYU Press, 2022).
