I. Introduction: The Trouble with Bureaucracy↑

As is well known, the Internal Revenue Service of the United States is the bureaucracy¹ responsible for applying the federal tax code. This is one of the most difficult tasks that could be imagined for any governmental organization, given the sheer size of the law and the complexities involved in applying it. One recent study estimated that federal tax laws and regulations are over ten million words in length. The entire King James translation of the Bible has around 800,000 words.²

Unsurprisingly, the IRS has grown in size along with the complexity of the tax code, even though in recent decades its workforce has seen some significant declines. The IRS is surely significant, particularly for those who violate the law, whether deliberately or inadvertently. Yet in what sense is the IRS, or any bureaucracy for that matter, politically significant. Is it not the case that the politics of tax policy takes place without any participation by the IRS—in the public realm, in the electoral realm, in the halls of Congress and the wings of the White House? Is it not the case that conflict and contention of politics is separate from the complex, though ultimately mundane tasks of administration?

The case of the IRS illustrates that the task of implementing law cannot always be separated from political questions and political controversies. Recently, the Federal Justice Department reached a negotiated settlement with over 428 conservative groups who claimed that their applications for tax-exempt status had been willfully mishandled by the IRS during the 2012 Presidential election cycle.³ As a consequence of IRS enforcement decisions, these groups were unable to attract donations, mobilize voters, or otherwise engage in the political process; in essence, they claimed that the IRS had deliberately “hacked” the American democratic process.⁴ As troubling as this particular incident might be, it may be part of a broader pattern. Policy analysts who have examined the IRS suggest that decisions about which organizations and individuals to target for investigation often varies, depending upon which party controls the Presidency. The targeting scandal suggests that the IRS can act in a politicized manner on its own; the broader pattern of IRS law enforcement suggests that bureaucracies, rather than being neutral enforcers of the law, are often used as tools for partisan political purposes.⁵

The IRS scandal raises some of the most fundamental questions about bureaucratic politics:

–How did the American bureaucracy develop over time?
–Why, or under what circumstances, are bureaucrats able to exercise power in a political manner?
–How can bureaucratic power be controlled and constrained by the other branches of government?
–Why is bureaucracy necessary, despite its potential for abuse, and are there ways to enjoy the advantages of bureaucratic government without suffering the disadvantages?

Bureaucracy and the Principal-Agent Problem

The word bureaucracy is similar to the word “democracy,” in that both words refer to kind of political power. Democracy is based on two Greek roots: “demos” for the people, and “kratos,”meaning rule or strength. Thus democracy means the rule or (interestingly) the strength of the people. The word “bureaucracy” is a combination of the French and the Greek with “bureau” meaning “desk.” So bureaucracy means “the rule of desks.” That is suggestive as well. We tend to think of bureaucracies as a kind of instrument or tool, whether in the private or public sector. Yet the word itself indicates that bureaucracy is a form of ruling, perhaps a kind of regime, like democracy, oligarchy, or aristocracy. But what does it mean to be ruled by “desks”? To what extent is the rule of desks compatible with the rule of the people?

We should consider the place of bureaucracy in political life by considering a very simple example of what “normal” policy-making looks like, and the role of bureaucrats within the normal policy process. Consider the case of traffic law. Traffic laws are determined by state legislatures. The executive branch bureaucrats responsible for enforcing these laws are state and local law enforcement officers. Judges and juries adjudicate individual cases to determine whether the law has been applied properly. The police do not play any role in creating the laws, but they exercise a considerable amount of discretion when applying it. A well-trained and well-disciplined force will apply the law in a reasonable and impartial manner. Other police forces may be lackadaisical, biased, or even corrupt in their approach to traffic laws. For the people who are subject to the law, controlling the executive branch bureaucracies responsible for enforcing law is at least as important as controlling the elected officials who create law. As we know from the case of police officers, bureaucracies that are insulated from all forms of citizen control can easily become abusive, or indifferent to the interests of the people they are meant to serve.

Controversies over policing in American politics are provide illuminating problems of controlling bureaucratic discretion, often referred to as the “principal-agent” problem. This problem occurs in any area of life—politics, business, even family relationships—where individuals or groups responsible for rules or decisions (the “principals”) must rely on some other group (the “agents”) to enforce the rule or implement the decision. How can the principals insure that their agents are acting appropriately? Police, for instance, are expected to be politically neutral, impartial law enforcers. In actual practice, or course, the on-the-ground decisions of “law enforcement professionals” can rooted in error, misjudgment, or prejudice. For instance, in the aftermath of Michael Brown’s killing by a St. Louis police officer in 2014, the Department of Justice released an extensive report on the abuse of police power in East St. Louis, Missouri, ranging from the “over-policing” of African American neighbourhoods, violation of constitutional rights, and extra-legal police violence.⁶ This is just one example of the simple fact that the meaning of “law” depends upon the decisions of those who implement it. Controlling the discretionary power of bureaucrats, whether local police officers or federal officials, is a never ending process; as long as bureaucracies exist, the “principal-agent” problem will exist.

The “principal-agent” problem of policing help to illustrate the problems of bureaucratic power as a whole, even though police officers are not really a paradigmatic example of the problem modern bureaucracy. Modern forms of bureaucratic power were created in order to overcome the inefficiencies of the Constitutional law-making process. The old, inefficient law making system, the old form of American constitutionalism, was thought by early 20th century progressive thinkers to be an obstacle to the creation of a modern state, in particular, a state that could serve the interests of the public as a whole.⁷ According to progressives, “the law” (the Constitution, but general legal principles as well) protected the interests of elites, and not the interests of the public. A new system of governance would enable bureaucrats to create and adjust rules in an efficient manner, using scientific expertise and expert discretion to serve the public will. Bureaucracy, according to this progressive view, is an essential component of true democracy, because bureaucracy allows the public will to be effective.

Progressive optimism about bureaucratic power may have been premature. It turns out that a more efficient form of “law-making” can, in some circumstances grant considerable power to the wealthiest and best organized (and those are often overlapping groups). The irony of bureaucratic policy making is that, while the modern administrative state was created to challenge elite dominance, there are ways in which the modern administrative state can actually increase interest group influence.⁸ This is obviously a contested point– the bureaucratic state is not simply the servant of organized interests or financial elites. Nevertheless, we will try to understand why the progressive assumption that “bureaucratic power” will serve the “public will” is incomplete at best, utterly naive at worst.

II. The Development of Bureaucracy in America↑

How can you control bureaucracies? One of the first answers to this question is “the spoils system.” The spoils system– from the saying “to the victor goes the spoils”—was a system for allocating government jobs through patronage. In the early years of the 19th century, federal offices were held by a relatively elite group, who had an almost proprietary interest in their positions; some officials, for instance, were able to pass down their positions to their sons.⁹ During the Jacksonian era (starting around 1828), the federal bureaucracy was “democratized;” positions in the federal government, instead of being held by a coterie of permanent bureaucrats, started to be based on “the rotation of office.” Under this system, when one party wins control of the national government, its own supporters and stalwarts fill the ranks of officialdom—at least in theory. In practice, many federal officials maintained their positions. After Andrew Jackson assumed the Presidency, no more than 20% of the federal civil service was actually replaced. Yet the principle was established that Presidents could replace federal officials (when necessary,) and make federal appointments on the basis of political expediency.¹⁰

What can be said in favour of the spoils system? Perhaps it was suitable for the times. It existed during a period when the problems of government were less complex, and thus the need for a permanent, expertly trained civil service was less urgent.¹¹ At no time in American development was “the state” absent from American life; from the very earliest days, Americans experienced a wide variety of regulations of economic activity, public behaviour, public health, and public safety. There was never really an era of complete laissez faire. However, the form of state power was very different in the first century of the nation’s existence. Regulation of conduct for the health, safety, morality, and welfare of the public took place at the state level. The issues that state-level regulations addressed were not so complicated that relevant decisions and rules could not be determined by legislators.¹² Almost any educated person, if given sufficient time to deliberate, research, and consult, can create appropriate rules for the running of cattle in the streets, or the proper procedure for trains when approaching a road crossing. But safety regulations for advanced pharmaceuticals? That is somewhat more complicated.

But were there any advantages that came from a de-centralized, (relatively) amateur, and partisan bureaucracy (as opposed to the modern form: centralized, professionalized, and independent)? This is actually a very complicated question, but I can suggest some general hypotheses. Patronage connected “the government” to party politics, and was part of the general “democratization” of the political system. It helped to maintain the experience of government as an activity of the people, as opposed to experiencing “the state” as an alien force that rules like an occupying army. Patronage based politics helped to forge links between local and state based parties and the national government, and given the limited scope of “expertise,” this may have justified “bureaucracy through patronage.”¹³

We can contrast this general approach to governance with some aspects of the European model, which was distinct from the USA even in the 19th century. The continental European vision of bureaucracy was articulated by early 19th century political thinkers such as Hegel, and (with much less enthusiasm) by early 20th century scholars such as Max Weber. Hegel argued that civil servants or bureaucrats transcended the particular interests of “civil society, and that in order to perform the function of implementing law, they had to be independent of the society that they governed.¹⁴ Max Weber tended to emphasize the role played by bureaucrats in exercising expert knowledge within complex modern economies; in contrast with than Hegel, Weber was concerned with the ways in which bureaucrats could pose a challenge to democratic forms of government. Given their technical expertise, independence, and sense of collective identity, it would be difficult to control the bureaucrats who claimed to be mere servants of the public will.¹⁵. The American political system, during the first century of its existence, placed much less value on the independence or autonomy of the permanent state; they wanted government to be dependent upon politics, they wanted government officials to be closely connected to the instruments of democratic control and influence.

An amateur, politicized, patronage based bureaucracy will have numerous defects, of course. The long period of civil service reform that began in the wake of the Civil War was propelled by the sense that “government by amateurs” was no longer possible in an era when the national government had more complex and technical responsibilities.¹⁶ It is also the case that civil servants who lack more or less permanent tenure will be subject to the temptations of corruption (though this does not apply to lecturers in political science, so please do not try to bribe me.) Over time, patronage was essentially eliminated from the federal civil service, a process that began with the Pendleton Act of 1883.¹⁷ The problems of patronage based politics was exemplified by the death of President James Garfield, who was assassinated by a disappointed office seeker in 1883. The Pendleton Act of 1883 was a response not only to the assassination, but also a result of the general sense that the government could no longer be conducted by well-meaning amateurs. Competitive examinations were introduced for federal positions; firing was prohibited except for cause; party-dues paying was no longer a condition for holding office. It established what we would recognize as a modern civil service, in fact well before Canada, where patronage had a longer life span.

By the New Deal era in the 1930s, the modern bureaucratic state was in place, and it was entrusted not only to oversee the social safety net, but to manage a wide range of economic activities and relationships. The theory of the New Deal had been articulated long before, for instance, in an essay by Woodrow Wilson entitled “The Science of Administration.” (written when Wilson was a political science professor at Princeton. ) Woodrow Wilson’s premise was quite simple: administration can be separated from politics. Wilson argued that an independent and professional civil service would merely implement the political will of the public. Once the public (through its representatives) has made a decision in regards to some goal or purpose, there is a non-political, technically correct way to determine how to reach that goal; an expert, apolitical civil service is therefore the best means for pursuing publicly determined ends.¹⁸

The development of American national bureaucracy in the late 19th century was significant, but relatively constrained. In particular, the demands for new forms of regulation were approached from a constitutional perspective. Legislators were aware of the problem of delegating unstructured legislative power to the executive branch, and thus they attempted to expand the scope of national regulation without expanding the scope of bureaucratic discretion. This proved to be very difficult, but not impossible. The example of the Interstate Commerce Act, which created the Interstate Commerce Commission, regarded by some as the first example of a “modern bureaucracy” based upon the delegation of legislative power from Congress to the executive¹⁹, can illustrate this point.

The Interstate Commerce Commission was created in order to regulate railroads, an industry that existed largely because of government subsidies (in the form of land grants); a variety of problems created calls for government regulation, particularly the practice of “pooling” (where railroads merged in order to limit price competition) and rate discrimination (e.g. between large industries and small farmers, or between long haul routes and short haul routes.)²⁰ Many reformers wanted to regulate railroads through traditional modes of law-making—that is, they wanted to make certain corporate railroad practices illegal, and then have those laws enforced in ordinary courts. The Interstate Commerce Commission was not really an alternative to this mode of regulation, but rather supplemented the normal policy process. The commission could investigate railroad practices to determine whether a legal violation had occurred, require that railroads submit reports on their activities, subpoena those same companies for information about their business and management practices, and issue ‘cease and desist’ orders (that had to be enforced in ordinary courts.) There were some ambiguities in the statutory mandates of the ICC, but in general its discretion was limited, and its powers were subject to review in ordinary courts of law.

The progressive vision of bureaucratic power was not really established until the early 20th century. This occurred in several stages. Theodore Roosevelt built upon 19th century civil service reforms by increasing the scope of merit-based hiring and providing clearer guidelines for salaries and promotions, reforms that were meant to reduce the influence of Congressionally-based political parties on the national state while increasing the ability of President to manage the executive branch. In 1906, Roosevelt and the Republican Congress passed the Hepburn Act, which transformed the Inter-State Commerce Commission into a genuinely modern bureaucracy, as envisioned by theorists of progressivism. The law, as initially conceived by Roosevelt and progressive Republican allies in Congress, granted the ICC both rate-setting powers as well as final adjudicatory powers; the final bill, though somewhat ambiguous, preserved a role for federal courts in overseeing ICC decisions.

Just as importantly, the number of national regulatory bureaucracies began to expand, and the powers of existing federal government powers were extended. The ICC was joined by the Federal Trade Commission (1913), which was meant to replace the existing American approach to business monopolies. Under the pre-existing Sherman Act, large-scale businesses could be prosecuted for “unfair business practices”; the FTC, in contrast, would regulate business conduct in order to prevent monopolies from developing in the first place. While it was structured like the ICC as an independent commission, the jurisdiction of the FTC extended throughout the economy as a whole, and thus the scope of its discretionary power was potentially much more significant.²¹

Even more significant, in the long run, was the creation of the Federal Reserve Board by the Federal Reserve Act of 1913. The Federal Reserve Board was created with the goal of maintaining a stable supply of credit and a stable dollar. As a federal bureaucracy, its structure is rather unusual. The federal reserve system consists of 12 regional banks, each with their own board. The central board of the federal reserve consists of members nominated by the President and confirmed by the Senate, who serve ten year terms. Through its ability to shape interest rates, the Federal Reserve exercised considerable power over the state of the American economy.

While the FTC and Federal Reserve Board were unusual, pre-existing elements of the federal bureaucracy extended their powers during this period as well. The Department of Agriculture, established just after the Civil War, was mostly concerned with gathering and disseminating information that was of interest to farmers and the agricultural sector; new laws such as the Pure Food and Drug Act extended the scope of the Department’s activities into the regulatory arena. Changes in market conditions, as well as public demands for federal regulation in the aftermath of widespread reporting on unsafe and unsanitary conditions in the food industry, were the main driving forces behind the creation of a federal “police power” over foodstuffs. Some scholars, not without reason, have even argued that federal powers of this kind were created because resistance by major economic interests were tepid; large scale agricultural and food related businesses may have suspected that they would benefit from more intrusive regulation, as they would be better able than their smaller competitor to respond to new regulatory demands. Technological changes were also at the root of new regulatory agencies that developed in the 1920s, well-after the peak of early progressivism in the T. Roosevelt and Woodrow Wilson administrations. The Radio Communications Act created the Federal Radio Commission, whose purpose was to assign access to radio wavelengths to various radio stations. This seemed to require the FRC to make difficult judgment calls about the “public interest, convenience, or necessity,” the closest thing to a legal standard in the act.

The New Deal era witnessed a great explosion of activity on the part of the national government, and was thus accompanied by a significant expansion of bureaucratic power. Some of the new federal initiatives had little long-lasting impact on the United States. The National Industrial Recovery Act, for instance, was declared invalid because it delegated legislative power from Congress to the executive branch with almost no actual legal standards to guide implementation of the law. Other bureaucratic initiatives proved to be far more enduring. Some raised few constitutional problems—the Social Security Administration, to name the most important, exercised relatively little independent discretion in fulfilling its duty to process claims for retirement payments. Similarly, the massive expansion of the military bureaucracy would create various kinds of political problems (particularly regarding spending), but the military did not exercise power independent of Congressional or Presidential control. More significant from a constitutional perspective were the large numbers of new regulatory bureaucracies that were modelled on the Federal Radio Commission, bureaucracies that were tasked with regulating important sectors of the economy without clear statutory guidelines. Some of the most important, and most enduring, of these New Deal agencies included the National Labor Relations Board (NLRB), the Securities and Exchange Commission (SEC), the Federal Communications Commission (FCC—which replaced the FRC), the Federal Power Commission, the Civil Aeronautics Board, the Federal Housing Administration, and so forth. Every new bureaucratic agency would face problems and controversies related to its own particular jurisdiction, but one problem was universal to all of them: given the discretionary nature of their duties, how is it possible to insure that these bureaucracies pursue the public interest? As we will see, Congress, the Presidency, and the federal courts would all struggle to deal with the problem of bureaucratic discretion.

The final major expansion of the federal bureaucratic power occurred during the period between 1964 and 1975, an era that spanned both the “Great Society” of Lyndon Johnson as well as the Nixon Administration. In comparison with the New Deal era, in which many bureaucracies were focussed on regulating specific national industries such as airlines or television, the new bureaucratic organizations of the 1960s and 1970s tended to focus on questions that affected broad areas of social and economic activity. The most significant legislative acts of Lyndon Johnson such as the Civil Rights Act and the Voting Rights Act; both would introduce new dimensions of bureaucratic power. For instance, the Civil Rights Act prohibited race or sex-based discrimination in the labor market; in practice, the meaning of discrimination had to be worked out by officials in the Equal Employment Opportunity Commission (EEOC), just as the meanings of “restraints of trade” or “unjust rates” had to be determined by the Federal Trade Commission and the Interstate Commerce Commission. New programs such as Medicare increased the scope of federal responsibilities in health care, while programs based upon federal grants (such as Medicaid or federal education grants) would intertwine the federal government with state-level administrators. Finally, the federal government established a variety of national regulatory agencies that extended federal responsibility over environmental policy, workplace safety, a broader range of consumer and product safety.

The basic structure of the American bureaucratic state has not changed since the development of nation-wide social and economic regulations in the 1970s. What has developed during that time are serious questions about how to control the bureaucratic state. In many cases, the debate over bureaucracy has fallen into familiar partisan patterns, with Republicans concerned with the scope of government power, and Democrats concerned with the failure of the bureaucratic state to ameliorate the abuses and inequalities of capitalism. Yet the issue of how to control the bureaucratic state in a democratic society is something of concern to individuals from across the political spectrum. To understand this issue, we will delve into the ways in which bureaucracies make and implement policy, the ways in which interest groups attempt to influence this process, and the ways in which Presidents, Congress, and the Courts attempt to direct bureaucratic power.

III. The Bureaucratic Process in the 21^st Century↑

Bureaucracies are organizations that exist in both the private and the public realm. What do these different types of bureaucracy have in common? In contrast with more informal modes of organization—such as a family, or perhaps a club– bureaucracies organize their regular activities in terms of official duties. The authority to discharge these duties is structured by rules that determine the means placed at the disposal of officials. General rules, in addition, are used to determine access to official positions—unlike a democratic legislature, where elections determine access to official positions. Modern government bureaucracies exist to put public policy into effect; police enforce criminal law and protect public safety; forest rangers… range the forests… teachers teach, or pretend to teach, and so on. Yet public bureaucracies differ from private bureaucracies in other important ways as well.

Government bureaucracies operate under constraints that do not exist in the private sphere, particularly with regards to the use of labour and capital (or, in general, any decision about the use or conservation of resources.) Consider the position of a public manager of a welfare office versus the manager at a McDonald’s restaurant. The public official will have much more difficulty firing a chronically absent or habitually drunk employee. In comparison with the owner of a small, independent restaurant, the McDonald’s manager might have less control over their choice of practices and procedures. The manager of a welfare office will face many more constraints. A government welfare office, even if it administers policies that were developed recently, is not distinctly modern—the manager and employees do not, for the most part, exercise political power (though like police officers, they may exercise some discretion in terms of how they implement rules.) What is distinctly “modern” about some government bureaucracies is the fusion of the traditional governing powers– the fusion of legislative power (the power to make general rules), executive power (the power to enforce rules) and judicial power (the power to adjudicate controversies over the application of rules.)

The Occupational Safety and Health Administration, or OSHA, helps to illustrate how modern bureaucracies fuse legislative, executive, and judicial powers within single institutions. OSHA was created during the Nixon administration, and it was tasked with enforcing the Occupational Health and Safety Act of 1970. Unlike the case of police officers, the various experts working within OSHA do not merely implement law. Rather, they must develop specific workplace standards, working within the framework created by Congress. Congress, through the Occupational Health and Safety Act, delegates its own legislative power to OSHA. The specific rules that govern workplace health and safety are not directly created by Congress, though as we will see, Congress has various ways in which in can still exercise control or influence over OSHA. The delegation of legislative power makes OSHA a peculiarly modern form of bureaucracy; while bureaucracies have always exercised discretion, the key dispute over bureaucratic politics today often arises from the fact that executive branch agencies actually create the rules that they enforce.²²

How exactly does the bureaucratic rule-making process work? The following is a simplified outline of what we might call “the bureaucratic process.”

Stage One: Establishing the Legislative Framework. Congress creates bureaucratic agencies, , and establishes their legal authority and jurisdiction. In the case of the Occupational Safety and Health administration, the underlying law is the “Occupational Health and Safety Act,” which established OSHA as an agency within the Department of Labor, and gave it the authority create the regulations various kinds of hazards within the workplace.

Stage Two: Rule-Making. Many bureaucracies must make “rules” or regulations that specify the general requirements of law. OSHA, for instance, must make specific determinations about what particular standards to establish, for almost incalculable number of workplaces. Legal and scientific experts working within OSHA are responsible for crafting particular rules, though they are subject to legal and political constraints. The most important legal constraint come from the Administrative Procedures Act of 1946, which requires all bureaucracies to subject new regulations to a “notice and comment” period, which allows interested parties to make their views known prior to the promulgation of new regulations.

Stage Three: Enforcement All agencies must determine how implement the rules that they are tasked with enforcing; each agency will approach the problems of enforcement in different ways. OSHA rules are enforced by the agency itself and by state-level regulatory bureaucracies. Enforcement occurs through workplace inspections, which can be initiated by employees who submit complaints. However, the primary locus for enforcement decisions comes from OSHA itself. To facilitate enforcement, OSHA divides the nation into 10 different regions, each of which has approximately 75 inspectors. State governments can choose to assume the enforcement of OSHA responsibilities, if they submit an acceptable plan to demonstrate that state enforcement will be at least as effective as federal enforcement. Enforcement itself involves decisions about when and how to inspect workplaces, as well as decisions about how to handle informal settlements.

Stage Four: Adjudication Many bureaucracies feature quasi-judicial tribunals (often referred to as administrative law judges) which review disputes over enforcement decisions. Disputes over the application of OSHA rules are head by Occupational Safety and Health Review Commission. Decisions of the commission can be appealed in federal courts, though some penalties are imposed without any possibility of review or appeal.

OSHA is thus a quintessentially “modern” bureaucracy, because of the way in which it exercises the three main aspects of government power (legislative, executive, and judicial.) In contrast with the U.S. Postal Service, OSHA does not merely implement decisions made by others; it must create the very rules that it implements. Unlike the Post Office, decisions about how to apply those rules involve a variety of discretionary decisions about what businesses to inspect, how to interpret rules, and whether to allow for informal settlements. Post officers may occasionally decide to evade cantankerous dogs instead of performing their duties, but the scope of their “executive discretion” is fairly limited. Finally, OSHA deals with highly complex and contentious issues that are either disputed in quasi-judicial bodies within OSHA, or within the federal courts. The U.S. Postal Service may be imperfect, but it has to go before administrative tribunals or courts only in rare circumstances. To understand how differing bureaucracies work, then, we have to understand their relationship with the other branches of government, as well the specific tasks that they aim to achieve.

IV. The Structure and Tasks of the American Federal Bureaucracy and its Relationship with Congress and the President↑

Bureaucratic organizations, whether in the public or private sector, tend to be both hierarchical and complex– hierarchical in the sense that there is a chain of command within the organization, complex in the sense that discretion can be exercised at various levels within the hierarchy. If bureaucracies merely implemented the choices made by Congress and the President, there would not be very much to say about them from a political perspective. Yet bureaucracies are not always mere transmission belts for the decisions of others. American bureaucracies are particularly complex because it is difficult to identify the top step in the hierarchical ladder. The President possesses the executive power, and thus appears to be the chief executive officer of the sprawling federal bureaucracy. Yet Congress passes laws and budgets, and thus has ultimate power over what bureaucracies can do, or even whether they continue to exist. While Congress may create the laws that establish bureaucracies, courts end up interpreting them, often at the behest of interest group litigants. While it is not quite correct to say that bureaucrats in Parliamentary systems serve only one set of masters (the Prime Minister and Cabinet) it is the case the American bureaucrats are subjected to a more confusing set of political masters.

The American national bureaucracy is organized into agencies and departments. Agencies are the basic organizational unit; there are about 800 of them. Most agencies are organized into departments, headed by cabinet secretaries. So, for instance, the Forest Service Agency is part of the United States Department of Agriculture, which reports to the Secretary of Agriculture in the Trump Administration. Not all agencies are part of departments; some agencies, such as the Environmental Protection Agency, are “free standing,” which means that they report directly to the President or a board. The Central Intelligence Agency fits into this category as well. In addition, there are independent regulatory commissions which are not directly under the control of the President.²³

Presidents and Congress have various ways of exercising control (or at least influencing) federal bureaucrats:

1.Law: Congress and the President possess the ultimate power to control federal bureaucracies, because they possess the ultimate power to shape the powers available to those bureaucracies, or even to eliminate them entirely. In addition to the laws that govern specific bureaucracies, Congress can also create various procedural requirements that bureaucratic agencies must follow. Examples include the already mentioned Administrative Procedures Act, the National Environmental Policy Act (which requires that agencies must submit an Environmental Impact Statement when taking actions that could have a significant effect upon the environment), and the Privacy Act (which requires that government agencies keep private information confidential.)

2.The Budget Process: Congress influences agency actions by making specific spending decisions, both at the authorization stage (which establishes the maximum amount that can be spent on a program) and at the appropriations stage (when Congress determines the specific amount of money that will be spent on a policy in a given year.) Congress authorizes spending on a particular program at its inception, though the length of authorization can vary; some programs are authorized for a number of years, some are permanent, and some are renewed on an annual basis. Usually, most established agencies are more concerned with the decisions of the appropriations committees, who (in theory, and often in practice) can determine agency priorities through funding decisions.

3.Personnel Presidents and Congress can exercise considerable influence over bureaucracies by controlling personnel. At the highest level, the cabinet officials that are nominated by the President and confirmed by Congress play a crucial role shaping how the federal bureaucracies operate. Consider, for instance, the role played by the Attorney General, who heads the Justice Department. The Attorney General can set policy priorities for federal prosecutors (U.S. Attorneys)²⁴ who are also appointed by the President and confirmed by the Senate. Though guided by Presidential directives and official department policies, the Attorney General and U.S. Attorneys about how to enforce the law, which issues to prioritize, and which individuals to prosecute. Robert Jackson, who served as Attorney General for Franklin Delano Roosevelt before being appointed to the Surpreme Court, described the task of federal prosecutors as follows:

Law enforcement is not automatic. It isn’t blind. One of the greatest difficulties of the position of prosecutor is that he must pick his cases, because no prosecutor can even, investigate all of the cases in which he receives complaints. If the Department of Justice were to make even a pretense of reaching every probable violation of federal law, ten times its present staff would be inadequate. We know that no local police force can strictly enforce the traffic laws, or it would arrest half the driving population on any given morning. What every prosecutor is practically required to do is to select the cases for prosecution and to select those in which the offense is the most flagrant, the public harm the greatest, and the proof the most certain.

Needless to say, the decisions made by federal prosecutors and many other high level federal bureaucrats have political dimensions, and it is for this reason that Presidents are able to appoint and remove them. President Trump, for instance, has made approximately 2000 “political appointments” within the executive branch. Other members of the vast federal bureaucracy are not subject to direct Presidential control, on the grounds that low-level federal bureaucrats do not exercise a political form of power.

The question of whether, or under what circumstances, the President and Congress will be able to exercise effective control over bureaucrat—and thus solve the “principal-agent” problem—is difficult to determine. One political scientist, James Q. Wilson, argued that the ability to control bureaucrats depends upon what the bureaucracies are trying to accomplish, and the ways which they go about accomplishing those goals. According to Wilson, we can conceive of bureaucracies in terms of the work that bureaucrats do on a day to day basis (their “output”) and the goals that they aim to achieve (“outcomes.”) The outputs and outcomes of bureaucracies are either easy to observe and measure, or difficult to observe or measure. This produces a schema of four different types of bureaucratic agencies:

Table 8.1: Bureaucratic Outputs and Outcomes

Production Agencies

Under some circumstances, bureaucracies are assigned goals (“outcomes”) that are relatively easy to define. In some cases, the means used to accomplish those goals (“outputs”) are relatively easy to measure and standardize. “Production agencies” exist when both of these conditions are present. Assuming that the bureaucrats in such agencies are relatively non-corrupt, it will be easy for Congress and the President to control what the bureaucrats do— a production agency might be large, it might be responsible for huge amounts of resources, but these types of agencies will usually be guided by fairly specific federal laws, and the bureaucrats working within the agencies will only exercise a minimal amount of discretion.

In the American context, the best example of a production agency is the Social Security Administration—the standards for receiving Social Security checks are set by Congress, the standards are clear and are not subject to much interpretation, and little independent judgment is acquired when applying those standards. In other cases, agencies that fall within the broad category of “production agency” might exercise a greater degree of discretion. The IRS, for instance, has to apply the federal tax code, a law that is far more complex and ambiguous than the Social Security Act and its various amendments. It is possible to observe and measure some aspects of the work performed by the IRS: the number of tax audits, for instance. Unlike the case of the Social Security Administration, however, bureaucratic discretion cannot be entirely eliminated from the IRS, even when the IRS agents themselves are not motivate by any particular political animus.²⁵

Procedural Agencies

Governments pursue many goals that, while they are legitimate subjects of state intervention, may be difficult to achieve. It may be difficult to know whether the goal has actually been achieved, it may be difficult to know whether the government actions were responsible for either the success or failure to achieve the goal, and there may be uncertainties and disagreements over what the goal actually is. Consider again the example of the Occupational Health and Safety Administration. This bureaucracy is responsible for insuring (amongst other things) a “healthy” and “safe” working environment for American citizens—a responsibility that gives OSHA jurisdiction over millions and millions of workplaces. Determining whether a workplace has had a detrimental effect upon the health of workers is no easy thing—health and wellness are affected by individual genetic predispositions, life-style choices, diet, the broader environment, and so on. It is therefore difficult to measure whether OSHA is doing an effective job, because of the multitude of various factors that shape “health.” Given the complex and over-determined nature of one its major goals, OSHA (and bureaucracies in similar situations) have a tendency to focus on the procedures used by its agents when evaluating them: the number of workplaces inspected, the number of citations issues, and so on.²⁶ Bureaucracies such as OSHA are “procedural agencies” because they pursue difficult to measure goals through relatively standardized procedures. Similarly, state employment agencies tend to micro-manage the actions of their employees (e.g. the kinds of services they offer to clients), for the simply reason that it is more or less impossible for the bureaucrats to know if they are achieving the goals of the agency. Employment is a consequence of a host of micro-level decisions (e.g. the initiative of the client) and macro-level factors (e.g. the state of the economy); there is no easy way to measure the actual effectiveness of the employment agencies themselves. If the results of bureaucratic actions cannot be easily measured, then the organization (and its political masters) will have a tendency to focus on the means employed by the organization. This is not possible in all circumstances, however.

Craft Organizations

For some types of bureaucratic organizations, it is easier to specify the outcomes that the organizations are meant to achieve, as opposed to the specific means that those organizations should use to achieve them. This is often because, for a variety of tasks, professional expertise and judgment are necessary, and they must be applied to a whole range of particular circumstances that are not easily predictable and not readily routinized. Bureaucracies operating in these circumstances are “craft” agencies, in that the success of the organizations depend upon the skills and talents (“craft”) of their workforce; political actors can specify the goals pursued by these agencies, but they can usually not determine the specific means that they use.

The American military (during war time) is an excellent example of a craft agency. The American military does not act independently of Congress and the Presidency, in that the political branches of the government determine missions performed by the military. While Congress creates certain rules and procedures that govern aspects of military conduct (e.g. regarding the treatment of prisoners), the military is largely free to determine the specific strategies and tactics that it employs in pursuit of the objectives that it has been assigned. When Presidents have attempted to “micro-manage” the details of military actions, the results have not been very reassuring. The most important controls that elected officials place upon the military are the decisions about when to engage in acts of war.

If the goals of a bureaucracy are relatively unambiguous, and if the actions required to achieve those goals are context-specific and dependent upon highly technical experts, Congress and the Presidency have an incentive to refrain from interfering in the day-to-day operations of that bureaucracy through detailed statutory controls. Aside from broad constitutional and legal guarantees that prohibit arbitrary and illegal actions, elected officials often rely on the norms of professional experts working within craft agencies —FBI investigators, police detectives, anti-trust lawyers in the Justice Department—because the goals of the agencies cannot be achieved without granting considerable flexibility to the bureaucrats who pursue them.

Coping Organizations

Bureaucrats are often tasked with achieving ambiguous outcomes, in circumstances where the possible methods are manifold, uncertain, and difficult to systematize. We can refer to these types of bureaucracies as “coping” organizations. Very different kinds of bureaucratic organizations can be found within this category; for instance, public schools, police departments, and the State Department are all examples of coping agencies. In all three cases, the bureaucrats are assigned goals that are determined by a host of factors. Teachers are given the task of educating students, but there is considerable disagreement over what it means to be educated; even assuming that there was agreement about the goal, teachers cannot possibly control all of the various personal and social factors that might determine whether any given individual actually acquires an education. Police officers are assigned the task of maintaining public order; state department officials aim to promote and protect the interests of the United States in foreign affairs; in both cases, there will also be disagreement about the meaning of the goal, and in both cases it is obvious that bureaucrats themselves cannot be entirely responsible for the success or failures of their organizations. Success or failure will often be the result of social, economic, or political factors that no one can control.

Measuring the effectiveness of coping organizations is thus extraordinarily difficult. This is easier to understand if we compare these bureaucracies to organizations operating in the private market. Consider the example of a private theatre company that performs original works only. As in the case of teachers, police officers, and State department officials, a theatre company pursues a goal that is inherently ambiguous—the production of successful plays. Different individuals and different audiences will define success in very different ways. More importantly, there is no way to completely standardize a theatrical production—you can employ people with experience and a record of success, but even highly trained theatre professionals can fail to achieve a successful production. Success or failure will often be the result of factors—such as the taste of theatre goers—that can be difficult to predict or anticipate. The theatre has one advantage over government “coping agencies” that operate under the cloud of ambiguity: the theatre company has a clear market signal—ticket sales—that can allow its manager to know when they are succeeding at finding an audience. It is also true that, in most instances, it will be easier for theatre companies to experiment with their resources and their personnel (e.g. by firing actors and directors who do not seem to be working out.) Public bureaucracies almost always lack clear market signals that can indicate whether they are failing or succeeding, and they almost always have limited control over their resources and personnel.

It is hard to operationalize or measure the goals of coping agencies, and it is hard (or rather, unwise) to dictate their procedures. The problems of coping agencies result from a host of idiosyncratic factors that simply cannot be controlled by setting up detailed procedures or standardized practices. Consider again the example of public schools, arguably the most important and the most criticized public bureaucracies in the United States. The quality of public schools depends upon leadership, particularly the ability of principals to recruit and retain good teachers and to maintain the morale of faculty and staff. Resources alone do not determine the quality of education.²⁷ Yet there is no magic way to produce good principals, and no way to standardize the intangible leadership practices that allow a person to be an effective principal. This may be because there are many different ways to be an effective manager of a public school, or it may be the case that strategies which work in one school will not work very well in others. Public officials are understandably frustrated with bureaucrats who fail to achieve goals that they are set, but the various attempts to standardize educational practices in order to achieve “best practices” have led to mediocre results at best, unintended consequences at worst.

When faced with complex problems that cannot be “solved” but can only be coped with, public officials are often tempted to ignore the irreducible complexity of the problems that they are trying to solve. For instance, it is possible to try to turn coping agencies into production agencies by changing the incentives of bureaucratic operators. For instance, police officers can be told to focus on ticketing, as opposed to maintaining order or engaging in the kind of informal problem solving that has long been a part of ordinary police-work. Yet the drive for measurement as a solution to the problem of ambiguous goals and difficult-to-standardize methods can lead to unusual outcomes. For instance, the American federal government has made several high-profile attempts to measure and standardize public school education, a process that has been incredibly expensive and more or less ineffective. There are, of course, a very extensive list of factors that could explain why increased federal spending on education has not led to improved educational outcomes. At least one possibility, however, is that the attempts to standardize educational practices might hamper innovative and effective teachers, instead of simply providing guidance and measurable objectives for less effective instructors and principals.²⁸

V. Iron Triangles and the Problem of Regulatory Capture↑

In the immediate aftermath of the New Deal, it quickly became apparent to a wide variety of historians, economists, and political scientists, that Woodrow Wilson’s vision of modern administration as a mere instrument of the general will was not reflected in many aspects of bureaucratic practice. In fact, bureaucracy introduced (or at least exacerbated) a problem in American political life that democracy was meant to solve with ease: the problem of minority faction. According to Federalist Paper #10, a faction is a part of the public, whether a minority or a majority, that pursues objects adverse to the rights of some or the permanent and aggregate interests of the community as a whole. The problem of minority faction is easily solved by representative democracy, or so Madison thought. But bureaucracy reintroduces the problem, through the form of iron triangles and issue networks.

According to the theory of “iron triangles,” bureaucratic decision making, rather than reflecting the will of Congress or the public interest, have a tendency to reflect the interests of a subset of the public, a subset of politicians, even a subset of private groups. Let us consider how this works through the example of the U.S. Food and Drug Administration (FDA). We begin with a decision by Congress to address a complex regulatory goal by delegating legislative power to the executive branch—in this case, the decision by Congress to vest the FDA with the power determine what food, drugs, and cosmetics are safe to sold to consumers within the United States. This power is not exercised by the executive branch alone. Congressional committees will oversee how this power is exercised.²⁹ Relevant interest groups will thus lobby both members of the Congressional committees and members of the bureaucracy in order to shape decisions, such as whether to approve a pharmaceutical. These three groups: the relevant federal agency, the relevant congressional committees which have jurisdiction over the agency, and the relevant interest groups (a regulated industry, for instance) compose the “iron triangle” of bureaucratic policy-making.³⁰

Bureaucratic capture occurs, in theory, when the targets of regulation (e.g. firms or industries) begin to exercise undue influence upon the regulators, aided and abetted by their allies in Congress. In its most extreme form, the close relationships between executive branch bureaucracy, congressional committees, and private interests allows regulations to advance the interests of a regulated industry, as opposed to insuring that private power does not harm the public interest. This situation is referred to as an iron triangle. Policy making in iron triangles creates the conditions for government by minority faction.

Here is another way of viewing the problem: bureaucratic policy-making may make it easier for well-organized, well-financed interests to engage in “rent-seeking” behaviour. The key problem of bureaucratic policy-making is that businesses can now more readily enlist the government to restrict their competitors, or to shape the law in their own favour in some other way. Most of the early studies or regulatory capture focussed on the ways in which influential firms were able to manage the regulatory process in order to limit competition within their industries.³¹ In cases of this kind, the regulated firms actually want strong regulations—particularly those that prevent new firms from entering the market. Another form of capture—what some have called “corrosive capture”—refers to a situation in which powerful industries use their influence to reduce regulatory burdens. For instance, during the Bush administration, national financial regulators prevented state level mortgage regulations from being applied to national financial institutions.³²

Policy-making through iron triangles is problematic because policy ismade by a sub-set of actors most directly affected by the policy (e.g. some members of Congress, bureaucrats, interest groups such as regulated industries) and not by representatives of society as a whole. Can Congress control this?Yes, but individual members want their own particular  projects that they want approved, and their own particular interests to pursue. In other words, bureaucratic politics is subject to the logrollingpolitics that occurs in the legislative process: legislators accept decisions that aren’t necessarily in the public interest, because they have their own goals that aren’t necessarily in the public interest. Presidents and political officials in the executive branch (as well as members of Congress) have limited time and limited expertise, and thus their oversight is often ineffectual. The public doesn’t care about any issue that doesn’t affect them directly, or any issue that is complicated, or doesn’t involve sex or war. In addition, we might add that there are additional incentives for politicians to accept regulatory capture: by helping favoured industries, or favoured firms, or favoured groups and individuals, politicians can hope for and even expect payback in the future. Perhaps this will take the form of campaign donations; perhaps it will take the form of offers of employment once their political careers are over.

Many scholars challenge the idea that “iron triangles” are able to “capture” bureaucratic policy-making. The contrary thesis is rooted in the idea of bureaucratic autonomy— the simple idea that bureaucrats are able to resist private interests when making decisions about the public interest. For instance, it is far from clear that pharmaceutical companies have been able to actually dominate decision making by the FDA or its congressional overseers. Rather than simply respond to the interests of “big Pharma,” the FDA is nestled within an issue network of consumer groups, health professionals, and scientific research communities, and is not simply beholden to a single set of interests. In practice, this has meant that the FDA has been able to impose very stringent requirements on pharmaceuticals; deregulation has often been in response to patient advocacy groups, not simply corporate lobbying.

VI. The Presidency, Congress, and the Bureaucracy: Methods of Influence↑

Given the vast power of bureaucracies, and given the potential for “iron triangles” to influence bureaucratic politics, it is particularly important to consider how they can be controlled or ar least influence by the other branches of government. Because Presidents can act unilaterally, they will often be better positioned than Congress to shape bureaucratic politics. In addition, Presidents have legislative power as well—very few laws can be passed without Presidential support, due to the veto power of the Presidency. While the veto power helps to maintain Presidential control over bureaucracy, the executive power and the appointment power are the most important tools that allow Presidents to shape and direct executive branch bureaucracies.

Presidents can try to control bureaucracies by appointing their ideological allies to key positions in executive branch bureaucracies. The problem is that political appointees will often lack the necessary expertise to fully understand and oversee the bureaucracies they are charged with controlling. Executive orders can also be used to direct the day to day operations of bureaucracies. President Reagan’s Executive Order 12,291 was one of the most significant executive orders in American history; this order required that all new regulation be subjected to cost-benefit analysis by the Office of Information and Regulatory Affairs, an office within the OMB (Office of Management and the Budget, the analytical brain-center of the White House.) As a result, if the OIRA found that the cost of any new regulation outweighed its benefits, then the regulation would not be implemented. Various other executive orders increased the role of the OMB in regulatory oversight.³³

Subsequent Presidents such as Clinton and Bush would modify Reagan’s executive order, but in its essentials it remained in place. President Obama, for instance used the office of the OIRA to shift regulatory standards; to oversee this task, he appointed Cass Sunstein, one of the most influential professors of law of the past 50 years, and someone who has long focused on developing a vision of administrative law that would combine administrative discretion and limits on Presidential discretion in administration. Sunstein argued that Presidents and their administrative czars (like Sunstein) should have the ability to shape the direction of regulatory policy, though he also wrote numerous articles criticizing judicial deference to agency discretion during the Reagan era. His appointment to the OIRA shows that from a procedural standpoint, the Obama administration had no great wish to “de-politicize” the administrative process.

Other Presidents have appointed “policy czars” to oversee the regulatory process, of course, but President Obama’s enthusiasm for selecting and promoting hand picked policy specialists to direct the growth of bureaucratic rule-making even caused some concern amongst members of his own party.³⁴ Senator Robert Byrd (D-WVa) expressed his discomfort in a letter to the President: “the rapid and easy accumulation of power by the White House staff can threaten the Constitutional system of checks and balances. At the worst, White House staff have taken direction and control of programmatic areas that are the statutory responsibility of Senate-confirmed officials.” Thus, President Obama followed in the footsteps of his Presidential forebears such as Reagan, using every means available to shape policy without having to use the cumbersome “Madisonian” constitution.

President Trump’s goal as “administrator in chief” is in theory easier than that of President Obama; while Obama wanted to use the federal bureaucracy to implement ambitious new policy initiatives (often without clear legal authorization) President Trump has claimed that he merely wants to “deconstruct” the administrative state (a phrase employed by his former advisor Steve Bannon.) To deconstruct the administrative state is not the same thing a destroying or removing (which would not only be difficult, but impossible.) Instead, the Trump administration seems intent on simply slowing down the expansion of government regulation—something that is far easier to do than removing existing bureaucracies or altering their statutory authority.

The Trump Presidency provides many illustrations of how an apparently weak President, hampered by investigations into allegations of criminal activity and collusion with foreign governments, divisions within his own party that have made it difficult to fulfill campaign promises, and approval ratings that are hovering in the low-40s, can nevertheless have a dramatic impact upon public policy through influence over the executive branch bureaucracy. Executive orders are one crucial mechanism through which Presidents can alter bureaucratic decision-making. President Trump expedited the approval process for the Keystone Pipeline, an oil pipeline project that had experienced significant delays under the Obama administration. In regards to immigration policy, both the underlying statutory basis of federal authority³⁵ and Supreme Court precedents³⁶ give the President considerable discretion choices regarding how immigration laws are to be enforced; this means that, in practice, Presidents can shape how bureaucracies such as the Immigration and Customs Enforcements service (ICE) perform their duties.³⁷ One should note that the Trump administration does not represent an aberration—in recent decades, all Presidents have used their influence over executive branch bureaucracies to directly reshape policy. ³⁸

The central problem of bureaucracy, as we have seen, is the problematic place of bureaucracy in a system of representative government. Once Congress delegates legislative power to administrative agencies, it is difficult for elected officials to control the actual laws that govern individual conduct. We might claim that the “representative character” of bureaucratic decision making is insured by Presidential control —- but even if you doubt the wisdom of allowing a single individual to reshape policy, you might not be comforted by the fact that the President has considerable difficulty overseeing the bureaucracy as a whole. Some scholars even suggest that bureaucracies should be granted even more independence from Presidential authority, because “bureaucrats” are more representative of the general public (in demographic and economic terms) than are elected officials; through such sophistry, unelected, unaccountable, and un-removable bureaucrats become the most democratic branch of government! But there seems to be no way to solve this problem– if you want a massive administrative state, it is difficult to avoid reducing the representative character of government.

However, it is important to note that Congress can still play a crucial role in bureaucratic politics. The most effective way for Congress and to control bureaucracies is through the laws that authorize policy making, though of course this is never easy, not least because Congress almost always requires Presidential support to change major pieces of legislation. New laws can help to shape bureaucratic discretion in a variety of ways. Consider the National Environmental Policy Act, of NEPA, still one of the most important environmental laws in the USA. One of the main consequences of NEPA was to force all government agencies to complete environmental impact statements when proposing new projects; the goal was to insure (as much as possible) that bureaucracies would take into account the broader public objective of environmental protection, as opposed to making decisions based solely on the concerns of their most immediate constituents and political supporters. Thus, NEPA, like the Administrative Procedures Act, reshaped bureaucratic power in crucial ways by imposing new procedural requirements.

The budgeting-funding process and the appropriations process are two of the most effective ways to assert political control over bureaucracies. The political branches assert control over bureaucracies by evaluating whether the agencies are achieving their goals in an effective manner; this is the significance of “cost-benefit” analysis and planning budgeting. The executive branch and Congress both use cost-benefit analysis as a means of overseeing and evaluating what bureaucracies do; a cost-benefit analysis of any given rule, regulation, or project forces bureaucrats to explain why they are doing what they are doing, and how it will benefit the public. Cost-benefit analysis has a considerable degree of uncertainty; it could only be an exact science if costs and benefits could always be measured in dollars and cents, and they often cannot. Nevertheless, the use of cost-benefit analyses as part of budget planning helps create a kind of dialogue between politicians and bureaucrats.

Congress has several important tools that it can use to control executive branch bureaucracies. Most fundamentally, Congress can control the scope of executive discretion through the legislative process—if Congress wishes to constrain bureaucratic power, it can craft specific rules that dictate what bureaucracies can and cannot do. The appropriations process also gives Congress an opportunity to shape bureaucratic power. Simply put, if Congress does not agree to provide funding for the enforcement of law and policy, then law and policy will not be enforced. Congress can therefore use the appropriations process to control what bureaucracies do, and this gives bureaucrats an incentive to follow the wishes of Congress (or, at least, the members of Congressional appropriations committees.) Similarly, some programs need to be re-authorized after a certain period of time; the re-authorization process gives Congress an opportunity to revisit earlier decisions about bureaucratic power and structure. Finally, Congress has the ability to conduct oversight of agencies (usually as a prelude to exercising their power through the legislative process, appropriations, and re-authorization.) Congressional oversight hearings give Congress the opportunity to determine exactly what bureaucracies are doing, and whether they are following the wishes of Congress.

Despite the tools that are available to Congress, it is important to understand why Presidents will always have an advantage when confronting Congress on the bureaucratic battlefield. The central problem for Congress is that Congress is not a unified institution; and we cannot really think of “Congress” as having well-defined interests. The temporary legislative coalitions that come together to create a new law or policy do have interests– they want Presidents from the opposing party to interfere with those interests as little as possible. The more precise and detailed Congress can be when creating its laws, the more certain it can be that Presidents will not have the ability to reshape policy. Yet when Congress has immense goals– re-shaping one sixth of the national economy through health care reform, for instance– it invariably relies on bureaucratic discretion—and the President’s powers of appointment and oversight give them greater ability than Congress to fine tune the details of bureaucratic implementation.

As we have seen, Presidents have several important advantages in the ongoing struggle over control of bureaucracy and bureaucratic policy-making. To begin with, “executive power” has been interpreted to encompass a broad range of unilateral Presidential action in regards to the bureaucracy: the President (time permitting, and he rarely has enough time) can force agencies to create, review and revise regulations, the President co-ordinates actions between agencies, and the President can remove agency leaders, to name only three of the most important examples. Congress may take some steps to restrain unilateral Presidential power– but keep in mind that Presidents are important players in the legislative process as well, and thus they can often shape law in ways that favour Presidential authority. Perhaps the most important reason for Presidential authority over bureaucracy comes from the very proliferation of law in the modern state. The President, the official charged with executing the law as a whole, must inevitably take primary responsibility for reconciling the various and often conflicting claims of the law.

Presidents tend to have an ongoing advantage in the struggle over bureaucracy because, as long as Congress wants to pursue complex objectives, it will be unable to create laws that limit or eliminate bureaucratic discretion. And Presidents will always have more ability to shape the discretionary decisions of bureaucrats than Congress. Furthermore, while different Presidents may disagree about many different things, they almost always agree that Presidents should be powerful. President Clinton, for instance, did not do away with the centralized regulatory review that President Reagan had established; he used it for much different purposes, and he was influenced by different ideological considerations, but he saw nothing wrong with Presidential power per se.

Vague delegations of legislative or rule-making power to bureaucracies are the source of immense Presidential power, as they allow Presidents to shape the meaning of law, without having to deal with the vexatious process created by the Constitution. Let me give you an additional example of this. When considering environmental policy in the aftermath of the 2010 elections, many American progressives rediscovered the virtues of unilateral executive power, as the election of a Republican- led House made it all but impossible for the Democrats to enact an ambitious agenda to control green house gases. Yet why not simply have the Environmental Protection Agency create new, more stringent environmental regulations based upon their already existing statutory authority?

The relevant law here is the Clean Air Act, which does not mention “green house gases.” When Congress created this law in the early 1970s, it was concerned with the very different problem of particulate matter in the air, not with the problem of greenhouse gasses like carbon dioxide. Still, the language of the Clean Air Act seemed capacious enough to encompass green house gasses. The most relevant statutory standards speak of the authority to create regulations to protect human health. While the legislators were probably thinking of pollutants that directly affect human health, the language of the statute is broad enough to encompass long-term, indirect threats, such as the threats associated with global warming. So we see here that the language of Congressional law seemed to give the EPA considerable discretion in terms of how it exercises its own legislative powers.

We should keep in mind, of course, that the existence of discretion in law does not mean the ability to use that discretion in fact. For instance, imagine that you were one of President Obama’s political advisors– that is, an aide primarily concerned with getting your boss re-elected. The year is 2011, would you advise the President to allow the EPA to ignore the normal legislative process, and enact a vast new form of government regulation based on the vague language of the Clean Air Act and its amendments? Even if you are an environmentalist, it isn’t clear that the answer is “yes.” If you are a progressive first and an environmentalist second, the answer is probably “no.” The President can control the shape of policy, but we should note that electoral considerations and public opinion still place considerable constraints on policy change.

Policy can be affected by discretionary decisions in other ways. Consider the impact of regulatory waivers. A regulatory waiver exempts or waives the application of an otherwise valid general rule to a particular individual or company. For instance, after the implementation of Obamacare in 2014,the department of Health and Human Services has issued 111 regulatory waivers to companies who claim that they will not be able to maintain health coverage for their employees if they are forced to comply with regulation under the new health plan. This is a good example of the “domestication” of prerogative power. What was once an extraordinary act of Presidential power– a decision to step back from the enforcement of the strict letter of the law, in order to address a temporary emergency or entirely unforeseen problem– is now integrated into the very structure of policy-making. The very existence of regulatory waivers illustrates how the notion of the rule of law is dead. General rules are not applied to those whose lawyers and lobbyists are clever enough to obtain a royal dispensation. Sometimes mercy is necessary to correct the excesses of a strict application of the law, and indeed that is why the American president has the power of pardon. But can regulatory waivers be understood as an exercise of mercy?

Given what the American public wants– a national government with the power to provide health, safety, and economic security, a national government that tries to fulfill what FDR called the “Second Bill of Rights,” the aspirations of the New Deal state– the American public must also accept a considerable degree of administrative discretion, and a considerable degree of unilateral Presidential authority to re-shape law. It is, however, easy to exaggerate the problems of bureaucratic discretion– this is because the scope of bureaucratic power and bureaucratic discretion depends upon what they bureaucrats are trying to do, and who they are trying to do it to.

VII. The Politics of Bureaucracy↑

The first political scientists who recognized some of the problems of bureaucratic iron triangles were of two minds about the outcome of the whole process. Or rather, there were two schools: the pluralists and the post-pluralists. These schools refer to general approaches to political science, though they are particularly useful when considering the politics of bureaucracy. The pluralists argued that while the structure of policy-making might have changed in the modern era, social forces continue to shape and constrain political outcomes. These scholars recognized that, contrary to Woodrow Wilson, administration was very much a political process; they just thought that the battle over bureaucratic policy making would reflect, over time, the relative balance of power in society as a whole. Post-pluralists attacked the New Deal state from the political left, on the grounds that modern bureaucracy undercut democratic accountability.

Works such as Ted Lowi’s The End of Liberalism illustrated a curiously American phenomenon: the critique of the bureaucratic state from the left, on the grounds that it is insufficiently democratic, that bureaucratic politics created a kind of “socialism for the organized,” which often meant socialism for the most well established economic actors.³⁹ The post-pluralists arguments were extraordinary influential in practice– this is one of the few instances where the work of policy-oriented political scientists had a real impact on political life; they shaped the developments of the administrative state that we have already discussed in this chapter. But how accurate were the arguments about bureaucratic “capture”? While it is easy enough to see the potential for iron triangles in theory, how prevalent are the iron triangles in practice? The answer, at least according to some scholars (perhaps we could call them neo-pluralists?) is that the problem of regulatory capture depends upon the political environment that agencies operate within.

The Political Environment of Bureaucratic Regulation

The scholar James Q. Wilson⁴⁰ argued that the problem of “capture” depends very much upon the political environment that any given bureaucracy operates within, and that environment can be understood in terms of kinds of costs imposed and benefits distributed by a bureaucracy. Costs and benefits can be understood in a basic sense: how does a bureaucratic decision affect the bottom line of a business or industry, or the pocketbooks of the public as a whole. In some instances costs and benefits can be symbolic. The NAACP attempts to influence the decisions of the Justice Department regarding the enforcement of voting rights laws, and these decisions only have political costs and benefits. Costs and benefits can be concentrated, in the sense that they can affect particular groups within society, or they can be diffuse– they can impose costs and distribute benefits through society as a whole. Wilson’s argument is that the arrangement of the costs imposed and benefits distributed by a bureaucracy will determine the kind of politics that animate any particular bureaucratic agency.

Costs, Benefits, and Bureaucratic Politics

Consider a situation where the benefits distributed by an agency are concentrated, but the costs are distributed throughout the society as a whole. Wilson argues that under these circumstances, we will have a situation where “capture” of bureaucratic agencies by the regulated industry is highly likely. He calls the politics of this environment “client politics,” which means that the bureaucracy will treat the object of its regulation as a client or customer. In this environment, the small group or industry receives a great deal from a government agency; benefits are high (for the industry), costs per person are relatively low. This create a situation where the regulated industry is likely to have a great deal of influence on its regulators: the regulated industry will devote time, attention, and resources to influencing the bureaucracy, because the regulated industry will face the consequences of bureaucratic decisions in a very direct way.

One interesting example of client politics is the relationship between the Civil Aeronautics Board and airlines, prior to de-regulation in the late 1970s and 1980s. The mission of The  Civil Aeronautics Board was to protect airlines against supposedly “destructive competition.” This meant, in practice, that the Board imposed rules that limited competition, and thus led to higher airline prices. The rules reflected the interests of the established economic actors, not the public as a whole, and for obvious reasons: the airlines did not want more competition, and they were willing to spend a great deal of time, energy, and money trying to influence the Civil Aeronautics Board. The public would benefit from more competition and lower prices, but the public might not understand the relationship between prices and competition, and most of the public has better things to do than lobby the Civil Aeronautics Board.

Bureaucratic inertia can be overcome in these circumstances. While economists disagree about many aspects of deregulation, this is one instance where no self-respecting economist could really make the argument that this arrangement made much sense from the perspective of the public interest. Deregulation took place, not under the Reagan administration, but under the auspices of President Carter. This illustrates that bureaucracies, even if they can be subject to capture, are also still subject to external controls, not least the expert judgment of regulators.

On the other hand, some client relationships are much more firmly embedded within the American government. Perhaps the most important example of well-established, politically resistant client relationships lies in the field of agricultural policy. In this area of policy, client agencies often have extremely close working relationships with the groups they regulate. Consider the Agricultural Stabilization and Conservation Service (part of the Department of Agriculture), which became the Farm Service Agency in 1994. This is a federal bureaucracy that oversees federal programs that, amongst other things, pays subsidies to farmers. The service is monitored by 3000 county committees composed of 3 farmers on each committee. Let us assume that, as in the case of the Civil Aeronautics Board, this creates a situation where bureaucracies might not be inclined to make decisions with only the public good in mind. The Farm Service Agency is likely to see the world from the perspective of its clients, and is in that sense “captured.” Why might agricultural policy, as represented by the Farm Service Agency, be more resistant to deregulation? Broader political support, and intense regional support, and bi-partisan support– all of these things help to maintain otherwise dubious “clientelistic” policies. Note P.J. O’Rourke’s great joke about agricultural policy in the United States: “In agricultural policy, there are policies to make prices go up, there are policies that make prices go down, and to balance it all out, there are policies that do nothing at all. ” (from his book Parliament of Whores.)

“Client politics” are not restricted to economic issues; for instance, enforcement priorities of the Civil Rights Commission were clearly influenced by civil rights groups. The basic idea is that segments of the public control or at least influence areas of public policy, because opposition is relatively diffuse and limited. The costs are not so great that they give anyone a particular incentive to mobilize to oppose them. This is the area of bureaucratic politics where “capture” is most likely, and while not inevitable, capture can be enduring.

Entrepreneurial Politics

Consider a situation where the costs of a policy are concentrated, but the benefits are diffuse. Wilson calls the politics of this environment “entrepreneurial,” because a great deal of enterprise is needed to create policies at all under these conditions. Consider safety regulations for automobiles. This is something that the public might have an abstract interest in, but it might not attract their attention; a “policy entrepreneur” is necessary to bring the issue to public attention, in order to sustain the desire for the creation of policies (which usually means, the creation of a bureaucratic apparatus to deal with the ongoing nature of the problem.) Policy entrepreneurs can be politicians with an axe to grind, politicians concerned with the public interest, or private citizens who seek to draw the public’s attention to political problems. In the case of automobile safety, the policy entrepreneur was the private citizen Ralph Nader (who literally wrote the book on the subject: Unsafe at Any Speed.)

Though the distribution of costs and benefits are reversed, entrepreneurial politics often end up resembling client politics; the potential for “capture” is high. Often, policies that impose concentrated costs and distribute diffuse benefits develop as a consequence of major public scandals. For example, the thalidomide scandals of the 1960s led to further reform at the FDA (the food and drug administration); the Enron scandal led to reforms of accounting practices, and so on.

What are the consequences of entrepreneurial politics, in relation to the issue of “capture”? The regulatory agencies are in a hostile relationship with the interests they regulate– unlike the case of client politics. The agencies will always face these enemies, but the allies who pushed to initiate the policies are likely to have limited attention spans. This is particularly true of that most erstwhile ally, public opinion. Scandals produce action, but in the absence of scandal, public attention will drift elsewhere. In such a situation, the agency becomes exposed to “capture”; the decision making processes of the agency are likely to be influenced  by the industries it regulates. Consider the problem of the FDA in relation to the pharmaceutical industry. Consider the beleaguered FDA scientist, charged with the task of evaluating the safety of the newest super-drug from Merck or Pfizer; he or she is overworked and underpaid, particularly in relationship to their private sector counterparts. The FDA interacts with the regulated industries on a daily basis, and there will be a temptation to follow the path of least resistance and give in to industry demands; this is not because of corruption, but merely because it will be difficult to challenge the determinations of the legions of well-paid Merck and Pfizer representatives. Note also the role of career incentives– and FDA scientist today can become a highly paid Merck scientist tomorrow.

Interest Group Politics

Now we turn to the political environment where the costs and benefits of a policy are both concentrated: this creates what Wilson calls “interest group politics.” An example is the Occupational Safety and Health Administration  (OSHA), which is typically lobbied by interest groups representing labour and management. OSHA has a difficult management task: whatever decisions they make, they are likely to anger one of the relevant groups. In an agency of this type, the political winds will shift a great deal; OSHA will be “captured” sometimes by labour, sometimes by management, depending who is in charge of the federal government. Yet the problem of “iron triangles” and “capture” is in a way less of a problem in this environment, because the contending groups always have a stake in keeping an eye on OSHA. Under Democratic administrations, management groups will use their allies in Congress, and in the press, in order to draw attention to decisions that they think unduly favor labor.

One way to think of “interest group politics” in bureaucracy is to think of it in terms of deliberation: when policies impose costs and benefits on very specific groups, bureaucracies can expect to be lobbied by groups with very divergent interests. If we are idealistic, we might think that this will lead to better overall public policy; if we are more cynical, we are likely to think that this simply means that there will be more political controversy surrounding these agencies.

In general, an agency like OSHA will tend to reflect the political preferences of its masters. But political ideology is not the only form of ideology; professions can have their ideological predispositions as well. Health professionals dominate OSHA, and the regulations they produce differ from those that would be produced by economists. Consider the problem of “noise in the workplace.” For health professionals, the answer to this problem is clear: create a quieter work place. For economists, the answer is to make workers wear earplugs. Which is the correct answer? I think that this is a useful example to consider, as it illustrates even further the limits of Woodrow Wilson’s conception of administrative expertise as something separate from politics; this is only true if there is one kind of expert for every kind of problem.

Majoritarian Politics

We come now to our final example: majoritarian politics. Here, we have an environment where the costs and benefits of a policy are both diffuse. In this environment, interest groups are not very active or influential, and thus the potential for capture is very low. Consider the example of the Sherman Anti-Trust Act, which led to the creation of the Anti-trust division of the Justice Department. Businesses did not really lobby for the creation of this policy; instead, it was created out of a vague sense that monopoly is a problem. Few firms are really afraid of being hammered; no one has a key interest in the program. The regulators responsible for determining anti-trust policy have an uncertain mandate, as the laws regarding monopolistic business practices are notoriously vague. This is why, when IBM was charged under anti-trust laws in the 1960s, the case took over two decades to conclude. The Department of Justice did not have to prove that IBM did “X, Y, and Z” Instead, it had to show that otherwise legal actions created a monopoly. An agency operating in a “majoritarian” environment can have a relatively free hand, depending upon how it exercises its power. For instance, the anti-trust division is in a relatively strong position as long as it uses its utterly vague powers to target individual firms and not entire industries (as they found out when they targeted used car dealers as an industry.)

Expertise, and the need for governments to have it, is the starting point of modern bureaucracies; while the political agenda of the 19th century could be implemented by amateurish patronage appointees, the public demands of the 20th century required highly trained elites who would serve politicians while being to some degree independent of them. Several changes in legal doctrine were necessary for modern bureaucracy to emerge: Congress had to be willing to delegate legislative power to the executive branch, and courts had to be willing to defer Congressional choices about what power to delegate, and bureaucratic decisions about how to use delegated power. Proponents of the administrative state argued that “horse and buggy” era notions about the rule of law, ex post facto laws, and other antiquated constitutional practices had to be sacrificed on the altar of progress. The consequences were of concern to progressives and classical liberals: the new bureaucratic forums were just as subject to private influence as was the older Constitutional system. By mid-20th century, it appeared that policy making in the new administrative state was often “captured” by “iron triangles” or “issue networks consisting of the regulating government agency, the regulated industry, the relevant congressional committees, and the broader world of policy intellectuals, activists, and interest groups. Rather than simply enable “evidence based” decision making, bureaucratic policy-making enabled new forms of interest group politics— while at the same illustrating that many administrative issues involve competing interests, and not just technical questions. The response to this problem was “proceduralism,” as embodied in laws like the Administrative Procedure Act with its notice and comment provisions. Judicial intervention followed soon after, as courts allowed interest groups to challenge factual findings of executive branch agencies. The ultimate consequence of this judicially managed hyper-pluralism was hyper-polarized bureaucratic politics and legal uncertainty. The bureaucratic state is now constrained by a variety conflicting legal standards that can be invoked to support whatever outcomes a judge might prefer; whether a judge chooses to invoke “deference” or “hard look” judicial review varies from issue to issue (at least in difficult cases.) The actual ability of bureaucracies to exercise power varies from issue to issue, and Congress certainly can constrain bureaucratic power when it chooses to do so; Congress’ continued willingness to delegate legislative power to the executive branch not only increases Presidential power, but actually enable the meaning of law to change depending upon the occupant of the oval office. Given what the public expects government to achieve, this is unavoidable; one of the most prescient observers of American democracy in the early 19th century was able to predict the outcome:

After having thus successively taken each member of the community in its powerful grasp and fashioned him at will, the supreme power then extends its arm over the whole community. It covers the surface of society with a network of small complicated rules, minute and uniform, through which the most original minds and the most energetic characters cannot penetrate, to rise above the crowd. The will of man is not shattered, but softened, bent, and guided; men are seldom forced by it to act, but they are constantly restrained from acting. Such a power does not destroy, but it prevents existence; it does not tyrannize, but it compresses, enervates, extinguishes, and stupefies a people, till each nation is reduced to nothing better than a flock of timid and industrious animals, of which the government is the shepherd.

I have always thought that servitude of the regular, quiet, and gentle kind which I have just described might be combined more easily than is commonly believed with some of the outward forms of freedom, and that it might even establish itself under the wing of the sovereignty of the people.

Our contemporaries are constantly excited by two conflicting passions: they want to be led, and they wish to remain free. As they cannot destroy either the one or the other of these contrary propensities, they strive to satisfy them both at once. They devise a sole, tutelary, and all-powerful form of government, but elected by the people. They combine the principle of centralization and that of popular sovereignty; this gives them a respite: they console themselves for being in tutelage by the reflection that they have chosen their own guardians. Every man allows himself to be put in leading-strings, because he sees that it is not a person or a class of persons, but the people at large who hold the end of his chain.

Alexis de Tocqueville, Democracy in America

Notes↑