Mathew D. McCubbins, Roger G. Noll, and Barry R. Weingast, “Administrative Procedures as Instruments of Political Control”
Abstract and Keywords
This chapter examines the perspective introduced by Mathew McCubbins, Roger Noll, and Barry Weingast (collectively called “McNollgast”) to explain the origins and effects of the administrative procedures employed by public bureaucracies in the formulation and implementation of public policy. Founded on concepts of positive political theory, this perspective essentially argues that Congress is quite effective at influencing bureaucratic agencies to pursue policies in its own interests, a theory known as “Congressional dominance.” The chapter reviews and contextualizes McNollgast’s seminal arguments, paying particular attention to their 1987 paper “Administrative Procedures as Instruments of Political Control.” It summarizes the content of the paper and discusses the nature of its originality, as well as the reasons for its importance. Finally, it analyses the paper’s key premises that lead McNollgast to interpret administrative procedures as devices to enhance political control.
Beginning in the 1980s, Mathew McCubbins, Roger Noll, and Barry Weingast—collectively, “McNollgast”—introduced a new perspective on the origins and effects of the procedures that public bureaucracies employ when making and implementing public policy. This perspective was built on concepts of positive political theory and held that major institutions of the legal and procedural environment of bureaucratic agencies were designed purposively by Congress to influence the choices of those agencies in their application of administrative discretion over public policy. Their work became a touchstone of the theory of “Congressional dominance,” the notion, in contrast to influential work on the topic in the 1970s, that Congress is quite effective at directing bureaucratic agencies to pursue policies in its own interests. It created a seismic shift in how political scientists understand the political environment of public bureaucracies and has long since percolated into the thinking of scholars of administrative law and public policy-making from a variety of disciplines. Scholars from myriad fields continue to extend, critique, and grapple with the arguments and insights that McNollgast presented over 25+ years in a series of articles.
Still the most cited of these is the original paper by McCubbins, Noll, and Weingast, “Administrative Procedures as Instruments of Political Control” (1987), (p. 466) published in the then-youthful Journal of Law, Economics, and Organization. At the time of writing, this paper has garnered nearly 2,000 citations on Google Scholar, at an accelerating rate over time. In this chapter I review and contextualize the seminal arguments of McNollgast, with special emphasis on this 1987 paper. I explain the content of the original 1987 article, explore the nature of its originality and reasons for its importance, and discuss its extraordinary fecundity in generating extensions as well as critiques.
McNollgast (1987): The Core Argument
This section lays out the logic of the argument in “Administrative Procedures as Instruments of Political Control.” The context of this argument in the literature prior to the article, and its influence on subsequent literatures, are laid out in subsequent sections.
The essential point of the original 1987 McNollgast paper is that administrative procedures can be reimagined as devices by which Congress enhances its control or influence over the policy decisions of the bureaucracy. This paper viewed bureaucracy-wide procedures, such as those created in the watershed Administrative Procedure Act (APA) of 1946, the National Environmental Policy Act of 1969, and the Freedom of Information and Government in the Sunshine Acts passed in the 1970s, through this lens. McNollgast’s related paper, “Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies” (1989, published in the Virginia Law Review) turned their core logic to understanding specific additions to or modifications of the general APA contained in the specific statutes which various agencies administer.
These papers do not present a specific model, either formal or verbal. They do not present formal hypotheses with tests or any structured empirical analysis. They do not present case-study analysis as scholars understand it today. What they do present is an interpretive lens through which to view the bureaucratic policy-making apparatus in the executive branch, and the compatibility of that apparatus with the democratic aspirations of the United States. The 1987 paper lays out several assumptions, unexceptionable in the positive political theory tradition (PPT) in which McNollgast worked (and continue to work), and works through the logic of how effectively legislators might be able to influence bureaucratic policy-making in light of those assumptions. Though the starting assumptions were in line with much of the PPT literature, the conclusions about bureaucratic control and interpretations of administrative structures were quite original, and at odds with much of the substantive work on these topics at the time. While there is no empirical analysis of a statistical or case-study variety, there are many examples of specific administrative procedures that are interpreted through the lens of the logical framework developed in the paper.
(p. 467) Logical Development of the Argument
There are several key premises in the original 1987 paper that lead to the conclusion that administrative procedures can be interpreted as devices to enhance political control. An important background assumption in this article and the entire McNollgast œuvre is that of rationality on the part of all actors involved in the institutional design and policy-making process. McNollgast stands as a signal exemplar of the PPT approach to the analysis of bureaucratic institutions, and it inherits this commitment to the rational choice postulate as an explanatory and interpretive device from that framework. In this context rationality implies that each political actor responds intelligibly and reliably to a given constellation of incentives created by its political-institutional environment, and that any political actor understands how any other political actor will respond to its incentives. More specifically, actors contemplate the choices available to other actors, understand how other actors evaluate the benefits and costs of various choices, and anticipate that other actors will choose the action most advantageous to themselves from their own point of view in any given situation.
Substantively, the key premise of McNollgast (1987) is that Congress is the architect of many pillars in the institutional environment of bureaucracies, and these affect bureaucratic incentives to choose among various policies, regulations, and implementation decisions. Though the paper does not articulate a specific model, embedded in its depiction of American policy-making is a sequence of choices by political actors that can be thought of as a proto-game. The sequence entails a choice by Congress of the legal-procedural institutions in which bureaucratic agencies act, and a subsequent choice by such an agency about the specifics of public policy or implementation. As Congress is a rational actor itself, we can therefore assume that it chooses the institutional environment for bureaucratic agencies that best suits its own purposes.
Those purposes, as stipulated by McNollgast, involve exerting influence or control over public policy. The Congressional incentive to exert this influence stems from the notion that policy outcomes affect the ability of members of Congress to hold their seats, and the common assumption in the PPT tradition that members of Congress wish above all to secure reelection. Given the complexity of modern policy, the “policy outcomes” of interest to members of Congress are not determined solely through legislation or any simple declaration of Congressional will. They are mediated by bureaucratic agencies, through the regulations they issue to give life to legislation and their implementation of statutes granting them regulatory or enforcement power.
Yet the ideal choices of bureaucratic agencies on these implementation, regulation, and enforcement decisions, left to their own devices, are unlikely to coincide with the ideal choices of members of Congress. The reason is that bureaucrats bring their own values to bear in making these decisions. These values may be informed by personal beliefs and preferences of bureaucrats, professional norms imbued through the specialized training bureaucrats receive, received wisdom within the bureaucratic agency, the perspective of interest groups with which bureaucrats frequently interact (p. 468) and on which they may depend for future employment, or even (but not necessarily) aversion to effort or maximization of agency “slack.” This implies that the uncontrolled decisions of bureaucratic agencies will not coincide with those that maximize the electoral fortunes of members of Congress. In short, as recognized in much literature prior to McNollgast, Congress faces a principal-agent problem with respect to the bureaucracy.
Inasmuch as electoral fortunes depend on policy outcomes, policy outcomes depend on decisions of bureaucratic agencies, and the uncontrolled decisions of bureaucratic agencies do not coincide with those preferred by Congress, Congress therefore has an incentive to influence or control the decisions of those agencies. From a positive standpoint, members of Congress, as rational actors, recognize this problem and attempt to solve it as best they can. One approach to controlling those decisions is to monitor them and apply sanctions when and if they fail to coincide with those preferred by Congress. If Congress’s preferred decisions are known and understood by bureaucratic agencies, then the monitor–sanction approach follows a more or less standard logic of rational deterrence theory. The legal-institutional environment of bureaucracies in the US is such that the monitoring and sanctioning can be executed either by Congress itself, through oversight, or through the courts, through judicial review of agency actions.
McNollgast contend that neither legislative oversight of bureaucratic agencies nor judicial review of agency actions are sufficient to realize the desired level of influence by Congress over bureaucratic policy decisions. While these channels of influence may be important, McNollgast contend that they are not alone up to the task. With respect to legislative oversight, McNollgast contend that both monitoring and sanctions are limited in their efficacy. Monitoring through legislative oversight is costly, in the sense that time spent on it is time taken away from other valuable activities such as fundraising, campaigning, legislating, and constituent service (cf. McCubbins and Schwartz 1984). Moreover, inasmuch as bureaucratic agencies possess more information about their activities than Congress (and its support agencies such as the Congressional Budget Office (CBO) and Government Accountability Office (GAO)), monitoring alone may be insufficient to reveal the difference between agency policy choices and those preferred by Congress. The problem is one of eliciting private information held by the agency.
Compounding the costs and inherent limitations of monitoring by Congress are the limits on and costs to sanctions Congress can impose. Criminal liability for bureaucratic noncompliance applies only in limited cases, and cutting budgets or programs, while perhaps costly for bureaucrats, is a double-edged sword that may be costly for Congress too, inasmuch as the programs in question are electorally useful. Moreover, McNollgast point out that tying up the legislative process with attempts to sanction bureaucrats involves undesirable delay and distraction.
Another channel of monitoring of bureaucratic policy decisions is judicial review. As long as bureaucratic agencies have made discretionary decisions with respect to policy or enforcement, federal courts have stood to review those actions when agencies are sued in federal court. Administrative law is the body of law stipulating the principles of this review. Constitutional guarantees of due process for individuals (p. 469) affected by bureaucratic decisions are one rationale for this review. In addition, the US Constitution does not permit Congress to delegate its power to make law (the “nondelegation doctrine”); thus, individuals can appeal to federal courts to ensure that agency decisions are made pursuant to specific statutory provisions enacted by Congress, and not reflective of extra-legal innovations made by agencies themselves. Federal courts can invalidate agency decisions that do not meet these standards of review.
Nevertheless, McNollgast argue that judicial review is also not a perfectly reliable mechanism of bureaucratic control from Congress’s point of view. First, legislative dictates are not always specific enough for a third party such as a court to determine whether an agency has in fact acted consistent with the wishes of Congress. Second, judges bring their own values to bear in judging, be they garden-variety ideological values of the sort other political actors are assumed to follow or legal norms inculcated by training and service in the legal profession. Either way, judges are not automatons that pursue and implement the will of Congress. Third, even if judges are aware of and wish to implement the will of Congress, they often cannot compel bureaucratic agencies to make a specific decision, and may instead simply remand a case for further consideration by the agency (e.g. when the evidence offered by an agency is insufficient to justify its decision in the eyes of a court).
Both legislative oversight and judicial review of agency actions are “ex post” mechanisms of control. They operate after the fact, after the relevant agency decisions are made. McNollgast recognized that Congress, in its search for influence, is not limited to these ex post mechanisms. Instead, they recognize, as did numerous prior scholars, that the procedures by which agencies render policy decisions can affect those decisions. These procedures affect both the set of actors that can present information to bureaucratic agencies and the extent to which agencies must consider that information in making decisions. In this way, these procedures affect the information presented to agencies about the consequences of their policy choices, both for the policy outcomes themselves and in terms of the political implications of their choices. Such procedures shape the incentives of agencies to render any particular decision in the first place. They are “ex ante” mechanisms of control.
McNollgast argue that administrative procedures ensure that bureaucracies obtain important information about the technical and political aspects of policy outcomes, and respond to this information. In the case of regulation or “rule-making”—the articulation by bureaucracies of general rules having the force of law—the Administrative Procedure Act of 1946 defines a presumptive set of procedures agencies must follow to issue rules that will pass muster in federal court. By far the most common approach to rule-making is “informal” or “notice and comment” rule-making. This requires agencies to provide public notice of their intent to issue a rule on a given topic in the Federal Register, solicit comments from interested parties about the rule in question, and (pursuant to judicial rulings on the topic, e.g. Citizens to Preserve Overton Park v. Volpe (1971)) produce a final rule (also published in the Federal Register) that is, at the least, not manifestly incompatible with the information presented by interested parties. If the agency fails to comply with these procedural requirements, its policy will reliably (p. 470) be contested in federal court by some (negatively) affected party, and that policy will be invalidated by the reviewing court.
This notice and comment process, while defined in procedural terms, is in fact deeply political. It requires agencies to assemble information on the politically relevant constellation of interests on a policy, to wit, those that are organized enough to submit comments. McNollgast contend that these interests are those which are also likely to be organized enough to matter in electoral politics. After all, the goal for a reelection-oriented Congress is not to maximize the net social benefits of a policy, but rather to identify and respond to those interests which will take note of a policy and act on it in the electoral arena. In essence, notice and comment rule-making is skewed toward those interests that have the wherewithal to make their preferences relevant to Congress. It is not an accident, McNollgast contend, that this is exactly the sort of information Congress would tap if it had to make fine-grained policy decisions itself. Since administrative procedures furnish this information to agencies and force their responsiveness to it, they induce agencies to act in a way more congruent with Congressional preferences than they would absent these administrative procedures. In addition, since agency adherence to these procedures is enforced in federal courts, the procedures outsource the costs of compliance with the will of Congress to the judiciary.
One of the most enduring and influential ideas laid out in McNollgast (1987) is that of “deck stacking.” This idea stems from the fact that administrative procedures structure the influence of various interests in agency policy-making. If a given set of interests is privileged in this process, then the outcomes of the process will be disproportionately responsive to those interests. If Congress wishes to deliver a flow of benefits to a specific set of interests, then it can simply arrange an administrative process in which those interests are given disproportionate weight in agency proceedings. Even if a statute, on its face, contains a broad delegation that an agency should “go forth and do good,” the procedural requirements empowering specific interests will tilt the decisions made by the agency toward those interests. By devising such procedures, Congress can deliver a flow of benefits to those interests without even knowing the specific outcomes they desire to achieve. In addition, this flow of benefits to favored interests will persist even after the legislative coalition seeking to confer them has faded from power. Regardless of the fate of coalitions in the legislature, an agency operating under procedures that provide an advantage to specific interests will continue to tilt its decisions in their favor. In this way, deck stacking allows coalitions in Congress to tailor policy outcomes in its interests, without even knowing the specific contours of the outcomes it desires, and without having to secure its own longevity in the legislature.
Another lasting insight from McNollgast (1987) stems from the realization that a policy or program’s enacting coalition in Congress may not simply wish to present a gift to a specific favored interest. What Congress does in hammering out the details of a policy, thus determining its incidence of benefits and costs, is balance the competing claims of various interests. As the political relevance and balance of those interests changes, Congress would wish for the tilt of benefits and costs among those interests to change as well. McNollgast argued that one function of administrative procedures is (p. 471) to replicate the conflict among interests to which Congress attends in the administrative process. As the organizational power of groups changes—a key dimension of their political relevance to Congress—so too will the force of their participation in administrative proceedings. If bureaucracies must respond to and change policy in light of these changes in the underlying political power of the interests with which they interact, then administrative procedures can induce bureaucracies to change policies over time in roughly the same way that the enacting coalition in Congress would want to change those policies, if it had the authority to do so. McNollgast refer to this aspect of administrative procedures as “autopilot,” because they induce relatively “automatic” (from Congress’s point of view) changes in policy in light of relevant political conditions. Autopilot means that Congress can attain its desired policy outcomes without dedicating additional time to legislative horse-trading. Moreover, combined with deck stacking, it means that an enacting coalition in Congress can deliver not just a flow of benefits over time consistent with its interests, but changes in those flows of benefits consistent with its interests, even after its ascendancy in Congress has lapsed.
While the theoretical logic of McNollgast (1987) is arguably best known for introducing the metaphors of deck stacking and autopilot to illustrate the power of ex ante controls in and of themselves, McNollgast also argue that ex ante controls are complements for ex post controls, in that the former amplify the effectiveness of the latter. Notice and comment rule-making requires the agency to announce its intent to regulate not only to interested parties, but to Congress as well. Requirements of a public record of comments, prohibitions on ex parte contacts between the agency and specific interests, and public disclosure requirements of the Freedom of Information Act and Government in the Sunshine Act mitigate the information asymmetry between Congress and agencies about the policy-making activities of the latter. These procedural requirements on the agency reduce the cost for Congress of acquiring information about agency activities.
Applications of the Argument
The logic just sketched provides the core theoretical framework offered by McNollgast (1987). The balance of their original 1987 paper applies this logic to interpret specific administrative procedures enacted by Congress. For example, the National Environmental Policy Act (NEPA) of 1969 required a large swath of agencies to consider the impacts of their policy decisions on the environment. This consideration was not even a twinkle in the eye of Congress in designing regulatory frameworks administered by agencies prior to the 1960s, yet in that decade the environmental movement gained considerable political ascendancy. Faced with this political change, Congress could have responded by changing each agency’s regulatory mandate in a piecemeal fashion to require attention to environmental impacts—a tall and prohibitively costly order. The approach in NEPA, by contrast, was to stipulate new requirements on agency policy-making in general. In this way, Congress designed a general administrative (p. 472) procedure that, at much lower cost, shifted agency decisions in the preferred direction by empowering a new interest to be heard in agency policy-making.
Likewise, Congress has tailored requirements for agencies to subsidize participation by specific interests, such as consumer groups or small businesses, in regulatory proceedings of agencies that often make rules effecting those interests. These “intervenor funding programs” increase the attention that agencies must pay to these interests. Intervenor programs can be turned on and off by Congress relatively easily as political circumstances dictate. Both NEPA and intervenor programs affect the enfranchisement of specific interests to which Congress wishes to pay special attention by structuring the interest group environment and information to which agencies must respond. These effects are implemented through legislative channels that do not require Congress itself to assess the merits of any specific argument of these interests; instead, that requirement is outsourced to the agencies.
The Originality and Importance of McNollgast
Overall, McNollgast’s 1987 paper and closely related 1989 paper present a penetrating and wide-ranging lens for interpreting the effects of administrative procedures. The metaphors of deck stacking and autopilot vividly suggest that Congress can design institutions to ensure adherence of policy choices to the preferences of Congress, without incurring many of the usual costs of ascertaining political and technical information necessary to make an informed decision. These institutions can effect this adherence, in fact, even after a legislative coalition responsible for their design loses its clout. In short, McNollgast’s analysis suggests that a legislative coalition need have neither the information nor the political power to enact its preferred policy at a given time, and yet that policy may be enacted anyway. What it needs instead are administrative procedures to direct an agent along its preferred path.
This argument has both beguiled and provoked scholars for more than 25 years. The reasons for the significance of this argument are both normative and positive. Normatively, the argument is important because it offers new hope for the compatibility of the administrative state with constitutional precepts of legislative direction of public policy, and ultimately, with at least some notions of democratic control of the state. For about as long as there has been an “administrative state” in the US, scholars have worried about its legitimacy. If unelected bureaucrats make policy decisions with the force of law, is Congress de facto delegating its legislative power? Even as we overlook concerns over the nondelegation doctrine, does Congress, by delegating so extensively, lose control over the contours of public policy?
The argument ebbs and flows, but scholars have never had difficulty finding justifications for the administrative state in constitutional terms, despite reasons for continued (p. 473) worry (and thus, continued scholarship). James M. Landis (1938) is but an early and forceful example. McNollgast’s take on the effects of administrative procedures offers an original vision of a specific channel by which Congress might ensure ultimate control over the content of policy, even as it grants broad, sometimes plenary authority over various policy areas to individual agencies. By reconciling extensive delegation and involvement in policy-making by bureaucrats with Congressional determination of policy, McNollgast provide a rationale for optimism in evaluating the legitimacy of bureaucratic policy-making (but see Stephenson (2008) for a general argument that tight political control may not justify this conclusion). In turn, the problem of ensuring not just legislative, but popular, control over policy-making returns to the classic problem of popular control of Congress. McNollgast have nothing to say about this problem, of course, but it is one with which every constitution of a representative government must grapple. McNollgast simply offer reason to hope that a shift to administrative government will not make the problem of popular control of policy worse, or at least not as much worse as has sometimes been feared (see also Calvert et al. 1989).
This normative thread of the legacy of McNollgast dovetails with the positive aspects of its performance. The first of the positive themes affected by McNollgast deals with the actual extent of legislative direction of bureaucracies. Through the 1970s, influential literatures in economics and political science argued that, as a matter of fact, executive branch bureaucracies were quite poorly controlled by Congress. In economics, Niskanen (1971) initiated an influential line of argument arguing that bureaucrats, endowed with preferences to maximize their budgets, private information about the costs of government programs, knowledge of the legislature’s demand for these programs, and some (poorly motivated) power to make take it or leave it offers of program budgets to Congress, are able to make programs far larger than Congress prefers, and correspondingly (in the respect his model is able to inform) poorly controlled by Congress. In political science, Lowi (1979) argued that broad delegations of policy-making authority to bureaucracies, languid legislative oversight of these bureaucracies, and capitulations of Congress as a whole to intense preferences in interest groups and Congressional (sub)committees, amounted to abdication by Congress of the responsibility and authority to govern. Needless to say, Lowi, like Niskanen, took a dim view of the prospects for Congress to exercise any meaningful control over agencies. Heclo (1977), though focused on interactions between career bureaucrats and political appointees within bureaucracies, rather than the bureaucracy and Congress, also cast doubt on the prospects for overhead control of the policy decisions of agencies. Heclo argued that information asymmetries, long tenure, and strong preferences all contribute to the ability of agency personnel to resist political direction from above.
McNollgast irrevocably changed the discussion about these issues. The paper centered the discussion of administrative institutional structure as a matter of legislative choice. Given that premise, an observer of these institutions, however byzantine or bizarre, must ask: what problem might this institution solve for Congress? Put differently: why would Congress design or passively tolerate bureaucratic institutions that significantly undermine its own goals? For instance, given how poorly Congress fares (p. 474) relative to the bureaucracy in Niskanen’s model—the clearest example of the genre to argue that not only does the bureaucracy run amok, but Congress knows it is the worse for it—the McNollgastian perspective insists that there must be more to this structure than meets the eye.
While this premise is foundational for the McNollgast analysis, it did not originate there. For instance, Fiorina (1981) explicitly proposes a germ of the same idea, arguing “Congress gives us the kind of bureaucracy it wants” (p. 333). The principal-agent metaphor had been asserted in this domain since at least 1975 (Mitnick 1975), and that metaphor trains attention on what the principal can do to mitigate (even if not completely eliminate) this problem. But the literature prior to McNollgast considered primarily ex post mechanisms of control to deal with this agency problem—mechanisms which the literature had, as had McNollgast, found too feeble to be very effective. What is original to McNollgast 1987 is their focus on specific aspects of the administrative procedures—ex ante mechanisms—as the object of Congressional choice, and their sustained application of PPT to explain numerous details pertinent to that focus. McNollgast did not simply raise the notion that Congress-as-principal might have strategies to deal with this agency problem; they identified a set of well-known institutional forms and procedures as examples of them.
The second positive theme affected by McNollgast deals with the legal analysis of administrative law. A major strand in this literature focused on the ability of administrative law to ensure due process for those affected by agency decisions, to prevent public decisions that were “arbitrary” or “capricious,” to ensure consistency of treatment of agencies before courts and individuals before courts as well as agencies, and the like (Stewart 1975). It is not a normative claim to assert that this is what administrative law exists to do, and that was the focus of much legal literature on the topic. McNollgast and sympathetic colleagues writing in law reviews made the politics of statutory administrative law—its conscious design with an eye to picking winners and losers in battles over policy—an integral part of the overall discussion in the legal literature on administrative law (Farber and Frickey 1991; Breyer et al. 1992; Kagan 2000; Bressman 2007; Stephenson 2008, 2010).
The McNollgast Tradition and Devleopment of Subsequent Literature
Scarcely a paragraph of the core theoretical argument in McNollgast (1987) has failed to generate subsequent literature elaborating, extending, testing, or critiquing its theoretical framework. Some of the papers and books elaborating specific elements of their argument have become recent classics in their own right. Much of this work has continued in the PPT tradition of McNollgast, and added explicit formal models that encapsulate new insights.
(p. 475) Extensions and Applications of the McNollgast Perspective
McNollgast postulated that administrative procedures map into the degree of constraint agencies face in their policy discretion, that agencies often have greater information about policy consequences than legislators (as did many other scholars), and that greater discretion would offer agencies greater scope to apply their information in policy choice. Epstein and O’Halloran (1994) formalized these arguments (elaborated in Epstein and O’Halloran 1999) by leveraging the canonical Holmström (1984) model of delegation. Epstein and O’Halloran themselves ushered in a new phase of analysis of administrative structures and delegation, because their delegation model more pointedly reveals the agency loss which principals face from delegation (to wit, it inherently allows agencies to pursue their own policy agendas in their own model) than did McNollgast.
Gailmard (2002) extended this model to consider the possibility, raised in McNollgast, that agencies might seek to evade the formal constraints of administrative procedures in pursuit of their own preferred policies; Volden (2002) extended it to include the role of the President as another, competing principal in the design of statutory procedures, with results that echo the point of Moe (1987) (and McCubbins et al. (1989), who present a spatial depiction of policy drift in a system with multiple principals) that this may attenuate the ability of Congress to induce agencies to hew strictly to its own goals. Ting (2001) developed a model to formalize McNollgast’s insight that applying sanctions to bureaucrats entails costs for Congress—in Ting’s case, the (crude) sanction is budget-cutting, and the (endogenous) cost is that legislators are constrained to consider either “good but wasteful” policies or “bad but efficient ones.” This result again crystallizes the agency loss of blunt and costly mechanisms of control. In addition, Huber and Shipan (2002) introduced a similar but distinct (i.e. not necessarily identical to Holmström 1984) model of delegation which formalizes a key idea in McNollgast, that fashioning constraints on agency discretion carries an opportunity cost for legislators, and the benefits of greater policy responsiveness must be balanced against this cost.
An important part of McNollgast’s argument about the effects of procedures such as notice and comment rule-making is that it helps agencies acquire information about the technical and political consequences of policy. But McNollgast took this acquisition of information as given (and subsequent literature, especially Epstein and O’Halloran (1994) and related work, assumed that the expertise about all relevant policy outcomes already existed in the bureaucracy), neglecting the micro-level incentives of agencies to acquire information and its possible trade-offs. Bawn (1995) began to consider these trade-offs. Articulating a model that assumed that administrative procedures which facilitate political control by Congress come at the expense of agency expertise, and vice versa, Bawn showed that Congress may have to consider expertise of agencies and political control of agencies as competing goals, not necessarily both perfectly attainable from one set of administrative structures.
(p. 476) Gailmard and Patty (2007) considered the microfoundations of agency incentives to acquire policy information. The very rationale for the development of expertise, they argue, may be for agencies to tilt public policy in a direction consistent with the agency’s own values. Expertise that must be applied in pursuit of some principal’s vision of good policy may not be valuable to the agency, and thus not as attractive to acquire. Gailmard and Patty formalize this argument, showing that bureaucratic discretion over policy—and therefore Congressional deference to (rather than control of) bureaucrats may be a necessary (and implicit) “incentive payment” for information acquisition. In addition to establishing why there might be a trade-off between political control and expertise, this paper implicitly suggests that Congress has an interest in influencing important bureaucratic decisions besides the policy choice highlighted in McNollgast. To wit, bureaucrats’ investment in information is also a decision, and it is in the interest of Congress to influence that too.
Gailmard and Patty (2012) elaborate this argument and point out that in many cases the key purpose of bureaucracies to Congress is to elicit information from or share information with political actors besides Congress itself, such as interest groups and the President. In such cases (e.g. eliciting information from regulated interests, or sharing information with the president about national security or economic policy), Congress’s ideal bureaucratic agency may be unresponsive to Congress itself, and responsive to these other actors instead. The administrative structures that best serve the interests of Congress, therefore, may be ones that undermine its own political control and intensify its disagreements with bureaucratic agents (see also Boehmke et al. 2006). While Gailmard and Patty (2012) argue that Congress may not “control the bureaucracy” in any sense obvious to observers, their point is still squarely consistent with McNollgast in that it argues that administrative structures effecting this lack of control are the best ones Congress may be able to design for its own purposes. Congress may simply benefit from imperfect control over agencies, and if so, we should expect to see bureaucratic institutions consistent with that.
Another theoretical extension relates to the complementarity between ex ante and ex post controls claimed by McNollgast. This introduces a tension not explored in their article, but touched on in subsequent literature. If ex post controls are effective at influencing bureaucratic policy-making, and the original, enacting coalition does not exercise these ex post controls (because it has faded from power), then some other coalition is necessarily empowered to influence bureaucratic policy-making. In this sense, the complementarity between ex ante and ex post controls is a double-edged sword from the perspective of the Congressional coalition that chooses ex ante procedures. This original coalition might wish to enact procedures that ensure a flow of bureaucratic policy decisions consistent with its own interests, regardless of the interests of future Congresses. This provides a rationale for “red tape” in bureaucracies, raising the difficulty of exerting influence by future Congresses (Horn 1995). Yet this perspective is naturally consistent with the overall thrust of McNollgast, inasmuch as it interprets administrative structures as designed to ensure policy choices consistent with the will of an enacting coalition. The issue is simply that the enacting coalition (p. 477) may wish to commit future, possibly different coalitions in Congress to follow its will, and this might rationalize structures that make oversight and ex post control more difficult, not less.
A large stream of papers has attempted to operationalize McNollgast in empirically testable terms. Though some of these papers find important cases in which operationalizations of McNollgast’s core theory do not explain their findings, on the whole this literature has found that McNollgast’s arguments significantly help in explaining the contours of administrative procedure. Some of this literature focuses on bureaucracy-wide procedures, as analyzed in McNollgast (1987); other papers focus on agency-specific administrative procedures, as explored in more detail in McNollgast (1989). This literature is large, and space precludes a full survey; only some of the more important contributions are considered here.
Bawn (1997) argued that legislators on committees with oversight jurisdiction over agencies would be less inclined to support ex ante controls over agencies (because they themselves have better access to ex post controls) than legislators not on those committees, and found this expectation borne out in legislation on delegation to the Environmental Protection Agency. Shipan (1997) tested the proposition, albeit closer to McNollgast (1989) (because it deals with procedural design of a specific agency, not bureaucracy-wide structures), that Congress chose judicial review provisions for the Federal Communications Commission in the 1930s with an eye to influencing ultimate policy decisions. He found that members allied with interest groups who believed the FCC would pursue their goals itself fought to curb judicial review, while those allied with interests threatened by the FCC fought for expansion of judicial review. Balla and Wright (2001) applied McNollgastian ideas to the analysis of advisory committees to federal agencies (as structured by the Federal Advisory Committees Act), arguing that these committees affect the flow of political and technical information from interest groups. They find, consistent with McNollgast (1987), that interests which were active in the legislative debate over a bill were more likely to be represented in an advisory committee to the EPA on the same issue. Potoski and Woods (2001) field an original survey of pollution control agencies at the state level and find that, in the view of these agencies and consistent with McNollgast (1987), administrative procedures have a significant effect on interest group and legislative influence. Wood and Bohte (2004) develop a theory of procedural design in the spirit of McNollgast, but focusing in particular on conflicts between the enacting coalition in Congress and potential future coalitions. They argue, in particular, that as enacting coalitions expect these conflicts to grow, they should be more likely to enact procedures that raise the difficulty for future legislative coalitions to influence agency policy choices, and find statistical results consistent with this expectation over a large span of time and range of agencies in the federal government. MacDonald (p. 478) (2010) originated the analysis of “limitation riders” in spending bills, which preclude agencies from spending money for a specific purpose in the fiscal year to which the rider applies, as another avenue of legislative control over the policy decisions of specific agencies. MacDonald showed that these riders are more likely to be deployed when agency policy choices stray too far from the preferences of legislative majorities, demonstrating that they are important tool of procedural control over specific agencies.
On the other hand, Balla (1998) and Spence (1999) found more limited evidence of political control arising from procedural design. Balla (1998) found that notice and comment procedures in the Health Care Financing Administration did not induce it to be more attentive to intended beneficiaries of a policy change than to the traditional interest of physicians. Spence (1999) found mixed evidence for the effects of administrative procedures and structures on decisions of the Federal Energy Regulatory Commission (FERC): although these ex ante control devices did have an effect, FERC was at times successful in resisting political control by Congress. Spence’s findings amplify the point to be raised shortly, that the preferences of the bureaucracy itself cannot be neglected.
The logic and applications of the McNollgast perspective has generated its share of criticism. Their original papers appeared alongside commentaries and critiques published in the same journal issues. Terry Moe (1987) published a critique of the “Congressional dominance” school, including but not limited to McNollgast (1987). With respect to McNollgast (1987), Moe contended that the concept of control is ambiguous, and McNollgast may overstate the prospects for attaining it. While McNollgast (1987) is replete with references to a central concept in the principal-agent literature, that principals in asymmetric information problems with agents rarely achieve their most desired results, this is a notion honored more in the breach than the observance in the original McNollgast paper. By the time the paper reaches the fine-grained analysis of administrative structures, the thread of agency loss is difficult to locate; it seems at least in their verbal expression that administrative procedures may well ensure that Congress’s most preferred policy is not only achieved, but that this occurs at lower cost than Congress would incur by doing the job itself. It is not clear what losses Congress faces in this depiction, and there is reason for suspicion of any principal-agent analysis implying this conclusion.
Another issue raised by Moe (and numerous other scholars, e.g. Hammond and Knott (1996), and indeed by McNollgast themselves in 1987) is that bureaucratic agency problems involve multiple principals. Multiple principals (or common agency) problems typically involve more agency loss for principals than bilateral agency problems, so McNollgast’s effective reduction of the problem to one between (a unitary) Congress and an agency may overstate Congress’s prospects for control. In particular, (p. 479) the original McNollgast paper does not devote substantial attention to interactions between agencies and the President. Since Presidents have developed formidable tools of influence over the bureaucracy (Moe 1985) and bureaucratic structure (Lewis 2003), such as regulatory review by the Office of Management and Budget (Wiseman 2009) and veto power over changes in legislative delegations or other statutes to control agencies (Volden 2002), neglecting the President overstates the potential for unilateral control of bureaucratic policy-making by Congress. Similarly, federal courts exercise their own influence over agency policy decisions (Canes-Wrone 2003), further complicating the prospects for unilateral control by other principals (see also Whitford (2005) and Gailmard (2009)). One implication of this multi-principal interaction may be the existence of space for bureaucrats to articulate their own values and put their stamp on the programs they administer (Krause 1999). In addition, exertions of control by one principal may significantly alter the mix of control tactics pursued by others (Ferejohn and Shipan 1990).
Beyond the space for “uncontrolled” bureaucratic action carved out by multiple principal problems, several scholars have argued that McNollgast understates the role of the bureaucracy itself in public policy decisions. Brehm and Gates (1999) identify a significant effect of bureaucratic culture on policy and enforcement decisions. Krause (1999), Carpenter (2001), and Moe (2006) contend that bureaucratic preferences may affect legislative preferences, not only vice versa. Waterman and Meier (1998) contend more generally that the premise of goal conflict between principals and agents may be overstated. This does not critique the McNollgast logic in situations where there is goal conflict, but it does suggest that the scope of the argument may be more limited than McNollgast claimed.
McNollgast mention several times that one benefit of administrative procedures from Congress’s point of view is that they make it difficult for bureaucrats to assemble coalitions of interest groups in pursuit of their own interests as against those of Congress. But Carpenter’s (2001) epic presentation of bureaucratic network and expertise development in the progressive era suggests that this is more than a possibility. Carpenter meticulously documents how bureaucrats used both expertise and networks with interest groups to show how they achieved not just discretion over policy (as might be granted and controlled by Congress), but autonomy—the ability to pursue their own vision of good public policy even when explicitly at odds with that of key coalitions in Congress. To be sure, the episodes Carpenter depicts occurred decades before the creation of the administrative structures which McNollgast feature as, in part, designed to prevent these occurrences. Yet it is not clear how these structures would inoculate Congress from the effects of the bureaucratic networking Carpenter depicts, for none of it is precluded by modern administrative procedures. Moe (2006) presents an analysis, analytically quite different but substantively similar to Carpenter’s in this respect, demonstrating contemporary bureaucratic interest group mobilization to influence the very preferences Congress brings to bear in legislative–bureaucratic interaction. In short, important work subsequent to McNollgast has indicated that the issue of bureaucratic interaction with (p. 480) interest groups to insulate the pursuit of their objectives from Congressional control is not straightforward to overcome with administrative procedures. Nevertheless, this work is consistent with one version of McNollgast’s point, which is that these problems would be more severe for Congress without current administrative procedures.
In light of Gailmard and Patty (2012), it is not clear that evidence of bureaucratic influence on policy or even autonomy is inconsistent with McNollgast, or at least a slight expansion of it. Principals can benefit from committing to leave policy-making authority to agents (because it provides an incentive to invest in expertise) even when their actions conflict with Congressional preferences, and can benefit from committing to make bureaucratic agents responsive to other political actors besides Congress itself. So findings of bureaucratic autonomy or influence of non-legislative factors on bureaucratic decisions are not necessarily inconsistent with the view that bureaucratic procedures promote the ultimate interest of Congress. On the one hand, this flexibility of the McNollgast perspective in its broadest terms raises questions of whether it is falsifiable even in principle. On the other hand, McNollgast, like much of rational choice theory in general, is important because it suggests the possibility that decisions which seem to undermine some actors’ interests may actually be compatible with them. This is important if for no other reason than that it provides a conceptual lens through which to interpret observed arrangements.
McNollgast (1987) is important because it identifies the foundations of bureaucratic policy choice in the preferences of Congress, and it articulates an original and provocative argument about how administrative procedures establish that foundation. Several scholars before McNollgast had begun to make the first point (e.g. Fiorina 1981; Weingast and Moran 1983; McCubbins and Schwartz 1984; McCubbins 1985). But before McNollgast, scholars were either agnostic on the institutional sources of congressional influence (especially Weingast and Moran) or focused exclusively on ex post mechanisms (especially McCubbins and Schwartz) that, though perhaps more effective than previously thought, are still limited in their efficacy. By focusing on administrative procedures as ex ante controls and identifying new functions for them such as deck stacking and autopilot, McNollgast (1987) offered a richer account of strategies available to Congress than had come before, and demonstrated conceptually how these tools could work. In addition, McNollgast gave a wholly new theoretical overlay on statutory aspects of administrative law, inducing a new understanding of these institutions. McNollgast has continued to be important not only because scholars seek to clarify, extend, and critique their ideas, but because those ideas provided a new lens through which to interpret administrative structures and the determinants of agency policy.
Balla, Steven J. 1998. “Administrative Procedures and Political Control of the Bureaucracy.” American Political Science Review, 92(3): 663–73.Find this resource:
Balla, Steven J., and John R. Wright. 2001. “Interest Groups, Advisory Committees, and Congressional Control of the Bureaucracy.” American Journal of Political Science 45(4): 799–812.Find this resource:
Bawn, Kathleen. 1995. “Political Control versus Expertise: Congressional Choices about Administrative Procedures.” American Political Science Review, 89(1): 62–73.Find this resource:
Bawn, Kathleen. 1997. “Choosing Strategies to Control the Bureaucracy: Statutory Constraints, Oversight, and the Committee System.” Journal of Law, Economics, and Organization, 13(1): 101–26.Find this resource:
Boehmke, Frederick, Sean Gailmard, and John Patty. 2006. “Whose Ear to Bend? Information Sources and Venue Choice in Policy Making.” Quarterly Journal of Political Science, 1(2): 139–69.Find this resource:
Brehm, John, and Scott Gates. 1999. Working, Shirking, and Sabotage: Bureaucratic Response to a Democratic Public. Ann Arbor: University of Michigan Press.Find this resource:
Bressman, Lisa Schultz. 2007. “Procedures as Politics in Administrative Law.” Columbia Law Review, 107(8): 1749–821.Find this resource:
Breyer, Stephen G., Richard B. Stewart, Cass R. Sunstein, and Matthew L. Spitzer. 1992. “Administrative Law and Regulatory Policy: Problems.” Text, and Cases, 3: 19–64.Find this resource:
Calvert, Randall, Mathew McCubbins, and Barry Weingast. 1989. “A Theory of Political Control and Agency Discretion.” American Journal of Political Science, 33: 588–611.Find this resource:
Canes-Wrone, Brandice. 2003. “Bureaucratic Decisions and the Composition of the Lower Courts.” American Journal of Political Science, 47(2): 205–14.Find this resource:
Carpenter, Daniel P. 2001. The Forging of Bureaucratic Autonomy: Reputations, Networks, and Policy Innovation in Executive Agencies, 1862–1928. Princeton: Princeton University Press.Find this resource:
Epstein, David, and Sharyn O’Halloran. 1994. “Administrative Procedures, Information, and Agency Discretion: Slack vs. Flexibility.” American Journal of Political Science, 38(3): 697–722.Find this resource:
Epstein, David, and Sharyn O’Halloran. 1999. Delegating Powers: A Transaction Cost Politics Approach to Policy Making Under Separate Powers. New York: Cambridge University Press.Find this resource:
Farber, Daniel A., and Philip P. Frickey. 1991. Law and Public Choice: A Critical Introduction. Chicago: University of Chicago Press.Find this resource:
Ferejohn, John, and Charles Shipan. 1990. “Congressional Influence on Bureaucracy.” Journal of Law, Economics, and Organization, 6(1): 1–20.Find this resource:
Fiorina, Morris. 1981. “Congressional Control of the Bureaucracy: A Mismatch of Incentives and Capabilities.” In Lawrence C. Dodd and Bruce I. Oppenheimer (eds), Congress Reconsidered, 3rd edn. Washington, DC: CQ Press, 332–48.Find this resource:
Gailmard, Sean. 2002. “Expertise, Subversion, and Bureaucratic Discretion.” Journal of Law, Economics, and Organization, 18(2): 536–55.Find this resource:
Gailmard, Sean. 2009. “Oversight and Agency Problems in Legislative-Bureaucratic Interaction.” Journal of Theoretical Politics, 121(2): 161–86.Find this resource:
Gailmard, Sean, and John W. Patty. 2007. “Slackers and Zealots: Civil Service, Policy Discretion, and Bureaucratic Expertise.” American Journal of Political Science, 51(4): 873–89.Find this resource:
Gailmard, Sean, and John W. Patty. 2012. “Formal Models of Bureaucracy.” Annual Review of Political Science, 15: 353–77.Find this resource:
(p. 482) Hammond, Thomas H., and Jack H. Knott. 1996. “Who Controls the Bureaucracy? Presidential Power, Congressional Dominance, Legal Constraints, and Bureaucratic Autonomy in a Model of Multi-Institutional Policy-Making.” Journal of Law, Economics, and Organization, 12(1): 119–66.Find this resource:
Heclo, Hugh. 1977. A Government of Strangers: Executive Politics in Washington. Washington, DC: Brookings Institution Press.Find this resource:
Holmström, Bengt. 1984. “On the Theory of Delegation.” In Marcel Boyer and Richard Khilstrom (eds), Bayesian Models in Economic Theory. New York: North Holland, 115–42.Find this resource:
Horn, Murray. 1995. The Political Economy of Public Administration: Institutional Choice in the Public Sector. Cambridge: Cambridge University Press.Find this resource:
Huber, John D., and Charles R. Shipan. 2002. Deliberate Discretion? The Institutional Foundations of Bureaucratic Autonomy. New York: Cambridge University Press.Find this resource:
Kagan, Elena. 2000. “Presidential Administration.” Harvard Law Review, 114: 2245.Find this resource:
Krause, George A. 1999. A Two-Way Street: The Institutional Dynamics of the Modern Administrative State. Pittsburgh: University of Pittsburgh Press.Find this resource:
Landis, James M. 1938. The Administrative Process. New Haven: Yale University Press.Find this resource:
Lewis, David E. 2003. Presidents and the Politics of Agency Design. Stanford, CA: Stanford University Press.Find this resource:
Lowi, Theodore J. 1979. The End of Liberalism, 2nd edn. New York: Norton Press.Find this resource:
McCubbins, Mathew D. 1985. “The Legislative Design of Regulatory Structure.” American Journal of Political Science, 29: 721–48.Find this resource:
McCubbins, Mathew, and Tom Schwartz. 1984. “Congressional Oversight Overlooked: Police Patrols versus Fire Alarms.” American Journal of Political Science, 28: 165–79.Find this resource:
McCubbins, Mathew, Roger Noll, and Barry Weingast. 1987. “Administrative Procedures as Instruments of Political Control.” Journal of Law, Economics, and Organization, 3(2): 243–77.Find this resource:
McCubbins, Mathew, Roger Noll, and Barry Weingast. 1989. “Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies.” Virginia Law Review, 75(2): 431–89.Find this resource:
MacDonald, Jason A. 2010. “Limitation Riders and Congressional Influence over Bureaucratic Policy Decisions.” American Political Science Review, 104(4): 766–82.Find this resource:
Mitnick, Barry M. 1975. “The Theory of Agency.” Public Choice, 24(1): 27–42.Find this resource:
Moe, Terry M. 1985. “The Politicized Presidency.” In John Chubb and Paul Peterson (eds), The New Direction in American Politics. Washington, DC: Brookings Institution, 235–72.Find this resource:
Moe, Terry M. 1987. “An Assessment of the Positive Theory of ‘Congressional Dominance’.” Legislative Studies Quarterly, 12 (Nov.): 475–520.Find this resource:
Moe, Terry M. 2006. “Political Control and the Power of the Agent.” Journal of Law, Economics, and Organization, 22(1): 1–29.Find this resource:
Niskanen, William. 1971. Bureaucracy and Representative Government. Chicago: Aldine.Find this resource:
Potoski, Matthew, and Neal D. Woods. 2001. “Designing State Clean Air Agencies: Administrative Procedures and Bureaucratic Autonomy.” Journal of Public Administration Research and Theory, 11(2): 203–22.Find this resource:
Shipan, Charles. 1997. Designing Judicial Review. Ann Arbor: University of Michigan Press.Find this resource:
Spence, David B. 1999. “Managing Delegation Ex Ante: Using Law to Steer Administrative Agencies.” Journal of Legal Studies, 28(2): 413–59.Find this resource:
Stephenson, Matthew C. 2008. “Optimal Political Control of the Bureaucracy.” Michigan Law Review, 107: 53–5.Find this resource:
(p. 483) Stephenson, Matthew C. 2010. “Information Acquisition and Institutional Design.” Harvard Law Review, 124: 1422.Find this resource:
Stewart, Richard B. 1975. “The Reformation of American Administrative Law.” Harvard Law Review, 88(8): 1667–813.Find this resource:
Ting, Michael M. 2001. “The ‘Power of the Purse’ and its Implications for Bureaucratic Policy-Making.” Public Choice, 106(3–4): 243–74.Find this resource:
Volden, Craig. 2002. “A Formal Model of the Politics of Delegation in a Separation of Powers System.” American Journal of Political Science, 46(1): 111–33.Find this resource:
Waterman, Richard W., and Kenneth J. Meier. 1998. “Principal-Agent Models: An Expansion?” Journal of Public Administration Research and Theory, 8(2): 173–202.Find this resource:
Weingast, Barry R., and Mark J. Moran. 1983. “Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission.” Journal of Political Economy, 91(5): 765–800.Find this resource:
Whitford, Andrew B. 2005. “The Pursuit of Political Control by Multiple Principals.” Journal of Politics, 67(1): 28–49.Find this resource:
Wiseman, Alan E. 2009. “Delegation and Positive-Sum Bureaucracies.” Journal of Politics, 71(3): 998–1014.Find this resource:
Wood, B. Dan, and John Bohte. 2004. “Political Transaction Costs and the Politics of Administrative Design.” Journal of Politics, 66(1): 176–202.Find this resource: