At first glance, the call for democratic administration is somewhat paradoxical. Democracy cannot be understood as the Athenian polis or, less ambitiously, as citizens voting for representatives who in turn govern by majority vote. Powers are delegated to bureaucracies precisely because parliaments do not want to be in the business of legislating minutiae and deciding on the case-by-case application of the law. Administration accountable to democratic institutions at the same time as it is speedy, expert, and fair is the more exact way of putting what liberal democracies strive for when designing their bureaucracies.22
Accountability, fairness, expertise, and speed figure differently in adjudication and rulemaking, the two major forms of government administration.23 It is convenient to think of an agency as a court for some purposes and a legislature for others. When an agency makes a determination of individual liability based on past or present facts it is obliged to follow trial-type procedures because of their value as fact-finding tools and guarantees of individual liberties in the face of state action. Judicial review tends to be demanding because even though the matter has been delegated in the interest of efficiency to a specialized agency, it is one that would traditionally have been handled by courts and therefore judicial notions of fair play and due process weigh heavily. In administration through adjudication, fairness ranks high, expertise less so, and speed and accountability are of equal importance.
In rulemaking, the class of government action to which Community implementing rules belong, the administration is generally allowed to resort to different, less cumbersome procedures. When an agency establishes future rights and liabilities for a class of individuals or other regulated entities, the relevant facts are mostly scientific or social scientific. Consequently, scientific experiments and epidemiological and statistical studies are generally more useful fact-finding tools than witness cross-examination and other trial-type procedures.24 Further, since the decision does not implicate the use of state coercion at a single individual's expense but rather represents a trade off between different socioeconomic interests that will be widely felt, participation cannot conceived as a right to push the state to the wall and test fully enforcement of the law in that instance. Rather it should be framed as the right of those who will be affected to give views and information that, together with a host of other considerations, aid government decisionmaking. This is not a matter of individual justice, as are participation rights in, say, Commission infringement proceedings against companies accused of violating Community competition law. Nevertheless, fairness demands that those who will be expected to comply with the rule have some opportunity for participation. Here, judicial review is generally narrow in scope because judges are neither trained to evaluate the scientific evidence nor institutionally suited to make the socioeconomic trade offs that rulemaking entails. To return to the list of values involved in the design of good bureaucracy, in rulemaking expertise ranks high, accountability to democratic institutions and speed retain their importance, and fairness is redefined as the right of interested parties to give views and information.
Through participation rights, the public or, more precisely, interest groups, also contribute valuable information to rulemaking. They and their members can be expected to have considerable experience in the field and therefore can inform administrators as to the nature and extent of the regulatory problem as well as the expected costs and benefits of various regulatory options.25 In other words, interest groups add to expertise in bureaucratic decisionmaking. Although some of the information interest groups make available can be expected to be slanted, the problem can be handled by including as wide a spectrum of interests as possible. Further, through their relationship with members, access to the media, and legislative lobbying activities, interest groups contribute to public debate on regulatory issues and therefore render bureaucrats more accountable to the general public.26 Finally, interest group participation can improve compliance. To the extent that interest organizations command the loyalty of their members, their inclusion in the rulemaking process gives some assurance that their members will acquiesce in agency policy choices.
European lawmakers are, for obvious reasons, significantly influenced by their national experiences in their efforts to render Community institutions more democratic. In all member states, this experience has been one of parliamentary government. As I argue in this section, however, parliamentary and separation of powers systems guarantee accountability and public participation in administrative rulemaking differently. I argue that, because the European Community is a separation of powers system, European lawmakers should look to the experience of other such governments, namely the United States.
In parliamentary and separation of powers systems, accountability in administrative rulemaking is achieved through different practices and institutional arrangements. Bureaucratic accountability can be conceived of as a classic principal-agent problem.27 Even though this is a term drawn from American literature in law and economics, there is nothing very new about thinking of a legislature as a principal that uses the state bureaucracy as an agent. As with all agents, the bureaucracy might not do its principal's bidding and might neglect to carry out statutes as lawmakers intended. The legislative principal, therefore, uses different techniques to monitor and control bureaucrats. This is not only a question of what legislatures typically do but what they are constitutionally obliged to do. A legislature that controls execution too carefully might violate separation of powers principles, but one that does not control enough might very well be charged with an unconstitutional abdication of power because of the ease with which execution can become lawmaking.
In parliamentary systems, the lawmaking principal -- the party or coalition of parties that won the majority of seats in parliament and goes on to form the government -- directly controls state administration. In what is known as party government, a relatively simple chain of command extends from voters to civil servants. The people elect representatives -- normally more for the party they represent than for their personal politics -- to sit in parliament. The party or coalition of parties that captures a majority of seats in parliament then forms the government, which in turn commands the state bureaucracy. The same party or group of parties controls both parliament and the government and therefore even though separation of powers theory would have parliaments legislate and governments execute, in practice the two functions tend to blur. Parliamentary majorities not only pass laws but also keep an eye on implementation through their relationship with the executive and, vice versa, the government is not only responsible for administering the law but also proposes legislation and pushes it through parliament.
Government cabinets, not parliaments, are the key players in the bureaucratic control game because they are ideally located, by virtue of their central position in both legislative politics and administration, to ensure that the same political agenda behind lawmaking will also guide implementation.28 The civil servants who staff a ministry answer, through a hierarchical chain of command, to the minister, cabinet, prime minister, and majority party or coalition of parties. When it comes time to draft rules, therefore, bureaucrats are supervised by the same politicians who crafted the enabling statute. Accountability is achieved largely through the informal politics of government cabinet guidance, monitoring, and punishment.
Parliaments and courts take a backseat to governments in assuring accountability in parliamentary systems. Parliaments wield a number of institutional tools to keep the administration in check. Yet, albeit with significant national variations, these tools are weak and parliaments do not make heavy use of them. Questions to members of government are a feeble means of directing administration. The power to veto or amend executive rules issued pursuant to statutory provisions is seldom used because it is a drastic measure only appropriate for acts that clearly demonstrate disrespect for legislative mandates. Finally, the power to approve the budget, although a highly persuasive institutional stick, is rarely used in practice; as in most other policy areas, the government is charged with proposing the budget and ensuring that the necessary coalition for its passage exists in parliament.
Courts also play a secondary role in making sure bureaucracies do lawmakers' bidding.29 When they review administrative rules, they guarantee that civil servants stick to the statutory terms they have been charged with implementing. They see their role as protecting parliament's lawmaking powers and ensuring that bureaucrats respect statutory guidelines. But courts are careful, once satisfied that bureaucrats have kept within those limits, to step back and permit the full exercise of discretion conferred upon the executive branch. When, as is generally the case, guidelines are vague and must be inferred from the statute's broader objectives as well as models of rational administrative action, courts are reluctant to intervene. Here too national differences abound: while judicial review in France and Germany is considered to be quite strict, it is much less so in England. Generally speaking, however, courts in European parliamentary systems police only for departures from statutory commands, and flagrant ones at that, because they assume that administrators will be kept on a tight leash by governments.
In separation of powers governments, by contrast, a hierarchical chain of command from the legislative branch to the civil servants responsible for rulemaking does not exist. In the United States, one of the major examples of a separation of powers system, three government bodies, each of which is elected by different constituencies and for different terms of office, legislate: the House of Representatives, the Senate, and the President. The executive power, however, is the sole responsibility of the President.
In a separation of powers system, accountability or, in other words, legislative control of the executive branch, is not guaranteed by the appointments and removal power, as in parliamentary systems, where government cabinets stand or fall based on the number of votes they command in parliament. Based on the American experience, it appears that, instead, the legislative branch relies on independent information-gathering as well as administrative procedure to influence executive branch rulemaking. Mathew McCubbins, Roger Noll, and Barry Weingast, a team of political scientists who call themselves "McNollgast," have explored the full range of techniques that Congress (lawmaking principal) uses to control the federal bureaucracy (the agent).30 McNollgast explain that control may be exercised through punishment (e.g., cutting off agency funding) and reward, which operate in the wake of administration policy choices, or through monitoring, a more direct and constant form of supervision. Congress monitors through permanent oversight activities (oversight committees and the Budget and General Accounting Offices) and ad hoc hearings triggered by constituency complaints or other events. Congress can also monitor bureaucrats with administrative procedure and judicial review. By giving the public access to information about agency activities and rights to participate in agency proceedings and seek judicial review, lawmakers enlist interest groups in the monitoring exercise. Interest groups pressure administrators to observe statutory commands directly, through their participation in agency proceedings and court challenges, and indirectly, by complaining to legislators.
Although McNollgast do not argue that administrative procedure as a control instrument, is unique to separation of powers governments, the hypothesis is consistent with their theoretical and historical account of American administrative procedure. McNollgast assert that control tools are necessary to overcome "political drift," a phenomenon that is much more likely in a system in which one party can win the legislative branch and another party the executive branch. Further, McNollgast trace the enactment of the Administrative Procedure Act to rivalry between a Democratic Congress and a possibly Republican President.31 They maintain that the Administrative Procedure Act was passed in 1946 because, at the time, the Democrats in Congress were afraid that a Republican would win the Presidency and fail to administer New Deal programs and policies as Democrats wished. Such rivalry can only occur in a separation of powers government. In addition, the argument that administrative procedure is characteristic of separation of powers systems is consistent with the observations of numerous comparativists.32
Because the European Community is, in many respects, a separation of powers system, comitology reformers can learn from the American experience with institutional mechanisms designed to achieve legislative control over executive branch policymaking. In Brussels, legislative powers are exercised by the Council and Parliament acting on a Commission proposal and rulemaking power is generally delegated to the Commission. Members of Parliament are directly elected for five year terms. The national bureaucrats, ministers, and heads of government in the Council are appointed by their national governments and parliaments. The Commission is staffed by civil servants and headed by a President and Commissioners nominated by the member states and approved by Parliament for a five-year term. Thus the Commission, when it exercises rulemaking power, should be held accountable to both the Council and Parliament, even though, once appointed, the Commission can, and indeed is under a duty to, act independently for its five-year term of office.
In the United States, the administrative procedure of notice and comment rulemaking permits Congress to exercise control when rulemaking power is delegated to the executive branch. As will be discussed in depth below, notice and comment rulemaking requires federal agencies to disclose significant amounts of information about proposed rules, allows individuals to submit comments to agencies about proposed rules, and gives those individuals the right to go to court to enforce their participation rights in the agency process as well as to obtain substantive review of final rules.
According to McNollgast, rulemaking procedure enables lawmakers to hold bureaucrats accountable in a number of ways. First, the information generated in the course of rulemaking and judicial review allows lawmakers to follow the agency process and react, if need be, before a rule goes into effect.33 Because American agencies are required to announce proposed rules well in advance, solicit the views of private actors with strong incentives to contest agency action, and explain rules in language comprehensible to the general public, Congress is fully informed about the rulemaking process. If Congress objects to the direction taken by an agency, it has ample time to act (e.g., threaten to cut funding) before a rule is enacted. Second, McNollgast argue that interest groups and courts function as third-party monitors in the principal-agent relationship between Congress on the one hand and agencies on the other. By allowing the public to participate in agency rulemaking and bring court challenges, Congress and the President ensure that the coalition of interests that originally backed a piece of legislation will control its implementation as well. By giving courts the power to strike rules that deviate from enabling legislation, Congress secures agency adherence to its policy choices. Although McNollgast acknowledge that interest groups and courts will sometimes do their own bidding -- and themselves become a control problem -- as long as this form of monitoring is less costly than more direct means of control, Congress will continue to support notice and comment.
It would be a mistake to take McNollgast's theory of administrative procedure as reflective of how notice and comment works in practice. Congressional control through administrative procedure is certainly not mathematic.34 Congressional staffers are unlikely to follow esoteric agency proceedings carefully, time and circumstance may very well lead one coalition of interests to press for passage of a law and another one to support subsequent regulations, and courts tend to have a mind of their own, even though they work from statutory text and legislative history. McNollgast's theory, however, points to one basic feature of American administrative procedure. In a system that disperses authority by entrusting legislative and executive powers to independent government branches, administrative procedure facilitates accountability. An administrative process that generates copious and intelligible public information about regulatory change and is open to interest groups who themselves have access to legislative branch politicians and courts is undoubtedly more accountable than one where information is scarce and debate is limited to the community of politicians and bureaucrats that staff the executive branch.
Although the form and extent of interest group influence over policy implementation vary greatly in Europe's parliamentary systems, everywhere interests play their part.35 In most countries, interest group first must win insider status within the policy community of party politicians and civil servants, after which they are informally consulted on everything from major legislative proposals to technical implementing rules.36 Note two features of this system of interest group representation. First, as a result of the strong and unitary nature of parliamentary governments, state actors can and do wield a heavy hand in selecting which interests will have a say in policymaking.37 For instance, in Germany, employers' associations and trade unions have historically had privileged access to the state in areas as diverse as industrial, energy, and health care policy.38 In France, central administration tends to rely upon individual firms, the major employers' organization, professional associations and well-respected local figures (notables) for information and consensus-building.39 In Britain, certain organizations such as the National Farmers' Union enjoy a monopoly of representation and close relations with the relevant government ministries.40 Second, again related to the structure of parliamentary systems, the same set of party politicians and civil servants are responsible for policy formulation and implementation and therefore once an interest group is recognized as legitimate, it is relatively sure of access to policymaking from start to finish. Interest groups rely on informal processes, not on legal guarantees, to influence policymakers. Indeed, to the extent that procedural rights do exist in parliamentary systems, they are a nuisance to such interest groups because they allow outsider groups without significant political clout to influence public administrators.
In separation of powers governments, the public relies far more heavily on legal process to influence policy implementation than it does in parliamentary systems. To take the American case again, a wide array of interest groups participates in the legislative process because political power is shared and there exist many points of access: senators, members of the House, the President, and political parties. In the United States, it is more difficult to shut out certain groups than in parliamentary systems because of the fragmented nature of government authority. Interest groups constantly compete among themselves for influence, in what is called a pluralist system of interest representation.41 They take part in lawmaking by shaping public opinion, making contributions to powerful members of Congress, the President, and political parties, and delivering votes.
The same pluralist constellation of interest groups influences policy implementation, largely as a result of administrative procedure. Even if an interest group does not happen to be an important constituency of the President, it can still influence agency action by exercising its legally guaranteed right to participate in the administrative process and seeking judicial review. In other words, the President and her executive branch political appointees cannot favor certain groups at the expense of others, which might have more clout in Congress, because all interest groups are legally guaranteed access to the administrative process. To borrow from the public choice literature that will be discussed in greater depth below, administrative procedure reduces the risk of systematic capture by executive branch constituencies.
To turn specifically to administration by rulemaking, when American agencies issue rules, they must inform the general public -- not just select constituencies -- in advance and consider comments that all interest groups are free to submit. If an agency decides to disregard comments, it must explain why. Participants that are dissatisfied with the results of the process may challenge rules in court. Through notice and comment rulemaking, therefore, the American system guarantees pluralist interest group participation when the legislative branch delegates rulemaking power to regulators.
As in the United States, the relationship between interest groups and government in the Community tends towards the pluralist model.42 This is somewhat surprising given that traditionally, in the individual member states, relations between government and society have been highly managed and have tended toward the corporatist model. There are many explanations for this development. First, a constitutional system of independent branches and shared powers creates a large number of access points for social groups and consequently allows for a wider range of participation. A Commission civil servant charged with protecting the supranational interest, a Council minister who represents the national interest, and a member of Parliament with affiliations to a national party and electorate, are likely to have very different ideas as to which interests should count and to what degree. Consensus is especially difficult to achieve when policymakers come out of different national traditions of interest representation. To take a clear example, in Germany and Scandinavia trade unions participate regularly in policy formulation and implementation, while they are excluded for the most part in the United Kingdom and France. Without a system of strong parties and central administration, however, the diverse interests that influence Community policy formulation can very easily be excluded from policy implementation. The Commission alone, supervised only loosely by the Council and even less so by the Parliament, is charged with rulemaking and therefore the interests that had a say in the lawmaking process, through a national minister or a European parliamentarian, will not necessarily have one in implementation. The informal mechanisms for involving interest groups in policy implementation that work in parliamentary systems because of the fusion of legislative and executive powers and the close relationship between parties and government ministries, are inadequate in the Community's separation of powers system. Rather, European comitology reformers should look to the American experience with the administrative procedure as a possible means of enabling interest groups to participate in administrative rulemaking.
22 See Renaud Dehousse, Citizens' Rights and the Reform of Comitology Procedures: The Case for a Pluralist Approach, EUI Working Paper RSC No. 98/4, at 9-10 (defining values legitimating Community administration as legislative mandate, accountability, due process, expertise, and efficiency and pointing to tensions among them).
23 As should be clear from the discussion that follows, whether administrative action is considered adjudication or rulemaking depends on the nature of the decision being made and not the label given by the agency.
24 I do not wish to suggest, however, that cross-examination has no place in rulemaking. It can be used, for instance, to clarify positions taken by experts in the written record that are not fully explained or appear contradictory.
25 See Peter Schuck, "Against (And For) Madison: An Essay in Praise of Factions," 15 Yale L. & Pol'y Rev. 553, 583-86 (1997).
26 See id. at 580-83.
27 See Mathew D. McCubbins, Roger G. Noll & Barry R. Weingast, "Administrative Procedures as Instruments of Political Control," 3 J. L. Econ. & Org. 243, 247 (1987) (citing a number of authors for the proposition that bureaucratic compliance with legislative policy choices is a principal-agent problem); Mathew D. McCubbins, Roger G. Noll & Barry R. Weingast, "Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies," 75 Va. L. Rev. 431, 440-45 (1989).
28 See generally James Q. Wilson, Bureaucracy 295-312 (1989) (describing control of bureaucratic discretion in parliamentary regimes as primarily a matter for executives and majority parties and only incidentally for parliaments and courts); Susan Rose-Ackerman, Controlling Environmental Policy: The Limits of Public Law in Germany and the United States 8-9 (1995) (discussing accountability in Germany).
29 See generally Frits Stroink, "Judicial control of the administration's discretionary powers (le bilan executif--judge administratif)," in Judicial Control 81 (Rob Bakker et al. eds., 1995) (describing administrative law in England, the Netherlands, Germany, and France and characterizing English courts as the most deferential and German courts as the least); P.P. Craig, Administrative Law 438 (1993) (describing English standard of review).
30 Mathew D. McCubbins, Roger G. Noll, & Barry R. Weingast, "Administrative Procedures as Instruments of Political Control," 3 J. L. Econ. & Org. 243 (1987).
31 See McNollgast, The Political Origins of the Administrative Procedure Act, 15 J.L. Econ. & Org. 180 (1999).
32 See Bruce Ackerman, The New Separation of Powers, 113 Harv. L. Rev. 633 (2000); Terry M. Moe & Michael Caldwell, "The Institutional Foundations of Democratic Government: A Comparison of Presidential and Parliamentary Systems," J. Institutional & Theoretical Econ. 1991 (1994).
33 Many rules contain long lead times for compliance and therefore they may be reviewed in court before coming into effect, allowing political actors to react to issues raised in the course of the judicial proceedings.
34 See generally Jerry L. Mashaw, "Explaining Administrative Process: Normative, Positive, and Critical Stories of Legal Development," 6 J.L. Econ. & Org. 267, 280-84 (Special Issue 1990) (criticizing principal-agent theory on grounds that it is logically flawed, for the legislature does not necessarily have an interest in ensuring the same coalition that supported legislation influences rulemaking and, more generally, that its empirical claims are non-falsifiable).
35 French government is generally characterized as closed to interest influence and ready to impose policy choices upon social groups, German government as open to interest groups and anxious to obtain social consensus, and Great Britain as somewhere in between. See Policy Styles in Western Europe (Jeremy Richardson ed., 1982) (situating Western European political systems on consensus relationship-imposition relationship axis).
36 Curiously, even though the political science literature on interest group involvement in policy implementation is extensive, most administrative law continues to operate on the assumption of a direct chain of command from statute to administrative act and limited room for administrative discretion. See, e.g., Jürgen Schwarze, European Administrative Law 261-94 (1992) (surveying statutory constraints and administrative discretion in member states).
37 See Graham K. Wilson, Interest Groups 138-42 (1990).
Kenneth Dyson, "West Germany: The Search for a Rationalist Consensus"
in Policy Styles in Western Europe (Jeremy Richardson ed., 1982).
Germany, Austria, Denmark, and Sweden are considered neo-corporatist states,
meaning that government both privileges certain interests and relies heavily on
those interests in policy formulation and implementation. Philippe Schmitter,
the well-known political scientist credited as one of the first to analyze the
return in Europe to corporatist forms of state-society relations in the 1970s,
has developed a handy definition:
Corporatism can be defined as a system of interest representation in which the constituent units are organized into a limited number of singular, compulsory, non-competitive, hierarchically ordered and functionally predetermined categories, recognized or licensed (if not created) by the state and granted a deliberate representational monopoly within their respective categories in exchange for observing certain governmentally imposed controls on their selection of leaders and articulation of demands and supports.
Philippe C. Schmitter, "Still the Century of Corporatism?", 36 Review of Politics 85, 93-94 (1974).
39 See Jack Hayward, "Mobilising Private Interests in the Service of Public Ambitions: The Salient Element in the Dual French Policy Style," in Policy Styles in Western Europe 118-27 (Jeremy Richardson ed., 1982).
40 See Graham K. Wilson, Interest Groups 77 (1990).
41 See, e.g., Graham K. Wilson, Interest Groups 5 (1990).
Justin Greenwood & Mark Aspinwall, "Conceptualising Collective Action
in the European Union," in Collective Action in the European Union 1, 2
(Justin Greenwood & Mark Aspinwall eds., 1998); Helen Wallace &
Alasdair R. Young, "Introduction," in Participation and Policy-Making in
the European Union (Helen Wallace & Alasdair R. Young eds., 1997); Michael
J. Gorges, Euro-Corporatism: Interest Intermediation in the European Community
8 (1996); Sonia Mazey & Jeremy Richardson, "Introduction: Transference of
Power, Decision Rules, and Rules of the Game" in Lobbying in the
European Community 3 (Sonia Mazey & Jeremy Richardson eds., 1993). Here I
borrow another convenient definition from Schmitter:
Pluralism can be defined as a system of interest representation in which the constituent units are organized into an unspecified number of multiple, voluntary, competitive, non-hierarchically ordered and self-determined (as to type or scope of interest) categories which are not specially licensed, recognized, subsidized, created or otherwise controlled by the state and which do not exercise a monopoly of representational activity within their respective categories.
Philippe C. Schmitter, "Still the Century of Corporatism?," 36 Review of Politics 85, 96 (1974).