The Target of Criticism and Two Visions of Improvement

Reform needs in public administration have been interpreted by means of three generic diagnoses and prescriptions with roots in, respectively, public law, market economics, and democratic politics. Here, the focus is on how effective help is assumed to depend on the organization of public administration and the interaction of administrators and citizens.

The target of criticism is the “classical” Weberian bureaucracy or “Old Public Administration” rooted in continental European public law and the territorial state as a sovereign and autonomous legal entity of its own. The defining activity is implementation of the law. Legitimacy is based on the idea that tasks are technical and problem-solving in nature: to identify a logically correct solution by interpreting rules and facts, or by applying expert causal knowledge to achieve pre-determined objectives. Will formation is exogenous to administration. The institutional core is the merit-based bureaucracy serving as a tool of higher authority—elected leaders, constitutional principles, and professional standards. Binding authority within a specified jurisdiction is claimed through a fourfold rule-bound hierarchical relation of command: between citizens and elected representatives, democratic legislation and administration, within administration, and between administration and citizens as subjects (Weber 1978, Ch. 11). Citizens then are both authors and subjects of law. Administrators are ruledriven, neutral agents acting with integrity on the basis of principles, methods, codes of appropriate conduct, and a public service ethos.⁴

Johan P. Olsen. The 17th Annual John Gaus Lecture, American Political Science Association, Philadelphia 29 August, 2003.

Two reform claims have had a prominent role in international debates since the end of the 1970s. One vision of how citizens can be better helped assumes a paradigmatic shift “from Old Public Administration to New Public Management” (Dunleavy and Hood 1994). A second (and partly overlapping) vision assumes a shift from hierarchical government to a new mode of network governance—“Governance without Government” among and within states (Rosenau and Czempiel 1992; Kickert and van Vught 1995; Peters 1996; O'Toole 1997a, 1997b; Rhodes 1997; Kickert and Stillman 1999; Frederickson and Smith 2003).

“New Public Management,” rooted in market economics and private management (Hood 1991; Pollitt and Bouckaert 2000; Christensen and Lægreid 2001), presents a global diagnosis and prescription: a centrally organized and rulebound public administration is outdated and NPM represents an “inevitable shift” toward a more advanced administration.⁵

All participants are self-interested actors governed by price and incentive. Managers have wide operational discretion and hands-off relations to ministries.⁷

Network governance reforms, like NPM, diagnose and prescribe a move away from hierarchical government. Networks between government and society are assumed to increase in number and importance and a world of networks is emerging (O'Toole 1997b, 46–7). The defining activity of administration is a political process of building support and mobilizing resources. Cooperation takes the form of common problem solving with regard to objective challenges society faces and involves conflict resolution. The legitimacy of networks is based on the acceptability of both the processes used and the substantive outcomes achieved.

The institutional characteristics of a “network” are less agreed-upon than those of “hierarchy” and “market,” but most versions of network governance describe a configuration of voluntary interdependent public and private actors involved in partnerships and joint ventures. Coordination takes place without a dominant center imposing a structure of hierarchical authority in the service of a predetermined objective (Joerges 2002, 3).

Participants with different motivations and resources are involved in bargaining, deliberation, and appeals to common norms. They have to convince and motivate each other. The reach of public administration is expanded but neither elected leaders nor administrators can expect to exercise command or compel compliance by virtue of their formal position. Dichotomies such as state-society, public-private, politics-administration, and expert-layman become obscure.⁸

Actors are (possibly) self-reflective and interaction and experiences can lead to the learning of new preferences and identities. The democratic quality of a network depends on the knowledge of participants and to what extent the network involves bargaining among powerful stakeholders or free deliberation among all affected citizens, voluntary associations, and public institutions (Kickert and van Vught 1995; O'Toole 1997a, 1997b; Rhodes 1997; Frederickson 1999; Raadschelders and Rutgers 1999; Frederickson and Smith 2003). The organization of administration reflects constellations of power in society. Administrative dynamics, including a fragmentation of the institutions of government, reflect changing power-relations and attempts to change existing power balances in society.⁹

Partial Prescriptions, and not so New

How likely are these reform visions to give effective help to citizens? There are good reasons to doubt that any of them will provide a panacea to the perceived malaise of public administration and democratic government.

First, both reform visions, as well as the Weberian bureaucratic prescription, assume a single set of principles for organizing public administration is functionally and normatively superior and that over time one form will replace the others and result in convergence. This view contrasts with the observation that administrative practice and theoretical ideas have been closely linked to the territory, borders, institutions, history, and culture of specific nation states. Administrations with different identities and experiences have been called the “solid bedrock for nationalism” (Nizzo 2001, 2). Long, strong, and varied institutional histories, with different trajectories of state- and nation-building, provide an institutional context that produces variations in the motivation and capability for institutional change and counteracts global or European convergence (Kickert and Vught 1995; Olsen and Peters 1996; Kickert 1997; Kickert and Stillman 1999; Rokkan 1999; Stillman 1999; Pollitt and Bouckaert 2000; Olsen 2003a; Rugge 2003).

Actually, what are presented as universal prescriptions for public administration are partial, time- and space-bound interpretations. Each perspective highlights specific components of the system of administration found in democratic polities, reflects a development in a specific time period, or is associated with a particular reform ideology. What is the institutional centerpiece in one order or reform ideology is an auxiliary institution in other orders and ideologies.

Secondly, the newness of the core ideas of the “new” perspectives is overdone. For example, the functionality of the state has always depended on nonhierarchical techniques of administration and the avoidance of command and force (Möllers 2003). The horror stories of lazy, inhumane, or powerful bureaucracies turning their political masters into dilettantes have been around for some time. Likewise, the tension between a legalbureaucratic and a market-managerial approach to public administration, or the propagation of private business administration as an exemplary model for the public sector, is hardly new (Gaus 1947; Waldo 1948; Mosher 1975b). Interdependencies and networks in the governmentsociety interface are also well known, and have been described for several European countries (Schmitter and Lehmbruch 1981; Olsen 1983; Katzenstein 1985; Rothstein 1996). In brief, many reform proposals are “repackaged versions of ideas that have been in public administration since its beginnings” (Hood 1996: 268). “New” approaches frequently rehash old ideas while the fundamental issues remain remarkably stable (Kettl 1993, 408).

Furthermore, complaints about public administration have not disappeared after decades of reform, suggesting that reformers have not been completely successful (Brunsson and Olsen 1993; Bekke, Kickert, and Kooiman 1995; Bekke, Perry, and Toonen 1996; Frederickson 1996; Kettl 1996; Olsen and Peters 1996; Peters 1996; Aberbach 2003; Peters and Pierre 2003). A possible consequence is that the weaknesses of the “new” perspectives will be rediscovered. Stories about bureaucratic failure will be supplemented with studies of how the use of markets and prices in public administration may create political and social inequality and unrest, rather than help citizens. Studies of public-private networks may come to highlight accountability problems and how insight, access, and influence are skewed among participants, how administration may be co-opting affected groups, and how organized interest may capture public agencies. Traces of such cyclical patterns in diagnosis and prescription can already be discerned.

Rediscovering Historical-institutional Context

The OECD and the World Bank are important institutions for diffusing administrative reform ideas. Although there have been competing projects, the cluster of ideas that dominated the OECD made the organization “a hothouse for reformist economists” during the 1980s and 1990s (Halligan 1996, 297). A core assumption was that public administration had to adapt to a globalized economy in order to serve the economy better, improve international competitiveness, arrest national decline, and reduce fiscal stress, budget deficits, and public debt.¹⁰

Reports from the World Bank also argued that a paradigmatic shift was needed, from a centralized, hierarchical, rule-driven administration to management and markets. The reform was presented as a generic medicine. Reformers insisted that they knew what had to be done:

In recent years, the belief in a universal cure has weakened. The World Bank—partly based on experiences with system breakdowns and weak, inefficient, illegitimate and unstable “quasi states” in developing countries (Jackson 1990)—expresses the need for in-depth understanding of the specific situation in individual countries. Few answers are right under all circumstances and no single imported recipe will do. Administrative reform must be matched carefully with the needs, traditions, and resources of each political system (World Bank 1997; 2000).

Even when it is observed that OECD countries cope with similar problems and share a repertoire of responses, it is (sometimes) argued that each country faces its own policy challenges, unique history, constitutional and political systems, and cultural and social circumstances. Therefore, there is no single model for effective public management (For example, OECD 1997, 7, 15, 31). The impact of competition among overlapping, quasi-autonomous agencies with wide discretion for managers is now seen as also involving problems. For example, a recent OECD report is concerned, not with how such forms spur innovation and adaptation and improve performance, but with how to ensure political coordination, policy consistency, and a coherent public service; how to develop less ambiguous roles and responsibilities and guarantee political accountability; and how to protect the public interest when highly political questions such as food safety and radioactive waste are left to autonomous experts in autonomous agencies (OECD 2002, 9, 21–22).

In sum, the two institutions have, to some degree, re-learned the old lesson of the importance of historicalinstitutional context. “Good administration” hinges on context dependent purposes, values, interests, capabilities, and institutional traditions (Dahl 1947; March 1997). Citizens in different contexts are likely to expect different things from government and be differently able and willing to provide administration with resources.

Functional Adaptation Across Jurisdictional Boundaries

A key type of problem targeted in the two reform visions is bureaucratic rigidity and resistance to change: an administration lacking initiative, unable or unwilling to improve performance by adapting to new circumstances. Consistent with American tradition, with reform ideas rooted in the big cities (Mosher 1975a, 8), George Frederickson, in his Gaus lecture, argued that the U.S. metropolitan area is a splendid site for understanding the changes underway in the wider context of public administration and the best empirical referent for developing theory about administration (Frederickson 1999).

In Frederickson's study of the Kansas City metropolitan area, key characteristics of the administrative context are found to be: high interdependency, fuzzy borders, declining salience of formal jurisdictions, and a significant erosion of the capability of dealing with social and economic issues within a single jurisdiction. Many problems transcend individual jurisdictional boundaries and important decisions that affect the represented are not controlled by those who represent them. A special system for problem solving and coordination has developed as a response. Networks of administrative units have evolved across jurisdictions. Functional specialists are key actors and linchpins, and legitimacy is based on professional claims to neutral expertise and not formal authority. Central authority is replaced by voluntary cooperation and “diplomacy.”

In contrast, politicians have taken little interest in such regional conjunctions. Fredrickson's most surprising finding is the absence of political influence. While administrators become involved in interdependent, joint problem solving; politics in terms of campaigns, elections, and offices remains jurisdictional and less interdependent. Politicians are stuck in old jurisdictions and incentive systems at the same time that existing borders are becoming less relevant for problem solving and representation. Many citizens also have little commitment to the jurisdictions in which they live and work.

A lesson from the Kansas City case is that public administration has successfully adapted to the new challenges and is de facto representing an inchoate public beyond the boundaries of a single jurisdiction. Furthermore, the dynamics observed are not forced upon administration by political authority, competition, and market signals, or organizational participation, as assumed by the three generic accounts of “good administration.” Public sector professionals cooperate on the basis of a shared professional identity, conceptions of the common good, and codes of appropriate behavior—institutional factors creating dynamics and not stasis, as is often argued. Politics, on the other hand, exhibit more inertia than administration. In so far as a new regional polity has developed, it is an administrative polity. Still, non-hierarchical administrative networks are seen to require the support of hierarchical institutions— they cannot be “distinctly at odds with” political leaders and jurisdictions (Frederickson 1999, 15–16).

The next case supports both observations: that public administration is not a laggard but a key to understanding inter-jurisdictional cooperation and change, and that there are limits to autonomous administrative adaptation and innovation in democratic settings.

A Quest for Democracy Beyond the Territorial State

The European Union is the best example of an administrative development transcending the framework of the sovereign territorial state and moving towards a post-Westphalian order (Kohler-Kohler and Eising 1999; Weiler 1999; Cowles, Caporaso, and Risse 2001; Egeberg 2001; Hooghe and Marks 2001). Up to the Treaty on European Union (Maastricht 1992) the EU/EC context exhibited several similarities with the Kansas City metropolitan area: high interdependence among states, more porous national borders, and reduced salience of formal jurisdictions and erosion of individual state problemsolving capacity. Executive and administrative adaptation and institutionalization, largely based on functional legitimacy, was far ahead of the integration of legislatures, political parties, social movements, and a common public space. There was fusion of administrations across levels of government and jurisdictions, for example through a complex committee system (Rometsch and Wessel 1996; Egeberg, Schaefer, and Trondal 2003).

Developments in the EU reflect that the Union has been a meeting ground for a variety of administrative traditions and schools of thought. In contrast to hypotheses of global or European convergence on a single superior model, EUleaders have assumed that member states can compete in a single market and implement common legislation while organizing domestic administrations differently. The basic rule has been that member states arrange their administrative structures as they see fit. Nevertheless, the EU has been torn between the ideology of national control and the need for uniform implementation of Community law,¹¹

The EU, unlike the Kansas City metropolitan area, has moved toward formalization and constitutionalization and the constitution recently drafted by the Convention on the Future of the Union has established “good administration” as a constitutional right. Good administration is here seen to require more transparency and better justification of institutions and policies to citizens. Beyond that, the concept has different meanings to different actors and the Union has not developed a common, coherent administrative policy (Sverdrup 2002). The EU is based on legal integration and continental public law. Hence, bureaucratic organization, hierarchies and rules are not seen as outmoded.

Nevertheless, markets and networks are also considered important. For example, the “open method of coordination” has been portrayed as a new mode of EU network governance featuring voluntary cooperation, benchmarking, and soft-law.¹³

The EU has both encouraged administrative change and allowed resilience. Change has taken place through several parallel processes, from “grand decisions” to incremental modifications of practice (Moravcsik 1998; Kohler-Koch and Eising 1999; Wallace 2001). Yet, several factors have discouraged convergence. Member states have been reluctant to give up control over domestic administrations. Different models have worked well in parallel and administrations formally organized as bureaucracies have been able to adapt to new circumstances within existing structural parameters. Significant change in administrative attention, interaction, and resource allocation has occurred without much change in formal-legal structures (Cowles, Caporaso, and Risse 2001; Knill 2001; Olsen 2003a; Jacobsson, Lægreid, and Pedersen 2003).

A key characteristic of post-Maastricht developments has, however, been the quest for democratic legitimacy. EU citizens hold high expectations but distrust Union institutions. Therefore, it is argued, a “democratic deficit” has to be corrected and institutions brought closer to citizens. Europe “must be returned to politics, to citizens, to public debate.” For the further development of the Union, efficient aggregation of existing preferences is not enough. Leaders and citizens have to examine national and institutional belongings and explore a European spirit and perspective.¹⁵

In general, the more the Union has moved from being a special-purpose organization with limited tasks, responsibilities, and powers toward a full-blown institutionalized polity, the more visible the limits of the principle of governance for the people based on functional legitimacy and indirect democratic legitimacy derived from the member states. The democratic quality of the Union's system of governance is seen to depend on direct democratic legitimacy and a stronger role for citizen-based institutions. Still, while the Convention has discussed the democratic life of the Union, it has happened with modest popular involvement. Most citizens have been ignorant about the discussion of the purposes and foundations of European government.¹⁶

A lesson is that in a multi-level, multi-centered polity such as the EU, the vision of an administration organized according to a single scheme and serving a single common authority cannot be upheld.¹⁸

As a result, the Union's problems of both democratic legitimacy and variations among domestic administrative models have become increasingly evident. The voluntary and subtle Kansas City model appears to have temporarily masked these problems, but they are problems nonetheless. Seen together, the two cases raise questions about the conditions under which administrative reform is a fairly autonomous process and the conditions under which it is overwhelmed by wider political processes. They also raise questions about where administrative success criteria and trust in institutions and agents come from and how such beliefs change.

The Complexity of Expectations

Trust in institutions and actors tends to reduce the demand for citizen participation. As a corollary, the more tasks and powers are transferred to government, the less agreement on tasks, objectives, procedures, and rules, and the less trust in institutions and actors there is, the more likely are demands for representation and participation, or for a reduction of the public agenda and protection against political intervention. Consider these two hypotheses: first, the more a single, shared, and stable objective, such as winning a war or improving economic competitiveness, has a privileged position in a polity, the more likely it is that decision making are left to nonmajoritarian, “guardian” institutions and experts. The criterion of good administration is then based on the ability to solve problems in terms of shared goals in an efficient and coherent way. Secondly, where there is agreement on stable rules for coping with enduring conflicts, tasks and competence are likely to be delegated to non-majoritarian institutions and agents. Hence, conflicts, and even crises, are dealt with in routine, predictable, and acceptable ways.

Much of the time, however, public administration operates in a complex ecology of institutions, actors, values, beliefs, interests, cleavages, and powers that do not provide clear competences, rules, objectives, or incentives. Individual citizens, elected politicians, judges, experts, organized groups, or mass media are likely to hold different and changing—not coherent and stable—concepts of “good administration.” They are likely to want administration to serve and be accountable to many, changing, and not necessarily consistent concerns. For example, administrators may be expected to:

• (a) ensure administrative loyalty to the incumbent government, and yet be politically neutral and loyal to all legitimately elected governments and the principle of representative government;
• (b) be rule-driven and the guardians of the constitution, the rule of law principle, and an impersonal legal arrangement, without ending up in excessive formalism;
• (c) be effective policy entrepreneurs taking initiative and getting problems solved but also be accountable to elected leaders and the people;
• (d) act as experts and the users of the best scientific knowledge available—an administration based on merit, professional competence, loyal to the principle of enlightened government—but avoid technocracy;
• (e) be cost-conscious managers governed by the principle of economic efficiency, adapting to changing circumstances and consumer demands, using markets and price mechanisms, without producing social inequality, exclusion, and protest;
• (f) protect specific constituencies and the principle that affected parties should be heard, without giving privileges to strongly organized interests.

Cycles and Long-Term Changes

Each of these expectations is a possible source of legitimacy as well as criticism and tensions between them are a source of administrative dynamics (Kaufman 1956; Jacobsen 1960; Orren and Skowronek 1994; Olsen 2001a). For example, political ideologies express trust in, and fear of, different institutions, actors, and resources. They express different views about the desirable power balance between institutions and whether and how different resources should be regulated. They have different trust in the common (wo)man and elected representatives, bureaucrats, judges, experts, leaders of organized interests, and business leaders. Some are afraid of majority based institutions, numbers, and majority power. Others are afraid of markets and monetary power, administrations and bureaucratic and technocratic power, courts and the power of judges, science-based institutions and expert power, or corporative arrangements and organizational power. Some want government and administration to be driven by public opinion while others prefer a more constrained role of (short term variations in) public opinion.²⁰

There are also fairly stable differences across nation states. For example, in the U.S. there is, compared to Norway, less willingness to permit government a strong role in the steering of society (Christensen and Peters 1999, 163). Yet, within fairly stable ideological and national differences there are cycles of attention and trust. There have, for instance, been cycles of confidence in the government's ability to direct society for the good of citizens, as well as the desirability of such intervention—cycles that have impacted and been impacted by changing inter-institutional power balances (Mosher 1975b; Kooiman 1993; Vught 1995). Likewise, the neutrality of expert knowledge has been challenged repeatedly, changing conceptions of the proper relations among public officials, experts, and citizens and the proper intertwining of expertise and lay judgment (March 1997, 8).

Discussion about and struggle over the adequacies and deficiencies of various institutions of government, and the power balance between them, are core processes in democratic governing. Throughout history conflicting expectations have been reflected in struggles over the size, organization, competences, resources, recruitment, and outcomes of public administration, as well as how relations to other institutions, interest groups, and individuals should be organized. Cycles in attention and trust have been common because a political order involves a variety of institutions defending different causal and normative beliefs and interests and because institutions have non-synchronized dynamics, so that imbalances and collisions are commonplace (Orren and Skowronek 1994; Olsen 2001a). Due to sequential attention, division of labor, and local rationality (Cyert and March 1963), different concerns are activated over time and across institutions. Any prescription based on hegemonic aspirations and the universalization of a single concern is likely to foster criticism and countervailing forces.

Therefore, what is presented as new trends in a short time perspective often reflects shifting cycles of attention and prioritization among familiar administrative components. As the mix of concerns change so do conceptions of good administration and administrators. During periods of transition, conceptions of the exemplary administration are challenged and can be dramatically redefined. Conventional wisdom becomes heresy. Virtues are turned into vices. Old expertise is scrapped and new types of knowledge, skills, and training are demanded. Trust in institutions disappears or emerges. Organizational structures, roles, and cultures are branded illegitimate and new ones are legitimized (Kaufman 1956; Jacobsen 1964; March and Olsen 1989).

Theorizing public administration then requires an understanding of how balances are struck and administrations find their place in a political order. Three time frames are relevant. First, under what conditions do shocks and crisis situations (such as wars, economic collapse, September 11) create enduring changes in belief-systems about what constitute good public administration—changes in success criteria as well as beliefs in what institutions and agents can be trusted? Secondly, under what conditions does change follow from long-term interactions in settings where old borders and institutions, as well as explanations and justifications of citizens' relations to institutions, are challenged and new ones are evolving?²¹

Subjects of Theory Development

First, public administration theory may benefit from taking into account the observation of a great diversity in human motivation and modes of action. Actors are driven by habit, emotion, coercion, interpretation of internalized rules, and principles, as well as calculated expected utility and incentive structures. Human character is variable and changeable, not universal and constant.

For example, New Public Management assumes self-interested, utility maximizing actors. “Old Public Administration” assumes administrators socialized into an ethos of rule following and public service. Actors are, however, constituted both by their interests, by which they evaluate expected consequences, and by the rules embedded in their identities and institutions. They try to calculate consequences and follow rules, and the relationship between the two is often subtle. Therefore, rather than assuming a single dominant behavioral logic, we may explore behavioral logics as complementary. We may inquire where and how different logics of actions and a sense of administrative identity and role is developed, lost, and redefined. We may also seek a better understanding of the interrelationship between strategic and rule-driven action and the conditions under which one logic comes to dominate another.²³

Secondly, public administration theory may also benefit from taking into account the observed diversity of organized settings and types of collectivities and social relationships within which administrators operate, and their different types of impact. Hierarchies, markets, and networks are commonplace in modern democracies and institutions have a role in coercion, in managing exchange, in redistribution, in building an administrative culture, and in developing constitutive structures for the sustenance of civic virtue and democratic politics (March and Olsen 1995, 245). Institutions provide opportunity and incentive structures regulating behavior and impacting transaction costs. They also constitute and transform actors by shaping their identities and mentalities through deliberation and socialization, and forms of government have historically been assessed according to their ability to foster the virtue and intelligence of the community (Mill 1962, 30–35).

When organized settings are seen as interdependent, supplementary, and competing, understanding is not likely to be furthered by a single set of assumptions. The success of one institution depends on the organization and functioning of a larger configuration of institutions, differently organized. Furthermore, identification can be a powerful motivator, yet identities can seldom be decreed from above. A challenge is to understand variations in the capacity and legitimacy to develop democratic officials and citizens with a sense of community, civility, and common good through differently organized administrative processes (Gawthrop 1998; Simon 2000; Heclo 2002).

Thirdly, public administration theory may provide a better understanding of continuity and change if, rather than assuming structural choice, it is observed that administrative life continuously achieves and loses structure through a variety of processes that may interact in complicated ways. Administrative change is a study in path dependencies as well as path departures. Authority is achieved, maintained, and lost. The basic units are constituted and reconstituted, and so are their relationships in terms of integration and disintegration, formalization and de-formalization, centralization and decentralization, politicization and de-politicization, bureaucratization and de-bureaucratization, professionalization and de-professionalization.

Change may be driven by functionalinstrumental concerns, by a commitment to principles, and by shifting cycles of attention and cycles of trust in institutions and agents. Yet, attempts to reform public administration implies intervention in large-scale, complex configurations of institutions with pre-existing identities, structures, internal dynamics, and resources and it can not always be assumed that leaders simply choose structures.²⁴

Helping Through Cooperation

John Gaus was right. Citizens need help. Direct participation by the people in administrative processes contributes to government for the people but only under some conditions. Citizens depend on the democratic quality of the institutions and actors that affect their life chances. They need institutions and agents that act reliably and with competence and integrity on the basis of agreed-upon, publicly known, and fairly stable principles and rules, standards and objectives.

Gaus was also right in arguing that any modern theory of public administration has to be a theory of politics and, as he implied, any modern theory of politics has also to be a theory of administration. In democracies, the distinction between administration and politics is as basic to legitimating administration as the two are difficult to separate in practice. A challenge for the discipline is to specify the conditions under which various institutions, processes, and agents provide help to citizens.

Doing so, students of public administration have to recognize that what citizens want from government, what they are willing to contribute in terms of time and resources, their beliefs about “good administration,” and their trust in institutions and agents, are formed partly endogenously in administrative processes. In order to effectively help citizens in such a complex world, public administration may have to cooperate more than before beyond specific national settings and across schools of thought.