Resumé

A governmental tradition is a set of beliefs about the institutions and history of government. In this article I argue the Anglo-Saxon governmental tradition interprets public sector reform differently to the Rechtsstaat, participation tradition of Denmark, leading to different aims, measures and outcomes. In the Introduction, I define NPM arguing that is has become everything is, therefore a meaningless term. I identify six dimensions to public sector reform: privatization, marketization, corporate management, regulation, decentralization and political control. In section 2, I describe the six dimensions of public sector reform in Britain and Denmark. section 3,1 explain the idea of a governmental tradition and argue the idea is essential to understanding the differences between Britain and Denmark. In section 4, I compare British and Danish governmental traditions, arguing the key differences lie in beliefs about the constitution, and state-civil society relations. Finally, I provide a summary explanation the differences and argue that traditions not only shape the aims, measures and outcomes of public sector reform but also lead to different interpretations of reform and its dilemmas. In Britain, the key dilemma concerns central steering capacity. In Denmark, the main dilemma is democratic accountability.

Introduction

We live in an era when, allegedly, governments reinvent themselves. At the forefront of these revolutionary reforms is the new public management (NPM). In this article I describe how such reforms differ in Britain and Denmark and explain why.

NPM refers to a focus on management, not policy, and on performance appraisal and efficiency; disaggregating public bureaucracies into agencies which deal with one another on a user pay basis; the use of quasi-markets and contracting-out to foster competition; cost-cutting; and a style of management which emphasizes, among other things, output targets, limited term contracts, monetary incentives and freedom to manage (see Hood 1991 from whom I paraphrase this list). It is said to be a global phenomenon.It a policy ambition for international organizations like the

---

* R. A. W. Rhodes, Department of Politics, University of Newcastle. Newcastle-upon-Tyne NEI 7RU, UK. E-mail: r.a.w.rhodes^ncl.ac.uk

OECD (1995) and the World Bank (1992). However, it rapidly became clear, to paraphrase Aaron Wildavsky, 'if NPM is everything maybe it is nothing.' The label now covers all types of public sector reform; it excludes nothing. So, the differences are of far greater interest than the similarities.

Christensen & Lægreid (1998, 470-71; also citing Kickert 1997 and Olsen & Peters 1996) identify a sharp distinction between Anglo-American reforms Continental reforms. They conclude 'NPM is not an integrated and consistent theory, but rather a loose collection of diverse doctrines, principles and measures which are partly in opposition to one another.' In a similar vein, Hood (1995) abandons his early characterization of NPM, suggesting the case for NPM as a global paradigm was overstated. These several commentators remark that the ideas of NPM are internally contradictory, for hierarchic steering and egalitarian empowerment in the same breath. Also, as we will see, the similarities in NPM are superficial, masking significant underlying differences. The trend to NPM is not uniform; example, traditional public administration persists in the EU Commission and Germany. The aims and results of NPM differ. In the UK, NPM aimed to create the minimalist state. In Denmark, it aimed to protect the state (Ministry of Finance 1993). The language of NPM obscures differences. NPM covers agencification in Britain, but not in Australia; regional devolution in France but not in Britain. Moreover, terms such as decentralization refer to deconcentration in an organization or devolution to sub-national governments. Several of the individual parts of NPM are not new; for example, performance measurement. Distinctiveness lies in the package not in the parts, but there is no uniform, agreed package. Finally, the meaning of NPM has changed; for example, in Britain the early focus was cost-cutting and efficiency but it gave way to a concern for the consumer. the label NPM masks far more than it reveals. The important questions now concern identifying and explaining the differences between the public sector reforms of the 1980s and 19905. I come not to bury NPM but to praise national distinctiveness.

The countries were selected for three reasons: conventionally they are seen as examples of different governmental traditions (see section 3 below); both have public sector reform programs (see section 2 below); and there is a body of existing research for me to re-analyze. In a single article it is not possible to cover more than two countries or two traditions in the necessary detail. Any description and analysis of public sector reform in Britain and Denmark must specify the dimensions along which these reforms varied. The existing literature suggests there are six key dimensions: privatization, marketization, corporate management, regulation, decentralization and political control (see, for example, Aucoin 1995; Hood 1991 and 1996; Pollitt 1993; Pollitt & Summa 1997; Rhodes 1997a; Wright 1994). I will

define each dimension as I discuss the British and Danish reforms. I will not use the generic term NPM. When corporate management is also known as managerialism and, to heighten confusion, managerialism is also known as NPM, it is time to drop both managerialism and NPM. I will refer to public sector reform, and when necessary, specify the dimension.²

Trends in Administrative Reform in Britain and Denmark 3

This section does not aim to be comprehensive, but to describe the main
trends in Britain and Denmark in the 19905.

Privatization

Privatization refers to the sale of public assets to the private sector. Whether the policy was a success in Britain, and one can argue about its several objectives and the extent to which they were achieved, there can be no doubt about the scale of the program. The government sold over 50 major businesses and reduced the state-owned sector of industry by some two thirds, raising some £64 billion to pay for tax cuts. The scale of privatization will necessarily decline under New Labour simply because so many public assets have been sold already. However, it is not ruled out on ideological grounds; for example it sold 60 percent of the equity in the Commonwealth Development Corporation.

In Denmark there is great support for the welfare state and the Conservative coalition's proposals for privatization in 1982 'provoked a major public outcry against this ideological crusade' (Jensen 1998a, 57), even though it limited itself to vague proposals for contracting-out. The Danes have no tradition of nationalization and, therefore, have less to sell. In 1997, the state held shares in 46 companies with an asset value of DKK 133 billion and had a controlling interest in 16 (Greve & Jespersen 1998, 12). Of the several sales, that of Tele Denmark was the largest share deal in Danish history, but any further large privatization is unlikely (Jensen 1998a, 57). Privatization in Denmark is neither a question of party ideology or of systematic policy, but of pragmatic budgetary calculations; of whether the sale solves the organization's budgetary problems or makes the government's budget round easier (Greve 1997; Jensen 1998a, 57).

If privatization policy is low-key, corporatization, especially creating state-owned enterprises (SOEs), has become a large and distinctive reform. Between 1975 and 1995, SOEs tripled in number (Christensen 1998). The aim underlying SOEs is: to give management freedom to manage while politicians remain responsible for general policy; and to make SOEs more

competitive by exposing them to market pressures and by recruiting private sector managers (but see Jensen 1998a, 58 for a discussion of the mix of motives). The 46 SOEs cover such industries as, for example, gas, air travel, steel, shipping and Copenhagen airport. Some SOEs have been privatized - for example, computing, army clothing and telecommunications - but most remain profitably state owned. The post office, electricity and the underground rail system will be the next significant 'corporatizations.' Their degree of autonomy varies. The policy remains pragmatic. To employ Jensen's (1998a, 58) graphic aphorism: 'the Conservative government of the 1980s had a privatisation policy without privatisation, the current government privatises ... without any explicit policy.'

Marketization

Marketization refers to the use of market mechanisms in the delivery of public services. The term covers mainly contracting-out (for example, compulsory tendering in local government); quasi-markets in the guise of the purchaser-provider split (for example, in health); and experiments voucher schemes (for example, education).

Previous Conservative governments forced local and health authorities to contract-out and vigorously marketized public services. New Labour will curb marketization. The purchaser-provider split in health and nursery vouchers will be abolished. However, there was no promised moratorium on contracting-out. New Labour is no enemy of the private sector and public-private partnerships will continue and expand; for example using the head teachers of private schools to improve poorly performing state schools. There has been a shift from the party dogmatism of Conservative governments to Labour pragmatism: marketization 'has a role to play . . . not out of dogmatism but out of pragmatism, because we want the best value for money' (Clark 1997).

In Denmark, although there has been much urging to contract-out, there has only been a small increase in contracting for goods and services from 17 percent to 20 percent in central government and no increase in local government (Greve 1998, 12). Indeed, the center reduced its capacity to control contracting-out by decentralizing more powers to local government below). Contracting-out has a distinct style in Denmark, described by Greve (1997, 1 and 12) as negotiating over the objectives of the policy, not just its implementation, and contracts are 'sites for negotiation' (Greve 1998, 16). There has been no 'grand unambiguous leap from "state" to "market"' (Jensen 1998a, 59). And let us not forget, contracting-out is nothing new. The public sector has used the private sector to provide 'technical' services such as refuse collection and cleaning for decades.

Corporate Management

Corporate management refers to introducing private sector management in the public sector. It stresses: hands-on, professional management; explicit standards and measures of performance; managing by results; value for money; and more recently closeness to the consumer. There are many variations this list and most have their roots in a preference for private sector management styles, stressing performance measurement in general and output measurement in particular (see for example Hood 1991; Pollitt 1993).

Corporate management has a long history in the UK where the reforms, with marketization, were used to break the alleged power of the public sector unions. Its present-day hallmark is the '3Es' of economy, efficiency and effectiveness with economy ever present. This emphasis shifted to making services transparent to their consumers through such mechanisms as the Citizen's Charter (see Pollitt 1993). New Labour's policy is pragmatic, the government supporting initiatives such as 'Benchmarking' (which aims to measure and improve quality) and 'Investors in People' (which sets standards for human resources management). Both continue the mini-tradition of importing private sector management practices into the public sector. As Theakston (1998a) concludes, 'Labour knows it cannot afford to halt the management revolution, given the need to squeeze as much as possible from every public sector pound spent.'⁴

Corporate management in Denmark covers management by objectives and performance-related pay (PRP). For example, PRP was introduced for senior executives combining measures of organizational outputs, strategic development and human resources development. The bonuses can amount to 15 percent of an individual's salary (Jensen 1998a, 61).

Before marveling at the pace of change in the public sector, I want to stress, again, that management by objectives and performance-related pay are not new management tools even in the public sector, and there seems little new here to warrant talk of a new global paradigm. These techniques are the standard fare of management consultants and business schools the world over. Also, in Denmark, the reforms may pose challenges for the dominant workplace culture of high job security coupled with centralized pay negotiations and powerful unions, but they do not seek to emasculate it. Rather, as Jensen (1998a, 62) argues, the corporate management reforms are Janus-like, facing in two directions at the same time. They seek to reward individual excellence through PRP and yet to encourage collective responsibility in delivering high-quality services.

Regulation

As the boundaries of the state were redrawn in the 1980s, many governments,
including the Commission of the European Union, sought to strengthen their

ability to regulate and audit institutions, their policies and implementation of those policies. With the advent of privatization, governments substituted regulation for ownership, and so multiplied the watchdogs of the new private sector monopolies. The 'audit explosion' refers to all forms of internal and external regulation. It covers management and financial audit and evaluation with related quality assurance mechanisms and 'a distinct mentality of administrative control' which displaces trust and focuses on quantified, external, ex-post, expert forms of control (Power 1994,8-9).

The new regulatory bodies include: the Office of Electricity Regulation, the Office of Gas Supply, the Office of Water Services, and the Office of the Rail Regulator. They differ in size, ranging from the Broadcasting Standards Council with a budget of £0.4 million and a staff of 14; the Office of Telecommunications (Oftel) with a budget of £4.5 million and a staff of 120; to the National River Authority with a budget of £30 million and a staff of 6,500. New Labour provides more of the same, setting up a further 11 new regulatory bodies (Hood et al. 1998a).

The increase in internal regulation is also massive. Hood et al. (1998b, 62-64) identify 134 such bodies spending £776 million a year in running costs and incurring about the same outlay in compliance costs. Examples of such bodies range from the National Audit Office, the Prisons Inspectorate, the Parliamentary Commissioner for Administration, the Office of Public Service and the Higher Education Funding Council.

The strict financial climate saw the Treasury become more powerful. The Conservative government introduced several changes in budgeting and financial management to strengthen its control of total spending while ostensibly increasing financial delegation (Thain & Wright 1995). However, financial discretion was an illusion. The brute reality of everyday life was cuts in spending. Financial regulation lies at the core of the UK reform package. It remains so under New Labour; 'As always the Treasury is a powerful and übiquitous force' (Theakston 1998b; see also Gray & Jenkins 1998, 120-23).

In Denmark, there are regulatory bodies for electricity, telecommunications, and postal services, but only Telestyrelsen (the Danish National Telecom Agency) approximates to an 'independent' utility regulator. are conflicting trends. For local authorities, there has been a 'campaign' to increase 'municipal liberty' by simplifying the rules used by the center (Hansen 1997, 57). For central agencies, however, departments introduced 'new controllers' to strengthen internal regulation (Greve & Jespersen 1998, 16-18). Central ministries advertise for controllers and set up new control sections targeted at specific directorates. They introduce new audit tools; for example, six-monthly reports on their current economic situation, and enterprise accounts. Responsibility for national audit was transferred to the parliament, giving it greater independence.

The Ministry of Finance plays an important coordinating role which is not restricted to budget making. No central actor coordinates ministries. Administrative reform is negotiated and, therefore the Ministry seeks to steer indirectly. The sheer number of reports testifies to its commitment to reform. But no organization has the direct controls of ministries. Knudsen (1999) shows that coordination, whether through the Council of State, the Meeting of Ministers (or Cabinet), or the various types of cabinet committee on 'trust and mutual understanding between the politicians involved, and a high level of information and feeling of personal responsibility civil servants.' The Ministry of Finance and the Prime Minister's Office (PMO) are key cogs in this informal system; for example, even at Cabinet level 'the detailed negotiations are conducted - or at least supervised - by the Ministry of Finance and the PMO in order to control the politically sensitive issues.' But the style is indirect, not direct, control.

Decentralization

Decentralization encompasses both deconcentration and devolution. Deconcentration to the redistribution of administrative responsibilities in central government. Devolution refers to the exercise of political authority lay, elected, institutions within areas defined by community characteristics.

Deconcentration. In the UK, most of the reforms of the 1980s and 1990s sought to deconcentrate managerial authority; for example to agencies. Agencification means creating semi-autonomous agencies responsible operational management. The key notion is to separate operational management from the policy making core or central department so the agency has freedom to manage. It is a classical doctrine of public administration. January 1998, there were 138 agencies in Britain employing 77 per cent of the civil service (Cm 3889 1998) and the program was almost complete. New Labour accepted agencies after early doubts. There is no consensus about the extent of deconcentration to agencies. Day & Klein (1997) conclude that in the NHS policy has migrated from the core department to the agency. For the chief executives of agencies the reality of everyday life is constant interference by parent department, Treasury and Cabinet Office (Barker 1997). As one chief executive pointed out T am a civil servant and cannot say no' (Cabinet Office 1994, 31).

The White Paper on The Citizen's Charter (Cm 1599 1991) was prime minister John Major's 'Big Idea.' The key objectives were to improve the quality of public services and provide better value for money. The Citizen's Charter contains six principles: published explicit standards; full and accurateinformation running services; choice for the users of services;

courteous and helpful service; effective remedies; and efficient and economicaldelivery services. Sir Robin Butler (1993, 402) describes the Citizen's Charter as 'the culmination of the movement to output measurement.'Consumer now dominate producer interests: 'people power.' It is a little early for such an assessment. Pollitt (1993,187) is nearer the mark when he concludes that 'it is not so much a charter for citizen empowerment as managerialism with a human face." New Labour seeks to 'reinvent' the Citizen's Charter and has a touching faith in the capacity of information technology to improve the quality and responsiveness of local service delivery (Clark 1997). For example, the government has set up a People's Panel comprising 5,000 people from all walks of life to give their views on improving service delivery.

Agencies were created in Britain in 1988. They are, therefore, part of recent reforms and an example of deconcentration. Although the rationale for agencies in Denmark also stresses separating policy and management, they are a long-standing feature of the Danish administrative landscape.⁵ They figure in recent reforms, but not as an example of deconcentration (see below)

Devolution. Devolution is a distinctive policy of New Labour (Cm 3658 1997) which will have many consequences for the public sector in the UK as a whole, not just Scotland. It is too early to assess the outcomes, but some possible problems are already clear. Thus, the Scottish civil servants have transferred to the Parliament but, as staff of the Scottish Executive, continue to be part of a unified civil service. There are obvious tensions in such dual loyalties. The government recognizes there is a problem, asserting the 'ultimate loyalty of civil servants will remain to the Crown.' However, 'in practice the loyalty of individual civil servants will be to whichever administration they are serving' and they 'will continue to take their instructions from their departmental ministers' (Personnel Division, Cabinet Office, 'Devolution and the Civil Service,' Note 2 1998). It seems certain the UK civil service will become less unified.

After devolution, Scotland is embedded in a changed and changing set of institutional relationships. The Scottish Parliament sets the framework for other Scottish public bodies, including elected local government. It controls local authority spending. There are Scottish representatives in Brussels, complementing the role of the UK's permanent representation. The Scottish Parliament is obliged to implement EU legislation. The legislationis on the role of the civil service and ministerial accountability. There is little or no discussion of how the several levels of government will work together, especially when under different political control, beyond pious urgings for the Scottish Executive and the UK government to work closely together at both ministerial and official levels. The Welsh Assembly

is a more modest affair (Cm 3718 1997; and for an assessment of the likely
effects of devolution for Whitehall, see Hazell & Morris 1999).

Before 1972, local government could be described as part of the constitution. such claim would be made today. British central government has exerted ever-greater control over local authorities, especially their spending. Policy was centralizing even if the outcomes were not always as intended (Rhodes 1997a, ch. 6). After much prevarication, the New Labour government decided there would be no regional assemblies for England. It is also nervous about devolving power to local government because it is determined to keep control over public spending. The centerpiece of the government's White Paper on local government (Cm 4014 1998) is elected mayors and its theme is the renewal of democracy, but it displays a marked reluctance to empower local authorities.

There is a long-standing tradition of using Boards in Denmark, and they come in various guises; for example, boards which integrate experts and the representatives of interest groups with public administration and the boards of directors of specific SOEs. The 1980s and 1990s saw a new variety; user boards. They are used to integrate service users, managers and employees in education, child care, and care of the elderly (Jensen 1998a, 63-64; Sum 1998, 11-12). Also parents can set up a self-organized kindergarten or school and agree funding with the local authority. Finally, individuals have been given the right to choose between schools and hospitals. Such freedom of choice aims both to improve efficiency and to empower customers.

It is important to distinguish between clients, users, customers and citizens (Jensen 1998a, 63; Siirn 1998, 11). Clients depend on benefits. Customers shop in markets. Users have formal rights. Citizens have a broad participation covering both the community and individual services. The British Citizen's Charter claims to empower users but falls short of giving them rights, and marketizing public services gives customers only limited opportunities shop around. The Danish reforms also cover the customer and user roles. But there is nothing in the British reforms remotely comparable to either user boards or self-organized services. As, Sum (1998, 12) reports, Danes believe the reforms have increased their efficacy as citizens; 'Danish schools and day care institutions are open and participatory institutions.' The distinction between customers and citizens oversimplifies, but it captures an important difference between the two countries.

The comprehensive of reform Danish local authorities in the 1970s created 275 municipalities and 14 counties and devolved important functionssuch social security and social services. It was widely interpreted as strengthening local self-government by giving local authorities greater functional authority and fiscal leverage (Kjellberg 1988; Norton 1994). The reforms continued in the 1980s. For example, secondary schools and 3,000 kilometers of roads were transferred to the counties; central state regulationsof

lationsoflocal authorities were either removed or simplified; and the free commune experiment enabled local authorities to exempt themselves from existing legislation. The other side of the coin was tighter spending controls through specific, conditional grants; for example, towards the operating costs of hospitals. Hansen (1997, 62-3) argues that these changes are not a return to local self-government, but are creating a 'new municipality.' There has been a shift from 'the traditional, state-centered form of government to a more multi-centered and local forms of governance' (ibid., 45) which added a strong role in policy making to their pre-eminence in implementation.The 'are no longer limited to local adaptation of concretemeasures' are 'participants in the overall goal setting' (ibid., 65). In brief, Denmark introduced more pluralist service delivery and sought to strengthen participation in an already decentralized system.

Political Control

The efforts of ministers to reassert political control over the civil service is
a common feature of recent public sector reform (Aucoin 1995, 8-9).

In Britain, the Royal Institute of Public Administration (RIPA 1987, 43) concluded there had been 'personalization not politicization' in senior appointments under Margaret Thatcher. There may have been no overt party politicization of the higher civil service, but we lost 'institutional scepticism' (Plowden 1994, 104). The transition to New Labour went smoothly - there was only one 'early retirement.' The Prime Minister wrote to the Head of the Home Civil Service, Sir Robin Butler, congratulating him on the smooth transition from a Conservative to a Labour government. There was no night of the long knives for permanent secretaries. The new Head of the Home Civil Service, Sir Richard Wilson, is recognizably a mandarin.

New Labour appointed 53 political advisers compared with 38 under the previous government and threw petrol on the dying embers of the politicization debate. The most notable increase was in the No. 10 Policy Unit, from 8 to 18. It included no civil servants. Tony Blair appointed Jonathon Powell his chief of staff but abandoned plans to appoint him as his principal private secretary (a post normally filled by a civil servant). Instead an Order in Council gave him special executive powers (Theakston 1998a). Sir Richard Wilson conducted a review of the Cabinet Office which aimed to make it the corporate heart of the civil service and strengthen cross-departmental coordination. The changes should not be interpreted as politicization but as centralization. As Hennessy argues, the No. 10 Unit and the Cabinet Office drive government policy making and this proactive core is supported by more than 100 task forces, advisory groups and reviews designed to tackle, for example, the 'wicked issues' which cross departmentalboundaries.

mentalboundaries.So, 'No. 10 is omnipresent in the serious policy reviews'
(Hennessy 1998, 15).

In Denmark at the end of the 1970s there were 87 agencies (or directorates). However, the distinctive story to be told is of de-agencification and contract steering, both of which seek to strengthen political accountability. & Hansen (1995, 552) suggest that agencies or directorates faced persistent problems over administrative duplication and the policyadministration either agencies interfered in policy making, or ministers would not limit their interventions to general policy, or both. The result was a loss of political accountability and control. Reform sought to resolve this problem and so directorates, departments and ministries were merged.

Contract steering aimed to improve operational efficiency and to separate policy from day-to-day management, thereby increasing political control over policy through a results-based contract which defined the objectives, tasks, targets, rights and duties of the agency (Pedersen et al. 1997). About 40 state agencies are on contract. The contracts are not legally binding and last for two to four years. There are no sanctions for failing to meet targets. The agency gets freedom to manage, provided it meets its objectives within its budget. Ministers get a firmer grip on policy by clarifying agency objectives, and they can cancel the contract when they want. By 1998, there were 56 such contracts, and the innovation is judged a success by the Ministry of Finance, its chief sponsor, because clarifying objectives led to improvements in efficiency. Others are less sanguine; for example key performance were not in the contracts (see Drumaux in Pedersen et al. 1997; Greve & Jespersen 1998, 13).

As the Danish government decentralized, it also centralized. Jørgensen
& Hansen (1995, 554, 561) argue these changes 'can be interpreted as
attempts to regain ministerial control: to bridge politics and administration

- and to give priority to politics.' The problem confronting the Danish government was lack of accountability and control. So, ministers opted for the twin strategies of: 'bridging' the politics-administration gap by centralizing; 'buffering' by using management reforms to delimit the scope for ministerial intervention.

Explaining the Differences, Analyzing Governmental Traditions

How do we Explain the Differences?

There have been several attempts to explain the variations in public sector
reform in Western Europe (see Hood 1996; Naschold 1995; Pollitt &
Summa 1997; Rhodes 1997a, ch. 5; Wright 1994, 104-8). Hood (1996,

273-82) focuses on 'English awfulness,' right-wing party politics, poor economic performance, and scaling down big government. He finds none an acceptable explanation of the variations. Naschold (1995, 215-17) argues that the successful implementation of public sector 'modernization' depends on 'path dependency' (or 'historical traditions, cultural norms and establishedpractices'); mobilization by advocacy coalitions of administrativeand elites; the institutionalization of such coalitions; and influential meta-organizations and institutions that produce knowledge (and on path dependency see Hall & Taylor 1996, 941-42; and Hay & Wincott 1998, 955-56). She does not try to explain why the reforms differ in their aims, measures and outcomes. As a final example, Pollitt & Summa (1997, 13-15) suggest that economic pressures, the nature of the political system; administrative structures and party political doctrine explain variations.They an institutional approach, concluding that 'the most convincing explanations . . . appear to rest . . . upon the characteristics of the political and administrative systems already in place. . . . [T]hese characteristics . . . most significantly influenced what was possible in terms of the scope, process and speed of reform (see also Peters 1997, 262). I want to develop the argument that governmental traditions and ideas are central to understanding recent changes.

Governmental Traditions

A tradition is a set of beliefs someone picks up during a process of socialization. This heritage comes to each individual who, through their agency, then can modify and transform it even as they pass it on to others. This set of shared theories form the background against which we construct the world. Traditions are contingent, constantly evolving, necessarily in a historical context and consist of theories and narratives. Narratives are the form theories take in the human sciences; narratives are to the human sciences what theories are to the natural sciences (paraphrased from Bevir 1999). There is a sense, therefore, in which there is no NPM or public sector reform because all complex political objects are constructed in part by our prior theories of the world and the traditions of which they are part. How we understand reform depends, therefore, on the theories which we do use. Reform is a narrative interpreted through traditions

A governmental tradition is a set of inherited beliefs about the institutionsand of government (cf. Davis 1998, 158; Perez-Dias 1993, 7). Loughlin & Peters (1997, 46) distinguish between the Anglo-Saxon (no state) tradition; the Germanic Rechtsstaat tradition; the French (Napoleonic) tradition; and the Scandinavian tradition which mixes the Anglo-Saxon and Germanic. Thus, in the Germanic tradition, state and civil society are part of one organic whole; the state is a transcendent entity.

The Anglo-Saxon pluralist tradition draws a more distinct boundary between state and civil society with contract rather than natural law as the basis to the state. Civil servants have no constitutional position. The Napoleonic tradition sees the French state as the one and indivisible republic, exercising strong central authority to contain the hostile relations between state and civil society. The Scandinavian tradition is also 'organicist,' influencedby ideas of the Rechtsstaat tradition, but differs from the Germanic tradition in being a decentralized unitary state with a strong participation ethic. (In this paragraph I paraphrase Loughlin & Peters 1997,46-55).

So British governmental traditions interpret public sector reform differently Danish state traditions and, as a result, the reforms have different aims, measures and outcomes. I distinguish between beliefs about central constitutional structures; political-bureaucratic relations; and statecivil relations, especially government-interest group links (adapted from Christensen 1995). In other words, administrative reform in Britain and Denmark differs because of significant differences in these beliefs.

Comparing Traditions

Britain 6

Beliefs about the Constitution. The conventional wisdom is captured in the Westminster model which sees Britain as a unitary state characterized by: parliamentary sovereignty; strong cabinet government; ministerial accountability; party control of the executive (that is, prime minister, cabinet and the civil service); elaborate conventions for the conduct of parliamentary business; institutionalized opposition, and the rules of debate (Gamble 1990, 407). Even those who disagree with the Westminster view use its language to debate the issue (Bevir & Rhodes 1999a and b; Birch 1964, 237). At its simplest, parliamentary sovereignty means parliament can make or unmake any law with no fear of legal challenge from the courts. Ministers are accountable to parliament for the actions of their department. Collective accountability means ministers must publicly support cabinet decisions.

Three facets of the model are directly relevant to my analysis of public sector reform. First, a defining characteristic of the British governmental tradition is its strong executive. Birch's summary is most apt: 'government should have both the powers and the ability to provide strong leadership' (1964,245).

Second, there are no constitutional constraints on that executive beyond
those it chooses to accept. Once the government decides on a change, it can
force it through. Crick captures the point concisely:

The only restraints (on authority) are political. Governments are restrained by what they think the country will stand for come the general election, and they adhere to things like general elections because they prefer (whether out of ethics, habit or prudence, or all three) to settle disputes politically rather than despotically and coercively (Crick 1964, 16)

Finally, the first-past-the-post, two-party system produces working majorities
parliament which the government of the day can use to deliver
legislation on its manifesto promises.

Beliefs about Political-bureaucratic Relations. Before the middle of the
19th century:

The 'permanent civil service' . . . differed from its modern counterpart in three significant
ways. It was not permanent, it was not civil, and it was not a service (Parris 1969, 22).

Subsequently Britain developed a 'constitutional bureaucracy' or:

'an unpolitical civil service whose primary connexion is with the Crown, and which, while subordinated to party governments, is unaffected by their changes: the two permanent elements, the Crown and the civil service, which not by chance together left the political arena, supply the framework for the free play of parliamentary politics and governments' (Namier 1974, 14; cited in Parris 1969, 49).

In other words, it is permanent, politically neutral, anonymous, and discreet
the point of being secretive (Birch 1964, 241-42).

In Britain, the generalist, in function and by educational background, is
admired.

We believe that men who have been engaged, up to one or two and twenty, in studies which have no immediate connexion with the business of any profession, and of which the effect is merely to open, to invigorate, and to enrich the mind, will generally be found, in the business every profession, superior to men who have, at eighteen or nineteen, devoted themselves the special studies of their calling (Macaulay quoted in Parris 1969, 287).

The Conservative government's determination to introduce private sector management skills into the civil service is the modern way of criticizing the service for its lack of specialist skills. The percentage of recruits from Oxbridge with arts and humanities degrees has fallen but remains high at nearly 45 percent. Their role is shaped by their constitutional setting; cabinet responsibility and ministerial accountability to parliament cast them in the inescapably generalist-political role of supporting their nonspecialist

The generalist civil servant owes no duty to a higher authority than the minister. Sir Robert Armstrong's The Duties and Responsibilities of Civil Servants in relation to Ministers. Note by the Head of the Civil Service (25 February 1985 in Cabinet Office 1994) restates the constitutional belief that: 'the duty of the individual civil servant is first and foremost to the Minister of the Crown who is in charge of the department in which he or she is serving.'

Beliefs about State-Civil Society Relations. Greenleaf (1983a, 15-20) describes British political tradition as a dialectic between two opposing tendencies: libertarianism and collectivism. Libertarianism stresses four things: the basic importance of the individual; the limited role of government; dangers of concentrating power; and the rule of law. Its antithesis, collectivism stresses: the public good; social justice; and the idea of positive government.⁷ These strains exist in both political parties. They set the boundaries to political debate.

Norton (1991, ch. 2) stresses four features of the British political culture: empirical problem-solving; allegiant and deferential attitudes to the political cooperative making decisions, and individualistic yet trusting relations with compatriots and allies. This combination is 'the civic culture' (see also Kavanagh 1990, ch. 4; Rose 1985, ch. 4). And since 1963 it has changed only incrementally: 'My argument is simply stated: Although subject modification, the civic culture remains' (Norton 1991, 29-30). This stress on political stability is an integral part of the Westminster model (see Gamble 1990, 411; Harrison 1996, 5-6; Kavanagh 1990, 56; and Bevir & Rhodes 1999a for more citations).

The pattern of interest intermediation takes the form of functional policy
networks. Such networks:

• limit participation in policy making;
• define the roles of actors;

• decide which issues to include and exclude from the policy agenda;
• through the rules of the game, shape the behavior of actors;

• privilege certain interests not only by according them access but also
by favoring their preferred policy outcomes; and

• substitute private government for public accountability (see Marsh &
Rhodes 1992; Rhodes 1997a, ch. 1 and 3; Rhodes 1999, ch. 8).

In short, as Birch (1964, 245) argues the British political tradition places 'consistency, prudence and leadership' before accountability to Parliament and responsiveness to public opinion. So, the networks of group integration are exclusionary in membership. Britain is characterized by political oligopoly a passive political culture.

Denmark 8

Beliefs about the Constitution. Knudsen identifies the following features
of'the Scandinavian model':

Homogeneity, continuity, high level of organisation. And at the level of the public sector: A
strong state tradition, but combined with a high integration between state and society through
strong local governments, and through integration with the organisations, a tradition for

consensual democracy, tradition for multi-party systems with strong Social Democratic parties and agrarian parties, high welfare ambitions based on institutional rights linked to citizenship, a public sector with universal bureaucracies, a public administration with a low level of corruption and a comparatively high level of efficiency (Knudsen (1991, 96).

Denmark differs from other Scandinavian countries because of its absolutist The beliefs in parliamentary sovereignty, the direct accountability ministers and strong local governments have deep historical roots in the challenge to the monarchy. It also owes its strong bureaucratic and legal tradition to the influence of its neighbor Germany. So, there is a large measure of agreement about the main features of the Danish governmental tradition and three features are important for the analysis of public sector reform.⁹

First, multi-party coalitions of 3-5 parties produce a negotiative decision style. Although there have been nine changes of government in the 1980s and 19905, each worked within the overall idiom set by previous coalitions. There are many small negotiable cleavages, but few large ideological ones.

Second, Denmark has a system of ministerial government:

in which each Cabinet minister is politically and judicially responsible, has helped to make central government administration sectoral and rather alien to co-ordination procedures and bodies. Thus, the staff in the Prime Minister's Office is small in number, the formal power of the Ministry of Finance is limited, and the work procedures of the Cabinet are informal (OECD 1993, 63; see also Knudsen 1991, 75).

There is no centralized authority able to organize the government machine and 'Each department is ... closely connected to it agencies and externally organised interests' (OECD 1993, 63). However, as we shall see, both the Prime Minister's Office and the Ministry of Finance seek to steer indirectly (Jensen 1998a, 60; Knudsen 1999).

Third, under the 1849 Constitution, local authorities have the right 'to manage their own affairs under the control of the state.' They have a high degree of constitutional autonomy; described as 'spectacular' by Knudsen (1991, 74). The National Association of Municipalities and the Association of Danish Counties are of 'major importance' (OECD 1993, 63) and constitute 'the local government world' speaking for 'the local interest,' phrases which capture their standing. They bargain for local authorities with the central state. Their influence is long-standing. They were one of the bulwarks against absolute monarchy. Local authorities have a dual role: local-self government and local state administration, including collecting 58 per cent of all taxes. This dual constitutional status persists and central government must negotiate; it cannot command.

Beliefs about Political-Bureaucratic Relations. The defining characteristicof
Rechtsstaat is that it is a legal state vested with exceptional
authority but constrained by its own laws. Civil servants are not just public

employees, but personifications of state authority (Loughlin & Peters 1997, 46). The Rechtsstaat tradition 'emphasizes a relatively sharp distinctionbetween and administration, political-hierarchical control of the administration and administrative activities based on universal rule' (Christensen 1995, 19). So, Knudsen describes the dominant tradition as:

a highly bureaucratic, rule-bound and hierarchic structure designed to promote objective
decision making on the basis of appropriate laws or regulations in obedience to the current
government, whatever its political complexion (Knudsen 1991, 37).

However, this tradition is a product of the 19th century, and few Danes today would stress the importance of the Rechtsstaat tradition, although its influence persists in, for example, the fact that some 50 percent of top civil servants have law degrees. Rather they would point to the growth of the welfare state, the attendant need for professional expertise and the autonomy such professionals. Thus, there has been a marked increase in the numbers of top civil servants with degrees in the social sciences, most notably economics.

State-Civil Society Relations. Allegedly, two Danes stuck in a lift will
instantly create an association (Jensen 1998a, 63). Indisputably Denmark has
much associational politics (OECD 1993,63). This tradition has deep roots:

there has been a participatory tradition of involving ordinary citizens in politics that goes
back to the political and cultural self-organisation of the farmers of the 19th century and to
the ideas of the workers movement in the 20th century (Sum 1998, 13).

A further characteristic of the Scandinavian state tradition is that 'participation in public decision-making is more regularized and formalized, as reflected in concepts like integrated participation in government, political segmentation and corporatist features' (Christensen 1995, 23). It is also referred to as the 'negotiated economy' in which resources are allocated through institutional bargaining.' These negotiations are 'between "independent" where the relevant public authority is usually one of several participants.' The resulting agreements are not subject to legal sanctions but are 'discursive, politically or morally binding.' This negotiated economy has 'shattered' public authority and 'institutional arrangements have been constituted, consisting of numerous, functionally differentiated and mutually institutions, which coordinate actions through networks, reach decisions via the negotiating games and extract commitments from one another through politically and morally binding agreements' (Nielsen & Pedersen 1988, 81, 82 and 97). So Danish political culture is divided between citizen autonomy expressed through associations and the formal organized participation of the negotiated economy and the tension between a strong state and a strong emphasis on self-organization and individual rights.

Comparing Divergent Trends, Interpreting Distinctive Dilemmas

Summary

There is always the danger of reifying traditions when comparing them. For any country there will be multiple traditions, and the key question is who articulates which tradition. Thus, for Denmark, such notions as Rechtsstaat, parliamentary sovereignty, welfare state professionalism and ministerial accountability are open to many interpretations (and on Britain see Bevir & Rhodes 1998). I accept that, ideally, I should provide a more nuanced account of the several traditions and their proponents. The adage has it that 'the proof of the pudding is in the eating.' I may oversimplify, but my broad-brush comparison demonstrates the value of analyzing traditions, not preclude a subsequent unpacking of the traditions of each country, and suggests several ways of developing such an analysis.

Against a backcloth of fiscal pressure, the British government pushed through public sector reform shaped by a governmental tradition characterized strong executive leadership, no constitutional limits and twoparty Administrative reform is always political and all British governments since 1979 have sought to cut government spending and to exert effective control over the administrative machine. The Conservative government to reduce public spending by redrawing the boundaries of the state and evolved a clear set of political ideas to justify and 'sell' its various reform packages to the electorate. It attacked big government, local government waste, used markets to create more individual choice and campaigned for the consumer. Privatization and marketization pushed back the boundaries of the state and gave the government more control (over financial totals) over less (service delivery). There was an explosion of new regulatory bodies, internal to government as well as for external bodies, and a protracted drive to introduce corporate management, to reinforce political control. The New Labour government differs only in that it is more pragmatic implementing the Conservatives' reforms. The outcomes are open to debate. Public ownership was reduced dramatically, but it was accompanied by an audit and regulatory explosion. The assertion of political control led to no reduction in public expenditure, although (arguably) it slowed the rate of increase. Fragmenting service delivery systems multiplied networks and reduced the center's capacity to steer. The fate of the British government was that of all governments; unintended consequences.

The Danish government's response to fiscal stress was a 'revolution in slow motion' (Olsen 1983, 188). The aim was to preserve a popular welfare state by selected reforms aimed at getting better value for money. Public sector reform is difficult to distinguish from any other policy area; it was characterized by a negotiated consensus and a pragmatism which avoided

clear winners and losers. The choice of means was a technical matter, not ideological. So, privatization and marketization were but two options among many, to be used when there was agreement they were the best way forward. Other means, such as regulation, were scarcely used at all. The government sought greater control through corporate management reforms and the use of'flexible' contracts. It also distanced itself from problems by decentralizing to local government and citizens. Devolution was sought and won by the national associations of authorities, but the policy also bears a more cynical interpretation as 'off-loading to the periphery' because responsibilities were handed over without the necessary fiscal resources. The reform orthodoxy of the Ministry of Finance claims the reforms enhanced efficiency because, for example, agencies have clearer goals, but the outcomes are uncertain and ambiguous. Many changes continue 'the Danish tradition of foggy corporate governance in a different disguise' (Jensen 1998a, 65).

So, there are marked differences in the aims, measures and outcomes of the reforms; differences which can be explained by the different governmental of the two countries.¹⁰ These differences are summarized in Table 1.

Comparing Divergent Trends

Beliefs about the Constitution.

a. Strong executive vs. negotiated consensus

The point is obvious. The British tradition of majority party government
underpinning a strong executive means the government can drive through

its reforms, whereas such reforms have to be agreed by a multi-party coalitionin and then negotiated with other affected parties. Privatization illustrates the difference. Britain had a comprehensive, ideologicallydriven of reform designed to create the minimalist state. Privatization was the flagship policy. Privatization is a pragmatic policy in Denmark. Multiplying SOEs may be distinctive, but it is an intermediate solution to changing the boundaries of the public and private sector; a means of preserving the state.

b. Parliamentary sovereignty vs. constitutional state

Although parliamentary sovereignty is a shared constitutional principle, Britain's uncodified constitution means there are few if any constraints on Britain's strong executive. In Denmark, the historical strength of local government, entrenched in the constitution, means it can effectively resist central government and its powers have increased. The opposite is true in Britain where parliamentary sovereignty meant local authorities were subjected ever more stringent central controls. Contracting-out is one obvious example; it was imposed. In Denmark, the national associations of local government defended local government against this policy. It was not imposed. The center had to rely on example and persuasion. There was little or no increase in contracting for goods and services in local government. although relational contracting is spreading in Britain, on a hard-soft contracting scale, the two countries remain poles apart. The Danish approach is non-ideological and contracting is an invitation to negotiate, although it can still stir the political emotions of both Left and Right. In Britain contracting was a central belief of New Right ideology for two decades; a tool for creating the minimal state.

Beliefs about the Bureaucracy.

a. Party vs. minister

British ministers are powerful. As in Denmark, they are individually accountable to parliament. But they are always subject to party discipline and collective cabinet accountability. Public sector reform was not at the discretion of individual ministers. It was an ideological, party driven, coordinated There is no equivalent to the Danish tradition of independent The Danish system of ministerial government means effective public sector reform depends on political cooperation between ministers. Each minister can decide on the preferred reforms for her or his ministry. There is no overall control of the reform process. No politicalbureaucratic can work without trust and pragmatism, but they are the essential currency of coordination in Denmark.

b. Generalists vs. professional autonomy

Generalist civil servants in Britain are political-administrators. They firefight ministers to keep them out of trouble in parliament and elsewhere. They draw together and interpret specialist advice for ministers who are rarely experts in their field of responsibility. By tradition they are a source of 'institutional skepticism' about policies but, once the decision is made, their job is to give ministers what they want; nowadays they are described as 'can do' civil servants. So, they delivered public sector reform.

There are no generalist civil servants in Denmark. All are specialists, whether lawyers or the professional experts of the welfare state, and they play a key role in policy formulation and design as well as implementation, providing 'integrated advice' (Ministry of Finance 1998 cited in Jensen et al. 1999). So, public sector reform in Denmark is characterized by a pragmatic orientation' (Greve & Jespersen 1998, 14),ⁿ a technocraticrational' of the reforms (Jensen 1998a, 60); and driven by bureaucrats, not politicians, most notably the Ministry of Finance.¹² So, SOEs reflect the pragmatic, technocratic Danish tradition because they are an intermediate reform; neither privatized nor state run.

c. Freedom to manage vs. political control

Although this summary comparison emphasizes the differences between the two countries, analyzing traditions can also identify similarities. Thus, parliamentary sovereignty and ministerial accountability mean both governments a problem; bureaucratic accountability.

Politicians and top bureaucrats in both countries distinguish between policy and management, justifying the reforms with the argument that it gives managers the freedom to manage and deliver public services efficiently.The contrast is between British agencification and Danish de-agencification. So, British reforms sought to increase the freedom to manage, whereas Danish reforms sought to enhance political control, a course of action which seems to undermine the rationale of the reforms. Any such conclusion is misleading. Agency reform in both countries seeks to increase political control of the bureaucracy. NPM is sometimes said to dismantle hierarchy. But several strands clearly aim to reinforce hierarchicalcontrol. is one example; it is an attempt to make hierarchy work. Britain faces the same tension between deconcentration and political control. The (then) Conservative Home Secretary. Michael Howard, sacked Derek Lewis, chief executive of the Prison Service, who complained bitterly about the Home Secretary's extensive interference in operational matters. He alleged Howard "invented a new definition of the word "operational" which meant "difficult".' He commented that Howard's attempt to 'use the distinction between policy and operation was no more

than a political fig leaf which was so small as to be grossly indecent' (cited in Barker 1997). Control is an objective common to politicians in both countries whether direct as with de-agencification or indirect as with agency frameworkdocuments contract steering.

Beliefs about State and Civil Society, At the most general level, the difference between Britain and Denmark is between an allegiant-deferential or passive political tradition and a participative tradition characterized by associational politics. This difference obviously underpins the distinctive decentralizing strand in Danish public sector reform. I now unpack this explanation of the differences between the two countries.

a. Exclusive vs. inclusive

Public sector reform in Britain was an attack on collectivism and a reassertion the libertarian strand in the British governmental tradition. It attacked policy networks which it castigated as examples of producer and trade union power. Thus, Henney (1984, 380-81) sees networks as an example of the corporate state; 'the institutionalised exercise of political and economic power' by the various types of local authority, government, the unions and to a lesser extent business. Each network builds a 'cultural cocoon' rationalizing their interests with the public interest. Producers interests rule, and it is not OK. Privatization and marketization aimed to destroy the cocoons. Union participation was curbed, but many British networks are based on professional interests concerned with the allocation of resources by networks in welfare state services. They persist and, because of service fragmentation, have multiplied and grown stronger in a localized form. It is hard to identify equivalent trends in Denmark where formal institutionalized participation remains a characteristic of Danish democracy does not systematically seek to exclude important interests.

b. Consumer vs. citizen

As a tourist, there is always the danger I will become dewy-eyed about Danish associational life.¹³ Strengthening user and citizen roles in public sector service delivery is distinctive and to British eyes an exciting alternative New Right market prescriptions. The description 'self-organizing' is apt and the consumer reforms in Britain are no parallel. Such reforms are distinctively Scandinavian and there is no reason to associate them with the NPM which never envisaged democratization as a means of delivering services let alone improving efficiency.¹⁴ If other reforms were 'interpreted' through the lens of Danish political traditions, the citizen reforms are a product ofthat tradition.

Interpreting Distinctive Dilemmas

Differences in traditions, within and between countries, explain the different interpretations of reforms and these competing interpretations produce distinctive dilemmas. A dilemma arises for an individual or an institution when a new idea stands in opposition to an existing idea and so forces a reconsideration. Public sector reform, as interpreted in the two countries, throws up different dilemmas.¹⁵

In Britain, public sector reform was one of the motors behind the shift from government to governance because it multiplied networks (Rhodes 1997a). A central objective of reform was to increase political control, but the policy had many unintended consequences. How do you control these networks? This question confronts the dilemma of steering. Other phrases used in Britain to describe the same dilemma are joined-up government or holistic governance. Whatever phrase is preferred, the problem is always the same. How do you control independent organizations? The answer is not with the tools of corporate management which focus on intra- not interorganizational and management. The 'New Labour' version of the dilemma runs as follows. Government confronts 'wicked problems' which do not fit in with functional government based on central departments and their associated policy networks. Such functional government is costly, centralized, short-term, focuses on cure not prevention, lacks coordination, measures the wrong things and is not accountable. The solution is holistic government which will span departmental cages. The twelve recommendations holistic budgets designed around outcomes, not functions; cross-functional outcome measures; integrated information systems (for example, one stop shops); and culture, value for money and preventive audits (Perri 6 1997, 10-12 and ch. 4-7). In short, the solution of new means of central control will conflict with the negotiating style of joint decision making characteristic of networks, a point grasped by neither Conservative nor Labour governments (Rhodes 1997b).

In Denmark, the key dilemma posed by public sector reform is the challenge to democracy and accountability. The dilemma occurs in many guises. First, there is the democratic dilemma of the gray zone where the accountability and market competition conflict to the detriment of the former (Greve 1998; Jensen 1998a). Second, there is the dilemma of institutionalizedparticipation the 'Everyday Maker.' Sum (1998, 14) sees 'a split in the Danish political culture between the strong sense of autonomy of citizens in civil society' and the state 'as the medium for the public good" operating through the institutionalized participation. For Bang & Sorensen (1998, 20), the Everyday Maker acts out various roles - contractor, board member, leader - and the locus of her political activity has shifted from 'formal organizing to more informal networking' in the fragmented networksof

worksofgovernance. So, the Everyday Maker is an alternative to, and a potential rival of, formal participation. Third, the active citizen may be rare, even an endangered species. User boards involve only a small minority and conventional methods of service delivery predominate. Jensen (1998b) shows how the democratic experiment in Danish social housing is confoundedby fatalism of tenants and the lack of suitable democratic skills. So associational governance may be a distinctive feature of Danish public sector reform, but is it the defining characteristic? Finally, do selforganizingmethods service delivery work? There are some obvious questions. What does it costs? Is it open? Who is excluded? The evidence is contradictory. Danish schools may be open. However, pool steering, a pool of government money for which local actors can bid to do something about 'wicked issues,' experiences variable success which hinges on the capacity of actors to cooperate and create effective local policy networks. They can only do so if they avoid hard-core social problems and 'difficult' minorities. So, self-organization is not necessarily open or accountable, nor is it controllableby local authorities or the central state and this reduced capacity to steer matters when groups are excluded.

Concluding Remarks

In theory at least, throughout this article, I have been talking about one topic, the New Public Management. I have shown that the British and Danish governments introduced markedly different policies under this label. I explained the differences by the differing beliefs about the institutions history of government. Finally, I argued these differences mean the governments confront distinct dilemmas. So, the next phase of public sector reform will not start at the same point. In Britain the main problem is constructed as steering: as re-establishing control over networks. In Denmark, the key dilemma is constructed as democratizing governance through self-organizing citizens.

There are some obvious qualifications to this bald argument. First, in both countries, the reforms are in progress. We do not know the outcomes (Jørgensen 1999). Any interpretation is speculative to a greater or lesser degree.

Second, there are many traditions in each country, each providing other possible constructions of reform. In Denmark, there is the Ministry of Finance's NPM reform strategy; a coherent narrative of orderly change where the key dilemma is weak central coordination vs. ministerial autonomy.So, reforms provide 'a simple, coherent narrative that reinforces human belief that change can be domesticated and controlled' (Jensen 1998a, 65). There is the 'slow revolution' narrative which sees change continuouslytranslated

tinuouslytranslatedthrough the beliefs and actions of actors socialized into the traditions of Danish government. There is the democratic revolution narrative built around the Everyday Maker. An understanding of change starts with conflicting beliefs. So, beliefs about parliamentary sovereignty conflict with beliefs about professional expertise and autonomy which in turn conflict with beliefs about ministerial accountability. The beliefs about the freedom to manage in the public sector reforms simply point up these conflicts because they bump into beliefs about professional autonomy and ministerial accountability. Even the distinctive user reforms create dilemmasas 'Everyday Maker' confronts beliefs in representative democracy, especially the role of the elected representatives as spokesperson for the citizen. The different stories, the collision of ideas and the dilemmas posed by conflicting ideas become the catalyst for yet more change (and on Britain see Bevir & Rhodes 1998).

I do not want to argue for any one interpretation, only for the inevitability such interpretations and the key role of beliefs about government those interpretations. There is no such thing as NPM. The need to domesticate public sector reform, to make it orderly and comprehensible, may be understandable but it does not mean there is an essentialist account of such changes; there are only competing webs of interpretation.

---

NOTES

1. Revised version of an inaugural lecture delivered at the University of Copenhagen on Tuesday 24 November 1998. My thanks to the public administration group at the Institute of Political Science, University of Copenhagen who provided invaluable advice, references and translations. I could not have written the lecture without the help of Carsten Greve, Lotte Jensen, Torben Beck Jørgensen and Tim Knudsen. I would also like to thank Mark Bevir and Janice McMillan for help and advice.

2. This article is a specific illustration of a broader trend: the shift from government to governance. As governments contract-out services, sell agencies to the private sector, create quasi-markets, and decentralize functions to local authorities, they fragment associations and voluntary organizations. This process creates the conditions in which networks multiply. To deliver their services, organizations have to cooperate with one another; they are interdependent. In the UK, corporate management and marketization not create markets. Fragmentation led to greater inter-organizational dependence multiplied networks as an unintended consequence of public sector reform. So to explore public sector reform is to explore one of the motors behind the new governance; it is, in part, a consequence of public sector reform (on Britain see Rhodes 1997a and on Denmark see Bang & Sørensen 1998).

3. Rhodes 1997a, ch. 5-7 and Rhodes 1998 provide an account of public sector reform in the UK under both the Conservative government and New Labour since May 1997. See also: Gray and Jenkins 1998; Hennessy 1998; Riddell 1997 and Theakston 1998a and b. Equivalent accounts in English of Danish public sector reform can be found in: Greve 1997 and 1998; Greve & Jespersen 1998; Jensen 1998: Jørgensen & Hansen 1995; Jørgensen 1999. I have also used private information collected in my capacity as Director of the ESRCs Whitehall Programme. For both the UK and Denmark I con- centrate on the 19905. I have not attempted to be up-to-date because my concern is to describe and interpret trends, not to provide a diary of administrative events.

---

3. Rhodes 1997a, ch. 5-7 and Rhodes 1998 provide an account of public sector reform in the UK under both the Conservative government and New Labour since May 1997. See also: Gray and Jenkins 1998; Hennessy 1998; Riddell 1997 and Theakston 1998a and b. Equivalent accounts in English of Danish public sector reform can be found in: Greve 1997 and 1998; Greve & Jespersen 1998; Jensen 1998: Jørgensen & Hansen 1995; Jørgensen 1999. I have also used private information collected in my capacity as Director of the ESRCs Whitehall Programme. For both the UK and Denmark I con- centrate on the 19905. I have not attempted to be up-to-date because my concern is to describe and interpret trends, not to provide a diary of administrative events.

4. Initially, a White Paper on 'Better Government' was promised for autumn 1997. It vvould clean up politics; root out waste and inefficiency; improve regulation; listen to the people in developing services; provide greater transparency and accountability; use the potential of IT to deliver services, not just information; and break down institutional barriers to working together (Clark 1997). The Chancellor of the Duchy of Lancaster promised a radical initiative which would set out the vision of what government would look like in 15 years. It was repeatedly delayed. The deadline of May 1998, and the 'new deadline' of June or July 1998 came and went. I err on the side of understatement when I infer public sector reform was not a priority. The White Paper finally emerged in March 1999 (Cm 4310). A defining characteristic is its call for 'joined-up government,' or better coordination between central and local government and any other sector involved in service delivery, but the White Paper is not an exciting document. It is mainly a case of more of the same.

5. Danish agencies are also known as directorates and distinguished by keeping professional within the ministry, albeit with a special status. This contrasts with the Swedish model which stresses professional autonomy and locates professional administration outside the ministries; and the Norwegian model which oscillates between the two, seeking a balance between professional autonomy and political control (Christensen 1995, 30).

6. On the British governmental tradition my main sources are: Beer 1982; Birch 1964; and Greenleaf 1983a and b; and, to show that my summary has at least widespread currency, I have drawn on the popular textbooks by Kavanagh 1990, Norton 1991; and Rose 1985.

7. My view of tradition differs. His opposing tendencies are ahistorical. Although they come into being in the nineteenth century, they remain static, acting as fixed categories, ideal types, into which he forces individual thinkers and texts, even different parts of the one text or different utterances by the one thinker. Instances cannot be constructed by comparison with the features of a tradition. The UK governmental tradition has many interwoven strands (and for an example see Bevir & Rhodes 1998).

8. My main sources on the Danish governmental tradition are: Christensen 1995; Elder et al. 1982; Jensen 1998; Knudsen 1991, ch. 1, especially 89-97; Knudsen & Rothstein 1994; and OECD 1993.

9. A few examples to support this conclusion. Elder et al. (1982, ch. 1) treat Denmark as a consensual democracy characterized by high 'regime legitimacy' and 'widespread support for the principle of parliamentary government'; 'a low level of confiict about the actual exercise of power within that state'; and 'a high degree of concertation in the gestation of public policy.' Christensen (1995, 25) argues 'the Scandinavian state tradition ... is primarily characterized by homogeneity and consensus, . . . strong cultural norms of collectivity and equality, and specialized arenas.' So, 'The parliamentary . . . resulted in both strong executive and legislative powers, and a close, cooperative and consensus-oriented style, characterized by trust and mutual adjustment'(ibid., 13).

10. This article seeks to show how traditions shape reforms. Another way of making my case is to demonstrate the inadequacies of accounts of public sector reform which ignore the impact of governmental traditions. See for example: Anne Drumaux, 'Commentary,' in Pedersen et al. 1997. She argues for sanctions in contracts, criticizes the ministerial power to revoke contracts, insists the Ministry of Finance must contiuue to advise on and promote the reforms, and wants standardization of contract procedures. She completely ignores the constitutional position of ministers and the central role of trust and pragmatism in bargaining between ministers. She wants to change the Danish governmental tradition without even recognizing that it is a constraint on her proposals! For example, the role of the Ministry of Finance as the pilot of the reforms can be interpreted as an erosion of ministerial government. The underlying message of reform is the ur^ent need for central cooroinntion, not mimstenal autonomy. Wnyt chance this message will be welcomed by ministers seeking to re-assert political control?

---

11. A touch of irony - Kavanagh (1990, 66) comments that 'British politicians are noted for their pragmatism, a quality that is contrasted with the tendency for politicians in some Continental countries to be more theoretical and abstract in their thinking.' His comment does not quite capture the differences in public sector reform in Britain and Denmark!

12. All of these differences could be summed up using the idea of the developmental state. As Marquand (1988, 145) argues, Britain failed 'to adapt to the waves of technological and institutional innovation sweeping through the world economy' and did not become an adaptive, developmental state because of a political culture suffused with the values and assumptions of whiggery, above all with the central Lockean assumption that individual property rights are antecedent to society. In such a culture, the whole notion of public power, standing apart from private interests, was bound to be alien. Yet without that notion, it is hard to see how a developmental state, with the capacity to form a view of the direction the economy ought to take, and the will and moral authority to put its views into practice, can come into existence (ibid., 154). Denmark does have a developmental state. And note, Britain's distinctive political tradition once again lies at the heart of the explanation.

13. It is as well to ask what the link is between associations and democracy rather than assuming there is one. It can be argued that associational life presumes social and cultural homogeneity and that minorities, such as 'refugees,' are excluded. If so, being 'one of us' becomes a prerequisite of associational democracy, it is not open to all and" therefore, undemocratic.

14. My Danish colleagues commented that decentralization, especially self-organizing, was not part of the new public management. I agree it should not be subsumed under this label, but it is, for example by the OECD (1995, 8).

15. Rhodes (1997a, ch. 5) argues that public sector reform poses five dilemmas for British government: fragmentation-complexity, steering, disasters, accountability and ethics (and for an equivalent list for Denmark see Jørgensen 1999). I do not have the space to discuss all five. I focus on steering and accountability to show how traditions spotlight different dilemmas and construe them differently.

16. My analysis of both countries is basedon academic constructions of how other actors, mainly the political and bureaucratic elite, construct their world. Ideally, of course. I should tell the story through the eyes of the several political and bureaucratic actors, focusing on their beliefs and reasons; that is 'thick descriptions" of their world. But their version of the story is not available to me. So, instead I use academic account of how public managers responded to public sector reform; that is, academic 'constructions of other people's constructions of what they are up to' (Geertz 1973, 9. See also Bevir & Rhodes 1999a and Rhodes 1999, ch. 8).

REFERENCES

Aucoin, P. 1995. The New Public Management. Canada in Comparatixe Perspective. Quebec:
Institute for Research on Public Policy.

Bang, H. P. & Sorensen, E. 1998. The Everyday Maker: a New Challenge to Democratic
Governance,' paper to the ECPR Workshops, 26th Joint Sessions. University of Warwick.
23-28 March. Forthcoming in Administrative Theorv and Praxis. 21/3. 325—41.

Barker, A. 1997. 'Political Responsibility for UK Prisons - Ministers Escape Again." Public
Administration 76/1, 1-23.

Beer, S. 1982 [1965]. Modem Britisk Politics. London: Faber.

Bevir, M. 1999. The Logic of the History ofldeas. Cambridge: Cambridge University Press.

Bevir, M. & Rhodes, R. A. W. 1999a. 'Studying British government: reconstructing the
research agenda,' Britisk Journal of Politics and International Studies 1(2). 215-39.

Bevir, M. & Rhodes, R. A. W. 1999b. 'Narratives on British Government.' Rerne Francaise
de Science Politique, 49(3). 355-77.

Birch, A. H. 1964. Representative and Responsible Government. London: Allen & Unwin.

Christensen, T. 1995. 'The Scandinavian State Tradition and Public Administration,' Paper.
American Political Science Association, Chicago, 31 August-3 September.

Christensen, T. & Lægreid, P. 1998. 'Administrative Reform Policy: the case of Norway,'
International Revievof Administrative Sciences 64/3, 457-75.

Clark, D. 1997. 'Delivering Better Government from the Bottom Up.' Speech by the Chancellor
of the Duchy of Lancaster, 17 June, Queen Elizabeth Conference Centre, London.

Cm 1599 199l". The Citizen's Charter. Raising the Standard. London: HMSO.

Cm 4014 1998. Modem Local Government. In Touch with the People. London: The Stationery
Office.

Cm 4310 1999. Modernising Government. London: The Stationery Office.

Davis, G. 1998. 'Australian Administrative Tradition,' in Shafritz, J. M., ed., International
Encyclopaedia of Public Policy and Administration. Boulder, Colorado: Westview Press.

Dav, P. & Klein, R. 1997. Steering but not Rowing? The Transformation of the Department
of Health: a case study. Bristol: Policy Press.

Elder, N., Thomas, A. H. & Arter, D. 1982. The Consensual Democracies? The Government
and Politics of the Scandinavian States. Oxford, Martin Robertson.

Gamble, A. 1990. 'Theories of British Politics,' Political Studies 38, 404-20.

Grav. A. & Jenkins, B. 1998. 'New Labour, New Government? Change and Continuity in
Public Administration and Government 1997,' Parliamentary Affairs 5111, 111-30.

Greenleaf, W. H. 1983a. The British Political Tradition. Volume 1. The Rise of Collectivism.
London: Methuen.

Greenleaf, W. H. 1983b. The British Political Tradition. Volume 2. The Ideological Heritage.
London: Methuen.

Greve, C. 1997. 'Governance by Contract. Creating Public-Private partnerships in Denmark.'
Paper, European Group of Public Administration Annual Conference, 10-13 September,
Leuven, Belgium.

Greve, C. 1998. 'Exploring Contracts and Boards as Reinvented Institutions for Bridging the Gap Between Politics and Non-state Governance.' Paper, ECPR Joint Sessions, Workshop on 'Political Institutions: Beyond the State,' 23-27 March, University of Warwick, Coventry, UK.

Greve, C. & Jespersen, P. K. 1998. 'New Public Management and Its Critics. Alternative
Roads to Flexible Service Delivery to Citizens.' Paper, European Group of Public
Administration Annual Conference, 14-17 September, Paris.

Hall, P. & Taylor, R. 1996. 'Political Science and the Three Institutionalisms,' Political
Studies 44, 936-57.

Hansen, K. 1997. 'The Municipality Between Central State and Local Self-Government:
Towards a New Municipality,' Local Government Studies 2114, 44-69.

Harrison, B. 1996. The Transformation of British Politics. Oxford: Oxford University Press.

Hazell, R. & Morris, B. 1999. 'Machinery of Government: Whitehall,' in Hazell, R., ed.,
Constitutional Futures. Oxford: Oxford University Press.

Hennessy, P. 1998. 'The Blair Style of Government,' Government and Opposition 33/1,
3-20.

Henney, A. 1984. Inside Local Government. The Case for Radical Reform. London: Sinclair
Browne.

Hood, C. 1991. 'A Public Management for All Seasons?,' Public Administration 69, 3-19.

Hood, C. 1996. 'Exploring Variations in Public Management reforms of the 19805,' in Bekke,
H. A. G. M., Perry, J. L. & Toonen, T. A. J., eds., Civil Service Systems in Comparative
Perspective. Bloomington, Indiana University Press.

Hood, C, James, O. & Scott, C. 1998a. 'Regulation in Government: Has it increased, is it
increasing and should it be diminished?' Paper, ESRC Whitehall Programme Final Conference,
of Birmingham, 17-19 December.

Hood, C, James, 0., Jones, G. Scott, C, & Travers, T. 1998b. 'Regulation Inside Government:
the New Public Management Meets the Audit Explosion,' Public Money and
Management 18/2, 61-68.

Jensen, H. N., Jensen, L. & Knudsen, T. 1999. 'The Changing Role of the Danish Senior Civil
Service.' Paper, Workshop on 'The Changing Role of the Civil Service,' Nuffield College,
Oxford, 15-16 January.

Jensen, L. 1998a. 'Interpreting New Public Management: The Case of Denmark,' Australian
Journal of Public Administration 51I'4, 55-66.

Jensen, L. 1998b. 'Cultural Theory and Democratizing Functional Domains: the case of
Danish Housing,' Public Administration 76, 117-39.

Jørgensen, T. B. & Hansen, C.-A. 1995. 'Agencification and De-agencification in Danish
International Review of Administrative Sciences 61, 549-63.

Jørgensen, T. B. 1999. 'The Public Sector in an In-Between Time,' Public Administration 11,
forthcoming.

Kavanagh, D. 1990. British Politics: continuity and change. 2nd edition. Oxford: Oxford
University Press.

Kickert, W. J. M. 1997. 'Public Management in the United States and Europe,' in Kickert,
W. J. M., ed., Public Management and Administrative Reform in Europe. Aldershot: Edward
Elgar.

Kjellberg, F. 1988. 'Local Government and the Welfare State: Reorganisation in Scandinavia,'
Dente, B. & Kjellberg, F., eds., The Dynamics of Institutional Change. Local
Government Reorganisation in Western Democracies. London: Sage.

Knudsen, T. 1991. 'State Building in Scandinavia: Denmark in a Nordic Context,'
in Knudsen, T., ed., Welfare Administration in Denmark. Copenhagen: Ministry of
Finance.

Knudsen, T. 1999. 'How Informal Can You Be? The Case of Denmark?.' in Wright. V.,
Peters, B. G. & Rhodes, R. A. W., eds., Administering the Summit: administration of the
core executive in developed countries. London: Macmillan.

Knudsen, T. & Rothstein, B. 1994. 'State Building in Scandinavia," Comparative Politics.
January, 203-20.

Loughlin, J. & Peters, B. G. 1997. 'State Traditions, Administrative Reform and Regionalization,'
Keating, M. & Loughlin, J., eds.. The Political Economy of Regionalism.
London: Frank Cass.

Marquand, D. 1988. The Unprincipled Society. London: Cape.

Marsh, D. & Rhodes, R. A. W. Eds. 1992. Policy Networks in British Government. Oxford:
Clarendon Press.

Ministry of Finance 1993. Nyt syn pit den offentlige sektor. Copenhu.wn: Minisiryof Finanee.

Ministry of Finance 1998. Forholdet mellem minister og emhedsnnnid. Copcnluigcn. Betænkning
1354.

Namier, Sir L. 1974. Personalities and Power. Westport, Connecticut: Greenwood Press.

Nielsen, K. & Pedersen, O. K. 1988. 'The Negotiated Economy: Ideal and History," Scandinavian
Political Studies 11/2,79-101.

Norton, A. 1994. The International Handbook of Local and Regional Government. A Comparative
of Advanced Democracies. Aldershot: Edward Elgar.

Norton, P. 1991. The British Polity. 2nd edition. New York: Loncman.

OECD 1995. Coxeruance in Transition: Public Management Reform in OECD Countries. Paris:
OECD/PUMA.

Olsen, J. P. 1983. Organized Democracy. Political Institutions in a Welfare State - The Case
ofNorway. Oslo: Universitetsforlaget.

Olsen, J. P. & Peters, B. G. 1996. Lessons from Experience. Oslo: Scandinavian University
Press.

Parris, H. 1969. Constitutional Bureaucracy. London: Allen & Unwin.

Pedersen, P. K., Sørensen, H. D. & Vestergaard, J. B. 1997. 'The Contract Management
Project in Denmark,' in OECD, Benchmarking, Evaluation and Strategk Management.
Paris: OECD.

Perez-Diaz, V. M. 1993. The Return of Civil Society. Cambridge, Mass.: Harvard University
Press.

Perri 6. 1997. Holistic Government. London: Demos.

Peters, G. B. 1997. 'A North American Perspective on Administrative Modernisation in
Europe,' in Kickert, W. J. M., ed., Public Management and Administrative Reform in
Western Europe. Aldershot: Edward Elgar.

Plowden, W. 1994. Ministers and Mandarins. London: Institute for Public Policy Research.

Power, M. 1994. The Audit Explosion. London: DEMOS.

Rhodes, R. A. W. 1997a. Understanding Governance. Buckingham: Open University Press.

Rhodes, R. A. W. 1998. 'Different Roads to Unfamiliar Piaces: UK experience in Comparative
Australian Journal of Public Administration 57/4, 19-31.

Rhodes, R. A. W. 1999 [1981]. Control and Power in Central-Local Relations. Aldershot:
Dartmouth.

Riddell, P. 1997. 'Advising the Prime Minister.' Paper, ESRC Conference, 'Future Whitehall,'
London.

RIPA. 1987. Top Jobs in Whitehall Appointments and Promotions in the Senior Civil Service.
London: RIPA.

Rose, R. 1985. Politics in England: change andpersistence. 4th edition. London: Faber.

Thain, C. & Wright, M. 1995. The Treasury and Whitehall: the planning and control of public
expenditure 1976-1993. Oxford: Oxford University Press.

Theakston, K. 1998a 'New Labour, New Whitehall?,' Public Policy and Administration 13/1,
13-33.

Theakston, K. 1998b. 'Labour and the Civil Service,' in Taylor, G., ed., The Impact of New
Labour. London: Macmillan.

World Bank. 1992. Governance and Development. Washington, D.C.: World Bank.