Administrative Reform: Changing Leadership Roles?

Transcription

1 Administrative Reform: Changing Leadership Roles? TOM CHRISTENSEN* Three interwoven change elements characterize New Public Management: substantial horizontal and vertical specialization, substituting an integrated sector model for a fragmented functional model, and extensive use of contracts as part of a make the manager manage kind of incentive system. This article discusses the effects and implications of these reform elements on political-democratic processes in general, and on political, administrative, and public enterprise leadership roles more specifically. Examples from Norway and New Zealand illustrate the discussion. INTRODUCTION 1 Modern reforms of civil service like New Public Management (NPM) are quite often preoccupied with management and economic perspectives on the public sector (Boston, Martin, Pallot, and Walsh 1996). They focus on how to let and to make the managers manage, and on how to improve the efficiency and quality of public services (Kettl). Seldom do these reforms emphasize the broader questions of democracy: whether the reforms have effects that fundamentally influence democratic-political processes in a country, and whether they more specifically weaken or strengthen the central political control (Christensen and Lægreid 2001). The latter questions are addressed here through emphasizing the effects and implications of reforms on political and administrative roles or, more precisely, the effects on the role enactment and role relationships of political leaders, administrative leaders, and leaders of state commercial units. 2 I will concentrate on three main reform elements that are typical of, if not exclusively connected to, NPM: the increased and strong horizontal specialization in civil service between different tasks and functions, resulting from a principle of single-purpose organizational units (Boston et al. 1996, 75); the extensive use of structural devolution or vertical specialization between different levels and institutions, used, for example, to justify the establishment of a number of commercial state companies; and the use of contracts, both individual and institutional, to regulate activities and accountability between leadership levels and institutions. These three elements are related, of course, but they will mostly be discussed *University of Oslo Governance: An International Journal of Policy and Administration, Vol. 14, No. 4, October 2001 (pp ) Blackwell Publishers, 350 Main St., Malden, MA 02148, USA, and 108 Cowley Road, Oxford, OX4 1JF, UK. ISSN

2 458 TOM CHRISTENSEN separately. The theoretical discussion is based on the organization design literature, focusing on effects of using different organizing principles in public organizations (Gulick; Hammond; Hood and Jackson; Simon). Several central questions are posed. Do modern administrative reforms through a combination of increased fragmentation through horizontal structural differentiation, increased decentralization through structural devolution, and undermining of traditional hierarchical control through contracts result in a general decrease in the control of political leaders and increase in the influence of administrative leaders and state commercial leaders (Christensen and Lægreid 2001)? Do the reforms result in a change of each leadership role, without changing the overall pattern of influence very much? Or are the reforms having differentiated effects on political control, meaning more control over the core public sector but less control over noncore institutions like state commercial units (Boston et al. 1996)? The scholarly and political debate about the effects and implications of NPM presents some clear and contradictory arguments. These arguments will be discussed and related to examples from studies of actual effects on leadership roles of administrative reforms in countries where the reform principles have been used to different extents. Norway and New Zealand are the main countries discussed, but I will also touch upon other countries. MAIN FEATURES OF REFORM IN NEW ZEALAND AND NORWAY Beginning in 1984, New Zealand chose a radical reform path, impelled by a very threatening economic crisis and made possible by an elective dictatorship (Boston et al. 1996; Mulgan; Nagel 1998, 241; Schick 1996, 11). 3 The reforms were furthered by an alliance of economists, both inside and outside the civil service, and the entrepreneurship of Roger Douglas, the incoming Minister of Finance for the Labour government, who had a number of close supporters in the party (Bollard, 89; Goldfinch, , 183; Massey, 65 67). New Zealand started out with radical reforms in the economic sector, and then proceeded with public sector reforms and changes in the structure of government. These latter changes were characterized by introducing radical measures into the three variables on which this paper focuses. First, extensive application of the single-purpose principle of structuring the administrative apparatus created a much more differentiated structure (Boston et al. 1996, 58 66, 77 80). This differentiation was economically inspired, leading to the establishment of units for ownership, provision and purchasing of public services, and so on. Second, a stronger devolution of authority and functions, involving both more managerial autonomy and structural devolution, led to a smaller core executive apparatus, more Crown entities, and more state-owned enterprises, and eventually to privatization. Stronger horizontal and vertical specialization combined resulted in a less integrated administrative apparatus (Christensen and Lægreid 2001).

3 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 459 Third, New Zealand started to develop individual performance contracts for leaders (Martin 1998) as part of the introduction of a make the managers manage principle (Kettl), with a stronger use of incentives, a feature that is now dominant. Norway had few economic reasons to start radical reforms in the early 1980s, but a general pressure to do so existed, created by a combination of internal questions about the legitimacy of a large public sector, rising electoral support of conservative and liberal parties, and pressure from international organizations for reform (Christensen and Lægreid 2001; Olsen 1996). 4 The reform process started slowly in 1986 and 1987 with the presentation of modernization and renewal programs from a conservative and a labor government respectively, and it picked up speed in the 1990s (Lægreid and Roness 2000). The first reform measure to be introduced widely was a modified Management by Objective system (Christensen and Lægreid 1998b). Norway has been much less keen than New Zealand to pursue a radical single-purpose principle of structuring the central civil service. Elements of this exist, but the motivation has not generally been economic. During the 1990s, Norway gradually chose more devolution, also more reluctantly than New Zealand had done. More ministerial functions were moved to agencies and more agencies transformed to state-owned enterprises, but few activities were privatized (Lægreid and Roness 1998). Concerning contracts, Norway started a new leadership wage and contract system in the early 1990s, but it has been applied reluctantly and is seen as controversial (Lægreid 2001). The main changes in Norway are the principle of let the managers manage to a limited extent and the use of institutional contracts. Comparing the effects and implications of administrative reforms in New Zealand and Norway has some obvious advantages and disadvantages. The advantages include the fact that both countries are small and therefore relatively easy to research. They have some of the same state traditions, concerning both cultural values and the structure of central civil service. Both are parliamentary democracies. They also chose different paths in the 1980s and 1990s concerning administrative reforms an extreme and a cautious one, respectively and thus can serve as good contrasting cases for the three main reform elements at the center of this analysis. The disadvantages include the fact that the comparison here will not cover larger and more complex countries, those with presidential systems, countries that chose a more middle-of-the-road reform strategy, and so on. However, the aim of the comparison is mainly illustrative, not systematic. DOES INCREASED HORIZONTAL SPECIALIZATION CLARIFY FUNCTIONS AND ROLES? There is no one best way to organize public organizations, regardless of environmental conditions, types of goals, types of leadership, types of

4 460 TOM CHRISTENSEN tasks or policy, cultural traditions, demographic features, and so on. However, not all ways of organizing work equally well to fulfill specific goals (Gulick). This leaves the field open for a competition between different organizing principles (Hood and Jackson). These principles have a tendency to come in waves or gain popularity in certain periods and then disappear or blend together with other principles, permitting new or redefined ones to come forward (Downs and Larkey; Light; Røvik). The discussion about horizontal specialization in government is an old one. Luther Gulick did the most influential early work in the field. Gulick advocates quite clearly that tasks in a public organization should be divided according to a principle of homogeneity, and the crucial point is then what principle of horizontal specialization should dominate. He (21 23, 31) says that the preference for and combination of principles depends on the goals one would like to achieve, leading to a specific policy content. Every specializing principle has both advantages and disadvantages. Gulick underscores that organizing based on purpose or sector is the best and dominant method for achieving certain major goals of large public organizations, because it is easy to use and understand. However, it can easily lead to too much centralization, overlapping purposes and problems using technical expertise. In contrast, specialization based on process or function which fully uses technical expertise, is often the most efficient, and creates good career opportunities. However, it can obstruct achievement of broad goals, because its focus is on how rather than what, and democratic control can be hindered if the experts are too powerful. Interestingly enough, these are all major points in the debate about the type of horizontal specialization in New Zealand. In many countries that have implemented NPM-like reforms, increased horizontal separation of administrative functions has gained popularity, often coupled with a clearer structural separation of political and administrative functions, i.e., some kind of small is beautiful functional model (Boston et al. 1996, 4, 72). Instead of a traditional integrative organizing of public functions, the goal is a sharp structural separation of tasks related to political advice, regulation and control, ownership, purchase and provision, and so on. New Zealand provides a typical example of this kind of structural reform, adopting the functional principle of single-purpose or single-mission organizations and divided responsibility while abandoning a traditional sectoral principle and inclusive responsibility (Boston et al. 1996, 13, 58 66, 81). This increased horizontal specialization represents something new in two respects. It subdivides bureaucracy to a greater extent than before. More importantly, it is based primarily on economic considerations and categories, paving the way for new public commercial units and, eventually, privatization elements, making it closely related to structural devolution. H. George Frederickson (266) stresses that the steering-rowing metaphor, found in reinventing government and other similar NPM

5 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 461 reforms, reintroduces the policy-administration dichotomy, but the rowing or doing part is now just service. One main argument for increased horizontal specialization is that it enhances efficiency and effectiveness by clarifying functions, avoiding overlap, making authority and lines of command more unambiguous, avoiding cross-subsidizing between services, avoiding provider capture, enhancing contestable service delivery, and so on (Boston et al. 1996, 4 5, 12 13, 77, 94). It is also unclear whether a sectoral model, with its many heterogeneous functions, can really develop an integrated culture. Arguments against the greater separation of different administrative functions primarily point to the problems of fragmentation (Boston et al. 1996, 13, 76). Services to users can end up fragmented. Policy advice can lack practical sensitivity. Collective public cultures can splinter and subcultures emerge, contact patterns and public expertise dissolve. Specialized units can gain some of their former functions, leading to more bureaucracy, not less (Christensen and Lægreid 2001). Fragmentation creates the need for more coordination, leading to more pressure on the government and cabinet, more demand for resources and attention, and eventually capacity problems. Jonathan Boston and colleagues (1996, 86 94) try to sum up the actual effects of the structural reforms in New Zealand, while emphasizing that there has been little relevant research and independent evaluation and that these structural reforms are complicated to evaluate. They point out some predictable effects of increased horizontal specialization, such as increased efficiency and productivity, better quality of services, and less influence from service delivery on policy recommendations all effects that are interconnected with those of vertical specialization and devolution. Increased horizontal specialization also seems to give leaders a more unambiguous sense of mission, and create fewer intraorganizational conflicts. However, Boston and colleagues also admit that some obvious control problems result from the substantial machinery of government changes. Numerous small departments and other organizational units create coordination and capacity problems; increasing conflicts occur between specialized units, some of which must report to too many reviewing and monitoring bodies; and consumers experience problems with fragmented service delivery. 5 Various other problems arise, including problems with separating policy and operational functions in some sectors (like defense) and with separating ownership and purchaser functions when government is the main provider, and some substantial costs related to the reorganization processes. This broad evaluation points out quite clearly both the efficiency and service advantages and the control and coordination disadvantages of such a substantial change of organizing principles, and makes clear that it is difficult to draw an overall conclusion.

6 462 TOM CHRISTENSEN The Norwegian central civil service has historically been characterized by a sectoral model containing relatively few ministries with broad and diverse functions (Christensen 1997b; Christensen and Roness). This has probably led to emphasis on an integrated cabinet with strong collegial features (Eriksen). The horizontal specialization of the subordinate agencies or directorates has become stronger, especially since World War II, creating options for state expansion and also problems with political control (Christensen 1997a; Rolland). Modern administrative reforms have not changed this general picture much, so the structure today is both less horizontally specialized and less oriented towards commercial functions than in New Zealand. However, during the 1990s Norway gradually moved towards applying these principles in a reluctant way. What are the implications for leadership roles of the changes experienced in horizontal specialization? One obvious implication is that NPM is creating more complexity in these roles, as can be seen in New Zealand. The roles of the political and administrative leadership are more separated, making the relationship more formal and creating potential tensions and problems of flow of information and trust (Christensen and Lægreid 2001). In addition, there are increasing problems with coordination and capacity due to increased specialization. More problems and actors are channeled directly towards the central political leadership, creating issues with span of control, attention and expertise. An economically defined increase in horizontal specialization probably also implies a diminished role for political leadership. Taken together, this could potentially undermine administrative leaders political control. However, single-purpose organizations can also create less ambiguous roles and steering, and as such create greater effectiveness. Increased efficiency can create more options for fulfilling political goals. And more exposure and hands-on experience with service delivery and effects of public policies could also be a plus for political leaders. However, such reforms can also create more problems of coordination for administrative leaders, who can be caught up in the cross-pressure between political leaders on the one hand and consumers and active leaders of state-owned enterprises on the other. The Norwegian experience with an integrated administrative model is that it creates a close and informal relationship between the political and administrative leadership, strengthening consensus and trust (Christensen and Peters). Integrated units can be administrative buffers, coordinating and balancing different interests and considerations and solving capacity problems for political leaders, as well as powerful instruments for control and implementation (Campbell and Halligan; Christensen 1997b; Pollitt). However, they can also increase administrative leaders strength and access, and the pressure such leaders can bring to bear on political leaders. It is not easy to evaluate whether this model, which has some obvious control and cultural advantages, is less effective and efficient than the disintegrated model associated with NPM. It is certainly vulnerable to

7 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 463 partiality (Christensen and Peters), and it cannot fulfill the demands of the disintegrated model concerning formality and transparency. However, a study of civil servants reaction to NPM in Norway seems to indicate that it is possible to combine the two models, even though this might appear to make life more complicated. Administrative leaders are both open to new reforms in moderate versions and strongly supportive of political loyalty and traditional professional considerations (Christensen and Lægreid 1998a). DOES STRUCTURAL DEVOLUTION UNDERMINE POLITICAL CONTROL? Structural devolution or vertical specialization implies that authority and responsibility are decentralized or delegated to lower levels, institutions, or positions in the civil service (Christensen and Lægreid 2001). This can be done by differentiating between organizational forms inside core civil service or through vertical interorganizational specialization for example, by establishing noncore units like state-owned companies (Lægreid and Roness 1998). What is typical for NPM-related reforms is not that the organizational forms of structural devolution are new, but that the speed of this process has increased considerably in the last decade and that relatively more noncore units have been established (Grønlie and Nagel; Grønlie and Selle). I will focus first on this latter part of the structural devolution, which is often called the separation of politics and business. Broadly speaking, structural devolution is a question of differentiating between governmental organizational forms. The main idea is that different state functions and tasks need different ways of organizing, or, more precisely, the political leadership needs to control activities to different degrees (NOU, 5). The more politically crucial certain administrative functions are, the closer the ties to the political leaders should be. This principle has been adopted widely all over the world, by, for example, creating megadepartments close to the political leadership (Campbell and Halligan; Stewart and Kimber). The less politically important the functions are, or the greater weight given to other considerations, the more devolution is appropriate for example, establishing agencies with routine tasks, calling on specialized competence a distance from the leadership, or keeping commercial companies at arm s length (Christensen 1997b; Hogwood; Rhodes). One important point is that, of course, no objective answer to what is politically important exists; this changes all the time (Olsen 1983). What is typical for the NPM-related reforms in many countries is that the traditional debate about the differentiated structuring of the state has been almost forgotten, and the focus has been primarily on the division between politics and business. There are many arguments for corporatization of the state and the establishment of state companies or enterprises for service production and delivery, and eventually also for privatization. Some of them are closely related to arguments for changing the horizontal specialization

8 464 TOM CHRISTENSEN from a sectoral to a functional model (Boston et al. 1996, 72 77). One main argument is that the efficiency and quality of services will increase substantially through devolution. Different versions of this argument exist, some broader than others. One holds that this efficiency must lead to a leaner state; another maintains that governments can save money for other collective purposes, even to give the executive political leaders strategic options for fulfilling popular policy goals. A second main argument for vertical disintegration, based on a certain environmentaldeterministic logic (Olsen 1992), is that commercial state units can more easily meet the inevitable increase in global competition because what is needed is more discretion, more freedom from central political control, and more commercial competence (Christensen and Lægreid 2001). A third group of arguments claims this type of structural devolution makes it far easier to separate commercial interests, on the one hand, and general and sector-specific political considerations, on the other. This is said to produce increased consciousness concerning the roles of different actors and institutions in the political-administrative system. The easier it is to differentiate and purify the different roles, the easier it is to balance different considerations, to handle accountability questions, and to judge the preconditions for political and administrative control. There are also different versions of this argument. One underscores that the government has to concentrate on its core functions, making it necessary to restructure, rebuild, and scale back the welfare state somewhat to keep it strong against future attacks. This has been called an argument for a defensive state (Grønlie, 79 80) and is quite often found among social-democratic parties in Europe that have added some neoliberal elements to their policies. A second version of the argument holds that differentiating between politics and business in fact strengthens and helps maintain political control, because exercising such an option is essentially political control of a more general nature (Nagel 1997, 354). There are also numerous arguments against this type of structural devolution. One main argument is that governments in many countries have long, good traditions for conducting business activities in an integrated way and controlling the economy, i.e., balancing political and commercial considerations, and that this integration has contributed to fulfilling major goals in the development of the welfare states and securing resources for public purposes (Christensen and Lægreid 2001). Another is that this will substantially weaken political-democratic control in general (Christensen and Lægreid 2001). Political considerations will be weakened relative to the economical and judicial ones. The central collective capacity for control is decreased, and the scope of control becomes narrowed. Political signals are seen as more distant in the commercial units. In these units, collective purposes are seen, not as the most important, but as expenses for which the state must pay the units. 6 In general, politicians legitimacy seems to be decreasing.

9 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 465 A third set of arguments concerns the efficiency of devolved state commercial units. The skeptics say that increased efficiency is just something that is postulated and not proven by independent evaluations, and that there is a tendency to oversell the efficiency argument (Christensen and Lægreid 2001). What is more, it is difficult to measure efficiency in public companies, and even more tricky to compare public and private business in a benchmarking exercise. A fourth argument says that efficiency questions are seldom seen in a broader economic perspective. What about, for example, the negative side effects of competition, which can eventually increase public expenditures? One example of this can be found in the major deregulations and cutbacks in state postal and telecommunication services in many countries. These can create unemployment and problems of alternative work, especially in the peripheries, which means more people on public pensions, increased social inequality, and few benefits for consumers (Gerritsen; Stephens). How do we know that short-term efficiency really brings long-term efficiency? 7 And what actually happens to a public organization that is split into different commercial units? Will this save money because the enterprises are run more efficiently, or will each of these units expand and increase their administrative costs substantially? Finally, do state units that have been restructured and turned into more commercially oriented units really compete on a level playing field with private businesses, or are they at a disadvantage that will eventually turn out to be an argument for less political control and privatization? 8 I will not follow up on the complexity of all these pro and con arguments. Instead, I will concentrate on the effects of structural devolution on political control and leadership roles. In their comprehensive comparative study of five state-owned enterprises in New Zealand, Barry Spicer, David Emmanuel, and Michael Powell (203) conclude that the reforms have been successful in making enterprises far more commercialized and efficient. 9 This accords with but is somewhat more optimistic than the evaluation by Boston and colleagues (1996, 86 94, ) of the effects of increased vertical specialization and a more unambiguous division of politics and business in New Zealand. Their conclusions are closely related to the main arguments about the effects of increased horizontal specialization: the efficiency and quality of services and policy advice increase, the new organizational units become more focused, entrepreneurial cultures evolve, and so on. However, they are also aware that political control and coordination have become somewhat more problematic and lines of authority more ambiguous, especially for some Crown entities. 10 Grønlie underscores that devolution of public enterprises is nothing new in Norway. He compares the development towards devolution in the 1940s and 1950s with that of the 1980s and 1990s, and shows that the pattern of actors is quite different. The earlier period was characterized by the Labour Party s desire to use public companies in an expansive

10 466 TOM CHRISTENSEN state strategy, about which the Conservative Party was very skeptical. In the later period, conservative and neoliberal forces, accompanied by a political-ideological shift of climate and harsh criticism of the companies for losing money and being overregulated, were eager to establish more companies with greater freedom from central control (Grønlie and Nagel, 327). At this point the Labour Party was skeptical, but acquiesced reluctantly based on the argument that it would be possible to combine strong political control and devolution. Grønlie (78) concludes that this double strategy is difficult to defend and that the effect of the increased devolution seems to be an undermining of political control. 11 Belief in increased control through sharper forms of judicial control is also usually difficult to defend, because such forms seem to create more ambiguity. In a broad study of the state-owned companies in Norway, the country s Directorate of Public Management (21) showed that the control regime is complicated. It consists of laws, a control unit in the superior ministry, regulatory bodies, a yearly statement of corporate intent (less formal than in New Zealand) and reporting of results, the annual formal meeting between the owner (the sectoral minister) and the enterprises, and the activities of the board relating to the company. The report concluded that ownership control is passive and undifferentiated. It is passive because it uses a restrictive, formal definition of control, requiring it to take place only once a year rather than continuously. This feature leaves much room for freelancing by the Parliament (Storting). It is undifferentiated because it fails to take account of the various types of companies. Enterprises that should be guided by broad political considerations are handled in the same passive way as companies with sectorpolitical connections. One can summarize this report by speaking of a double weakening of political control, resulting from the increased distance to the political leadership and from a passive and undifferentiated use of the control instruments at hand. Hans Robert Zuna compares the working of state-owned enterprises in New Zealand and Norway, concluding that there are some typical differences. The SOEs in New Zealand are formally controlled by the Minister of Finance and a Minister of State-Owned Enterprises in the same department, with a connected administrative unit the Crown Company Monitoring and Advisory Unit (CCMAU) that should focus only on commercial interests (Spicer, Emmanuel, and Powell, 47). 12 In New Zealand, noncommercial interests in SOEs often meaning collective political interests must explicitly be paid for (Spicer, Emmanuel, and Powell, 16 20). Their Norwegian counterparts are controlled by the sectoral ministries and should formally combine political (general and sectoral) and commercial considerations. The general effects of more extensive use of state-owned companies in the two countries, as compared to earlier periods, indicate that political control has decreased. However, the nature of this control is different in the two countries. The control in New Zealand focuses narrowly on financial oversight and is more coordinated across

11 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 467 companies; the control in Norway is passive concerning both political and commercial aspects, but more sector-oriented than in New Zealand. My main conclusion here is that the division of politics and business through structural devolution generally undermines the central political control, weakening the executive political leadership and strengthening the leaders of the state commercial units. Studies in Norway and New Zealand show that state commercial activities are about to be moved out of the realm of politics in both countries, even though the control is organized differently. It seems to be generally accepted that, if one permits the establishment of modern state-owned companies, they must be granted a lot of commercial discretion, meaning less political intervention, which often seems to be nearly illegitimate (Christensen and Lægreid 2001). Even though it is taken for granted in most countries that state-owned companies should have considerable discretion for reasons of efficiency, there is still a lot of agonizing over the question. Paradoxically enough, it is also often accepted that they should be controlled by different means. The major point of contention, however, arises over how this control should be organized, and to what degree the particular solutions chosen will likely shape the leadership roles. One way to organize the division between politics and business is to say that the crucial role of politicians in the executive is to take the policy decision to establish state companies and enterprises, but that apart from this they should have a passive ownership role (Boston et al. 1996, 8). The central capacity to control the companies is therefore cut back, and the crucial question becomes whether they are earning money. This typifies the situation in New Zealand. Another approach is to say that the political leaders take the policy decisions about companies, but leave the control of these units to the administrative leaders, who have to build up capacity and a corporate culture, and to a regulatory body and often also to a professional board. This is fairly similar to the situation in Norway today. It is a rather complex way of organizing, probably increasing the influence of administrative leaders, but also creating tension between them and the regulatory bodies. A third method of organization and a more active one for the political leadership is to say that they remain the foremost representatives of the companies owner and should have an active relationship to the companies. This could be done through formal and informal means for example, laws and regulations, continuous formal and informal dialogue, interference in special questions, and forcing the companies to be socially responsible (Boston 1999a, 9). This way of organizing, traditionally employed in both New Zealand and Norway and still seen in some sectors in Norway today, is also rather complicated. It generally leaves both the political and administrative leaders with a stronger role, but the involvement of political leaders tends to vary with type of government and political context.

12 468 TOM CHRISTENSEN DOES THE USE OF CONTRACTS IMPROVE TRANSPARENCY AND CONTROL? The use of contracts has increased quite a lot in many countries, especially those that have emphasized a combination of extensive horizontal and vertical specialization (devolution) and the principle of making the managers manage (Christensen and Lægreid 2001; Kettl). The main idea behind using contracts is that they are new coordination and control instruments, to be substituted for the traditional hierarchical command structure and aiming at more efficiency through incentives. The central precondition for the use of contracts is transparency that the control and service delivery is more open for scrutiny, the goals are unambiguous, the monitoring system already in place, and the incentive system clearly defined (Martin 1995, 40). But will a contract system change central political, administrative, and institutional leadership roles? Will such a system further political control and responsibility, or quite the opposite, lead to a more differentiated control? According to economic theory, contracts can be of different types (Martin 1995, 39). One is the spot contract: short term, informal, and typical for many transactions in markets. A second type is the classic, formal-legal, binding contract. The third is the relational type of contract. This is long-term, informal, and based on mutual understanding and needs, and is not a real contract. Some types of contracts under NPM-like reforms are intended to be similar to the classic type, but also to contain elements of the relational type and to have a quasicontractualist nature (Boston 1999a, 11). The contracts used in relation to NPM reforms seem more specifically to be of two types, the institutional and the individual. There are three different kinds of institutional contracts. First, there are the more general contracts, which regulate the relationship between the political leadership as owner and subordinate levels and institutions as producers. These contracts state major aims and use of resources as policy constraints, but the subordinate units are given discretion to manage and use the instruments necessary to implement policies, although they are obligated to enter into a dialogue with their superiors and to report back about the main results of their activities. These contracts can either be relatively broad and loose or more specific and formal. Examples include annual purchase agreements between ministers and departments and the statement of corporate intent that regulates the relationship between the ministries and SOEs in New Zealand (Boston 1999a, 6). A second type of institutional contract regulates the relationship between providers and purchasers. These contracts can either be strictly inside the public sector or involve private actors (contracting out). This type of contract differs from the first one because of an explicit competitive element in getting the contracts. Potential problems for the political leadership include too few providers and high transaction costs, strategic behavior on the part of the providers, ambiguity concerning fulfilling the contracts, and even private monopolies (Boston et al. 1996, 23).

13 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 469 A third type of institutional contract is one with the users or customers, like the citizens or service charters used in the U.K. and New Zealand, the putting the customer first measures in the U.S., and the customer quality systems in Sweden (Aberbach and Rockman, 11; Boston et al. 1996, ; Gustafsson and Svensson, ; Rhodes, 45). These contracts try to state, often not very specifically, the rights of the users and the obligations of the service-delivering institutions, such as promises about the time taken to handle cases and applications. Individual contracts can be of two different types, or a combination of the two. One is more traditional, specifying the various terms of employment. The other, typically more connected to a competitive labor market and NPM reforms, is the short-term performance contract found in new leadership wage systems, used most extensively for leaders of public companies but also for administrative chief executives. This type of contract is characterized by relatively explicit goals and expected results, and should be individualized (Hood, ). One main reason for using such contracts is to increase efficiency through a differentiated system of rewards and punishments i.e., to make the manager manage (Kettl). The transparency criterion concerning the use of contracts rests on certain requirements that could be problematic to fulfill (Martin 1995, 40 41). One is the specification of unambiguous goals and results, leading to questions about whether contracts should include all types of public activities or only those where it is possible to quantify goals and results. And what are the eventual effects of trying to quantify traditionally qualitative activities? Another problematic requirement is that actors bound by contracts are assumed to be honest and give unbiased information about their activities. But how easy is it to prevent individuals from acting strategically out of self-interest, as presumed in economic thinking to gain rewards and avoid punishment, with the result that collective interests are pushed into the background? A third question is related to political responsibility. Are public contracts primarily vehicles for placing responsibility or accountability unambiguously on one party either political leadership, bureaucratic leaders, or leaders of commercial entities or do these contracts entail very complex responsibility conditions? For example, will political leaders experiences be that they bear the formal responsibility and take the blame while not having a commensurate ability to influence public decision processes (Brunsson)? Or do contracts lead to the burdens of using sanctions ex post, instead of placing responsibility firmly ex ante? Two additional questions relate to whether contracts will push economic values into the foreground and disregard other values such as rights and fairness, and whether the use of contracts could end in rigidity and misuse of resources, because more attention is paid to renegotiating contracts than to solving real problems and adjusting to changing conditions (Martin 1995, 41).

14 470 TOM CHRISTENSEN What, then, are the arguments for and against some of the main types of contracts? Arguments for contracting out for institutional contracts regulating the relationship between purchaser and provider, including private providers are primarily that they further efficiency, as shown in some comprehensive review studies (Domberger and Jensen; Domberger and Rimmer; Vickers and Yarrow; Winston). Also, combining competition and tight control supposedly enhances accountability. Arguments against doing so are that such contracts split up public organizations that are functioning well, thereby weakening public expertise and resulting in more resources used, and that they create more conflicts between actors. They also have a tendency to treat most public obligations as production goals, thereby leading to potential goal displacement and problems of measuring (Gregory 1995, 58). In addition, they are said to create private monopolies and cartels, leading to a private exploitation of public resources. Finally, trust between political leaders and administrative or institutional leaders may diminish because of opportunistic behavior (Boston 1995, 98). One major argument for using individual short-term performance contracts holds that they strengthen political control, because the recruitment may have been politicized and it is easier to get the administrative and institutional leaders to be loyal through the use of incentives, disregarding some traditional merit principles (Boston, Martin, Pallot, and Walsh 1991; Campbell and Halligan, 31). Another argument is that these contracts allow one to clearly specify the accountability of managers, follow up more easily on the results obtained, and eventually lay the blame on the administrative leaders. Arguments emphasizing negative effects of individual contracts cite evidence that political control is undermined and made more ambiguous. New types of administrative and corporate leaders will dominate the political sphere, leaders who are far less concerned with political signals, less loyal, and less committed to the public service ethos (Boston et al. 1996, 87; Boston 1999a, 8; Gregory 1995, 59 60; Gregory 1998a; Martin 1994). Responsibility and trust in a traditional and culturally integrated control system, based on appropriateness, is replaced by a more complex accountability system, based on self-interest and incentives (Christensen and Røvik; Day and Klein; March; March and Olsen; McLeay; Schick 1996). The contracts will often be administered and controlled by administrative leaders, not elected officials (Christensen and Lægreid 2001). Conflicts between institutions, groups, and individuals will probably increase. And the organizations for the civil servants will be substantially weakened, as individual considerations are emphasized (Harbridge, Walsh, and Crawford). The use of contracts is an integral part of the administrative reforms in New Zealand, often labeled government by contract (Schick 1998), but it is of far less significance for most civil servants in Norway (Christensen and Lægreid 1998b). John Martin (1995, 43 50) analyzes the use of contracts in New Zealand in two different cases and draws some more

15 ADMINISTRATIVE REFORM: CHANGING LEADERSHIP ROLES? 471 general conclusions. The first case involves a private health provider, and the second one a performance agreement and service charter between the Civil Aviation Authority and the ministry. Martin indicates that both cases involve advantages related to efficiency and the unambiguous definition of the services. However, he worries about accountability, and sees a tendency towards more secrecy than before. More broadly, he also questions the costs of reorganization for the sake of introducing contracts. Is it really altogether better, for example, to split up a traditional integrated system with hierarchical lines of command into a downsized ministry, a statutory body (CAA), a board, and a contract system? In a comparative analysis of leadership contracts in the U.K. and New Zealand, Christopher Hood ( ) notes that one reason for introducing leadership contracts in the two countries was to try to avoid political blame and divert more attention to the administrative leaders, an intention that one can debate as being a good one for increasing political control. Evidence of such effects exists, but contracts also tend to expose the politicians to more criticism. A third possible effect of contracts is that the accountability becomes more ambiguous, as in the sad case of Cave Creek, where many people were killed when a viewing platform collapsed (Boston 1999a, 11; Gregory 1998b). Contracts like these can also end up producing more rigidity, because it is hard to determine whether the leaders are fulfilling their contracts and unclear whether they are accountable for something outside the contracts. Lægreid (1999) compares New Zealand and Norway with regard to the salary development and individual performance contracts for chief executive administrative leaders. He underscores that the systems in the two countries, established in 1988 and 1990 respectively, differ in many respects, even though both have individual contracts specifying goals, evaluating performance, and using incentives. The leadership contract system in New Zealand started out going in a more radical direction than the Norwegian one and has since been extended, while the Norwegian one has become more limited over time. The system in New Zealand is more formal, more performance- and incentive-based, and more related to the private sector. It uses fixed-term contracts and affords more discretion and flexibility, while the Norwegian system encompasses fewer leadership positions, allows less scope for pay inequality, and employs leaders on a permanent basis. A look at the effects of the systems shows both similarities and differences between Norway and New Zealand. Both systems have produced a pay increase in real terms and increases in internal wage inequality. Both countries have problems attracting people from the private sector through the contracts systems, and the gap between public and private leaders salaries is growing in both countries. Both countries have problems specifying goals and measuring performance. While this is more accepted as a ritual in Norway, it is seen as problematic in New Zealand, where considerable energy is spent on renegotiating contracts. The new

16 472 TOM CHRISTENSEN contracts system in Norway seems to have changed little in the relationship between political and administrative leaders, a relationship based much on shared values, mutual trust, and responsibility (Christensen and Lægreid 1998a). By contrast, the effects in New Zealand are characterized by increased policy discretion for chief executives and more stress on managerial accountability, but also problems of ambiguity and lack of trust. Summing up, how does the use of contracts influence leadership roles and their relationship? The effects and implications seem to point in different directions, even in the same country. First and very much in line with NPM objectives contracts do in fact give the political leaders a potent control instrument, making it easier for them to fulfill political goals and priorities. Features that support this view include the general effects of success in specifying goals, policy instruments, and results through contracts. Features such as creating loyalty between political and administrative leaders and wage differentials as a result of contracts, which are claimed to raise efficiency, at least in New Zealand, also belong here. My main argument is that there are other, more typical effects of the use of contracts, which lead to an undermining of political control and increased influence of administrative and institutional leaders. The use of contracts, whether institutional or individual, gives administrative leaders a greater role, something that may benefit political leaders by relieving capacity problems, but which also potentially undermines their authority. One important aspect of this is that the traditional culturally based responsibility element is partly replaced by administrative and institutional leaders being held formally accountable to the political leaders (Gregory 1995). This seems to result in tension, lack of trust, ambiguity of accountability, the problems of many hands (Thompson), and reduced involvement of administrative leaders, rather than the opposite, idealized outcomes (Christensen and Lægreid 2001). In addition, it is possible that political leaders eventually see the use of contracts as a burden, because it requires them to spend more resources on various control instruments and thus creates capacity problems, leaving few gains. A survey in Norway (Christensen and Lægreid 1998a) has shown that leaders would rather participate in political processes with short-term perspectives and carry out some symbolic exercises now and then; they prefer not to use contracts extensively; and they would rather stick to the traditional methods of control. CONCLUSION In combination, the three elements of NPM discussed above increased horizontal and vertical specialization and more extensive use of contracts closely resemble what Olsen (1988, ) labels the supermarket state model. This contrasts with the traditional centralized state model,