ARENA Working Papers
WP 98/15

Causes of Bureaucratic Autonomy: The Impact of Organizational and Cultural Factors*

Morten Egeberg
Department of Political Science, University of Oslo

Introduction

This paper deals with bureaucratic autonomy and how this autonomy is related to certain institutional conditions. By `bureaucratic autonomy' is meant to what extent officials, in their actual behaviour, are constrained by the interests and actions of other actors, like the political leadership, interest groups, news media, and international organizations. To what degree are the recruitment and careers of officials encapsulated from political interventions? To what extent are bureaucrats, when developing policy proposals and implementing public policies, dependent on signals and demands from their political superiors, organized interests, and international organizations? Does formal structure, for instance the amount of `agencification', make a difference in this respect? To what extent are cultural norms legitimating bureaucratic autonomy, or, on the contrary, the primacy of political masters, the access of interest groups and the public to administrative decision arenas? Thus, our theme is one of enduring scholarly and practical interest, both in terms of the empirical questions that it addresses, and as regards the normative issues that it raises (Mill 1878; Wilson 1887; Weber 1970; Campbell 1988).

Due to historical circumstances, the `institutional borderline' between politics and administration has been drawn differently in Europe and the U.S.A. Broadly speaking, the bureaucracy was often well established in Europe when democracy and parliamentarism made their entrance (Raadschelders and Rutgers 1996). The situation was quite opposite in the U.S. where democracy preceded the development of a professional administration (Derlien 1996 ). Thus, in Europe, elected politicians had to conquer the top positions of the executive branch, i.e. the cabinet positions, until then often occupied by senior officials. However, politicians didn't really take over genuine administrative posts, although the political party affiliation of civil servants aspiring for top level positions is taken into consideration in many countries on the European continent (Page 1992). In the U.S., on the other hand, elected politicians and their appointees had to be pushed upwards in the administrative hierarchy in order to allow professional bureaucrats to hold senior administrative posts (Wilson 1887). In spite of this `Europeanization' effort, higher level jobs may still, to a considerable extent, be filled by people recruited on political grounds from outside the ranks of the civil service (Peters 1995). Thus, the `institutional borderline' between politics and administration is drawn through the administrative apparatus, probably implying that the career service in the U.S. is more constrained than its European counterparts.

However, also in the relatively young democracies in southern Europe (Greece, Portugal and Spain), an extensive politicization of the higher echelons of the administrative hierarchy has been observed (Auer et al. 1996). This may be interpreted as deliberate attempts at curtailing discretion within bureaucracies that, by and large, were inherited from the preceding authoritarian and repressive regimes due to bureaucracy's `functional indispensability'.

In this paper, it is focused on the amount of administrative autonomy in a small, parliamentary democracy (Norway), and how this autonomy is supposed to vary within one system, conditional upon organizational and cultural factors. The paper examines the degree of actual bureaucratic self-determination as far as recruitment and promotion decisions are concerned, as well as regarding policy-making and implementation in general. It tries to relate the amount of autonomy to internal organizational characteristics like vertical specialization, or `agencification' (Kickert and Beck Jørgensen 1995), to external organizational features like corporative arrangements and transnational administrative networks, and, finally, to cultural norms, like those emphasizing the desirability of professional integrity and neutrality.

I will proceed from here by discussing the theoretical relevance of institutional factors, like organizational structure and procedural norms, for understanding and explaining the actual decision behaviour of officials. Then I will present the sources of data on which the empirical part rests. I then go on to present organizational arrangements and cultural norms that are supposed to regulate the relationships between the Norwegian central administration and the political leadership, interest groups, news media and the international community. Finally, in the empirical part, it will be shown whether these normative structures manifest themselves in actual behaviour.

Institutional Factors and Behaviour: Some Theoretical Considerations

Institutionalists of all kinds seem to agree that actors' strategies, or the ways in which they want to achieve their goals, are affected by their organizational or institutional context. Opinions differ, however, concerning to what extent actors' preferences and interests are reformulated too (Rothstein 1996). Those adhering to what has been denoted `sociological institutionalism' have most consistently argued that preferences and interests can be endogenously shaped; i.e. furnished by the organizational or institutional structures in which actors are embedded (March and Olsen 1989). Given that individual persons are characterized by limited cognitive capacities and bounded rationality, institutions may be looked upon as structures that simplify decision situations that might otherwise have been too complex and incomprehensible (Simon 1965). This may be done by providing an administrative milieu that focuses a decision-maker's attention on certain problems and solutions, while others are legitimately excluded from consideration. The structure thus constrain choices, but, at the same time it creates and enhances action capacity in certain directions. Human rationality may thus be described as institutionalized, embedded or contextualized (Simon et al. 1950; March and Simon 1993; March and Olsen 1996). The private interests (and other `irrelevant' concerns) of bureaucrats are supposed to be `neutralized' by organizational rewards like pay and career prospects, by a logic of appropriateness, and by supervisors, and thus effectively separated from decision situations in which organizational members act in their capacity as officials (Lægreid and Olsen 1984; Scharpf 1986; March and Olsen 1989; Egeberg 1995).

It may be worth noticing that in the new institutional wave, and particularly in its rational choice variant, Selznick's original distinction between `institutions' and `organizations' has been left out. Thus, most rules are considered to be of an `institutional' character. According to Selznick, however, `institutions' are infused with value beyond the technical requirements of the task at hand (Selznick 1957:17). This could mean that people relate to `institutions' in a deeper, more personal and `internal' way than they do to `organizations'. Thus, the potential for resocializing individuals, in the sense of reshaping their interests and values without having to rely extensively on `external' control mechanisms, may be considerably greater in `institutions' than in `organizations'. It should be remembered that government organizations quite often may be characterized as highly institutionalized entities.

From a rational choice perspective, aspects of bureaucratic autonomy is not primarily related to organizational and cultural norms, but rather to what officials would find to their advantage. For instance, while institutionalists would attribute officials' neutrality in actual policy-making to administrative culture, or to organizational devices like `agencification', rational choice theorists would rather point to how the degree of alternation between the parties in power could affect the behaviour of civil servants in this respect (Hammond 1996). In systems which have seen frequent alternation of parties in and out of power, officials would find it to their advantage to adopt `a studied neutrality about policies and parties, and make much of their willingness to work with whatever party is currently in power' (Hammond 1996, 146). Under these circumstances, to retain any influence with future governments, officials have to keep incumbent governments at a certain distance. For systems with one party in power for a long period of time, on the other hand, the assumption is that officials will tend to adopt the policy views of this party. `Recruitment, socialization, promotion, and development of policy proposals will all have to be adapted to the demands of the dominant party' (Hammond 1996, 146-47).

Data and Method

The empirical material on which this study is based, stems mainly from three surveys conducted among Norwegian central government officials in 1976, 1986 and 1996. These surveys span a huge range of topics, of which only a few are touched upon on this occasion. At all three points in time, questionnaires were distributed to all officials in the ministries (core departments) engaged in administrative and advisory duties which require a university education (category A staff). Those who had occupied their positions for less than a year were excluded. In 1976 the response rate was 72 per cent, which gave us a total of 784 respondents (Lægreid and Olsen 1978). In 1986, 1185 (72%) responded (Egeberg 1989), and, finally, in 1996, 1482 officials filled in the questionnaires, giving us once more a response rate of 72 (Egeberg and Trondal 1997; Christensen and Lægreid 1998). In addition, in 1986 and 1996, category A staff personnel in the central government agencies subordinated to the ministries were included in two separate surveys. Because of the considerable size of these `directorates', we sampled randomly one-third of the relevant population. That gave us in 1986 1072 respondents (68%) (Egeberg 1989). In 1996, 64 per cent of the agency officials completed the forms (1025 persons) (Egeberg and Trondal 1997; Christensen and Lægreid 1998). The representativeness of the respondents was found to be satisfactory at both administrative levels. The survey data will be used to reveal the concerns and actors that the officials consider to be important in the actual decision making process. They will also be drawn on for describing the organizational linkages that exist to important parts of the bureaucracy's environments.

In addition, a material on the careers of Norwegian central government top civil servants will be applied. Those at work in 1936, 1956, 1976 and 1996 were all included in the study (538 persons). Based on the `State Directory', the career each official had had through the fifteen years preceding his/her appointment as a top civil servant was recorded (Nerland 1997). These data can show us to what extent the careers of senior officials have been interrupted by partisan assignments, like becoming members of parliament or junior ministers. Although we thus may get an indication of the role of party politics in recruitment and promotion processes, these data can not rule out the possibility that partisan concerns may have played a role even if the officials have stayed in the services all the time.

Organizational and Cultural Norms Regulating Bureaucratic Autonomy in Norway

Norms regulating the bureaucrat-politician relationship

Taking solely the formal structure of the central government as one's point of departure, few indices of bureaucratic autonomy can be derived. Ultimately, the cabinet decides, and, within each ministry, the minister has the last word, provided that his/her choices can not be deemed illegal. This follows from the principle of ministerial responsibility in a parliamentary system. In Norway, the minister's formal superiority remains unchallenged even as regards decisions on recruitment, promotion and dismissal of career officials. Provided that the applicants meet the formal criteria for being recruited to the higher civil service (e.g. a university degree), the minister is entitled to choose among the candidates. He or she is also given the right to dismiss civil servants in the ministries, provided that their tasks can be said to have disappeared.

However, among the factors that are supposed to constrain the exercise of ministerial discretion, are cultural norms embedded in the governance system. One of the basic norms asserts that positions in the ministries should be considered as permanent, provided that the incumbents behave appropriately, and want to stay. It follows from this that recruitment and promotion should be non-partisan in character since future political leaderships, potentially originating from a competing political party, have to inherit the existing personnel. Further, on these grounds, career officials are expected to behave neutrally in order to achieve acceptance from alternating `principals', but, simultaneously, to comply with directives and signals stemming from their current political superiors. To reconcile these contending role expectations, officials should be attributed a considerable amount of professional autonomy. They should be allowed to, and even encouraged to, elaborate on alternative courses of action, and advocate particular solutions on the basis of knowledge. This delicate balancing of loyalty, neutrality and professional autonomy is at the core of the British and Scandinavian civil servant role (Jacobsen 1960; Page 1992). Political instructions are to be implemented with energy, but without enthusiasm. One could assert that this role conception expresses a very well developed normative argument that takes on board some of the complexities that characterize the governance of modern societies. `Good governance' has to reconcile the needs for political agency, continuity, appropriate procedures, professional and scientifically based knowledge, and taking into account the concerns of minorities and especially affected parties. In this respect, the principal-agent perspective on the politician-bureaucrat relationship, considered as a normative theory, appears to be rather primitive and inadequately developed indeed.

There is, however, an organizational device too that may be of special importance as far as bureaucratic autonomy is concerned. Given our theoretical point of departure, officials in government agencies that are organizationally separated from cabinet level departments, can be expected to assign less weight to signals from the minister and his or her political advisors than their colleagues within cabinet level departments. When agency officials exercise discretion, professional and technical concerns may come to dominate (Egeberg 1994). Thus, even if ministers remain responsible for most decisions taken in the directorates, including the appointment of top level executives, `agencification' may underpin autonomy norms to a considerable degree. In Norway, there are at the moment approximately 60 agencies of this kind. Over time, numerous cabinet level departments have been transformed into separate directorates, but agencies have also been integrated, or reintegrated, into core ministries, illustrating the design potential of this organizational variable.

From an institutional perspective then, we would, in our case, expect recruitment and promotion decisions to happen without much political intervention in this century. From a rationalist point of view, the same prediction can be advanced as regards the periods before 1940 and after 1965, during which frequent alternation of parties in and out of power took place. As regards the in between years 1945-65, however, we should, from a rationalist perspective, expect more politicization. This period, during which the Labour Party remained in power, has, by historians, been characterized as the time of the `single party state' (Seip 1963). Regarding the policy process in general, we should, from an institutional perspective, expect that the officials strive to balance and reconcile the loyalty and professional integrity concerns. A somewhat similar prediction could be made from a rationalist point of view as regards the period for which data are available on this topic (1986-96). `Professional integrity' should, however, here be replaced by `neutrality'.

Norms regulating bureaucracy-interest group relationships

The actual amount of bureaucratic autonomy may also depend on the degree to which interest groups are accepted as legitimate participants in the policy process, and, in case, to what extent this is formalized by corporative arrangements linked to the administrative apparatus. At least in this century, societal groups that are seen as especially affected by certain public policies, have been invited to give their opinions on relevant policy proposals. As far as the preparation of secondary (complementary) legislation is concerned, cultural norms legitimating group involvement (`hearing') have been underpinned by administrative law (1967). One interpretation of this codification is that the bureaucracy's large amount of discretion as regards the drafting of supplementary legislation should be partly circumscribed by interest group participation (Christensen and Egeberg 1979). Although these formal rules have not been questioned, non-socialist governments in the 1980s, influenced by New Public Management ideas and neo-liberalism, launched reform programmes that were highly critical about the role of interest organizations in public policy-making (Olsen 1991; Lægreid and Roness 1997). The organizations were blamed for their contribution to the growth of public expenditures. It was further argued that individual citizens and firms should be `re-empowered' at the expense of their organizations. The government's attack on corporatism was followed up in practice by starting downsizing the collegial system of governance. This committee system consists of study groups preparing new policies and legislation (temporary committees), advisory groups, and boards at the top of agencies (permanent committees). Over time, interest groups have been represented on approximately half of these collegial bodies. In the 80s and 90s, they were, however, represented more often on committees that were dissolved than on committees that were established (Nordby 1995; 1996). Table 1 shows the development of the committee system with publicly appointed members under the Norwegian central government.

Table 1

The number of committees under the Norwegian central government at different points in time.

Source: Nordby 1996, 289.

Table 1 reveals that the number of temporary as well as permanent groups reached a peak point in the 70s, after which it started to decline considerably. (Scholarly too, the 70s represented `the age of corporatism', cf. Schmitter and Lembruch 1979.)

Table 2 shows that the functions taken care of by the `official' committee system have

become increasingly `internalized' into the administration itself. Between 1976 and 1996, the proportion of ministry officials who have attended meetings in public committees during a year have decreased significantly (even in absolute numbers). In the same period, there has been a steady growth as regards their participation on `internal' groups, i.e. intra- or inter-ministerial committees consisting of officials only.

Table 2

Proportions of the officials who have participated on committees during a year, at different points in time. Percentages. (Ministry surveys.)

Thus, our hypothesis is that, due to political, cultural and organizational reasons, the constraints that corporatism imposes on actual bureaucratic decision-making have decreased from the 70s on. Although it is here focused on to what extent corporative arrangements might circumscribe the administration's autonomy, it should also be pointed out that interest group involvement may benefit bureaucratic actors by providing information, predictability, and political support.

In addition to this `external' type of corporatism, an `internal' variant may as well contribute to curtailing the room for bureaucratic manoeuvring. In the 70s, concerns were mounting as regards the role of work place democracy in the services. The struggle resulted in an agreement (1980) between the government and the unions of the civil servants guaranteeing the unions rights of participation and co-determination in reorganization processes (Roness 1993). The unions had in the 70s achieved some co-determination regarding recruitment of personnel beneath the level of head of division, and, in 1958, bargaining rights as far as wage determination is concerned. Thus, as regards decision-making related to personnel and organizational matters, we would expect that `internal' interest groups are assigned considerable weight in the decision process. It might have happened, however, that the political initiative taken in the 80s in order to limit the access of external groups, could have contributed to a growing scepticism towards group involvement in general.

Norms regulating the bureaucracy-media relationship

Bureaucratic transparency became a topic high on the agenda in the 60s. In order to enable the news media and the public to monitor and oversee administrative decision-making to a greater extent, a new law was passed in 1970. In the empirical part we will touch upon the role of the news media as seen by the officials, at different points in time. In parts of the literature, modern mass media have been attributed an almost devastating impact on reasoned will formation in society, and on ministers' ability to give direction to the administrative branch. Having a huge influence on the agenda and, thus, on politicians' allocation of time and energy, news media are seen as threatening the appropriate functioning of representative democracy itself (cf. for instance Fox and Miller 1995).

Norms regulating the bureaucracy-international actors relationship

The amount of actual administrative independence may also depend on the degree to which national executives are organizationally linked to international organizations and governments of other countries. Norway was a member of 17 international governmental organizations in 1910, 58 in 1950, and about a hundred today. Table 3 shows the proportions of central governments officials who, at three points in time, are involved in international affairs. Since it is rather obvious that institutions like the Foreign Ministry is heavily engaged, we focus here on the 'domestic' services.

Table 3

Proportions of the officials in `domestic services'* who work (`to some extent, or more') on international matters. Percentages. (Ministry and agency surveys.)

*: Ministries of Foreign Affairs and Defence, and the Directorate of Development Aid are excluded.

The numbers illustrate clearly the increasing internationalization of the national bureaucracy. The European Economic Area (EEA) agreement (1994), making Norway a part of `the internal market' of the European Union, represents a significant step further in this direction (Egeberg and Trondal 1997; Sverdrup 1998). According to this agreement, Norway is expected to adopt and transpose most EU legislation related to the economy. In practice, all ministries are involved to some extent. Almost a half (45%) of the officials in the `domestic' services are affected `to some extent, or more'. All of these officials will certainly not consider themselves directly engaged in international affairs (cf. table 3). They report, however, that their issue areas are affected by the agreement. On this background, we will expect international actors to be considered important in actual decision-making processes in the national bureaucracy, and, increasingly so. This assumption may be particularly pertinent to the legislative process.

Although it is here focused on the possible constraints that international organizational linkages may impose on bureaucratic autonomy, it should be made clear that these relationships, like those between the administration and various interest groups, could represent resources as well. It has, for instance, been argued that the high level of bureaucratic interaction within the EU has strengthened the roles of the national executives at the expense of national parliaments (Wessels and Rometsch 1996).

Bureaucratic Autonomy in Actual Decision Making

We will start by looking at to what extent administrative career patterns can be seen as encapsulated from political processes. One indicator could be to what degree political and bureaucratic careers are clearly separated. Table 4 reveals, at different points in time, the proportions of top civil servants who have occupied central political positions prior to their appointments as top officials.

We find the largest proportion with a political background among the top civil servants in 1976 (11%). At the level of director general in the ministries, very few people have had a political career previously (2-8%). In general, leaders of agencies more often have political experiences than their counterparts in the ministries. However, in 1996, as many as 44 per cent of the secretaries general had occupied a political position prior to their appointments. Because of the low number (16 persons), we should, however, not `overdramatize' this proportion.

Table 4

Proportions of top officials who have been ministers, under secretaries of state, political advisors for the minister, or parliamentarians at different points in time. Percentages.

In order to assess the role of politics here, we have, in addition, to draw on two important observations in Nerland's (1997) study. First, he found that, in 1996, almost all of those with a political background combined this involvement with a bureaucratic career. Over time, those who have occupied political positions have become increasingly similar to `main stream' bureaucrats, both as regards career patterns and educational backgrounds. Thus, it is hard to say what difference political affiliations really make for persons' advancements. Second, he discovered that almost a half (9 out of 19) of those with a political past (in 1996) had been appointed by a government originating from a political party other than their own. Thus, if political experience has been assigned any weight when candidates have been selected for administrative leadership positions, it seems as if such experience is considered valuable independent of any particular partisan affiliation.

As pointed out earlier in this paper, these data can not rule out the possibility that partisan concerns may have played a role even if the officials have stayed in the services all the time. However, Lægreid (1989), in a study based on a survey among decision-makers involved in recruitment processes, found few indications of partisan concerns. Neither during the period 1945-65, in which the Labour Party was dominant (`the single party state'), are there clear signs of that the application of merit criteria has been relaxed significantly (cf. table 4). This is confirmed in an in-depth study of the Ministry of Finance, the main pillar of social democratic planning philosophy and practice in the post-war era. Between 1945 and 1965 only a couple of appointments were disputed on political grounds (Lie 1995).

We now turn to decision making in general, and focus on how bureaucrats manage to cope with highly competing role expectations in practice. Table 5 reveals the proportions of the officials who attach importance to signals from the minister, stress professional or `technical' concerns, and pay heed to signals from users and clientele when they exercise discretion. In the ministries, we find a quite stable juxtaposing of political and professional concerns. This very strong emphasis on professional concerns, and, though more moderate, sensitivity to user interests, probably express the existence of a considerable amount of bureaucratic autonomy in relation to the

Table 5

Proportions of the officials who see various signals and concerns as important when exercising discretion. Percentages. (Ministry and agency surveys.)

*: Incl. only those who exercise discretion `to some extent, or more'.

political level (Christensen 1991). However, politics is a reality for career officials working in ministries, and, probably more so in systems without political appointees in the administrative ranks. Thus, Aberbach and Rockman (1997) show that US federal career executives tend to focus more exclusively on the legal and technical aspects of their role than their counterparts in Norway. We see, however, from our data, that `agencification', or the organizational separation of directorates from cabinet level departments, seems to enhance the amount of administrative independence from the political leadership of the ministry. This is part of a general pattern observed across countries (Egeberg 1994). On the other hand, our data show that personnel at this level tend to be more sensitive to user concerns. Other studies, based on interviews with Norwegian ministers from different parties, confirm this picture of the officials' role behaviour. It was quite exceptional that the loyalty and the professional integrity of the services were questioned (Eriksen 1988). The shrinking sensitivity to user concerns was not expected, given the political and cultural re-orientation in the 80s. One explanation for this could be that consumer and clientele interests might be more difficult to detect subsequent to the down-sizing of the corporative structures. In a non-market situation, consumer preferences have to be organized to get heard , and this fact may have been underestimated by the government.

Table 6 reveals the officials' assessment of the importance of other actors when decisions are made within their own issue area.

Table 6

Proportions of the officials who consider other actors (listed beneath) to be important when decisions are made within their own issue area. Percentages. (Ministry surveys.)

Consistent with what we have seen in table 5, the relationship to the political leadership, here, the cabinet, seems to be one of the stable elements. As expected, the proportion who consider interest groups to be important when decisions are made within their own issue area has dropped considerably, from 57 per cent in 1976 to 35 per cent in 1996. Within this broad `interest group category', civil servants' unions as well as `external' economic interest organizations follow this main pattern. A quite similar tendency can be observed in the agency survey data. However, while the bureaucrats thus seem to have become less dependent on interest groups, they have become significantly more dependent on international actors. If we leave out those in the Foreign Ministry, the proportion who consider international actors to be important has risen from 11 per cent in 1976 to 30 per cent in 1996. The decreasing role of corporatism doesn't seem to be directly linked to the internationalization wave: the officials' degree of involvement in international affairs is not negatively correlated with to what extent they consider interest groups to be important actors (Gamma=.08) (Egeberg and Trondal 1997). Finally, the news media are seen as only slightly more pivotal in the policy process over time. They don't seem to have reached the level of importance that parts of the literature have attributed to them.

Table 7 shows how the officials appraise the role of various actors and concerns, conditional upon the kind of functional area they are in. First, we observe that the officials obviously strive to balance the political agency and professional integrity concerns in most fields. There are a couple of understandable exceptions to this (result reporting, evaluation, and information services). It is not that obvious that so much emphasis is given to political signals when laws and regulations are applied on concrete cases. Equal treatment of similar cases could be said to preclude routine intervention from the political level.

Table 7

Proportions of the officials who consider the following actors, signals and concerns to be important, conditional upon their main functional area. Percentages. (Ministry survey 1996.)

Functional area:

Second, as expected, the drafting of new legislation is one of the areas in which external interest organizations are most involved. Third, we see that, though the unions of the civil servants are not that central in general, a relatively large proportion (63 per cent) consider them to be important when decisions are made on personnel and organizational issues. Fourth, also as expected, those working in the legislative area have to take into account the concerns and decisions of international actors to a greater extent than most other officials (49 per cent). In fact, we observe that the international environments, in addition to the political leadership, seem to impose the most salient constraints on bureaucratic action as regards the legislative process. In comparison, the budgetary process appears to unfold on a much more closed arena.

Conclusion

From a principal-agent perspective, the bureaucracy should only be an instrument at the principal's disposal. It could be argued, however, that this represents a highly underdeveloped normative position that doesn't take on board a wide range of concerns that are pertinent to what is perceived as `good governance'. A more elaborated argument would emphasize the administration's responsibility for bringing knowledge and expertise into the policy process, and for ensuring that appropriate procedures and the concerns of especially affected parties are adequately paid attention to. This paper is, however, an empirical-theoretical one: it aims at describing normative structures (organizational and cultural) that may impact on the room for bureaucratic manoeuvring, and, at showing to what extent these hypothesized relationships materialize in the data.

As far as recruitment and promotion decisions are concerned, cultural norms seem to circumscribe relatively effectively the formal superiority of the minister. This was the case also under the `single party state' 1945-65. From a rationalist point of view, one could have expected more politicization during this period. Regarding decision making in general, the Norwegian bureaucracy is clearly subject to political guidance, and more so in the ministries (core departments) than in the agencies (directorates). Due to this organizational separation, or `agencification', political signals seem to have more difficulty in penetrating this level of the administrative branch. In the ministries, while obviously under political direction, cultural norms may help to explain the very great emphasis that the officials give to professional and `technical' concerns, and also to user interests, when exercising discretion.

Corporatism came under pressure from the 80s on. Institutional arrangements were partly dismantled, and policy functions that had been taken care of by this system, like the drafting of new legislation and public policies in the economic area, became gradually `internalized' into the administration itself. This is reflected in the actual decision behaviour of the officials: they have increased their autonomy vis-a-vis interest groups of various kinds. During the same period, however, new international agreements and administrative linkages to international actors have created new dependencies that are reflected in actual policy making, particularly as regards the legislative process.

References

Aberbach, J.D. and B.A. Rockman. 1997. Back to the Future? Senior Federal Executives in the United States. Governance 10:323-349.

Auer, A., C. Demmke and R. Polet. 1996. Civil Services in the Europe of Fifteen: Current Situation and Prospects. Maastricht: European Institute of Public Administration.

Campbell, C. 1988. Review Article: The Political Roles of Senior Government Officials in Advanced Democracies. British Journal of Political Science 18:243-272.

Christensen, T. 1991. Bureaucratic Roles: Political Loyalty and Professional Autonomy. Scandinavian Political Studies 14:303-320.

__________. and M. Egeberg. 1979. Organized Group-Government Relations in Norway: On the Structured Selection of Participants, Problems, Solutions, and Choice Opportunities. Scandinavian Political Studies 2:239-259.

__________. and P. Lægreid. 1998. Den Moderne Forvaltning. Oslo: Tano Aschehoug.

Derlien, H.-U. 1996. The Politicization of Bureaucracy in Historical and Comparative Perspective. In eds. B.G. Peters and B.A. Rockman. Agenda for Excellence 2. Administering the State. Chatham: Chatham House.

Egeberg, M. 1989, ed. Institusjonspolitikk og Forvaltningsutvikling. Oslo: Tano Aschehoug.

_________. 1994. Bridging the Gap Between Theory and Practice: the Case of Administrative Policy. Governance 7:83-98.

_________. 1995. Bureaucrats as Public Policy-Makers and Their Self-Interests. Journal of Theoretical Politics 7:157-167.

________. and J. Trondal. 1997. Innenriksforvaltningens og den Offentlige Politikkens Internasjonalisering. In eds. T. Christensen and M. Egeberg. Forvaltningskunnskap. Oslo: Tano Aschehoug.

Eriksen, S. 1988. Herskap og Tjenere. Oslo: Tano Aschehoug.

Fox, C.J and H.T. Miller. 1995. Postmodern Public Administration. Toward Discourse. Thousand Oaks: Sage.

Hammond, T.H. 1996. Formal Theory and the Institutions of Governance. Governance 9:107-185.

Jacobsen. K.D. 1960. Lojalitet, Nøytralitet og Faglig Uavhengighet i Sentraladministrasjonen. Tidsskrift for Samfunnsforskning 1:231-248.

Kickert, W.J.M. and T. Beck Jørgensen. 1995. Introduction: Managerial Reform Trends in Western Europe. International Review of Administrative Sciences 61:499-510.

Lægreid, P. 1989. Rekrutteringspolitikk i Sentraladministrasjonen. Working Paper. Bergen: Norwegian Research Centre in Organization and Management.

-----------. and J.P. Olsen. 1978. Byråkrati og Beslutninger. Bergen: Scandinavian University Press.

-----------. and ----------. 1984. Top Civil Servants in Norway: Key Players - on Different Teams? In ed. E.N. Suleiman. Bureaucrats and Policy-Making. New York: Holmes & Meier.

-----------. and P.G. Roness. 1997. Political Parties, Bureaucracies and Corporatism. In eds. K. Strom and L. Svåsand. Challenges to Political Parties: the Case of Norway.

Michigan: The University of Michigan Press.

Lie, E. 1995. Ambisjon og Tradisjon. Finansdepartementet 1945-1965. Oslo: Scandinavian University Press.

March, J.G. and J.P. Olsen. 1989. Rediscovering Institutions. The Organizational Basis of Politics. New York: The Free Press.

------------. and ----------. 1996. Institutional Perspectives on Political Institutions. Governance 9:247-264.

------------. and H.A. Simon. 1993. Organizations. Oxford: Blackwell (1^st ed. 1958: Wiley.

Mill, J.S. 1878. Considerations on Representative Government. London: Longmans.

Nerland,S. 1997. Toppledere i Sentralforvaltningen med Politisk Bakgrunn. En Studie av Toppledernes Karrieremønster med Utgangspunkt i Årene 1936, 1956, 1976 og 1996. Thesis. Oslo: Department of Political Science.

Nordby, T. 1995. Norwegian Corporatism - With Emphasis on Recent Years' Development. In eds. T. Nordby, H. Lorentzen and M. Micheletti. Who Governs Whom? State and Organizations in Modern Societies. Oslo: Institute for Social Research.

-----------. 1996. Hvem Styrer Hvem? Statlige Utvalg som Maktorganer. Nytt Norsk Tidsskrift 13:281-294.

Olsen, J.P. 1991. Modernization Programs in Perspective: Institutional Analysis of Organizational Change. Governance 4:125-149.

Page, E.C. 1992. Political Authority and Bureaucratic Power. A Comparative Analysis. New York: Harvester - Wheatsheaf.

Peters, B.G. 1995. The Politics of Bureaucracy. New York: Longman.

Raadschelders, J.C.N. and M.R. Rutgers. 1996. The Evolution of Civil Service Systems. In eds. H.A.G.M. Bekke, J.L. Perry and T.A.J. Toonen. Civil Service Systems in Comparative Perspective. Bloomington: Indiana University Press.

Roness, P.G. Tenestemannsorganisasjonane si Rolle i Omforminga av Offentleg Sektor. Nordisk Administrativt Tidsskrift 74:264-278.

Rothstein, B. 1996. Political Institutions: An Overview. In eds. R.E. Goodin and H.-D. Klingemann. A New Handbook of Political Science. Oxford: Oxford University Press.

Scharpf, F. 1986. Policy Failure and Institutional Reform: Why Should Form Follow Function? International Social Science Journal 37:179-189.

Schmitter, P.C. and G. Lehmbruch. 1979, eds. Trends Toward Corporatist Intermediation. Beverly Hills: Sage.

Seip, J.A. 1963. Fra Embedsmannsstat til Ettpartistat og Andre Essays. Oslo: Scandinavian University Press.

Selznick, P. 1957. Leadership in Administration. A Sociological Interpretation. New York: Harper & Row.

Simon, H.A. 1965. Administrative Behavior. New York: The Free Press.

------------., V.A. Thompson and D.W. Smithburg. 1950. Public Administration. New York: Alfred A. Knopf.

Sverdrup, U. 1998. Norway: An Adaptive Non Member. In eds. K. Hanf and B. Soetendorp. Small States in Europe. London: Longman.

Weber, M. 1970. Bureaucracy. In eds. H.H. Gerth and C.W. Mills. From Max Weber. London: Routledge & Kegan Paul.

Wessels, W. and D. Rometsch. 1996. Conclusion: European Union and National Institutions. In eds. D. Rometsch and W. Wessels. The European Union and Member States. Towards Institutional Fusion? Manchester: Manchester University Press.

Wilson, W. 1887. The Study of Administration. Political Science Quarterly 2:197-222.

Footnotes

* Prepared for delivery at the Annual Meeting of the American Political Science Association, Boston Marriott Copley Place and Sheraton Boston Hotel and Towers, September 3-6, 1998. Copyright by the American Political Science Association.

[Date of publication in the ARENA Working Paper series: 15.08.1998]