Resumé

In this article, I attempt to explain the different roles of Swedish and Danish citizens when they encounter the welfare state. The case chosen is primary education, the area with the greatest variation. My thesis is that the differences are primarily explained by the different institutional legacies in the two countries, rather than the often-suggested alternatives, i.e., the different strength of the labor movements and the importance of Grundtvigian ideas in Denmark. I argue that in the first decades of the 20th century, Sweden and Denmark established two very different systems of school administration. Once established, these have shown a high, though by no means total, stability and the development has been path dependent. Due to the different institutional legacies, the same political ideas have often resulted in different decisions and measures in the two countries.

The Problem

Most observers think the Scandinavian welfare states are based on the same model. It is indeed true that the similarities between them are considerable, especially where the social insurance system is concerned. Social insurance has traditionally been the focus of welfare state research, even though the European concept of the "welfare state" also embraces education, health care, child care and elder care. In part, this reflects the fact that European researchers have been influenced by the connotations of the American concept of "welfare," so that the focus of research has been restricted to poor relief and its successor, social insurance. It also reflects the fact that many welfare state researchers have proceeded on Marxist assumptions, and therefore have been particularly interested in the effects of the welfare state on the labor market, i.e., on power relations between labor and capital (cf. Esping-Andersen 1990).

Feminist research on the welfare state has widened the scope of research in
this area. It has focused on the welfare state's care side, but since researchers

---

* Anders Lindbom, Department of Government. Uppsala University. Box 514. S-751-20 Uppsala, Sweden.

in this vein take a special interest in relations between the sexes, they have studied the welfare state primarily from the standpoint of its significance for women's prospects on the labor market (cf. Sainesbury 1994; Lewis 1993). This feminist perspective also gives the impression that the Scandinavian welfare states are based on one single model.

1 would like to go a step further. The reproductive sectors of the welfare state are important, not just because they offer women the opportunity to participate actively in the labor market, but also because they supply services that citizens consider extremely important. The Swedish Social Democrats, for example, have declared that such services, on account of their great importance, must be given priority over social insurance during this time of cutbacks.

The manner in which the production of these services is organized reflects different ways of looking at the role of citizens in their encounters with the welfare state. In this regard, the Danish and Swedish welfare states differ sharply from each other, despite the fact that policy makers in both countries have entertained far-reaching egalitarian ambitions. In Sweden, the role of citizens has been to passively accept, as clients, the centrally directed services offered. In Denmark, citizens have long enjoyed, as users and/or consumers, a considerable freedom of choice among producers; they have also been granted the right - in such crucial welfare areas as education, child care and (in part) elder care - to elect representatives to serve on user boards with considerable influence on operations. To simplify the matter, we could describe the Swedish welfare state as paternalistic and the Danish one as popular {or folkelig to use a favorite Danish expression).

The purpose of this article is to explain this difference in the area where it is most significant, namely primary education. The formal differences between Denmark and Sweden in this area have substantial effects on how much influence citizens feel they have on their own situation. 70 percent of Danish parents say that their opportunities to exert influence are good or very good, compared to only 4 percent of Swedish parents.¹ This is of particular interest, since no other welfare program is as far-reaching as primary education: It affects all citizens, five days a week for nine years. It is also ambitious when it comes to influencing individual behavior. Education is not just a service offered to the citizens, it is also an attempt to form them.

It has been claimed that Sv/eden and Denmark have developed different concepts of education. In Sweden, the school is seen as an extension of the state, while in Denmark it is seen as a state-supported extension of the home and local community (Lauglo 1980). The starting point for Swedish educational policy was the Utopian ideal of social engineering: a new, more rational, more enlightened and more engaged citizen would be created through the fostering of children by public expertise based in science (Hirdman 1989). The Danish view of the school's role in this regard was

cogently formulated by a one-time minister of education: "It is not the State which shall commission the People, but the People who shall commission the State" (Borgbjerg 1930, 1323). The former approach has a clearly paternalist basis; the latter is open to lay influence.

These differing views on school management were plainly manifested in contrasting decisions in 1989/90. The Danish parliament decided to grant school boards (skolebestyrelser) on which parent representatives hold a majority, enhanced rights to make decisions about school management. At the same time, any such institutionalized influence on the part of users was explicitly rejected by a broad majority of the Swedish parliament, which in fact abolished the obligation of the principal to inform and consult with parents. What can explain this variation in parent influence on school management?

Possible Explanations

In my forty or so interviews with politicians, officials and scholars, there are two recurring explanations that focus on specific differences between the countries in order to explain the divergent decisions. The first is based on the varying strength of the labor movement in the two countries, the second on the importance of Grundtvigianism in Denmark (cf. Telhaug & Tønnesen 1992).

The foremost claim of theories that focus on power resources, or more specifically on the strength of the labor movement, is to be able to explain the extent and character of the welfare state. Countries in which the working class is highly mobilized, and in which social democratic parties have held office for long periods have a relatively even distribution of income, a relatively low percentage of poor persons and a relatively low level of unemployment. They are also known for applying an "institutional" model of social policy (Korpi 1981; Esping-Andersen 1985).

Authors in this tradition have chosen to look almost exclusively at the social insurance system. It is well known, however, that the manner in which the welfare state is organized has political repercussions within the service producing sectors of the welfare state, i.e., education, child care, etc. Since social democracy has been stronger in Sweden than in Denmark, one hypothesis is that Swedish educational administration reflects the interests of the labor movement more strongly than Danish educational administration does. Since parent influence is often thought to favor the well-educated middle class, the interests of workers would be better served by a system of detailed central control. A centrally controlled school system would thus be better suited to achieving social democratic goals of equality than a more decentralized and parent-directed system like the Danish one.

Another common explanation for variations in welfare state organization focuses on differences in political culture (King 1973). Some scholars argue that the Danish and Swedish welfare states are founded on different ideas. According to this view, the impact of Grundtvigianism has been such that the Danish welfare state is based on a more liberal intellectual tradition than the Swedish (Østergård 1990). Two Danish former ministers of education have referred to Grundtvigian ideas to explain why Denmark's educational system differs in certain respects from those in the other Nordic countries (Andersen 1983; Haarder 1983).

The founder of Grundtvigianism was N.F.S. Grundtvig (1783-1872). This clergyman, hymn writer, parliamentarian, and school and church reformer occupies a central place in modern Danish history. Grundtvig was a member of the National Constitutional Assembly, and later he worked in parliament for freedom in educational and religious matters. Grundtvig took the view that the Word of God could be apprehended directly by lay people, and his ideas eventually suffused the Danish state church. He also considered it crucial that power over education be decentralized, so that it could not be used to promote any particular ideology or faith. Parent influence through exit and protest is therefore important in Denmark.

A discursive explanation proceeds from the assumption that political decision making takes place within a framework of prevailing political ideas and conceptions (a discourse), and that this discourse is decisive for what is regarded as a political problem, who can take part in the discussion on such problems, and what arguments are viewed as valid in the discussion (Hall 1989; Lægreid & Pedersen 1994). From such a perspective, the different ways of seeing the relation between users and welfare state in Sweden and in Denmark could be explained by the fact that Sweden lacks Denmark's Grundtvigian intellectual heritage.

It is my thesis that the above explanations are, if not mistaken, then at least insufficient. In my view, it is the institutional legacy of an administration that determines how general administrative notions are converted into measures in different countries. The alternative explanations are insufficient because "[njeither interests nor values have substantive meaning if abstracted from the institutional context in which humans define them" (Steinmo 1989, 502).

My thesis is inspired by "new institutionalism," which has been a vital research approach in social science since March & Olsen's "classic" article from 1984 (March & Olsen 1984). I share these authors' conviction that "the organization of political life makes a difference" (March & Olsen 1989, 1). However, I am critical of some of their theoretical assumptions, which risk leading them merely to postulate (rather than investigate) the influence of institutions on actors' decisions. Institutionalist theory further claims that institutions are characterized by considerable inertia. Olsen and others argue that this reflects how institutions determine how actors think. I would argue,

however, that an incremental course of development can have other causes
than the cognitive incapacity of the actors. Political actors are regarded here
as potentially autonomous and rational.

For one thing, political decision making is not primarily an inner mental condition on the part of an actor, but an interactive process among different actors whose interests are partly opposed to each other. "Because it usually is impossible to win agreement on large changes, restricting analysis to 'incremental' policy proposals that may be politically feasible is a way of conserving on scarce time and energy" (Lindblom & Woodhouse 1993, 27). Thus, it is often politically rational for actors to seek to reduce the severity of political conflict by treating the decision making process as a continuous learning process in which small changes are made in order to handle small problems. This can be more rational than constantly acting as if all decisions were of a "constitutional" character.

In addition, reality is such that new institutional elements cannot be introduced into existing institutional configurations any which way. It is certainly true that existing institutions are seldom entirely uniform, since they represent a kind of historical layering. But a certain coherence is necessary - otherwise chaos would prevail. One conceivable possibility would be for actors to carry out a series of follow-up changes within other parts of the institutional configuration; in that case, however, the risk of conflict would dramatically increase, and unforeseen effects could appear in other political areas.

This analytical framework also differs from that of March & Olsen in using a narrow concept of institutions to avoid classifying "everything" as institutions (leaving nothing to be investigated). The concept of institutions refers to the organization of formal political units; thus political institutions are regarded as consciously created. The analysis focuses on the interaction between actors/ideas and institutions. Since institutions defined in this manner are sufficient to answer my question, there is in my case no need to cast an unnecessarily broad conceptual net. The argument is more convincing if the independent variable is precisely defined.

To summarize the three explanations and clearly highlight what distinguishes them from each other, we can focus on how they portray the appearance and significance of preferences. The strength-of-labor thesis emphasizes the common interests of labor movements in all countries, which in turn conflict with the preferences of the political right. The varying relations of political strength thus explain the fact that different decisions are taken in different countries. The Grundtvigian explanation, on the other hand, claims that consensus on preferences prevails within each country, but vary between them. Different preferences thus directly explain the fact that different decisions are made in different nations. According to the institutional legacy thesis, it is not certain that preferences have to be decisive.

Divergent institutional legacies can mean that the same political ideas are
expressed in different decisions in different states.

The Genealogy of Educational Administration in Denmark and Sweden

The clerical origins of the educational system have left traces in both countries, and until the breakthrough of parliamentarism, the educational administration systems in the two countries were structured similarly (Ussing Olsen 1982; Tegborg 1969). But the process of democratization took diverging courses in the two countries, a fact that proved important for which administrative structures were then established. In Sweden, the farmers had been represented in the four-estate parliament for centuries. The Danish farmers, on the other hand, were forced to carry out a long struggle for their political rights, which created a widespread suspicion of the state. While the Swedish farmers turned to the state for protective tariffs against foreign competition, their Danish counterparts were able to compete on the world market and supported free trade. The experiences of Danish farmers with the state were thus negative and unlike the Swedish farmers, they did not depend on it. The result of these differences in the democratization process was that the popular movements in the Danish countryside were substantially more critical of the governing elite and of the state as such than their Swedish counterparts (Knudsen & Rothstein 1994). After the breakthrough of parliamentarism, the Swedish liberals accordingly undertook a heavy centralization of educational administration, an unthinkable policy for their Danish sister party.

The different administrative structures chosen early in the century (in Sweden in 1913, in Denmark in 1933) have since been decisive - even up to the close of the twentieth century - for the manner in which educational administration has been organized in the two countries. This does not mean that the administrative institutions in question are the same today as they were 60-70 years ago, but the many administrative policy decisions and substantial changes have not altered the fundamental features of the administrative structure (especially not those separating the two countries). As I will show, the changes in educational administration have had a stable direction in each country, and they have followed two different tracks. This dual character is well captured by North's concept of path dependency (North 1990), an approach that facilitates comparative historical analysis.

Path dependency gives rise to effects of the type seen in the story of the English tourist in Ireland, who asked the way to Tipperary and got the answer: "I wouldn't start from here." In other words, certain political "destinations" are unlikely given a certain institutional starting point. Therefore, the

different administrative institutions established early in the twentieth century strongly condition the solutions actors formulate to the administrative policy problems of today. My thesis is that the decisions in Denmark and Sweden in 1989/90 were made on the background of two different institutional legacies, which explains why the same administrative policy ideas have resulted in different, even opposite, measures in the two countries.

Hypothesis Testing The Institutional History

The incremental development of the administrative history of education
in Denmark and Sweden can be summarized in the following fashion (cf.
Lindbom 1995).

Denmark

The development of Danish school commissions into present-day school boards (skolebestyrelser) was marked by continuity, and modifications were made within the existing system of rules. To a large degree, the school commissions (skolekommissioner) of 1814 represented the formalization of prevailing practices with roots in the Reformation. It was decreed in 1867 that two commission members could be recruited from outside the local council; in 1899, it was stipulated that such members must have children of school age. Parent representatives were appointed by local councilors. A law passed in 1933 permitted the establishment of a new body, known as a parent council (forældreråd), which would work at the level of the individual school, rather than at the municipal level, as the school commissions did. Its members were chosen by the parents at each school. Parent councils had been proposed twenty years earlier by a royal commission, and had been tested successfully in Southern Jutland.³ Parent councils were given no real powers or tasks, nor were they made mandatory. They were transformed into school boards (skolenævn) in 1949 and given greater decision powers. In districts without school boards, the parent interest was represented in the school commission, with parent representatives comprising nearly half its members.

In 1961, a proposal for mandatory school boards was rejected by virtually all political parties. Eight years later, however, they all agreed that such a reform ought to be carried out. The reason for the switch was municipal amalgamations which were part of a major local government reform that came into effect at the same time as the school boards. There was widespread fear that the distance between parents and the local authority leadership would become too great.

Therefore, it was obvious to use known institutions, and there were already
legal provisions for school boards - they were mandatory in Copenhagen and

common in Southern Jutland. In many rural municipalities, school boards
would simply replace the former school commissions, although with reduced
decision powers (Ussing Olsen 1982, 21 lf.).

Once school boards were made mandatory in 1970, the development that culminated in present-day school boards took off. Through later changes in the School Management Act (1974, 1978), the school boards were granted increased decision powers, among other things in economic matters. Thus, the powers of the individual school boards approached those of the school commissions. This created demands to abolish the commissions, which happened in 1990. Their functions were transferred in part to the new school boards (skolebestyrelser), in part to the local councils' school committees.

Thus the user bodies have changed greatly with the passage of time. The school commissions of 1814 were supervisory bodies, and as such placed under regional supervisory bodies and the Ministry of Education. Since 1990, school boards have been governance bodies without supervisory functions and make decisions on the operation of individual schools within the framework set by the local council and the ministry.

Sweden

The establishment in 1914 of the Elementary School Board made for a different development in Sweden. The Elementary School Board was a driving factor behind the strong centralization of educational administration. The question of establishing a user body never made it on to the agenda during the era of centralization. At the end of the 19605, however, ideas of grassroots democracy made a strong impact on the public debate. In 1974, an investigative commission on education launched a proposal for a user body endowed with decision powers. The idea of allowing students and parents to become committee members provoked local government criticism which demanded that "the interests of society" be represented in the body. Also the teacher unions sharply criticized the proposal for not being in accordance with existing labor law provisions. This criticism has recurred many times.

Despite this, the Minister of Education proposed that a user body be established, and the parliament endorsed this principle. The Co-determination Act (Medbestämmandelagen) presumably required that the user body would be incompetent in employee matters. A new working group was appointed, but its proposal was subjected to the same criticism as the earlier proposal. Again legislation was postponed, and an investigative commission appointed. A conflict with the Ministry of Municipalities produced further problems, and the result was a very watered down decision. In 1981, principals were enjoined to keep student and parent representatives informed on a continuous basis regarding school operations, and to confer with these representatives on

important questions. Other investigative commissions were established later,
but the interest in parent influence clearly subsided.

Why then was no user body established? Opposition from the teacher unions and the Swedish Association of Local Authorities proved sufficient to block legislation to institute user influence. The provisions of labor and local government laws also produced great difficulties for attempts to insert a law establishing a user body into already existing legislation. It proved difficult to clarify the separation between the tasks of the user organ and the areas designated for negotiation in the Co-determination Act. Another issue was whether a user body could be established through special legislation, or whether the Local Government Act prevented this. Since the two sets of legal provisions pertain to different ministries, this was very important, especially since the Ministry of Municipalities opposed any special provision for the educational system.

At the end of the 1980s, state control over the school system was seen as part of the problem, rather than as part of the solution. In this perspective, deregulation of user influence in the 1990s becomes an element in the "local governmentalization" of the educational system. New state regulations were virtually unthinkable. While formal parent influence was strengthened at this time in Denmark, central regulations on such matters disappeared in Sweden. This can seem paradoxical, but the riddle is solved if we look to the institutional history.

The Conventional Explanations

The above institutional histories illustrate how the course of development has been path dependent and incremental, which supports my thesis. I analyze the alternative explanations below, and indicate a number of anomalies to which they give rise.

The Varying Strength of the Labor Movement

According to this hypothesis, Social Democrats in the two countries are expected to counteract, or at least not strengthen, parent influence. The Danish Social Democrats, however, have contributed actively to institutionalizing parent influence through user bodies. The parent councils provided for in 1933 were actually pushed by a Social Democratic minister, whose original proposal went much further than the final agreement with the Agrarian Liberals. Likewise, several years before the bourgeois government proposed a bill in 1969 making school boards mandatory, the Social Democrats had declared their support for such an arrangement. In addition, a social democratic proposal preceded the new Primary School Act's enhancement of the school boards' decision powers in 1989.⁴ Thus, on

several occasions, the Danish Social Democrats have shown that they do not
find parent influence in primary education to be against their interests.

Nor are the Swedish Social Democrats unequivocally opposed to parent influence in primary education. A Social Democratic minister of education chose not to launch any concrete proposal for a user body in 1976, yet her bill laid down that some sort of body with user representation would be established. The Minister of Public Administration between 1982 and 1988 tried to make user influence an important element in the "modernization" of the public sector, but the result was negligible. The bourgeois parties been in any great hurry to introduce parent influence either. They considered the question when they held office from 1976 to 1982, but never introduced a bill. The question was also low on their agenda when they held office again from 1991 to 1994.

Attitudes toward parent influence are determined not by party affiliation but by sector affiliation. Between 1970 and 1985, user bodies were analyzed by several investigative commissions, some appointed by the Ministry of Education, others by the Ministry of Municipalities. The former produced proposals for the establishment of user bodies, while the latter criticized these same proposals and pushed other reforms. Politicians' institutional affiliation is more important than their ideology for their standpoint on this question. In contrast to politicians dealing with local government affairs, educational politicians have been favorably disposed to reforms of this type. What qualifies as an anomaly for the strength-of-labor theory can thus be explained institutionally.

In both countries, the consensus on the development of educational administration has been very strong. This fact contradicts the strength-oflabor explanation, which is implicitly based on the notion that different classes (and thus different parties) have opposing interests. The fact that preferences regarding administrative structures differ more between the countries than between the parties is an anomaly for this approach, but the puzzle is solved when we take the impact of the institutional legacy on party interests into account.

As a result of the different institutional legacies, the interpretation of private schools diverges in Sweden and Denmark. For a long time, the public school in both countries was a school for the poor; the rich avoided it and sent their children to private schools. Since 1899, however, state grants have been allocated to the so-called Danish free schools (friskoler), thus changing their character over time. As a Social Democratic minister of education commented:

Private schools in Denmark did not become expensive schools for the upper class. Instead, through substantial state financial support, they became an alternative to the public school for parents desiring a school with a particular world view or educational orientation. (Andersen 1983, 90).

In general, private schools in Sweden did not have the same popular basis as the Grundtvigian free schools. Instead, private schools were primarily used by the privileged classes in the cities (SOU 1981:34). Private schools in Sweden have thus become strongly identified with social segregation. These different institutional legacies are more important for the behavior of the social democratic parties than their common ideology.⁵

The Importance of Grundtvigianism

The mechanisms through which Grundtvigian discourse affects political decision making are not specified by the champions of this explanatory approach in any great detail. They point to the correlations between the discourse and the policy pursued. However, political behavior that accords with Grundtvigian notions is not necessarily caused by these ideas. A study of the content of the Danish primary school system concludes that: "Today it is clear that themes that could reasonably be said to contain Grundtvigian educational notions are now interpreted on the basis of psychological and educational ideas that would have been altogether foreign to Grundtvig the historical person" (Thestrup Pedersen 1978). Danish public schools also contain elements that are contrary to Grundtvigian ideas: Although Grundtvig firmly opposed examinations, final examinations {afgangsprøver) have been used all along in the Danish public school system. Such examinations were abolished long ago in Sweden and Norway, which does not make these school systems Grundtvigian.

Another point against this explanation is the fact that, in Sweden, a decision was made to introduce user bodies into primary schools in 1976. This decision shows that the institutional restrictions in question do not concern thought, speech, or even decisions, but rather what reforms are possible to implement. The norm that the users (parents) should have influence on the school is found in both countries. The variation appears when the abstract norm is to be made operational, i.e., converted into institutional arrangements. In this operationalization, other norms in the two countries concerning the structures and functions of the public sector must be considered, e.g., that the central government must formulate basic requirements in order to ensure equal education in the entire country, that the local authority is responsible for implementation, and that teachers are entitled to co-determination and influence founded in their educational expertise. The actual shape of educational management is the result of attempts by politicians to combine these partially conflicting norms on the background of existing administrative structures.

By studying the decisions that have shaped parent influence, we can judge
whether the political actors have acted in accordance with the premises of the
explanation in question. When parent influence was institutionalized in 1933,

the Danish Social Democrats were the driving force, even though Grundtvigian ideas had a relatively hard time penetrating that party in particular (cf. above), and the Agrarian Liberals, the party with the strongest historical Grundtvigian connections, in faet weakened the parent influence that the Social Democratic proposal was aiming for.⁶ Consequently, Grundtvigianism cannot explain the establishment of parent influence in Denmark.

Nor can it explain the fact that parent influence has grown over time. It would be wrong to say that Grundtvigianism has become stronger with the passage of time, and that parent influence has developed accordingly. On the contrary, Danish society has become progressively more secular, which would indicate that Grundtvigianism is weaker today than it was earlier. The growth of parent influence is, therefore, an anomaly for this explanation.

The crux of the matter is the premise of the discursive explanation, namely that ideas explain political practice. The theory cannot examine the inbetween steps without weakening its explanatory potential, and therefore it cannot pay due heed to the difficulties of transforming ideas into political practice. If we take our point of departure in institutional theory, we will not be surprised to learn that the development of parent influence has proceeded in steps. That the actors usually "choose" to change existing institutions on the margins is exactly what the theory claims. Finally, I will present the institutional mechanisms that impel actors to behave in this way.

Mechanisms of the Institutional Legacy

The concept of "mechanisms" calls up images of gearwheels and mechanical notions of causality. The point is that good explanations should account for how a certain explanatory factor has influenced the outcome in a given empirical case (Elster 1989). That is, they should not simply state that A caused B, but give a closer account of the manner in which this happened (cf. Skocpol 1992).

Institutionalized Action Capacity Provides Power To .

Actors' access to alternative courses of action is strongly influenced by the institutional context (Shepsle 1989). Existing institutions provide actors with a certain type of action capacity and offer potential solutions to the problems with which politics is constantly confronted. In many cases, "problems" do not make it on to the political agenda unless the institutional "solutions" are available (Weir & Skocpol 1985). When these solutions vary, the political measures can vary as well.

The existing forms of consultation were important for the way the state
chose to legitimize a given program. Since it is much easier to make small

incremental changes in an institution than to establish a new one, the state acted differently in Denmark and in Sweden when grassroots democracy emerged high on the agenda. There was a basis in Denmark for the establishment of user-democratic bodies which was lacking in Sweden; as a result, different reforms were chosen in the two countries. The Danes followed their path and developed user bodies; in Sweden, meanwhile, community councils {kommundelsnämnder) were preferred as the strategy for grassroots-democratic reform. As a result of the institutional history, grassroots influence in Denmark has been interpreted in terms of user affiliation, and not, as in Sweden, in terms of geographical proximity.

Institutions and Interests

Preferences/interests are typically grouped along either class or party lines. As shown above, however, institutionalized interest has been more important than political party ideology for the attitude on user influence taken by representatives in investigative commissions and parliamentary committees. When institutions change, their impact on how actors interpret their interests often emerges. As mentioned earlier, a result of the 1970 Local Government Reform in Denmark was mandatory school boards. Prior to the reform, the National Association of Local Authorities and the teacher union had opposed school boards, but afterwards their attitude changed.

Thus, the National Association of Local Governments has sought to strengthen the school boards, especially in economic matters. Through strengthened school boards, the Association hoped to shut down the school commissions, which it regarded as economically irresponsible. Unlike the district council and its committees, the school boards operate within given economic constraints, and they are sometimes thought capable of disciplining users. Likewise, the Danish Union of Teachers has not demanded closure of school boards since 1970, but has instead demanded stronger teacher representation. In sum, ever since school boards were made mandatory, they have become a given 'structural' parameter for organizations with a vested interest in primary education. They have therefore focused their efforts on adapting the institution to their own purposes rather than on shutting it down.

While the main reason no user body was established in Sweden during the 1970s and 1980s had to do with the strong opposition from local government interests and teacher organizations, the Danish organizations have interpreted their interests differently and acted accordingly. This reflects the fact that the organizations in question operate within different institutional contexts. The Swedish organizations still operated in the "same" context as their Danish sister organizations had done before 1970, and therefore their opposition to user bodies is not surprising.

Learning (Policies Affect Politics)

Politics is traditionally thought to concern conflict and power. But politics is something more: uncertainty and insufficient knowledge mean that it is also a matter of collective learning and problem solving. However, what one learns depends on what one does, since one's experiences are dependent on earlier measures (Heclo 1974). Therefore, political reforms are generally modifications of earlier policies, rather than total re-organizations in which a new program is built from scratch. This calls for a historical perspective on administrative policy decisions (cf. Skocpol 1992).

Problems are often unintended consequences of the existing administrative structure. This is clearest during periods of relative stability when adjustment occurs gradually. But also more radical reforms, such as those in Sweden in the late 1980s, clearly demonstrate this path-dependent type of learning. There was broad consensus that state control over the educational system was part of the problem rather than part of the solution. Therefore, state regulation of parents' rights to influence became impossible; on the contrary, such control over the organization of local authorities was to be removed. Thus, local decision making was deregulated and decentralized to the local government level. In Denmark, by contrast, the central government decentralized responsibilities and functions to the school level and gave the user boards further decision power.

Institutions Affect the Power of Actors Over .

After school boards became mandatory in 1970, membership in the Danish parents' association increased sharply, and a positive spiral subsequently characterized the development of this organization. Due to its increased strength, the organization was invited, together with the National Association of Municipalities and the Danish Union of Teachers to participate in the negotiations conducted by the Ministry of Education on the school management reform of 1989. Thus, institutionalized influence can give rise to a self-reinforcing pattern of development. By contrast, the Swedish parent organizations have never experienced such a take-off (cf. Rothstein 1992a).

Above I have shown that the institutional legacy thesis proves capable not only of explaining the overarching question of this article regarding the difference between parent influence in Danish and Swedish educational administration, but also of explaining the behavior of the actors at different points in the historical process (cf. George & McKeown 1985).

Another example is the right granted to Swedish local authorities in 1996 to delegate decision powers to boards with user representation in the public school system. This reform may seem to contradict the thesis of this article; however, my thesis is not that reforms are impossible, but that they are

difficult to carry out if they awaken resistance from powerful interests. These user boards are not, as in earlier proposals, mandatory, but optional, and local authorities can determine the functions and responsibilities granted to the user boards, and to some extent their composition. Furthermore, the introduction of user boards in Swedish schools merely has the form of an experiment. If this is the first step towards introducing user boards in Swedish schools, it is a small step. This accords with the thesis of institutionalist theory that institutional change is normally incremental.

The Interplay of Explanatory Variables

The fact that I have described the development in Swedish and Danish
educational administration as path dependent does not mean that I regard
political actors as insignificant. North puts it this way:

If, however, the foregoing story sounds like an inevitable, foreordained account, it should not. At every step along the way there were choices - political and economic - that provided real alternatives. Path dependence is a way to narrow conceptually the choice set and link decision making through time. It is not a story of inevitability in which the past neatly predicts the future (North 1990, 98f.).

Existing institutional structures are the result of decisions with varying objectives taken at different times. They are not uniform, but are bearers of many different "futures." Often this indeterminacy yields a space for varying "rational" choices of incremental change, which can produce large effects with the passage of time. When undertaking an institutionalist interpretation of history, therefore, one must avoid retrospective rationalization. Long-term historical processes are seldom planned, but rather incremental and unpredictable. However, one can (almost) always extract a historical "line of development."

Interests and ideas have been important for the course of institutional development examined in this article. It is the interplay between political ideas and institutional structures that explains the comparative development of Swedish and Danish educational administration in the 1900s. Grundtvigian ideas combined with the experiences of the farmers' constitutional struggle have had a great impact on how Danish Agrarian Liberals view the role of the state in society, and thus on the shape of the educational administration. The interests of actors have played a more subordinate role in the development of Swedish and Danish school administration. The political consensus has been significant in both countries, and conflicts among different interests have played a minor role. To a certain extent, however, the significance of existing administrative structures depends on power relations. In the 1920s and 19305, the Danish Social Democrats sought to strengthen the influence of the state over the educational system, but how far they wished to go is unclear.

In sum, political institutions are not a causa ultima, but they do furnish the specific conditions under which actors interpret their interests and act. In this way, they contribute to both continuity and change in politics (Thelen & Steinmo 1992). The point of interest lies in analyzing how the institutional legacy interacts with the power, interests and ideas of the actors, rather than in formulating a reductionist institutionalist explanation. Indeed, while I see institutions as intentionally created, and at the same time as determining the action capacities of actors, I am inclined to regard the debate on whether actors or institutions provide the definitive explanation as unfruitful (Boudon 1986; Taylor 1989).

General Applicability

One limitation of my analysis of the citizens' role in the welfare state is that I have only studied a part of the problem, i.e., the role of parents in the administration of primary education. I would argue that this has been a strength up to now, rather than a limitation. By systematically following the course of development within one policy sector, I have avoided treating the empirical data as a collection of disconnected and convenient examples, and been forced to confront anomalies. However, it may be wise to take a closer look at how citizens' encounters with the welfare state have been structured in other welfare sectors. That the data are presented somewhat rhapsodically cannot be avoided.

In the 1980s, both Denmark and Sweden faced a new and deadly problem, AIDS. How have the two countries chosen to handle the problem? In Sweden, the Communicable Diseases Act (Smittskyddslagen) is based on (a) registration of infected persons, (b) a series of regulations specifying how infected persons are to behave, and (c) the possibility of subjecting infected persons to compulsory confinement for an indefinite period if they do not obey. In Denmark, by contrast, the fight against AIDS has proceeded on a voluntary and anonymous basis. Is this a coincidence? I do not believe so. To answer this question with greater assurance, we would need to study the institutional history of the regulations regarding contagious/venereal diseases. In the absence of such a broad investigation, however, the above account still serves to illustrate what seems to be systematic differences in viewpoint regarding the rights and duties of the citizen, and how and when the state may use its coercive power (Knudsen 1993).

One of many examples is alcohol policy. The difference between the two countries is sale of alcohol, and the attitude is more liberal in Denmark. Between 1923 and 1956, Sweden's restrictive alcohol policy with ration books featured an allocation based on examinations in each individual case. Such examinations could be quite thorough, to the extent that personal rights

were violated (Rothstein 1992b). The aversion to paternalism and coercion
has been stronger in Denmark.

Also in more service-focused areas of the welfare state are citizen roles different in the two countries. In Sweden, there are no statutory and mandatory user bodies in any welfare sector. Local authorities were recently enabled to delegate decision powers user bodies, but because the law was passed only in 1994, we do not know how significant its impact will be. On the one hand, there are many signs of increasing interest in new forms of citizen participation; on the other, there is the strong norm that the voice of the people should be channeled through political parties and traditional popular movements. However, the limited use since 1986/87 of conditioned delegation, i.e., local authorities' right to delegate decision power to a civil servant on the condition that said official agrees with "user representatives" (SOU 1993:90), indicates that user influence may not become widespread.

Thus, user influence is not particularly advanced in Sweden, in contrast to the situation in Denmark, where user boards are not restricted to primary education. In 1970, they were introduced in secondary education, and in 1994 in child care. These boards enjoy management rights, and they are (numerically, at least) dominated by users. Also in other sectors, many local authorities have chosen to establish different forms of user influence (Wadskjær 1995).

In sum, we find systematic differences between the roles assigned to citizens in Denmark and Sweden in their encounters with the welfare state. To simplify the matter, the relatively liberal welfare state of Denmark regards those employing its services as users, while the relatively paternalistic Swedish welfare state sees them as clients. I have shown that, in the educational area, these conceptions are derived from long-standing administrative-historical traditions. It would seem a reasonable hypothesis that the same applies to other welfare sectors too.

---

NOTES

1. The exact values should not be taken too literally, since the formulation of the questions and the reply scales were different in the two countries. But the difference between the countries is clear beyond any doubt. The figures are basedon raw data from the Swedish and Danish civic studies, which are reported in Petersson et. al. (1989), and Andersen et. al. (1993).

---

2. I see no reason to incorporate into the definition of rationality the absence of information costs, i.e., the assumption that the actors have complete information and that their notions of reality are objectively correct (cf. Boudon 1989). An actor can therefore have good reason (subjectively) to behave in an, objectively speaking, "irrational" manner. The preferences of the actors are influenced by their institutional context; if, however, the "preference change" is conscious, it is fully compatible with assumptions of rationality (cf. Elster 1986, 15).

3. In 1920, the Danish parliament passed a provisional law on primary education in Southern Jutland, which had been German until the end of World War I. A school commission was to be appointed in each municipality. However, a large number of small municipalities remained from the German period, which meant that in practice a school commission was set up for each school. Experiences with the small school commissions were so positive that they were kept when municipal amalgamations were carried out in 1923 (Lehmann 1930, 61-71).

4. The Primary School Act replaced the School Management Act and the former Primary School Act.

5. While the Danish Social Democrats accept private schools today, it was another matter during the first half of this century. The 1923 party congress declared that the educational system ought to be organized by the state with assistance from the municipalities - a formulation leaving no room for private schools. Their acceptance of private schools today reflects the fact that segregation has been limited, but it is doubtful whether they regard free school choice as a right. For instance, the party congress has decided to work for an arrangement that will allow municipalities to prevent the establishment of private schools if they threaten the economy of the public schools.

6. One conceivable institutional explanation for this paradox is that the Agrarian Liberals feared (and the Social Democrats hoped) that the parent councils would strengthen local forces who demanded an improved, but also costlier, system of education. For this reason, it is no wonder that the Agrarian Liberals took a negative view of the social democratic proposal, although they had been champions of increased parent influence for many years. Several questions were thus interwoven in a single game (cf. Tsebelis 1990).

ACKNOWLEDGMENTS

The author would like to thank Peter Mayers and Annette Andersen who have transformed my original manuscript into readable English, HSFR that paid for this extravaganza, and Bo Rothstein who was my doctoral advisor when I made the study. I would also like to thank the three anonymous referees and the editors who made some valuable suggestions for improving the manuscript.

REFERENCES

Andersen, J., eta 1.1993. Medborgerskab. Demokrati og politisk deltagelse. Herning: Systime.

Borgbjerg, F. 1930. Folketingstidende 1930/31 F. Copenhagen.

Boudon, R. 1989. "Subjective Rationality and the Explanation of Social Behaviour." Max
Institut fur Gesellschaftsforschung Köln, Discussion Paper 89/6.

Cranil, M. 1994. Decentralisering og selvforvaltning i folkeskolen: de første års erfaringer
med folkeskolens nye styrelsesbestemmelser. København: Undervisningsministeriet, folkeskoleafdelingen.

Elster, J. 1986. "Introduction," in Elster, J., ed., Rational Choice. Oxford: Basil Blackwell.

Elster, J. 1989. The Cement of Society: A Study of Social Order. Cambridge: Cambridge
University Press.

Esping-Andersen, G. 1985. Politics Against Markets. New Jersey: Princeton University Press.

European Journal of Political Research. 1993. Special Issue, Political Data 1945-90, Party
Government in 20 Democracies 24:1.

George, A. & McKeown, T. J. 1985. "Case Studies and Theories of Organizational Decision
Making," in Advances in Information Processing in Organizations vol. 2. Santa Barbara:
JAI Press.

Grønnegård Christensen, J. & Ibsen, M. 1994. Bureaukrati og bureaukrater. Herning: Systime.

Hall, P. A. 1989. "Conclusion: The Politics of Keynesian Ideas," in Hall, P. A., ed., The
Political Power of Economic Ideas. New Jersey: Princeton University Press.

Heclo, H. 1974. Modem Social Politics in Britain and Sweden. New Haven & London: Yale
University Press.

Hirdman, Y. 1989. Att lägga livet till rätta. Stockholm: Carlssons.

King, A. 1973. "Ideas, Institutions, and the Policies of Governments: A Comparative
Analysis." British Journal of Political Science 3:3.

King, G., Keohane, R. & Verba, S. 1994. Designing Social Inquiry. Scientific Inference in
Qualitative Research. New Jersey: Princeton University Press.

Knudsen, T. & Rothstein, B. 1994. "State Building in Scandinavia." Comparative Politics
26:2.

Knudsen, T. 1993. Den danske stat i Europa. København: Jurist- og Økonomforbundets forlag.

Korpi, W. 1981. Den demokratiska klasskampen. Svensk politik i ett jåmforande perspektiv.
Stockholm: Tiden.

Lauglo, J. 1980. "Participatory Committees and Contrasting Administrative Styles in
Scandinavian School Governance," in Baron, G., ed., The Politics of School Government.
1980. Oxford: Pergamon Press.

Lehmann, H. 1930. Haandbog i Lovgivningen om den danske folkeskole. København: Nordisk
forlag. (3rd ed.).

Lewis, J. ed. 1993. Women and Social Policies in Europe: Work, Family and the State.
Aldershot, Elgar.

Lindblom, C. E. & Woodhouse, E. J. 1993. The Policy-making Process. Eaglewood: Prentice
Hall. (3rd ed.).

Lindbom, A. 1995. Medborgarskapet i välfärdsstaten. Föräldrainflytande i skandinavisk
grundskola. Stockholm: Almqvist & Wiksell International.

Lægreid, P. & Pedersen, O. K. (eds.) 1994. Förvaltningspolitik i Norden. København: Juristog
Økonomforbundets forlag.

March, J. G. & Olsen, J. P. 1984. "The New Institutionalism: Organizational Factors in
Political Life," American Political Science Review 78:3, 734-49.

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

Nordiska Ministerrådet. 1990. Skola och skolforvaltning i Norden. København: Nord 1990:8.

North, D. C. 1990. Institutions, Institutional Change and Economic Perfonnance. Cambridge
Universtity Press.

Petersson, 0., Westholm, A. & Blomberg, G. 1989. Medborgarnas Makt. Stockholm:
Carlssons.

Rothstein, B. 1992a. "Labor-Market Institutions and Working-Class Strength," in Steinmo.
Thelen & Longstreth, eds., Structuring Politics: Historical Institutionalism in Comparative
Analysis. Cambridge: University Press.

Rothstein, B. 1992b. Den korporativa staten. Intresseorganisationer och statsförvaltning i
svensk politik. Stockholm: Norstedts.

Sainesbury, D. ed. 1994. Gendering Welfare States. London: SAGE.

Shepsle, K. 1989. "Studying Institutions: Some Lessons From the Rational Choice Approach,"
Journal of Theoretical Politics 2.

Skocpol, T. 1992. Protecting Soldiers and Mothers: Political Origins of Social Policy in the
United States. Cambridge Ma.: Harvard University Press.

SOU 1981:34.

SOU 1993:90.

Steinmo, S. 1989. "Political Institutions and Tax Policy in the United States, Sweden and
Britain," World Politics 41:4.

Sørensen, E. 1995. "Brugerindflydelse og demokrati: Brugerindflydelsens rolle i udviklingen
af det danske demokrati," Politico 27:1.

Taylor, M. 1989. "Structure, Culture and Action in the Explanation of Social Change," Politics
& Society 2.

Tegborg, L. 1969. Folkskolans sekularisering 1895-1909: upplösning av del administrativa

sambandet mellan folkskola och kyrka i Sverige. Stockholm: Foreningen for svensk
undervisningshistoria.

Telhaug, A. O. & Tønnesen, R. T. 1992. Dansk utdanningspolitikk under Bertel Hoarder
1982-92. Ny liberalistisk og nykonservativ skoletenkning. Oslo: Universitetsforlaget.

Thelen, K. & Steinmo, S. 1992. "Historical Institutionalism in Comparative Politics," in
Steinmo, Thelen & Longstreth, eds., Structuring Politics: Historical Institutionalism in

Comparative Analysis. Cambridge: Cambridge University Press.
Thestrup Pedersen, E. 1978. "Grundtvigs betydning for den danske folkeskole," in Årbog for
Dansk Skolehistorie 1978. København: Selskabet for Dansk Skolehistorie.

Tsebelis, G. 1990. Nested Games: Rational Choice in Comparative Politics. Berkeley:
University of California.

Ussing Olsen, P. 1982. Skolestyrets udvikling. Vejle: Kroghs.

Wadskjær, H. 1995. "Learning User Influence. The Constitution of User-Subjects in a Differentiated Welfare State," in Bogason, P., ed., New Models of Local Political Organizing: Local Government Fragmentation in Scandinavia. New York: NOVA Science Publisher.

Weir, M. & Skocpol, T. 1985. "State Structures and the Possibilities for 'Keynesian' Responses to the Great Depression in Sweden, Britain, and the United States," in Evans, P., Rueschemayer, D. & Skocpol, T., eds., Bringing the State Back in. Cambridge University Press.

Østergård, U. 1990. "Peasants and Danes: Danish National Identity And Political Culture."
Arhus, Working Paper 75.