Resumé

This article starts from the assumption that some of the theoretical work used to explain welfare-state expansion can be used 'in reverse' to explain privatization, here seen as welfare state contraction, i.e., the transfer from the public to the private sector of the responsibility for certain activities involved in welfare provision. To this end, variants of the 'power resources' approach are examined. Finding that the 'labour movement' thesis does not have a strong predictive value, I then discuss the 'game theoretical' variant of the 'power resources' approach. Its usefulness for predicting privatization patterns seems limited because of the reductionism built into it. Following the neo-Institutionalists' argument that actor preferences and strategies both reflect and form institutional arrangements of the welfare state, some institutional typologies are developed which are relevant to the course and patterns of privatization. An ideal-type dichotomy between 'pluralist' and 'corporatist' institutions is discussed, in connection with a further delineation of decision-making, financing, and implementing In conclusion, an effort is made to combine actor-oriented and institutional approaches to formulate some preliminary predictions about when and where certain patterns of privatization will occur.

Introduction

There is by now a host of theories to explain the growth of the modern welfare state. From the 'pluralist industrial society' school comes the idea of a 'logic of industrialism'; welfare-state growth is a secular and universal evolutionary process intimately linked to economic growth, industrialization, and urbanization but not to politics (Wilensky 1975). According to the 'functionalist neo-Marxist' line of thinking, the welfare state grows mainly as a repressive social-control mechanism to legitimize the position of the dominant capitalist class in society (Olson 1982). The 'popular protest' school views welfare service development as an effect of elite concessions to policy demands made outside representative, parliamentary channels, for example through strikes, demonstrations, or riots (Piven and Cloward 1971). Within the 'power resources' approach, we find at least two explanations. The 'labour movement' thesis states that welfare services grow as a result of left-wing parties and workingclassorganizations politically more powerful than parties and groups representing capital (Korpi 1978; Stephens 1979; Castles 1982;

Esping-Andersen 1985). A more 'game theoretical' variant holds that public welfare programs grow as a result of decisions of interdependent actors seeking power positions to enhance longer-term policy objectives (Korpi 1987). To followers of the 'state autonomy' approach, a central role in welfare-state expansion is played by public officials, using their degree of autonomy from interest groups to enhance their self-interest by pursuing policies which enlarge the public sector (Niskanen 1971; Skocpol 1985).

However, the 'privatization boom' of the last decade poses some crucial problems for social scientists engaged in efforts to explain welfare-state expansion. Take the 'pluralist industrial society' school; if public welfare and economic growth are parellel trends, how are we to explain the continued increase in privatization measures in many welfare states during the 1980s, despite the re-emergence of positive growth in their economies? And even if many industrialized states experienced a downturn in the business cycle coupled with heavy fiscal stress at the turn of the decade, this has not unfolded in parallel welfare-state contractions (ef. Henig et al. 1988, 457). Furthermore, what about 'popular protest'?; has privatization expanded because of widespread strikes and demonstrations demanding cut-downs in welfare-state services during the 1980s? Also, the 'labour movement' theory meets with problems here; for if public-sector approaches to welfare provision are particularly visible in states where Left parties and trade unions hold a crucial power position, how are we to account for the vivid interest in, and actual adoption of, privatization alternatives in such countries? Finally, what about the 'functionalist neo- Marxist' explanation: should privatization be seen as evidence that the use of state repression has become less necessary for capitalists in the 1980s, amidst increased unemployment and labour-market unrest? Or should privatization be seen as a corroboration of another neo-Marxist thesis, i.e. that the welfare state has finally reached its limits, where any further expansion would threaten the functions of the capitalist economic system (Gough 1979; cf. Offe 1984, 153).

In this paper, I follow two paths in discussing possible explanations of the phenomenon of privatization. First, I assess some of the most wellknown which view welfare-state expansion as a result of the rational actions of individuals or collectivities. Could any of them be used 'in reverse', i.e. to explain the welfare-state contraction inherent in the withdrawal of public-sector responsibility from some of the activities involved in welfare-service provision?

Second, rational actors will face different restrictions and possibilities from earlier policy decisions, institutional arrangements, and the like. Therefore, I examine different institutional typologies of the welfare state, in order to enable us to make predictions about the patterns of privatization emanating from the interplay between actors and institutional settings.

The Meaning of 'Privatization'

First of all, however, we ought to define the dependent variable. What should be meant by 'privatization'? There is an abundance of suggestions for defining this concept. To some, it is the introduction of market principles in the distribution of welfare services (Forrest & Williams 1984, 1167; Shackleton 1984, 59; Adams 1987,129). To others, privatization represents a move from collective to individual levels of service provision, saying that the decisive criterion is the 'decollectivization' of welfare services (Kemeny 1980, 347; Harloe & Paris 1984, 77 f.). Finally, some use the technique of shifting responsibility from public to private as the criterion for defining privatization (Heald 1984, 38 ff.; Lorentzen 1987, 263 ff.).

Elsewhere, I have suggested that for the comparative study of privatization a fruitful way of defining privatization may be to look at the location of responsibility and type of activity (Lundqvist 1988, 12). Privatization should be seen as 'actions taken by actors legitimately representing the public sector to transfer the hitherto public responsibility for a certain activity away from the public and into the private sector'. This relocation of responsibility thus becomes the main criterion for delimiting 'privatization' in a general policy sense, regardless of policy area. If the focus is on policies for welfare-service provision, we find that several main activities are involved. The search for privatization patterns thus necessitates a classification of such activities. Drawing from a lot of different suggestions (see, for example, Kielland 1986, 201 f.; Leat 1986, 291 f.; Le Grand & Robinson 1984, 4 ff.; Kristensen 1987a, 223), I suggest that this classification should include regulation, financing, and production of services. In this way, one gets a matrix showing several different patterns of privatization (see Fig. 1). It is to the possibilities of explaining different privatization patterns in welfare programs that this paper is devoted.

Actor Explanations to Welfare-State Development

As a general point of departure, the 'power resources' theory of welfarestate points out that in all societies, there is a particular balance between the state and the market, i.e. between different criteria for welfare distribution. These are identified as buying power in the market and citizen rights in politics. This balance between public and private solutions to welfare problems is interpreted as a result of how different actors and collectivities on each side of the labour-capital divide have mobilized and used their respective power resources to affect it (Korpi 1987, 10).

The approach rests heavily on rationalistic assumptions at the same time
as it seeks to provide a long-term historical explanation for the different

patterns of public-welfare expansion in modern nations. Its basic theoretical
assumption is that actors and collectivities

which are weak in terms of their market resources can be expected to attempt to use their relatively more favourable positions in terms of political resources to affect the conditions for and outcomes of distributive conflicts in society because the potential gains, which come to an actor from limiting the sphere of market operations, and market criteria in distribution, will increase with decreasing market capacities of the actor (Korpi 1987, 9 ff.).

Empirically, the labour movement and the parties to the Left have been the main representatives of the classes weak in market resources. In countries where they have outgrown right-wing parties and capitalist organizations their power resources, and thus gained political control, there is a tendency for public-welfare programs to increase somewhat more than where 'market-strong' actors and collectivities have remained more powerful in politics (Pampel and Williamson 1985, 784). It should be noted that where parties are not organized according to class but along other lines, such as the Dutch confessional parties, the tendency towards welfarestate reflects the relative strength of 'market-weak' groups within the ruling party or parties (van Kersbergen and Becker 1988).

The theoretically derived connection between increased power resources for actors and collectivities representing 'market-weak' segments of society on the one hand, and welfare state expansion on the other (which some say also exist empirically - see Shalev 1983; Skocpol & Amenta 1986,140), has important implications for the efforts of generating hypotheses about

privatization of welfare services. First of all, it builds on the assumption that actors on each side of the labour-capital divide seek political power, not just for power's sake, but to use it to adjust the boundary between state and market in such a way as to maximize the utility of their constituencies. It follows that actors representing 'market-weak' classes will use their power resources to expand the 'non-market' provision of welfare to almost every walk of life. In the extreme case, every service will be publicly provided: if the collective actors representing 'market-weak' groups have an overwhelminglystrong position, there will thus be no move towards privatization in the sense defined above.

It furthermore follows that actors representing 'market-strong' segments in society will use their political power to expand the private provision of welfare as far as possible. In the extreme case, all welfare will be privately provided; where the 'market-strong' collective actors hold a strong, longterm of political power, there will in the end be nothing left to privatize in welfare.

Brought to the extreme in both directions, the original 'power resources' approach leads to a 'privatization paradox': nothing can be or is left to be privatized. However, such an application of the 'power resources' theory seems too unsophisticated: we simply cannot postulate this linear relationship the actors' political power position and welfare-state development. 'market-strong' collective actors have adopted publicwelfare and those representing the 'market-weak' strata have sometimes supported moving programs out of the public sector.

Is there any way to accommodate such cases within the 'power resources'
approach? There is at least one line of reasoning worth discussing here.
According to Korpi (1987, 12) the

power resources approach generates, instead, what in essence amounts to a game theoretical perspective on the analysis of interdependent actors, where the decisions made and the strategies adopted by each actor are seen as affected by the actor's perceptions of the relative power and probable choices of other actors.

This qualification of the hypothesis about a linear relationship between increased political power for working-class organizations and Left parties on the one hand, and public-welfare expansion on the other, has important implications for a predictive theory of privatization. It modifies the assumption actors seek political power only to promote their long-term goals with regard to the proper demarcation between state and market. To a considerable extent, their actions can also be explained as calculated moves to gain or retain power in a competitive political environment. As Quadagno (1987, 116) puts it; when 'organized labour attempts to implement socialist goals through competitive political parties, it is constrained by the need to expand the party base beyond the working class'.

This version of the 'power resources' approach can thus be used to explain 'socialization' or 'privatization' moves which seem contradictory to what our assumptions about the links between 'market-strong' or marketweak' bases and the ensuing ideological view of proper state-market boundaries in welfare provision would have us predict. However, if every decision concerning the boundary between public and private welfare provision could be explained as calculated moves to gain or retain political power (cf., for example, Rimlinger 1971; Ascher 1987, 47 f.), then we seem to end up in the morass of scientific reductionism. Whatever the differences found in the patterns of privatization, they can be explained by the powerseeking of political actors. Such a research strategy does not promise to be very successful.

On the other hand, this should not be read as an all-out dismissal of Korpi's modification. Indeed, power calculi may be of utmost importance. What I am suggesting here is that his game-theoretical variant of the 'power resources' approach must be seen in an institutional context, if any fruitful propositions are to be gained from it. It does not take much imagination to see that power calculi may play a much more central role in a powerbalance than in a context in which either the 'market-strong' or the 'market-weak' party constellations have a very strong and stable majority in a parliamentary system. Thus, the efforts to build a predictive theory of privatization have to combine the rational-actor approach with an institutional

An Institutionalist View of Welfare-State Development

Regardless of whether they cater for the interests of 'market-weak' or 'market-strong' segments in society, goal-seeking politicians are all acting within an institutional framework. Since the relationship between preferences results is more indeterminate and indirect in politics than in economics, the political institutional framework will probably have an independent and perhaps decisive influence on the relationship between preferences and the content of political decisions, as well as on the way those decisions are implemented (Kristensen 1987b, 37). At the same time, however, it is also clear that rational political and bureaucratic actors, through their endeavours to expand or contain the welfare state, are also building institutions. As Levi (1987, 687) puts it:

At the heart of the rational choice approach to institutional change is the investigation of how rational individuals create and maintain institutions through their choices and of what the unintended consequences of their choices are. Equally important is how institutions delimit future choices.

Therefore,

theories of political science must be conscious of the interplay among political institutions, the wider society, and individual actors. How do institutions generate or block change? How do they influence individual preferences, expectations, and resources? (Olsen 1985, 5; my translation).

If we see privatization as a questioning of the institutions of the welfare state, how do we identify and characterize the institutional forms relevant to political actors promoting or fighting privatization? And how is this done in a way most fruitful for our explanatory efforts, i.e. to help us see the interplay between institutions and actors? Neo-Institutionalists say that there is an institutional separation between state and society, and 'state autonomy' scholars call for 'bringing the state back in'. However, do these scholars always provide a view of institutions which directs attention to their interplay with actors? When the 'state' is interpreted mainly as the activities of the state bureaucrats, it leads to a focus not on institutions, but on the goal-oriented behaviour of self-interested bureaucrats (cf. Dunleavy 1986). Along another line of thinking, the 'state' is interpreted as the legacy of earlier policy; prior state actions shape future goals, the primary linkage being the 'political learning' of bureaucrats and politicians (cf. Heclo 1975). A third interpretation of the 'state' seems less inclined to reduce institutions to behaviour: those who study historical variations in state structures contend that such institutional features as democratization, bureaucratization, and electoral systems have an independent influence on the timing and content of welfare state-related proposals and policies (see, for example, Skocpol 1980; March & Olsen 1984; Torgersen 1984).

The last of these three interpretations points to some general institutional features found in most political systems. Again, however, if we see privatization a questioning of welfare-state institutions, we must look for dimensions of those institutions which have a particular relevance for the way political actors go about dealing with that issue. Further, we know that the development of the modern welfare state has taken different routes in different countries, and we cannot therefore assume a priori that all such states are identical with regard to institutional forms.

The distinction between 'pluralist' and 'corporatist' welfare states is important here. Seen as ideal types, they differ in motives for, and forms of, state intervention in the market as well as in institutional structures. Three domensions are particularly important in separating the ideal types. First, there is type of policy: we find the well-known distinction between supplementary and comprehensive welfare policies (see, for example, Donnisonand 1982). Second, there is type of implementation. Ruggie (1984) asserts that the 'pluralist' ideal type is characterized by incremental

institution-building, leading to a fragmented and weakly interrelated implementation structure. This is because welfare policies and programs are established ad hoc, i.e. only when market dysfunctions have produced inequalities which can no longer be politically ignored. The result is policies supplementing those who are marginalized in the market. In contrast, the ideal-type 'corporatist' welfare state is built on a premeditated vision of equality, and comprehensive policies and implementing agencies are developed in accordance with that vision. There is an institutionalization of equality, and a provision of welfare services as a 'right', regardless of the recipients' strength in the market.

Third, there is the position of organized interests. In the ideal-type 'corporatist' welfare state, governments grant a form of public status to 'recognized' interests, meaning that the state deliberately gives a representational to an organization within a particular category of interests, in return for some control over that organization's interest articulation 1974,93; Crouch 1983,453 f.). Furthermore, such welfarestate use both laws and formal or informal agreements to transfer implementation responsibilities for a certain public policy to those organizations which have been 'recognized' as the legitimate ones in that particular policy field (Cawson 1978, 184). In typical 'pluralist' welfare states, on the other hand, organized interests may at best come out ahead in the competition for power positions in such informal structures as 'iron triangles' (Jordan 1981).

To sum up: three dimensions of the modern welfare state seem most relevant to our efforts of predicting how privatization patterns diverge among countries because of the interplay between goal-seeking political actors and institutional features (ef. Henig et al. 1988, 459 ff.). From the literature, we have distilled two ideal types, the 'pluralist' and the 'corporatist' welfare states. The 'pluralist' variant is characterized by supplementary a fragmented and weak implementing bureaucracy, and a sharp divide between the state and (competing) organized interests. The 'corporatist' welfare state promotes comprehensive policies, builds strong and coordinated implementing structures, thereby incorporating organized interests which are granted some form of monopoly of representation their respective fields. Before going further, let us look at some of the implications for the interplay between these institutional ideal types and the rational actors representing 'market-weak' or marketstrong' classes.

In the 'pluralist' welfare state, institutions will not be much of a hindrance to privatization proposals. Earlier interventions are typically very modest, and limited to covering up for market failures by providing supplementary benefits to well-defined groups. The implementing structure is fragmented. Since the target groups are 'marginalized' people with little or no organizationalstrength,

izationalstrength,there will also be little organized interest support for such welfare institutions should they come under heavy political attack (cf. Skocpol 1984). Consequently, privatization proposals may meet little hindrance in such an institutional context.

In contrast, the 'corporatist' institutional structure may prove a formidable factor to welfare-state contraction. The comprehensiveness welfare policies creates large constituencies for public services. Institutionally, the 'corporatist' arrangement means joint stateorganization for the order and progress of policy; the recognized' will have a stake in this type of ordered development, and will thus support institutions in which they hold key positions. Since the recognition of an organization is dependent on its sharing objectives in common with the government, it follows that proposals which threaten the 'recognized' interest organizations' position within the institutional structure will meet with strong opposition. This will particularly be the case where a 'market-weak' party has so far run 'comprehensive' welfare programs in close collaboration with 'recognized' interests such as professional and unions.

However, the institutional typologies offered so far are very broad and general. They seem more fruitful for explaining the extent and speed of privatization than the patterns of that phenomenon. To enable such predictions, a more detailed institutional perspective is necessary. One way of getting to these more detailed features could be to separate between the decision-making, financing, and implementing institutions within the political system (cf. Kristensen 1987b, 38 ff.). Under decision-making, we must pay particular attention to how and where such power is institutionalized. we dealing with a two-party or multi-party parliamentary system? Is there a precarious balance, or are either the 'market-weak' or the 'market-strong' parties enjoying a stable majority? Finally, how is policy-making on welfare organized with respect to the position of organized interests; is there competitive pluralism or monopolistic representation?

Given what has been said so far about the character of 'pluralist' and 'corporatist' institutional relationships, we may find another 'paradox of privatization'. Patterns of privatized decision-making, in the form of regulation, probably easier to establish in a 'corporatist' institutional setting. This is because ruling parties trust the interests to which they have afforded 'recognized' status more than they would do lobbying groups in a competitive institutional setting, where control is less easy to establish. The question - to be dealt with below - is whether this should or should not be dubbed 'privatization'.

In an institutional perspective, it is also important to find out how services
are financed. When a welfare service is paid for by income taxes, and
constitutes a highly visible item in the state budget, it may be more open

to attack than another service, financed through some sort of indirect tax or off-budget and 'hidden' among other fiscal items (cf. Schwartz 1987, 166 ff.). In terms of privatization, it is certainly also important to find out how payment for a service is connected to the delivery of that service. Evidently, it may be easier to privatize services where there is a rather direct relation of this kind, than with ones where the connection between payment and service is remote (Dunleavy 1986, 20). It is also of importance to trace the routes of money: is it given to one large line bureaucracy, which also delivers the service, or does it pass one or several transfer bureau layers before reaching the implementing agency? As Dunleavy (1986, 20) notes, if a bureaucratic institution is concerned with all his budget types - core, bureau, and program - there is a close connection between the bureaucrat's self-interest and the survival of his or her bureau. Thus, attacks on public financing attempted in such institutional settings may meet with difficulties.

Obviously, an important feature about implementing agencies is how they are integrated with decision-making and financing institutions: the more integration, the less vulnerability to attack. Another important feature of implementing institutions is the way in which they are connected to, and supported by, organized interests in society. As was pointed out above, ideal-type 'pluralist' welfare-state institutions render services only to marginalized', thus most often unorganized groups in society. If decisionmaking, and implementation functions are separated institutionally, if furthermore the financing is very 'visible' with a cleavage between payers and receivers, then the relatively 'unprotected' implementing agencies will be a natural target for attacks from those promoting non-public solutions. On the other hand, if the implementing institutions are supported by organized interests who also participate as 'recognized' parties in implementation, then such institutions will probably ward off these attacks more easily (cf. Ascher 1987, 265).

Towards a Predictive Theory of Privatization?

This paper began with the assumption that some of the theoretical work used to explain welfare-state expansion can be used 'in reverse' to explain privatization, when privatization is defined as the transfer from the public to the private sector of the responsibility for certain activities involved in welfare provision, i.e. as welfare state contraction. To this end, I examined different variants of the 'power resources' approach. The 'labour movement' thesis does not seem to have a fully predictive value. Then, I discussed the 'game theoretical' variant of the 'power resources' approach, finding its usefulness for predicting privatization limited because of the reductionism built into it.

Following the neo-Institutionalists' idea that actor preferences and strategies not only formed by, but are also reflected in, the institutional arrangements of the welfare state, I furthermore tried to develop institutional which should be of relevance to the development and patterns of privatization. In my view, the ideal-type dichotomy between 'pluralist' and 'corporatist' welfare states may be fruitful, especially in connection with a further delineation of decision-making, financing, and implementing institutional arrangements. Depending on how all these institutions differ, the extent and patterns of privatization will differ, even if the actor strategies predicted by the 'power resources' thesis are pointing towards a similarity of these patterns.

What remains is to bring these different theories and typologies together to find out if their combination could yield some testable predictions about the development and patterns of privatization policies in different nations. One way of doing this is to develop 'scenarios'; starting from the ideal-type dichotomy between 'pluralist' and 'corporatist' welfare states with their various institutional arrangements for decision-making, financing, and implementation, we try to predict how these interplay with the behaviour and strategies of 'market-weak' and 'market-strong' political parties under different 'power resources' conditions.

Starting with the ideal-type setting of the 'pluralist' welfare state, we have characterized it as institutionally weak and fragmented, with a clear divide between state and society, and with organized interests competitively lobbying for politicians' attention. Furthermore, alliances between welfare providers in the state apparatus and their weakly organized, 'marginalized' clients are rare. If there is a stable political majority for the party or parties representing 'market-strong' segments in society, we may either be left with the paradox of having nothing left to privatize because of earlier such decisions, or find a lively privatization process under way. But what pattern could be predicted? Assuming the commitment to market solutions and the 'market-strong' power base of the ruling political party as given (cf. Ascher 1987, 48), we may predict that production of welfare services will be transferred to firms competing in the market. This is so because such parties believe that competition between providers will result in more efficiency, i.e. more services at lower costs. As was said earlier, such privatization is even more likely where implementing institutions are separateentities where a rather direct relation can be established between payment and delivery of the service (cf. Ascher 1987). The state will also shed responsibilities for financing the services. On the other hand, we may predict a continued public regulation. This is because the separation between the state-governing party on the one hand, and the firms competing in the market on the other, makes ideological 'recognition' and trust less reliable as control mechanisms than in a 'corporatist' setting (cf. Heald and

Thomas 1986, 63). We may call this expected pattern one of franchise (for
the origin of this and following labels, cf. Kristensen 1987a, 224 ff.).

If we have a competitive balance between 'market-weak' and marketstrong' in a 'pluralist' institutional setting, then we may predict that the policies toward privatization will be more marked by power calculi than when either party - or group of parties - enjoys a strong majority. From the 'market-strong' parties, we may expect some moderation in their privatization strategies. In particular, it could be predicted that such parties will opt for privatizing welfare-service production not only to firms competing the market, but also to voluntary organizations. They will do so in order to gain votes from crucial constituencies who are 'boundary cases' in terms of market strength. This also means that a 'market-strong' majority will leave much of the public financing intact. We can label this alternative contracting out.

In such a balanced power situation, a party representing 'market-weak' segments will presumably opt for leaving regulation of certain services to such organized interests which - in a future 'corporatist' structure - would be their allies as 'recognized' interests. By such privatization, the marketweak' can hope to build a stronger power base, enabling the party to pursue a future, more public-oriented mix which is more satisfying to its core groups, i.e. the classes weak in power resources in the market. We can also predict a clearer preference for public financing than is the case with the competing 'market-strong' party. Furthermore, the strategic calculus of the 'market-weak' party will also point towards the positive 'power resources' effects of transferring some of the production activities to potentially 'recognizable' interests. We may thus expect a pattern of deregulation and quasi-'contracting out\ although this time for corporatist control rather than for competitive efficiency.

A 'market-weak' party reaching a strong majority in a 'pluralist' welfare state will most probably have gained this position as a result of voter dissatisfaction with present welfare conditions. We can therefore predict very few moves towards privatization; they would go against the fundamentalinterests those voting that party into power. Rather, we expect welfare reforms which transfer public money to the 'market-weak', in combination with the establishment of stronger government agencies. At the same time, the incoming political majority may begin moving towards a 'corporatist' structure by drawing 'recognized' interests closer to the state, giving them responsibilities in both the production and regulation of welfare services (cf. Rein and Rainwater 1986, 39). This will particularly be the case in welfare sectors important to potential future voters. By giving such organized interests self-regulating powers, they may be won over as a secured power resource. While this may seem like a strengthening of the

public sector rather than privatization, it still represents a mixed publicprivatepattern:
may call it general transfers and self-regulation.

Although it may be borne by a strong voter-urge for more freedom of choice, a 'market-strong' party coming into a majority position in a hitherto 'corporatist' welfare state will be facing tremendous institutional obstacles when trying to privatize welfare programs. The 'recognized' interests which are sympathetic to the 'corporatist' welfare institutions are closely intertwined political and bureaucratic institutions, and will fight against privatization proposals which could weaken their 'power resources' or threaten the public programs from which their members benefit (cf. Ascher 1987, 265). We can thus predict that a 'market-strong' party will try to break up the 'corporatist' structure by privatizing the production of welfare services where there is a direct relation between payment and services, and where the organized interests have a weaker institutional position. Privatizing production will bring about competition, and break the monopoly 'recognized' interests. At the same time, we can predict a mixed pattern of regulative activities: where 'recognized' interests have been strong, there will be 're-regulation' to enable government control; where they have been weak, we can expect less regulation. We therefore expect a combined pattern of contracting out, franchise, and deregulation.

When there is a balance between 'market-strong' and 'market-weak' parties in a 'corporatist' institutional setting, privatization will follow different The 'market-strong' party or parties will try to break up the structure even more eagerly than when they enjoy a strong majority: in this way, it may be possible to achieve such a majority. Thus, welfareservice will be privatized to achieve a more pluralistic competition. will, however, be kept public, especially for services that are particularly cherished by 'market-strong' voters. Regulation will be brought more firmly into the realm of public decision-making, to diminish the 'power resource' base of the 'recognized' interests. We thus expect a pattern of franchise and vouchers.

For 'market-weak' parties, the situation calls for increasing their 'power resources' base by giving more responsibility to 'recognized' allied interests, keeping as much as possible of the welfare publicly financed to calm 'market-weak' voters, and trusting allied interests with self-regulating powers. Again, we expect a pattern of self-regulation in combination with transfers.

What was just said points to the problem of how to treat privatization in the 'corporatist' welfare state, particularly when a 'market-weak' party enjoys a stable majority with regard to political 'power resources'. Here, we will find 'recognized' interests well entrenched in the bureacratic institutions,or public functions in close cooperation with the marketweak'party the bureaucracy. Although giving authority to 'recognized'

interests is formally equivalent to privatization, the incorporation of these interests into state programs in reality lends them a quasi-public character. This type of 'privatization' could equally well be seen as a form of decentralization'to interests, who can be trusted because of their close ideological affiliation with the 'market-weak' party's policies. We may particularly expect to find production of welfare services allocated to such trusted interest organizations, who also enjoy a considerable amount of self-regulation. However, we can be rather sure that financing is kept within the sovereign control of the governing party: indeed, public financing may be seen as the key to the 'corporatist' system. In the well-entrenched 'corporatist' welfare state, we may expect a public-private mix characterizedby transfer payments and (non-profit) service productionby interests.

We thus arrive at a number of predictions about dominant privatization
patterns resulting from the interplay between institutions and actors.

Admittedly, this effort to link the phenomenon of privatization to more general theories about the development and fate of the modern welfare state is rather sketchy - it needs further development and refining. However, the work to formulate predictive theories about privatization is a necessary

one. The present dominance of normative and highly ideological studies concentrated on proving the superior efficiency of privatization compared to public-sector provision should give way to a more nuanced and theoretically more sophisticated way of looking at the subject.

ACKNOWLEDGEMENTS

Research for this article was made possible by a grant from the Bank of Sweden Tercentenary Foundation. Thanks are due to Stephan Schmidt for incisive, yet inciting comments on earlier drafts. I am also indebted to participants in the NSIBR Annual Workshop on Housing Policy, the ECPR Workshop on Privatization, and the Swedish Political Association's Workshop on Comparative Politics, for comments and suggestions.

REFERENCES

Adams, C. T. 1987. 'The Politics of Privatization', in Turner, 8., Kemeny, J. and
Lundqvist, L. J., eds., Between State and Market: Housing in the Post-industrial Era.
Stockholm: Almqvist & Wiksell International, pp. 129-157.

Ascher, K. 1987. The Politics of Privatization. Contracting out Public Services. London:
Macmillan.

Castles, F. G. 1982. 'The Impact of Parties on Public Expenditure', in Castles, F. G.,
ed., The Impact of Parties: Politics and Policies in Democratic Capitalist States. London:
Sage, pp. 21-96.

Cawson, A. 1978. 'Pluralism, Corporatism and the Role of the State', Government and
Opposition 13, 178-198.

Crouch, C. 1983. 'Pluralism and the New Corporatism: A Rejoinder', Political Studies
XXXI, 452-460.

Donnison, D. & Ungerson, C. 1982. Housing Policy. Harmondsworth: Penguin.

Dunleavy, P. 1986. 'Explaining the Privatization Boom: Public Choice versus Radical
Approaches', Public Administration (London) 64, 13-34.

Esping-Andersen, G. 1985. Politics Against Markets. The Social Democratic Road to
Power. Princeton, NJ: Princeton University Press.

Forrest, R. & Williams, P. 1984. 'Commodification and Housing: Emerging Issues and
Contradictions', Environment and Planning A, 16, 1163-1180.

Gough, I. 1979. The Political Economy of the Welfare State. London: Macmillan.

Harloe, M. & Paris, C. 1984. 'The Decollectivization of Consumption: Housing and Local Government Finance in England and Wales, 1979-1981', in Szelenyi, 1., ed., Cities in Recession, Critical Responses to the Urban Policies of New Right. London and Beverly Hills: Sage Publications, Sage Studies in International Sociology, vol. 30, pp. 70-98.

Heald, D. 1984. 'Privatization: Analysing its Appeal and Limitations', Fiscal Studies 5,
36-46.

Heald, D. & Thomas, D. 1986. 'Privatization as Theology', Public Policy and Administration
1, 46-66.

Heclo, H. 1974. Modern Social Politics in Britain and Sweden. New Haven, CT: Yale
University Press.

Henig, J. R., Hamnett, C & Feigenbaum, H. R. 1988. 'The Politics of Privatization:
A Comparative Perspective', Governance: An International Journal of Policy and
Administration 1, 442-468.

Jordan, A. G. 1981. 'Iron Triangles, Woolly Corporatism and Elastic Nets: Images of the
Policy Process', Journal of Public Policy 1, 95-123.

Kemeny, J. (1980) 'Home Ownership and Privatization', International Journal of Urban
and Regional Research 4, 372-388.

van Kersbergen, K. & Becker, U. 1988. 'The Netherlands: A Passive Social Democratic

Welfare State in a Christian Democratic Ruled Society', Journal of Social Policy 17,
477-499.

Kielland, E. 1986. 'Privatisering - troll med mange hoder', Tidsskrift for samfunnsforskning
27, 195-216.

Korpi, W. (1978) The Working Class in "Welfare Capitalism. London: Routledge.

Korpi, W. (1987) 'Power, Politics, and State Autonomy in the Development of Social
Citizenship: Social Rights During Sickness in 18 OECD Countries'. Stockholm: Swedish
Institute for Social Research (mimeo).

Kristensen, O. P. 1987a. 'Privatization', in Jan Kooiman, J. & Eliasen, K. A., eds.,
Managing Public Organizations. Lessons from Contemporary European Experience.
London and Beverly Hills: Sage Publications, pp. 223-238.

Kristensen, O. P. 1987b. Vaeksten i den offentlige sektor. Institutioner og politik.
Kobenhavn: Jurist- og Okonomforbundets Forlag.

Leat, D. 1986. 'Privatization and Voluntarization', Quarterly Journal of Social Affairs 2,
285-320.

Le Grand, J. & Robinson, R. 1984. 'Privatization and the Welfare State: An Introduction',
in Le Grand, J. & Robinson, R., eds., Privatization and the Welfare State. London:
Allen & Unwin, pp. 1-18.

Levi, M. 1987. 'Theories of Historical and Institutional Change', PS XX, 684-688.

Lorentzen, H. 1987. 'Privatisering: Begreppets innehold og implikasjoner', Tidsskrift for
samfunnsforskning 28, 259-280.

Lundqvist, L. J. 1988. 'Privatization: Towards a Concept for Comparative Policy Analysis',
Journal of Public Policy 8, 1-19.

March, J. G. & Olsen, J. P. 1984. 'The New Institutionalism: Organizational Factors in
Political Life', American Political Science Review 78, 734-749.

Niskanen, W. A. 1971. Bureacracy and Representative Government. Chicago: Aldine-
Atherton.

Offe, C. 1984. Contradictions of the Welfare State. London: Hutchinson.

Olsen, J. P. 1985. 'Nyinstitusjonalismen og statsvitenskapen', Statsvetenskaplig Tidskrift
88. 1-14.

Olson, L. 1982. The Political Economy of the Welfare State. New York: Columbia
University Press.

Pampel, F. C. & Williamson, J. B. 1985. 'Age Structure, Politics, and Cross-National
Patterns of Public Pension Expenditures', American Sociological Review 50, 782-799.

Piven, F. F. & Cloward, R. A. 1971. Regulating the Poor. New York: Vintage.

Ouadagno, J. 1987. 'Theories of the Welfare State', in Scott, W. R. & Short, J. F., eds.,
Annual Review of Sociology 13. Palo Alto: Annual Reviews, pp. 109-128.

Rein, M. & Rainwater, L. 1986. 'The Institutions of Social Protection', in Rein, M. &
Rainwater, L., eds., Public/Private Interplay in Social Protection. A Comparative Study.
Armonk, NY and London: M. E. Sharpe, pp. 25-56.

Rimlinger, G. 1971. Welfare Policy and Industrialization in Europe, America, and Russia.
New York: Wiley.

Ruggie, M. 1984. The State and Working Women. A Comparative Study of Britain and
Sweden. Princeton, NJ: Princeton University Press.

Schmitter, P. 1974. 'Still the Century of Corporatism?', Review of Politics 36, 85-131.

Schwartz, N. H. 1987. 'The Relation of Politics to the Instruments of Housing Policy', in
Turner, 8., Kemeny, J. & Lundqvist, L. J., eds., Between State and Market: Housing
in the Post-industrial Era. Stockholm: Almqvist & Wiksell, pp. 156-185.

Shackleton, J. R. 1984. 'Privatization: The Case Examined', National Westminster Bank
Quaterly Review, May, 59-73.

Shalev, M. 1983. 'The Social Democratic Model and Beyond: Two Generations of
Comparative Social Research on the Welfare State', Comparative Social Research 6,
315-352.

Skocpol, T. 1980. 'Political Response to Capitalist Crisis: Neo-Marxist Theories of the
State and the Case of New Deal', Political Sociology 10, 155-201.

Skocpol, T. 1984. 'America's Incomplete Welfare State: The Limits of New Deal Reforms

Skocpol, T. 1985. 'Bringing the State Back In: Strategies of Analysis in Current Research',
in Evans, P. 8., Rueschmeyer, D. and Skocpol, Th., eds., Bringing the State Back In.
New York: Cambridge University Press, pp. 3-43.

Skocpol, T. & Amenta, E. 1986. 'States and Social Policies', in Turner, H. & Short, J.
F., eds., Annual Review of Sociology 12. Palo Alto: Annual Reviews, pp. 131-157.

Stephens, J. D. 1979. The Transition from Capitalism to Socialism. London: Macmillan.
Torgersen, U. 1984. 'Kalenderen som statsmakt', Samtiden 2, 35-38.

Wilensky, H. 1975. The Welfare State and Equality: Structural and Ideological Roots of
Public Expenditures. Berkeley: University of California Press.