This article was downloaded by: [Vrije Universiteit, Library] On: 1 June 2011 Access details: Access Details: [subscription number 907218019] Publisher Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK Australian Journal of Political Science Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713404457 Political Stability in Divided Societies: A Rational-Institutional Explanation Hans Keman Online publication date: 09 June 2010 To cite this Article Keman, Hans(1999) 'Political Stability in Divided Societies: A Rational-Institutional Explanation', Australian Journal of Political Science, 34: 2, 249 — 268 To link to this Article: DOI: 10.1080/10361149950407 URL: http://dx.doi.org/10.1080/10361149950407 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Australian Journal of Political Science, Vol. 34, No. 2, pp. 249±268 Political Stability in Divided Societies: A Rational±Institutional Explanation H ANS K E MAN * Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 Vrije Universiteit Amsterdam Analyses of consociationalism and corporatism are based, more often than not, on descriptive case studies. These concepts are considered to explain the stability of the political systems, where one would expect that con¯ict and stalemate in dem ocratic decision-m aking were the rule rather than the exception. In this article an attempt is made to develop a more general argument, based on the idea of a structure-induced equilibrium , to analyse institutional adaptation and change in plural societies. Central to this approach, which I label `rational institutionalism’, are the concepts of `room to manoeuvre’ for political and societal actors, on the one hand, and the `feasibility of political choice’ allowing for m ore space to develop positive sum results for all involved, on the other hand. This argument is ®rst elaborated a priori and then applied to Dutch consociationalism and corporatism. These empirical illustrations demonstrate that a logic of interaction has been developed in the Netherlands, which is not a feature of the Dutch case alone, but can be understood as a form of institutional adaptation and change. This institutional approach can be considered as an analytical instrum ent to explain political stability under differing circumstances cross-nationally and across tim e in democracies. A central paradox of parliamentary democracy is its ability to combine ¯exibility and change with an underlying stability. Various comparative explanations have been advanced to account for this (see, for instance, Daalder and Mair 1983; Lane and Ersson 1994). Many of them are ¯awed, however, by their failure to systematically relate the type and occurrence of political action to those institutions that affect collective decision-making in pluralist democracies (Shepsle 1995, 281±2; Scharpf 1998; Olsen 1998). Since the 1970s two concepts have been developed in the ®eld of comparative politics, which attempt to break away from the m ainstream literature on the causes and consequences of political stability and democratic perform ance in plural societies (eg Lipset 1963; Dahl 1971; Powell 1982; Almond, Powell and Mundt 1993). These challenge the general argument that plural societies, characterised by strong political divisiveness, would produce unstable politics and volatile types of governance. It is argued instead that patterns of institutionalised behaviour could * This article has been discussed at the Netherlands Institute of Advanced Studies (NIAS) and the Research School for the Social Sciences (RSSS/Canberra). I am grateful for the helpful comments of the discussants and in particular for the comments of Ian Budge (Essex) and Philippe C. Schmitter (EUI, Florence). 1036-1146/99/020249-20 Ó 1999 Australasian Political Studies Association Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 250 H . KEMA N and would emerge under these circum stances that counteract the feared ill-effects of societal heterogeneity and political divisiveness. These alternative explanations incorporate concepts like consociationalism (Lijphart 1975, 1977; Daalder 1974; Steiner 1974; Lehmbruch 1967) and corporatism (Lehm bruch 1979; Schmitter 1981; Katzenstein 1985; Scharpf 1987). These concepts claim that pluralism is not the only, nor even the best, explanation of the institutional perform ance of democracies. On the contrary, it is argued that the pure concept of democratic pluralism is inadequate to understand the political behaviour of the relevant actors and consequent democratic performance of socially divided societies (Offe 1979; W eaver and Rockm an 1993; Keman 1997b). Consociationalism and corporatism attempt to explain the relation between societal con¯ict and political consensus by means of institutionalised patterns of collective choice. Institutions are de®ned as sets of rules which occur in social practice in the form of appropriate m odes of conduct for political and societal actors. This is particularly important if and when the actors involved are having con¯icts that m ay well lead to recurrent behaviour that com plies to those rules and jeopardise political stability (Ostrom 1990; Scharpf 1998). These institutions, ie the form al and informal `rules of the political game’, have emerged under conditions of societal strife which tended to produce a stalemate between the con¯icting actors in a parliamentary democracy (Steinm o, Thelan and Longstreth 1992, 2; Keman 1997b, 10). Managing these political crises appeared not only possible but in due course produced standard operating procedures accepted by m ost actors involved, over and over again (L ijphart 1977; March and Olsen 1989). It was exactly such developments which have been explained by the concepts of consociationalism and corporatism. However, both models of collective decision-making are merely developed from descriptive analyses of speci®c cases (eg The Netherlands, Belgium , Austria, Sweden and Switzerland; see, for instance, Rothstein, in Steinm o, Thelen and Longstreth 1992, 51±3; Tsebelis 1990, 162±3; Keman 1993a, 157±9). There has been some doubt, therefore, whether or not these explanations can travel across countries and time, as has been claimed by their protagonists (in particular by Lijphart 1977, 1989; Lehm bruch 1984). One of the aims of this paper is therefore to discover the extent to which consociationalism and corporatism , orÐ m ore broadlyÐ systems of associational interest intermediation (Lehm bruch 1998) work under changing circum stances and differing conditions and are still capable of furthering consensus where con¯ict appears inevitable in pluralist democracies. This requires us to develop a more systematic approach to analyse the underlying m echanisms that may well explain the effective operation of consociationalism and corporatism across countries and time. W e need to develop a conceptual fram ework of collective decision-m aking regarding societal problems-cum-con¯icts that is not primarily based on pathdependent accounts alone. Path dependency is an approach to demonstrate the discrete working of institutions that have historically emerged in a (set of) countries. Such an analysis is useful to understand the constraints for action and subsequent choice within a system per se (see Steinmo, Thelen and Longstreth 1992, 28; Putnam 1993, 7±8). Instead m y point of departure will be that collective actors tend to pursue societal interests rationally through actions that are shaped by existing or evolving institutions. Hence, political action is not only considered as being determined by form al rules, but is also in¯uenced by the preferences of the POLITICAL STABILITY IN DIVIDED SOCIETIES 251 Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 actor involved. In addition to rules-for-use, one must take into account the existing or emerging rules-in-use (see Laver 1986; Ostrom 1990; HeÂritier 1998). This kind of `rational institutionalism’ m ay well account for the paradox of con¯ict and consensus in various societies with differently institutionalised m odes of decision-making (Shepsle 1995; Scharpf 1998). In what follows I shall ®rst develop m y argum ent concerning rational institutionalism by discussing brie¯y the principles of `rational choice’ and `new institutionalism’ as an analytical approach to political action. I shall then introduce a conceptual framework which is intended to capture the relationships between actors in con¯ict and the consensus that is achieved. This raises, and hopefully answers, questions as to how and to what extent institutions matter in terms of problem-solving. Finally I shall try to demonstrate that such a general approach is applicable by analysing the Netherlands under conditions of consociationalism and corporatism. The Paradox of Con¯ict and Consensus: Political Decisions and Societal Problem-solving One of the central questions in political science is about the way in which societal con¯icts can be constructively channelled into problem-solving. Con¯icts appear to the actors concerned as a zero-sum game. If societal con¯icts are indeed `solved’ in an unilateral and one-sided fashion (for instance, by ignoring substantial minorities due to simple majority voting), this will in the long run, according to the logic of game theory, inevitably lead to a situation of a sub-optimal outcomes for all participants (albeit, of course, in a different degree across the population in a society). In form al political theory this situation has been m odelled by the (well-known) Prisoner’s Dilemm a. Even where situations are not so con¯ictual sub-optimal outcom es may emerge as a result of `free riding’. A consequence of political action based on the self-interest of collective actors (Ostrom 1990, 38; Scharpf 1998, 46±50; Laver 1986). This explanation underlines the so-called micro/macro-paradox of politics and society: rational actors pursuing their interests by m eans of societal interaction whilst being dependent on others, m ust act strategically within the rules of the political game to achieve their individual utility. However, this process tends to yield optimal results (instead of m aximum pay-offs) for all actors involved.1 This paradox becom es apparent in societal con¯icts, if and when politicised (ie forming part of the political agenda subject to the democratic rules of decision-making). By de®nition, the public outcome cannot satisfy all the preferences of the individuals or groups concerned. In other words, political decision-m aking is inevitably characterised by the m icro/m acro-paradox. It is literally paradoxical, since the democratic polity can only produce public goods (measures and rewards) to satisfy individual desires (Taylor 1987; Ostrom 1990, 15±18; Laver 1997). Yet, the 1 Note that the use of the term `optimum’ is different here from what is meant by a Pareto-optimal solution, ie the `social welfare’ function in social choice literature (see Mueller 1989, 384 ff.), nor does it point to the application of (economic) norms of `ef®ciency’ and `effectivity’ as employed by the Virginian school approach to problems of `public choice’ (see Lane 1993, 150 ff.). Essential to our understanding is whether or not the actual outcome of the political process is acceptable to all concerned, rather than that it represents the perfect result one can theoretically construct, either on the m icro-level (ie social choice) or on the macro-level (ie public choice). As far as `choice’ is involved, we mean political choices shaping mandatory policies in a society by means of public goods (see Scharpf 1998). Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 252 H . KEMA N existence and working of form al and inform al democratic institutions, which aid consensus formation, are essential for understanding to what extent this is possible or not. An example of this effect of institutions is the majority rule governing decisionmaking in parliament. This could lead to stalemates in decision-making, eg `voting cycles’ in democracies (Mueller 1989, 63±89). In practice, these situations have been avoided in two ways: (1) by means of com prom ises among decision-m aking actors (hence, exchange in order to cooperate); (2) by adapting form al procedures, introducing additional rules concerning agenda-setting and rank-ordering of issues to decide on.2 Consociationalism must be seen as a process of redistributing certain policy preferences by m eans of trading off. The underlying m echanism is an ability to avoid a stalemate or voting cycles without sub-optimal outcom es for the actors involved. Another example of the paradoxical effects of democratic institutions is given by Mancur Olson. If, indeed, political and societal actors behave rationally, then it follows that individuals will not participate in collective action if they expect that the related pay-offs will come their way anyhow. This may well lead to less societal cooperation and to m ore self-interested behaviour (and often to free riding). According to Olson (1982) this is due, in part, to individual rationality and, in part, to the (inherent) working of democratic institutions. Avoiding such a situation would imply either a complete overhaul of the democratic institutions, or the incorporation of several `distributional coalitions’ into an encom passing institutional arrangement under the aegis of the democratic state (Olson 1986). In fact, the latter option is the only feasible one, and this impliesÐgiven the constitutional design of a liberal democracyÐthat the extent to which its decision rules allow for defection, obstruction and thus non-cooperation between societal actors and the state should be taken into account in terms of reducing and subsequently accomm odating this type of self-interested behaviour (cf Scharpf 1998, 56). This line of reasoning is exactly the foundation of those who see corporatism as associational interest intermediation. This model represents an alternative to the perverse consequences of the pluralist politics of interest intermediation as envisaged by Olson on the basis of exchange relations (Keman, Paloheimo and W hiteley 1987; Colom er 1996; Czada 1998). Hence, de®ciencies of democratic polities are in practice solved by adapting and extending institutions to reduce con¯ict in democratic society. Both consociationalism and corporatism can be seen as examples of additive institutionalisation. These m odes of collective decision-making are in fact a combination of formal rules, on the one hand, with inform al procedures, on the other. This observation underlines the need to develop empirically based m odels of exchange relations between societal and political actors which specify the patterns of interaction between institutions and the political choices made by actors. Taking the micro/macro-paradox as a point of departure with regard to societal problemsolving is a valuable approach in delineating the extent to which institutions m ay facilitate coordination and regulation of those problems. W ithout institutional devices, these situations cannot be solved adequately, the ill-effects of the m icro/ 2 It is precisely these practices which have been described in consociationalism by Lijphart (1975) and which laid the foundation of his model of consensus democracy as opposed to the `majoritarean’ types of democracy (Lijphart 1984). POLITICAL STABILITY IN DIVIDED SOCIETIES 253 macro-paradox of individual interests and preferences cannot be reduced, nor can the resulting societal con¯icts be resolved and subsequent political consensus be attained (Keman 1996). The ®rst assignm ent is then to develop a conceptual approach that speci®es how and to what extent interactions between institutions and actors are relevant in producing a stable political order from the process of political decision-making and related outcom es. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 The Interactions between Actors and Institutions The extent to which there is an institutionalised political order of society that is capable of producing a balance between con¯ict and consensus, which is enduring and produces a minimal loss to any individual or group within a society, is what I call a structure-induced equilibrium capable of explaining the relative stability of the political order. This concept, introduced by Shepsle (1995, 284; see also Colom er 1996, 4), points to a situation in which there is no alternative choice allowed by the rules of procedure and rules of the game available to the actors who possess veto power or voting power. In this way the existence and working of the rules of a political com munity shape the room to m anoeuvre for policy choices. The concept of a structure-induced equilibrium com bines institutional features of a polity with the strategic behaviour of actors who have m ore or less con¯icting preferences. Studying the development and working of political institutionsÐ which structure the behavioural patterns of the actors involvedÐis thus crucial to a better understanding of the kind of equilibrium that is tenable under conditions of strong societal con¯ict (HeÂritier 1998; Keman 1997a). Institutional arrangements regulate the behaviour of political and societal actors who are considered to be both rational and interdependent in a democratic polity. The concept of political order can be conceived as consisting of three dimensions: politics, polity and policy (Powell 1982; Scharpf 1987; Keman 1993b, 43±7). Politics is the political process: actors (m ostly aggregates of individuals organised in parties, associations, or interest groups) interact with each other when they have con¯icting interests or preferences regarding societal problems that are characterised by a de®ciency of self-regulation. The process of solving those problems becomes visible through the institutions that have emerged in order to facilitate con¯ict resolution. Institutions are meant to develop coalescence and to achieve a consensus among con¯icting actors through developing alternative preferences. These alternative preferences represent, in theory, the solution, ie that alternative which is acceptable to all concerned and enhances cooperation. Institutions m anifest themselves in the rules of the game in a society. This game is then directed by the existing fram ework of rules or the polity. Institutions are considered to be either form al, like, for instance, those enshrined in a constitution, which can be enforced by m eans of authority, or as informal rules, ie rules evolving over time are respected and followed as a code of conduct by m ost actors involved. As a result, non-compliance or defection leads to sub-optimal results for the player or the exclusion from the game. The ®nal result of this process can be observed by m eans of policy choices which re¯ect the extent to which con¯icts have been resolved (or not) in a way that it is conducive to an equilibrium. The urge to solve societal con¯icts and thus the need for effective rules which facilitate a political consensus among contestants can be understood as follows: ®rst, rules reduce uncertainty among the actors involved, hence they can act Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 254 H . KEMA N strategically with an eye to optimal solutions; second, rules provide room for exchange and bargaining for those problems that are described by Ostrom (1990) as `com mon pool requirements’ (in order to avoid a `tragedy of the com mons’) on the one hand, and as `public problems’ on the other; third, policy-making by m eans of consensus and concertation may well decrease external costs to a society and thus the perverse effects of free riding. The options chosen or decided upon for political action to solve a problem are what we shall call policy formation. This process is equivalent to what others would call state intervention or the `authoritative allocation of values in a society’. Actions of the state, or a related allocating agency, are in this conceptualisation of the `political’ viewed as relatively independent from societal interests (Skocpol 1985, 45). That is to say, political action, ie the relation between politics and policy-making, requires a degree of autonom y in order to be feasible and effective. It is precisely the existence and working of institutions which m akes this possible. This abstract description of the political process can be observed almost every day: for example, pollution, unemployment or crime are public problems which become recognised as political issues (by parties and the m edia). Given the nature of the problem and the societal response, on the one hand, and the position taken up by politicians, on the other, policy action will follow (nb: non-decisions or non-action are also policy formation!). Yet, the extent to which the `public’ recognises that this result is the outcome of the (democratic) political process is then also an indication of its democratic perform ance, ie the perception of an effective political order. This interactive process between political institutions and rational actors will be our theoretical point of departure in order to assess the way in which societal con¯icts and interdependent choices of actors reach the political arena and how subsequent decision-making takes place. In the next section I shall elaborate the underlying mechanisms which m otivate and drive actors into action as well as limit their actual scope for action. This refers to the concepts of `feasibility of political choice’ of actors, on the one hand, and the `institutional room to m anoeuvre’ of actors, on the other (Keman 1998). Both concepts are crucial for understanding how and when a structure-induced equilibrium is attainable. Rational Institutionalism and Political Action In recent times institutionalism has been revived by com parative analysis of the relation between politics and policy formation (March and Olsen 1989; Budge and Keman 1990; Scharpf 1987, 1992; Putnam 1993; Czada, HeÂritier and Keman 1998). This type of research often provides empirical evidence demonstrating that political institutions add signi®cantly to the understanding of the cross-national and inter-temporal variation in the governing capacity of democracies (Keman, Paloheimo and W hiteley 1987; Scharpf 1992; Steinm o, Thelen and Longstreth 1992; Weaver and Rockman 1993; Schm idt 1996). The question is whether political decisions and actions are more adequately understood in terms of intentional behaviour directed by individual preferences or as the outcom e of institutional arrangements and related procedures. On the level of politics and policies, our concern here, individually guided behaviour is not prim arily relevant. By de®nition political actors perform at the meso-level of society this level of (organised) action or forms of (political) participation. The Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 POLITICAL STABILITY IN DIVIDED SOCIETIES 255 individual voter counts less than the aggregated votes and the way they are translated into representation. The single demonstrator and his/her slogan has less of an effect than the extent to which the slogan is shared by others and is transformed into collective action that is relevant to the political arena. 3 This point of view also implies that we need ®rst to discuss the tenability of the notion of (individual) `preferences’ on the level of politics (L aver 1997, 20). First of all, preferences are not only m anifold, but are also shaped by social and econom ic conditions. Second, preferences are not `given’ but are strongly in¯uenced by their sociocultural environment (E lster 1979). Third, under conditions of societal con¯ict it is better to speak of `needs’ than of preferences (L aver 1986). In politics and the related process of political choice, individual `needs’ show up through the actions of organised actors. The various positions related to `needs’ and taken up by relevant collective actors with respect to societal con¯icts guide their behaviour and will drive them into political action. However, actors live in an interdependent world and subsequently have to take the behaviour of other actors into account. This implies that a political actor, representing a rationally de®ned and rank-ordered notion of `needs’ underlying societal con¯ict, will often forego a short-term gain in order to attain an agreement that is both enduring and closer to his/her own goals in view of improving public welfare (HeÂritier 1998). Political actors represent an aggregate of individual preferences and a collectively shared and rank-ordered de®ned set of (urgent) needs, as well as operate in an interdependent world ®lled with other rational actors, implies that political decisionmaking is characterised by the feasibility of political choice of each actor involved. Thus, optimal goal achievement directs behaviour in the political game of decisionmaking rather than m aximalisation of interests. 4 Institutions are seen in this light not just as constraints but also as opportunities to enhance stable outcomes. The ®rst concept essential for understanding the relation between con¯ict and consensus in terms of concerted political action (politics) and collective decisionmaking (policy-formation) is: the feasibility of political choice. Every (organised) political actor participating in collective decision-making will work from a rankordered set of `needs’ whose feasibility is, however, also dependent on the (rational) behaviour of other actors. If two (or more) parties compete in the same realm of `needs’ then two possibilities to com e to an agreem ent emerge (assuming a competitive situation): either on the basis of complementary or of parallel needs. 5 3 Hence there is a `collective action’ problem that precedes much of the successful organisation of political action (Olson 1965). While this may be a problem for the political actor per se, it is not our concern here. I investigate political actors and their interdependent relations and assume that this type of actor represents the intentions of its followers. See also Laver and Scho®eld (1990, Appendix A, 217±44). 4 This m eans by giving away m aximalisation that the concept of rationality becomes quite open-ended or even obscure, yet in my opinion the pure concept of economic rationality obscures most of political action. Unlike the notion of the `market’ where one can trade utilities on the level of the individual and on equal footing, the political `arena’ is fundamentally different: politics is about establishing a balance between con¯ict and consensus, which is not the same as an equilibrium that is construed by an `invisible hand’, but by regulation and the authoritative allocation of values rather than the anonym ous exchange of utilities (see also North 1990; Mueller 1989; Scharpf 1992). 5 Complementarity is a situation in which two functionally different actors demand identical measures to achieve a societal situation that m eets their respective needs (eg organised employees and em ployers requiring policy initiatives to restore or stimulate econom ic growth). Parallelism is the situation in which Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 256 H . KEMA N Complementary needs will further the form ation of strategic coalitions, whereas parallelism will enhance cooperative strategies. If these situations occur it may well enlarge the feasibility of political choices (Czada 1998). If not, then agreem ents are more dif®cult to reach and require extensive bargaining. The extent to which a situation of optimal decision-m aking and concurrent policy-form ation may occur is, however, dependent on the institutional room to manoeuvre available to the actors involved. A number of authors have pointed out, quite correctly, that political and societal actors operate within different arenas of decision-m aking in a political system. In a democratic society, for instance, parties are active in civil society and in parliament and government simultaneously. Interest groups, on the other hand, are mainly based in civil society and seek access to the political arena through various channels (eg the bureaucracy or advisory boards, etc; Putnam 1993; W oldendorp 1995) Hence, it is important to know which actor has access to which arena. Each arena has its own rules, which are almost always both of a form al and of an informal nature. These rule con®gurations de®ne the institutional room to manoeuvre of the actors involved. Moreover, as these actors often take on roles in more than one arena, this may well affect the relative opportunities of an actor, andÐ vice versaÐ limit others in their room to m anoeuvre as they too have to play multiple roles (this relates to the epiphenomenon of `nested games’; see Tsebelis 1990; Ostrom 1990). In short, a set of institutions, or rule con®gurations, can be considered as opportunity structures for action. Finally, all institutions are created by political actors at some point in time. They re¯ect the modes of interaction between the interdependent actors involved. Form al rules are here seen as conditions under which decision-m aking takes place and which set the limits for rational m an in m aximising his own `needs’. Inform al rules, on the other hand, de®ne the options available to each actor, given their potential for defection and vetoing. Together both types of rule indicate the room to manoeuvre available to each political actor involved from an interdependent perspective. Political institutions are thus de®ned as sets of rules that shape (at least in part) the strategic behaviour of interdependent political actors and facilitate the process of decision-making, inducing the furthering of public welfare. Rules de®ne the room to m anoeuvre for the participating actors, whose rank-ordered needs in turn de®ne the feasibility of political choice. Both these analytical concepts offer thus the possibility of investigating the opportunities and limitations available in a society to contain con¯ict and to create consensus. The more the `needs’ of the interdependent actors involved are of a complementary or parallel nature, the greater the likelihood of cooperation and com prom ise between them in solving con¯icts (Scharpf 1998; Keman 1997b). In the next section this rational institutionalist approach will be applied to the concepts of consociationalism and corporatism in the Netherlands. This exercise is meant to show that the conceptual framework discussed in this section can be applied to a concrete case. The analysis will demonstrate that the underlying Footnote continued separately organised actors with similar needs operate within the political system (eg political parties with a comparable stance on the welfare state, but within a different ordering of needs which are to be met ®rst). Both situations can occur simultaneously. POLITICAL STABILITY IN DIVIDED SOCIETIES 257 mechanisms that have produced a structure-induced equilibrium in the Netherlands have wider theoretical implications beyond simply explaining the Dutch case per se. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 Consociationalism and Corporatism Explained by Rational Institutionalism The extent to which the so-called paradox of con¯ict and consensus can be solved depends by and large on the way the relevant sociopolitical actors are able to operate rationally. This capability is in turn in¯uenced by the institutional environment in which the political action takes place. The two basic concepts of rational institutionalismÐthe `feasibility of political choice’ of the actors involved and their actual `room to manoeuvre’Ðwill be applied below to consociationalism and corporatism and in the Netherlands to demonstrate the viability of the national± institutionalist approach (Keman 1993a, 1996). The rationale of both modes of collective decision-m aking according to its protagonists (viz Daalder 1974; Lijphart 1977, 1989; Lehm bruch and Schmitter 1982; Lehm bruch 1998), is that it preserves the existing political order, ie stability, by means of democratic institutions, although the societal structure of the speci®c cases they study tends to generate and to continue con¯icts. They also argue that consociationalism and corporatism often perform better than many other types of parliamentary arrangements in terms of political stability and public welfare (see, for example, Lijphart 1977, ch. 4; Katzenstein 1985, ch. 1). I shall analyse consociationalism and corporatism a priori in terms of the political behaviour displayed, when con¯ict appears predom inant and governance by consensus almost impossible. Consociational Democracy In plural societies with several cleavages, the related political con¯icts cannot be overcom e easily by m eans of m ajoritarean rule. Therefore the existing political± institutional fram ework for decision-making needs to be adapted or even changed. If not, a situation would occur with `civil strife rather than democracy. W hat these societies need is a democratic regime that emphasises consensus instead of opposition that includes rather than excludes’ (Lijphart 1977, 23). However, the liberal democratic rule con®guration as it is organised by m eans of an electoral system and a m ode of collective decision-making (by means of simple m ajorities) appears to misrepresent a (societally segm ented) population and is often bound to m ake non-decisions as a result of voting cycles created by larger minorities in parliament. In short, the form al system is at odds with the way the existing cleavage structure and related political interests within the Netherlands are organised and represent their needs. This impliesÐ from a rational institutionalist perspectiveÐthat the sociopolitical actors (ie parties) are in need of room to manoeuvre to negotiate in such a way that the formal institutions do not constrain positive-sum actions. Yet each actor, interdependent as they are, must seek opportunities to pursue their feasible set of political choice in order to improve their own welfare as well as the public welfare simultaneously. In terms of `needs’, each and every actor thus must be able to recognise the com plementarity and parallelism in the `needs’ pursued by themselves and by other participating actors. In addition, each actor must be convinced that these preferences can somehow be realised in the (near) future in an optimal way. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 258 H . KEMA N Three conditions to facilitate this process can be discerned in the consociational model (L ijphart 1977; Lehmbruch 1984). First, cooperative behaviour can only be implemented by coalition governm ent, which is effectively controlled by m eans of mutual com mitments agreed upon beforehand. Although the form al rules to make coalitions work can in part be derived from the constitution, in a large part they depend on the resources of each actor (eg the share of the votes; the threat of non-cooperation and so on). This condition of coalition formation essentially limits the room to m anoeuvre of all participating parties at the level of parliamentary decision-making. In addition, it requires a certain span of control by governing parties of their parliamentary delegations, as well as over their electoral followers. In other words, depending on the weight of a contested `need’, a party m ust attempt to detect complementarity and parallelism at the level of decision-making and must at the same time remain true to its original preferences in order not to lose its power-base. Hence, credible comm itment and an iterated type of bargaining de®ne the feasibility of political choice (Scharpf 1998; Tsebelis 1990, 164±8). The second condition of consociationalism is the representation of m ost `minorities’ (segments) through bicameralism, functional or territorial decentralisation and quali®ed rules which safeguard certain collective rights of m inorities.6 These rules often limit the feasible set of choice open to all actors. This second condition links the organisation of collective choice to the structure of the state. Elinor Ostrom has correctly emphasised that its institutional design m ust be related to the operational procedures that transform the outcomes of collective decision-making into credible comm itments that can be monitored by all actors (here: m inorities) involved (Ostrom 1990, 51±4). The third condition of consociationalism is that democratic representation must mirror society by m eans of a proportional representation (PR) electoral system in order to include m ost relevant societal segm ents. If not, some of the politically organised societal cleavages would not be included in the politics of problemsolving. This formal prerequisite of consociationalism also has an inform al signi®cance: societal actors must be able to organise their own segm ent politically in order to govern by m eans of cooperative behaviour. This makes optimal decision-making m ore feasible. Here, then, we have an institutional framework which m ediates con¯icts in such a way that m any actors involved can play a role in reaching an optimal agreement. Necessary for the success of such an institutional setting is the recognition by the actors of the opportunity to pursue their `needs’ in an optimal way, ie collective choice must lead to discernible pay-offs for all actors involved. W hether this contention holds true can best be observed in a situation of genuine sociopolitical con¯ict. Let us therefore take a closer look at the ®rst amicable agreem ent to solve a set of sharp sociopolitical con¯icts in the Netherlands: the political stalemate of 1917 (see Lijphart 1975). This illustration demonstrates not only the potential of a rational institutionalist explanation of a seemingly isolated event, but also that consociationalism can be seen as a new structure-induced equilibrium produced by the politics of problem-solving (see also North 1990; Keman 1996, 1997b; Shepsle 1995; Colomer 1996). 6 These formal rules are developed as a part of the overall need for `checks and balances’. Often they tend to function as impediments, in particular if and when the needs are fundamentally contested (see Schmidt 1996; Braun 1989). POLITICAL STABILITY IN DIVIDED SOCIETIES 259 Table 1. Rank-ordering of needs of Labour, Liberal, Christian parties, before 1917 Needs Universal suffrage and PR Public versus private schooling Welfare regulation Labour 1 3 2 Liberal 2 1( 1 ) 3 Christian 3 1( 2 ) 2 Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 ( 1 ) 5 in favour of public schooling only; ( 2 ) 5 in favour of subsidised private schooling. Rank-order: 1 5 an urgent need; 2 5 important; 3 5 indifference. In 1917 there were three (blocks of) parties in the Dutch parliament, Labour, Liberal and Christians each representing equally sized societal m inorities. Each block was in pursuit of urgent societal `needs’ that could only be attained by m eans of state intervention (see Table 1). There was no m ajority party, nor did constructive exchanges appear feasible, owing to the fact that each party used its power of veto to prevent others ful®lling its need (ie producing voting cycles). The price paid by each and every party was a sub-optimal outcome. According to Lijphart (1975) the agreement eventually reached was due to the `com mon sense of the leadership’ of each `pillar’ and the overwhelming feeling that the national interest was at stake. Hence a bargain was struck at the elite level and this set a precedent for achieving compromises by m eans of the institutional conditionsÐconsociationalismÐspelled out in this section (Daalder 1974). However, the story told in rational institutionalist terms is different: each political actor representing one of the societal segm ents had a rational (self)interest to pursue their own `needs’ regardless of the national interest or accomm odation. Partly because of electoral motives, partly because of ideologically driven m otives (see also Strom 1990), this resulted in a parliamentary stalemate. Because none of the parties had a simple m ajority it was strategically rational for each party to compare its own `needs’ with those of the other two party blocks in order to ®nd out the degree of parallelism (see n. 6). Each actor knew they could not gain anything without form ing a (temporary) strategic coalition with one of the other parties in parliament (interdependence) to create a m ajority. All parties had a rational understanding of this problem and an optimal m ode of conduct developed, not only based on a one-off trade-off and pay-off, but also on the calculation that by ®nding an agreem ent other optimal pay-offs m ight be realised. The actual outcom e of this process can be traced by looking at the rank-ordering of the `needs’ of each party in the period before 1917, which involved strong party-political differences over the electoral law, the subsidisation of private (ie non-secular) schooling, and the degree of welfare statism (based on Lijphart 1975; Daalder 1987; Kem an 1993a) (see Table 1). No agreement was possible as long as the ordering as set out in Table 1 rem ained imm ovable and parliamentary m ajorities required two out of the three parties. Each party bloc was big enough to prevent a decision on matters other than their own preferred need, and hence there was always a threat of a strategic veto by a blocking coalition. Yet it is also obvious that som e of the `needs’ could only be m et if they were considered to be either complementary or parallel. Hardly any parallelism appears in the rank-orders by each party (schooling ranked high by Liberals and Christians but in an opposite direction, hence there is a m aximum distance) except regarding welfare regulation for Christians and Labour. The latter party became in this situation the pivot or broker because it reduced its demand Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 260 H . KEMA N with respect to universal suffrage to manhood suffrage (as was demanded by the Christian block). As no party was vehemently against welfare regulation, it was possible to develop a package deal in which the remaining sources of con¯ict could be handled by means of com prom ising behaviour. The Liberals had com plimentary goals with Labour on suffrage and the concom itant introduction of a proportional representation (PR) electoral system. Labour was m ildly indifferent to private schooling (which had nothing to do with elitist education in the Netherlands) and quite adamant about the extension of welfare regulation. In short, because complementary and som e degree of parallel needs existed between each party, temporary majorities could be form ed on each of the separate issues. The package deal represented the optimal pay-off for each party and the only feasible set of political choice as indicated in Table 1. Yet this outcome was only feasible owing to the fact that the form al rule of simple m ajority voting on separate issues could be circumvented in 1917 by m eans of the package deal. If not, the stalemate due to veto-voting in parliament would have continued. Had this situation been a `one-shot game’, consociationalism as a m ode of democratic decision-making would not have come into existence. However, owing to the introduction of universal suffrage on the basis of a PR electoral system (in 1919), the distribution of power resources among the political actors changed. And it was this constitutional change which turned the practice of consociationalism into a new set of rules. In other words, the compromise that emerged within the existing room for m anoeuvre brought about institutional change. The new set of rules-in-use created a different room for m anoeuvre for each political actor and led to re-iterated types of negotiation preceding the process of collective decision-making (Andeweg 1989; Kem an 1996). The actual story of the origins of Dutch consociationalism is less important than the institutional practices that emanated from it. The crucial point is that the participants were forced to behave rationally, realising the conditions of interdependence as well as recognising that optimal pay-offs could be achieved only in terms of identifying parallel and complimentary `needs’. Com m on sense did indeed prevail, but not because national interest per se, or elite leadership, were crucial factors. Instead, a new set of formal and inform al rules was developed. These operational rules appeared to work under the condition that no party lost m uch more than the others (hence avoiding a sub-optimal outcome for themselves and preventing m aximal results for others). Thus the stability of the political order was enhanced. By using concepts derived from rational institutionalism, it is possible to interpret the development of the consociational theme in a m ore structured way than hitherto done: the concepts of political room to m anoeuvre and feasibility of choice appear useful in helping us understand how contradictory situations of societal con¯ict can be turned into agreement and subsequent political consensus. By employing the notions of complementary and parallel ordering of `needs’ it is possible to detect the underlying mechanism that enables actors to achieve an equilibrium within the existing institutional context. As a result it can be argued that consociationalism is not a different type of democracy or a deviant case, but one of the existing forms of democratic regime, it is not a unique developm ent as most analysts of consociationalism (eg Lehmbruch 1967; Daalder 1974; Steiner 1974) tend to argue. POLITICAL STABILITY IN DIVIDED SOCIETIES 261 Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 Corporatism Unlike consociationalism, corporatism is a relatively recent development in the relations between politics and society. Although som e authors trace it from the 1930s or see it as an imm ediate postwar phenom enon (eg Katzenstein 1985) most consider it a political±institutional arrangement that has emerged since the worldwide stag¯ation crisis started (around the mid-1970s; Schm idt 1982; Czada 1987; Kem an 1988; Braun 1989). Corporatism is de®ned as a m ore or less loosely organised system of tripartite interest-representation aiming at concerted political action to avert the consequences of econom ic crisis (L ehmbruch and Schmitter 1982; Keman 1989; Braun 1989; Visser and Hemerijck 1997, 63±80). The three central actors are: government, organised capital and labour. W hat interests us here is not whether `corporatism’ exists or m atters politically (see, for example, Schm idt 1982; Therborn 1987; Lijphart and CreÂpaz 1991; W oldendorp 1995), but rather to what extent this type of interest-intermediation can be understood in terms of rational institutionalism, ie as a structure-induced response to econom ic conditions which prom otes an optimal level of problem-solving. As with consociationalism, corporatism entails a decision-m aking structure that is by and large de®ned by (inform al) codes of conduct. However, in contrast to consociationalism, it concerns two societal actors who are not primarily organised to operate politically. They rather operate on the borders of the parliamentary system. Under consociationalism, however vehement com petition between and among parties m ay be, party-political actors pursue similar overall goals: of®ceand vote-seeking, policy enactments and political resources (Strom 1990; Keman 1997b). In contrast, the m ain societal agents within a corporatist fram ework are `natural’ (class) enemies and have therefore fundamentally different urgent `needs’ but pursue similar measures. At the same time, they are strongly interdependent, since the one cannot function without the other (ie in a m arket economy). Hence, in order to avoid zero-sum outcom es, which would be detrimental to all participants, certain institutions are expected to direct rational behaviour and produce outcom es bene®cial to all of them (Scharpf 1992). Corporatism is thus an institutionalised form of socioeconomic con¯ict-regulation, and ought to be seen as a bargaining process involving material exchanges to enhance public welfare (ie a stable economy). The degree of success is then measured by the pay-offs to each participant from the ®nal agreement. This result should be devoid of zero-sum results and conducive to optimal outcomes (see also Scharpf 1987; Lehmbruch 1984; Braun 1989; Czada 1987; Woldendorp 1995; Keman 1996). Corporatism is therefore a variable which is predominantly dependent on the feasible policy choices open to each participant and the room to manoeuvre which each of them has. Let us therefore consider which rule con®gurations as well as sets of (negotiable) `needs’ can be identi®ed with respect to corporatism. Again, the Netherlands will serve as an illustration (Braun 1989; Woldendorp 1995; Keman 1997b). A basic condition of corporatism is that negotiations take place on a voluntary basis and that only those organised actors have access who can be trusted to represent the m ajority of their m embership (and oblige that m ajority to accept the outcom es of the negotiations). The more this is so, the m ore likely a corporatist arrangement is likely to emerge not just once, but repeatedly under the condition that pay-offs are by and large equally perceived. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 262 H . KEMA N Another condition relates to the `collective action’ problem: corporatism works under the condition of no-exclusion of those that are included. Every participant has a potential veto: unanimous decision-m aking is a logical consequence of the developm ent of concerted action (Axelrod 1984; Laver 1986). This implies that the agreement reached must be optimal for the actors involved. But it m ay be sub-optimal regarding the attainment or m aintenance of public welfare for the nation as a whole (compare the Olsonian argument with respect to distributional coalitions and corporatism: Olson 1986). A third condition is that the results of corporatist bargaining need to be sanctioned by parliamentary approval, which to some extent constrains the room to manoeuvre of party-government in the bargaining process. This practice implies a `politicisation’ of corporatism because it is also dependent on parliamentary consent and in som e cases this process has been form alised with respect to the role of corporatist institutions. 7 All in all the institutional m ode of corporatism is characterised by conditions that enhance cooperative behaviour between actors. It is made possible because actors have few exit opportunities. On the other hand, tripartism also enlarges the room to manoeuvre for the actors representing capital and labour as they are attributed a semi-public status (ie they are recognised as participants in the process of collective choice; see also Offe 1979; Kem an 1989). The role of representative government is important in this respect as a legitimising body and it can contribute to the bargaining process by providing public goods to further an optimal pay-off to the other actors (eg by m eans of m aking the agreem ent legally binding). As there are so few formal rules laying down the conditions for a collective effort, particularly in the ®eld of socioeconomic policy with adversarial `needs’, the active involvement of government is crucial. The feasible set of policy options available is often limited and thus the right conditions for problem-solving are vital. The feasibility of political choice depends on the objective econom ic situation (ie econom ic development) and the rank-ordering of (urgent) `needs’ as envisaged by each participant. First of all, the way the societal actors view the economic crisis is important, and their view will not only in¯uence their ordering of preferences, but also the extent to which they are considered as immutable. Second, the bargaining process can only be successful if all participants see it as an iterative game. This implies the expectation of all actors involved that the negotiations will be repeated over time which allows actors to take up and abandon positions in turn to ®nd agreement (Taylor 1987; Scharpf 1998). In contrast to the game that is played under consociational conditions, all actors have the same list of `needs’ but (almost by de®nition) differ on their functional weight. For instance, active labour-m arket policy has a different weight for capital, labour and party-governm ent. Employers’ organisations are mostly interested in an adequate (qualitative) labour supply, whereas trade unions tend to stress the effects of the existing rate of unemployment in terms of incom e. 8 Finally, governm ents use 7 In some countries, eg Austria and the Netherlands, this process has been formalised to some extent: in The Netherlands by laying down that government must take the advice of the Social Economic Council (until 1995), and in Austria through the ParitaÈtischen Kommission which gives binding advice on wage rates and price control (Woldendorp 1997). 8 In The Netherlands the average wage rate of an industrial worker is directly related to the level of unemployment bene®ts. Hence trade unions tend to safeguard their mem bers’ income by maintaining wage levels rather than by attempting to in¯uence the level of social security bene®ts directly. POLITICAL STABILITY IN DIVIDED SOCIETIES 263 Table 2. Rank-ordering of needs of government, capital and labour Needs Active labour-market policy Statutory wage rates Social security bene®ts Governm ent 1 3 2 Capital Labour 2 1( 2 ) 3 3 1( 1 ) 2 Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 ( 1 ) 5 in favour of higher wage rates; ( 2 ) 5 in favour of lower wage rates. Rank-order: 1 5 an urgent need; 2 5 important; 3 5 indifference. labour-market policy not only as an (indirect) m eans to prevent high levels of unemployment, but also to avoid a negative in¯uence on popularity (W hiteley 1986). The same line of reasoning can be developed for other itemsÐ for instance for economic-related `needs’, such as the developm ent of wage rates in relation to the extension of social security bene®ts (Braun 1989; Keman 1989; Visser and Hemerijck 1997). Given this com plex pattern of similar econom ic `needs’ with a different weight for each actor, the notions of parallelism and, in particular, complementarity becom e important as they will indicate the feasibility of viable agreem ents between the actors involved. It is important to note this, because the institutionalisation of corporatist arrangements focuses on a single issue problem, whereas consociationalism is a procedure to ®nd an equilibrium with respect to multifarious issues. Table 2 sets out a plausible rank-ordering of economic `needs’ (see also Visser and Hemerijck 1997; W oldendorp 1995). The ordering shown in Table 2 is highly adversarial regarding one item: the developm ent of statutory wage rates. This gives governments an important role in achieving agreem ent among the social partners. Yet, in order to m aintain the corporatist arrangement it is not only necessary to settle wage rates, but also to compensate the societal actors given their `needs’. It is obvious, following this example, that, again, the solution is dependent on a governm ent which formally controls the regulation of social security and can interfere in wage settlements in the Netherlands. This means that party-governm ent in the Netherlands is the only actor which can m ediate the con¯ict of interests between the other actors and develop a viable agreement. If the above holds true, how can the working of corporatism be explained as a m echanism to produce a structure-induced equilibrium by m eans of rational institutionalist argum ents? First, it m ust be noted that the actors involved are functionally different but fundamentally interdependent (de®ning the room to m anoeuvre). Corporatism is a rational strategy for all the actors involved in their search for optimal trade-offs. Instead of analysing Table 2 horizontally (as was done regarding the interactions between parties in consociationalism), it should be examined vertically (ie in regard to the feasibility of policy choices for each actor). The question is: to what extent is it possible for each actor to change their ordering of needs (the feasibility set) in order to reach an agreement which m akes no one worse off and is credible for its own members? The following exchange emerges and m akes cooperative behaviour possible: the need for higher wages can be m odi®ed by increasing social security bene®ts (as a kind of `social wage’). Active labour-m arket policy can be adjusted to balance demand and supply to the need of all actors, since it is by and large complementary for all actors. In actual fact, given that corporatist arrangements do work on the basis of give and take as well as taking it in turns to make Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 264 H . KEMA N concessions over time, it may be that a low graded `need’ (3) is an asset for an actor to bargain for an optimal agreem ent on a highly graded one (trading off). Therefore it may be expected that governments will forego direct constraints on wages and social security to ensure cooperation regarding labour-m arket policies. According to m any observers, Visser and Hemerijck (1997) for instance, Dutch governments behaved strategically in this way during the 1980s (when they were dominated by the Christian Democrats; see also Hemerijck and van Kersbergen 1997). It was only in the 1990s that the option of the active labour market became the cornerstone of the `Dutch m iracle’. If governm ent, the pivotal actor within a corporatist arrangement, does not follow such a strategy by de®ning the room to m anoeuvre, it will be confronted by a non-cooperative coalition. This has on occasions been the case in The Netherlands. Dutch governm ents attempted to reduce wage drift as well as curb social security bene®ts (in 1980/81 and again in 1989/90; Bran 1989; W oldendorp 1995). This resulted in a lack of concerted action and a steep increase in unemploym ent. Only after this policy direction was abandoned has general agreem ent again been possible in the NetherlandsÐ based on com plementarity on active labour-market policy and reducing social securityÐand concerted action feasible (Keman 1993a, 1996; Woldendorp 1995). Hence, instead of welfare for work, the shared or parallel need became jobs for welfare (Hemerijck and van Kersbergen 1997). The fam ous `Dutch m iracle’ which now seems to have cured the equally well-known `Dutch disease’ is the result of a reordering by `needs’ of the actors involved. Yet, and this is m y claim, it could not have occurred had there not been a rule con®guration (ie corporatism) that allowed for an alternative choice, ie a structure-induced equilibrium .9 T his process of exchange and substituting of the priorities of each actor can be considered as a rational abstention fro m m aximising one’s ow n needs. The interdependence of the social partners affects the feasibility of choice, which in turn is dictated by the room for m anoeuvre of each and everyone. Since defection is dif®cult, the room for m anoeuvre depends on the role played by governm ent (unless the social partners decide to cooperate regardless of their prioritised `needs’).T he extent to w hich corporatism w ill w ork is thus dependent on institutionalising tripartite collective decision-m aking. The actual w orking and continuity of such an arrangem ent depends on whether the bargaining gam e is perceived rationally w hen it com es to de®ning the exchange patterns and on whether institutionalised behaviour allow s for iterative plays rather than being a one-shot gam e. A gain, as with m y criticism of consociationalism, corporatism does not solely occur in certain countries due to speci®c historical circum stances (see, for exam ple, Katzenstein 1985) or to com m on sense and policy learning (eg Visser and Hem erijck 1997), but rather as a result of rational action within a ¯exible institutional setting allowing for optimal outcom es each and every time. In m y view, corporatism should be understood as a result of rational behaviour 9 Compare this logic with that of Scharpf (1998) where he explains an optimal outcome by means of the `Battle of the Sexes’ gam e, which he labels as a `problem-solving style of decision-making’ (pp. 56±8). However, in his view the equilibrium is only possible by repeating the game and to redistribute the pay-offs over time. In my view the institutional structure allows for changing outcomes during each game played. POLITICAL STABILITY IN DIVIDED SOCIETIES 265 taking into account the institutional context and circum stances allow ing for an optimal outcom e. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 Concluding Rem arks The descriptive analysis of consociationalism and corporatism has contributed to the understanding of how sharp societal con¯ict turns into a viable political consensus. Agreements that can be reached between adversarial or con¯icting actors depend on the feasibility of choice which, in turn, can be derived from the rank-ordering of the (urgent) `needs’ of each actor. In the case of consociationalism, where the room for m anoeuvre is more form ally de®ned, unrelated issues can be exchanged. Such exchanges can assist decision-making through cooperation in parliament, which depends on the degree of parallelism in their needs recognised by all parties during negotiations. This mutual understanding helps to ®nd workable and sustainable compromises and avoids stalemates. In the case of corporatism, the room for m anoeuvre is by and large de®ned by party-government, and the `needs’, as de®ned by the actors, are strongly interrelated. Here the exchange is more dependent on the ordering by each actor. Only certain combinations of `needs’ ordered by each actor on the basis of complementarity will lead to a package deal which is acceptable to all and which can be implemented by concerted action. By characterising actors in relation to their (urgent) `needs’ and considering institutions as intervening variables that shape the room to manoeuvre, we can understand the paradoxical relation between societal con¯icts and the development of political consensus as regards the feasibility of shared policy choices. In this article I have attempted to combine institutionalist argum ents with argum ents originating from rational choice theory. I consider institutions as particularly crucial in the study of politics, but only if analysed in relation to actors and their `needs’. Institutions come and (som etimes) go, and shape the political interactions in m odern society: the structure of the polity is therefore the principal intervening variable between politics and policies (Keman 1998; but see also Scharpf 1992; HeÂritier 1998; Laver 1997). By de®ning and elaborating on the basic elements and working of institutional arrangements with respect to democratic politics, I have identi®ed a plausible line of advance for research in this area. The two concepts that I have proposedÐ the actor’s room for m anoeuvre and his or her set of politically feasible choicesÐ represent a step forward. In particular they are useful when analysing under which circum stances a structure-induced equilibrium is m ore likely to develop (Scharpf 1998; Shepsle 1995). The other aim has been to relate m y a priori argum ents to existing concepts in political science: consociationalism and corporatism. These concepts are a challenge to any student of comparative politics, since both concepts appear able to explain paradoxical relationships which do not ®t general pluralist democratic theory. The question of how political stability com es about in the existing studies has been, however, too speci®c and contextually driven. Traditional approaches to consociationalism and corporatism are restricted by their descriptive analyses which are insuf®cient to explain the paradox of con¯ict and consensus in a more general fashion. They lack an overarching theoretical argument. This makes it dif®cult to apply the concepts to comparable cases, be it cross-nationally or over time. A rational institutional approach m ay well do a better job in this respect. 266 H . KEMA N Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 References Almond, G., G.B. Powell and R. Mundt. 1993. Comparative Politics. A Theoretical Framew ork. New York: Harper Collins. Andeweg, R.B. 1989. TweeÈerlei M inisterraad (inaugural address). Leiden: Rijksuniversiteit Leiden. Axelrod, R. 1984. The Evolution of Cooperation. New Jersey: Basic Books. Braun, D. 1989. Grenzen politischer Reguliering. Der W eg in die M assenarbeitslosigkeit. Wiesbaden: DuÈV. Budge, I. and H. Keman. 1990. Parties and Democracy. Government Formation and Government Functioning in Twenty States. Oxford: Oxford University Press. Colomer, J.M. ed. 1996. Political Institutions in Europe. London and New York: Routledge. Czada, R.M. 1987. `The Impact of Interest Politics on Flexible Adjustment Policies.’ In Coping with the Economic Crisis. Alternative Responses to Economic Recession in Advanced Industrial Societies, eds Kem an, H., H. Paloheimo and P.F. Whiteley. London: Sage. Czada, R.M. 1998. `Interest Groups, Self-interest and the Institutionalization of Political Action.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Kem an. Amsterdam: VU Press. Czada, R.M., A. HeÂritier and H. Keman, eds. 1998. Institutions and Political Choice: On the Limits of Rationality. Amsterdam: VU Press. Daalder, H. 1974. The Consociational Democracy Theme.’ World Politics XXVI/4: 604±21. Daalder, H. 1987. `The Dutch Party System: From Segmentation to Polarization and Then?’ In Party Systems in Denmark, Austria, Switzerland, the Netherlands and Belgium, ed. H. Daalder. London: Frances Pinter. Daalder, H. and P. Mair, eds. 1983. W estern European Party Systems. Continuity and Change. London: Sage. Dahl, R.A. 1971. Polyarchy. Participation and Opposition. New Haven and London: Yale University Press. Elster, J. 1979. Ulysses and the Syrens. Cambridge: Cambridge University Press. Hemerijck, A. and K. van Kersbergen. 1997. `Explaining the New Politics of the Welfare State in the Netherlands.’ Acta Politica 32: 258±80. HeÂritier, A. 1998. `Institutions, Interests, and Political Choice.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Kem an. Amsterdam: VU Press. Katzenstein, P.J. 1985. Small States in World Markets. Industrial Policy in Europe. Ithaca and London: Cornell University Press. Keman, H. 1988. The Development toward Surplus Welfare: Social Democratic Politics and Policies in Advanced Capitalist Democracies. 1965±1984. Am sterdam: CT Press. Keman, H. 1989. `EntsaÈulung, Interessenausgleich und SteuerungsfaÈhigkeit: Evolution und Revolution in der NiederlaÈndischen Politik.’ In Westlichen Demokratien und Interessenvermittlung, hrsg. R. Kleinfeld and D. Lutthardt. Hagen: FernuniversitaÈt-Kursbuch. Keman, H. 1993a. `Politik der Mitte in den Niederlanden: Konsens und Kooperation ohne Politikproduktion.’ In W estlichen Demokratien und Interessenvermittlung, hrsg. R. Kleinfeld and W. Luthardt. Marburg: SchuÈren, Marburg. Keman, H., ed. 1993b. Comparative Politics. New Directions in Theory and M ethod. Amsterdam: VU Press. Keman, H. 1996. `Political Institutions in the ªLow Countriesº: Confrontation and Coalescence in Segmented Societies.’ In Political Institutions in Europe, ed. I.J. Colomer. London: Routledge. Keman, H. 1997a `Approaches to the Analysis of Institutions.’ In Political Institutions and Public Policy. Perspectives on European Decision-making, eds B. Steunenberg and F. van Vught. Dordrecht: Kluwer. Keman, H. 1997b. The Politics of Problem-solving in Postwar Democracies. London: Macmillan. Keman, H. 1998. `Political Institutions and Public Governance.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Kem an: 109±34. Amsterdam: VU Press. Keman, H., H. Paloheimo and P.F. Whiteley, eds. 1987. Coping with the Economic Crisis. Alternative Responses to Economic Recession in Advanced Industrial Societies. London: Sage. Lane, J.-E. 1993. The Public Sector. Concepts, M odels and Aproaches, London: Sage. Lane, J.-E. and S.V. Errson. 1994. Politics and Society in Western Europe, 3rd ed. London: Sage. Laver, M. 1986. Social Choice and Public Policy. London: Basil Blackwell. Laver, M. 1997. Private Desires, Political Action. London: Sage. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 POLITICAL STABILITY IN DIVIDED SOCIETIES 267 Laver, M. and N. Scho®eld. 1990. Multiparty Government: The Politics of Government in Europe. Oxford: Oxford University Press. Lehmbruch, G. 1967. Proporzdemokratie: Politisches System und Politische Kultur in der Schweiz und in Oesterreich. TuÈbingen: Mohr. Lehmbruch, G. 1979. `Liberal Corporatism and Party Government.’ In Trends Toward Corporatist Intermediation, eds P.C. Schm itter and G. Lehmbruch. London: Sage. Lehmbruch, G. 1984. `Concertation and the Structure of Capitalist Networks.’ In Order and Con¯ict in Contemporary Capitalism, ed. J.H. Goldthorpe. Oxford: Clarendon Press. Lehmbruch, G. 1998. `The Organization of Society, Administrative Strategies and Policy Networks.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Keman. Amsterdam: VU Press. Lehmbruch, G. and P.C. Schmitter, eds. 1982. Patterns of Corporatist Policy-making. London: Sage. Lijphart, A. 1975. The Politics of Accommodation: Pluralism and Democracy in The Netherlands, 2nd ed. Berkeley: University of California Press. Lijphart, A. 1977. Democracy in Plural Societies: A Comparative Exploration. New Haven and London: Yale University Press. Lijphart, A. 1984. Democracies. Patterns of M ajoritarian and Consensus Government in 21 Countries. New Haven and London, Yale University Press. Lijphart, A. 1989. `From the Politics of Accom modation to Adversarial Politics in The Netherlands: A Reassessment.’ West European Politics 21(1): 139±53. Lijphart, A. and M.M.L. CreÂpaz. 1991. `Notes and Comments: Corporatism and Concensus Democracy in Eighteen Countries: Conceptual and Empirical Linkeages.’ British Journal of Political Science 21: 235±56. Lipset, S.M. 1963. Political Man. The Social Bases of Politics. New York: Doubleday Anchor Books. March, J.G. and J.P. Olsen. 1989. Rediscovering Institutions. New York: The Free Press. Mueller, D.C. 1989. Public Choice II. A Revised Edition. Cambridge: Cambridge University Press. North, D.C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. Offe, C. 1979. `ªUnregierbarkeitº. Zur Renaissance konservativer Krisentheorien.’ In StichwoÈrter zur `geistigen Situation der Zeit’, hrsg. J. Habermas. Frankfurt: Suhrkamp. Olsen, J.P. 1998. `Political Science and Organization Theory. Parallel Agendas but Mutual Disregard.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Kem an. Amsterdam: VU Press. Olson, M. 1982. The Rise and Decline of Nations. Cambridge, MA: Harvard University Press. Olson, M. 1986. `An Appreciation of the Tests and Criticisms.’ Scandinavian Political Studies 9(1): 65±80. Ostrom, E. 1990. Governing the Commons. The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press. Powell, G.B. Jr. 1982. Contemporary Democracies. Participation, Stability and Violence. Cambridge, MA: Harvard University Press. Putnam, R.D. 1993. Making Democracy W ork. Princeton: Princeton University Press. Scharpf, F.W. 1987. `A Game-theoretical Interpretation of In¯ation and Unem ployment in Western Europe.’ Journal of Public Policy 7(1): 227±58. Scharpf, F.W. 1992. Crisis and Choice in European Social Democracy. Ithaca and London: Cornell University Press. Scharpf, F.W. 1998. `Political Institutions, Decision Styles, and Policy Choices.’ In Institutions and Political Choice, eds R.M. Czada, A. HeÂritier and H. Kem an. Amsterdam: VU Press. Schmidt, M.G. 1982. `Does Corporatism Matter? Economic Crisis, Politics and Rates of Unem ployment in Capitalist Democracies in the 1970s.’ In Patterns of Corporatist Policy-making, eds G. Lehm bruch and P.C. Schmitter. London: Sage. Schmidt, M.G. 1996. `When Parties Matter: A Review of the Possibilities and Limits of Partisan In¯uence on Public Policy.’ European Journal of Political Research 30: 155±83. Schmitter, P.C. 1981. `Interest Intermediation and Regime Governability in Contemporary Western Europe.’ In Organizing Interests in W estern Europe. Pluralism, Corporatism and the Transformation of Politics, ed. S. Berger. Cambridge: Cambridge University Press. Shepsle, K.A. 1995. `Studying Institutions: Some Lessons from the Rational Choice Approach.’ In Political Science in History, eds J. Farr, J.S. Dryzek and S.T. Leonard. Cambridge: Cam bridge University Press. Downloaded By: [Vrije Universiteit, Library] At: 11:02 1 June 2011 268 H . KEMA N Skocpol, T. 1985. `Bringing the State Back In. Strategies of Analysis in Current Research.’ In Bringing the State Back In, eds P. Evans, D. Rueschemeyer and T. Skocpol. Cambridge: Cam bridge University Press. Steiner, J.A. 1974. Amicable Agreement versus M ajority Rule. Con¯ict Resolution in Switzerland. Chapel Hill: University of North Carolina Press. Steinmo, S., K. Thelen, F. Longstreth, eds. 1992. Structuring Politics. Historical Institutionalism in Comparative Perspective. Cambridge: Cambridge University Press. Strom, K. 1990. `A Behavioral Theory of Com petitive Political Parties.’ American Journal of Political Science 2: 565±98. Taylor, M. 1987. The Possibility of Cooperation. Cambridge: Cambridge University Press. Therborn, G. 1987. `Does Corporatism Matter? The Economic Crisis and Issues of Political Theory.’ Journal of Public Policy 7(3): 259±84. Tsebelis, G. 1990. Nested Games. Rational Choice in Comparative Politics. Berkeley, Los Angeles and Oxford: University of California Press. Visser, J. and A. Hemerijck. 1997. `A Dutch Miracle’: Job Growth, Welfare Reform and Corporatism in The Netherlands. Amsterdam: Amsterdam University Press. Weaver, R.K. and B.A. Rockman, eds. 1993. Do Institutions M atter? Government Capabilities and Abroad. Washington DC: The Brooking Institution. Whiteley, P.F. 1986. The Political Control of the Economy. London: Sage. Woldendorp, J.J. 1995. `Neo-corporatism as a Strategy for Con¯ict Regulation in The Netherlands. 1970±1990.’ Acta Politica XXX: 2. Woldendorp, J.J. 1997. `Corporatism and Socioeconomic Con¯ict-regulation.’ In The Politics of Problem-solving in Postwar Democracies, ed. H. Keman. London: Macmillan Press.