The problem of institutional change
It is commonplace to argue that political institutions are a source of inertia and resistance to change. Institutions are seen as excessively static and likely to remain on the same path unless some effort is made to divert them. It is also commonplace to claim that the ‘New Institutionalism’ as an approach to political life is not useful for making sense of institutional change, planned or not. The assumed inability to explain change is a result because the new institutionalism is overly structuralist and does not grant purposeful actors a proper role. The approach does not deal adequately with political agency, conflict and power asymmetries and can therefore not account for deliberate institutional design as a political instrument (Peters, Reference Peters1999a, Reference Petersb; Peters and Pierre, Reference Peters and Pierre2005). Yet the label ‘New Institutionalism’ is used for a variety of approaches that understand change differently and ‘[M]ost fundamentally, there is the question of whether or not change is recognized as an ordinary part of institutional life or as the exception to a rule of stability, and perhaps even hyperstability’ (Peters, Reference Peters1999a: 147).
The aim of this paper is not to take stock of competing approaches to institutional change. It is to use one specific institutional approach, with roots in studies of formal organizations, to explore how we may think about the mechanisms through which institutions arise, evolve, and decline and how new institutions replace or supplement older ones. What conditions are likely to sustain or undermine change, and what is the significance of existing institutional arrangements?
The paper calls attention to aspects usually neglected by approaches giving primacy to large-scale societal forces or deliberate design. Within the first, institutional arrangements are determined by the external environment through competitive pressure and selection stemming from advantageous traits and differential survival and growth. Within the second, institutional arrangements are malleable and a matter of choice and change is driven by actor strategies. In contrast, an institutional approach, as understood here, assumes that institutions are not pawns of external forces or obedient tools in the hands of some master. They have an internal life of their own, and developments are, to some degree, independent of external events and decisions. Change is an ordinary part of political life. It is rule-bound and takes place through standard processes, as institutions interpret and respond to experience through learning and adaptation. Yet such processes are not guaranteed to be ‘efficient’ in reaching an enduring equilibrium, and sometimes change is discontinuous.Footnote 1
Focus is on the relations between institutional characteristics and change in governmental institutions in modern democracies. Accounting for how and why institutions emerge and change, however, requires a rephrasing of the questions an institutional approach should aspire to answer. The task of democratic government is not to maximize change. It is to balance order and change, and the scholarly challenge is to account for how and why institutions remain stable as well as how and why they change.
Democratic government is, furthermore, made up of organized components with shifting relations to one another, and governing is structured collective action coordinating multiple actors, organizations, and resources. Knowledge about how formal organizations operate and change, therefore, is assumed to provide insight into the dynamics of governmental institutions. Formal organizations and formally organized institutions are conceived as collections of rules and standard operating procedures, pre-defined patterns of thought and action, including but not limited to legal rules and procedures, and resources.
First, the problem of change is reformulated and it is observed that institutions have a role in generating both order and change and in balancing the two. Second, the concepts of institution and institutionalization are elaborated. Third, institutional sources of change and continuity are explored. Fourth, some implications on how democratic change and order can be conceived are spelled out, and, finally, some future challenges are suggested.
Neither Newton nor Heraclitus
Portraying institutions as unable or unwilling to adapt to new contingencies and tasks has long historical roots. A standard argument has been that political institutions lag behind economic, technological, and social change – a claim that has been frequently repeated as a premise for reforms in the public sector during the last three decades. Contemporary societies emphasize the need for innovation and change, yet the main assumption, inspired by neo-classical economics, is that market competition is, and should be, the dominant mechanism of innovation (Fagerberg, Mowery, and Nelson, Reference Fagerberg, Mowery and Nelson2005).
The argumentation seems inspired by Newton’s first law: the law of inertia. Political institutions, like any material body that is at rest or is moving at a constant speed, will remain at rest or keep moving at a constant speed unless they are acted upon by a force. External forces decide the direction and magnitude of change, and in the absence of external forces, political institutions will cling to the status quo, or there are monotone and inevitable developments towards modernization, rationalization, democratization, bureaucratization, etc.
Historically, however, political thinking has been as much concerned with the conditions for legitimate order, authority, and rule, as with change (Immergut, Reference Immergut2006). Stability and ordered relationships have been viewed as a precarious achievement, always threatened by disorganization, entropy, and chaos that endanger life and property. The point of departure for this strand of analysis is closer to Heraclitus (540–480 BC) than to Newton. Everything is in flux under the pressure of shifting situations. Organizational arrangements are infinitely changeable and always in transition, and it is a Sisyphean job to create and maintain political order. Because it is as difficult to keep institutions constant as it is to change them, there is a need to explain continuity and smooth developments as well as radical transformations and abrupt breakdowns. Understanding order and change are two sides of the same coin and there is a need to know what processes and conditions may maintain or challenge the status quo.
Elements of order
An institutional approach assumes that political life is neither deterministic (caused by external forces and laws) nor random (governed by the laws of chance) and that political institutions are neither completely static nor in constant flux. In contrast with the heirs of Heraclitus, institutions are assumed to create elements of order and predictability in political life. Institutions organize actors, issues, and resources in or out of politics and structure patterns of political struggle (Schattschneider, Reference Schattschneider1960; Steinmo, Thelen, and Longstreeth, Reference Steinmo, Thelen and Longstreeth1992; Egeberg, Reference Egeberg2006). They make less likely pure temporal sorting where decision opportunities, actors, problems, and solutions flow together solely as a function of time (Cohen, March, and Olsen, Reference Cohen, March and Olsen1972, Reference Cohen, March and Olsen2007). In contrast to the heirs of Newton, political institutions are assumed to have dynamics of their own. The assumption, that institutional arrangements persist unless there are external chocks, underestimates both intra- and inter-institutional sources of change.
Imperfect processes
Through what processes, then, do institutions emerge and change? To what extent are forms of government a matter of choice (Mill, Reference Mill1962: 1)? Are societies of men capable of establishing good government from reflection and choice, or are they forever destined to depend on accident and force (Hamilton, Jay, and Madison, Reference Hamilton, Jay and Madison1964: 1)? To answer these questions, students of political institutions have borrowed metaphors from both engineering and biology.
Political engineering and rational design assume that institutions are deliberately created and reformed in order to achieve substantive ends. Some actors have a vision of a better society. They have a diagnosis of what is wrong and see institutions as partly causing the problems. They have a prescription for better ways of doing things and know how institutions should be changed in order to achieve better results. They also control the resources required to implement the prescription.
An alternative to rationalism – ‘the most remarkable intellectual fashion of post-Renaissance Europe’ (Oakeshott, Reference Oakeshott1991: 5) – is to see ‘living institutions’ as social organisms that evolve over time as an unplanned result of historical processes. Institutions grow as an artifact of interaction, cooperation, and competition and embody the experience and normative and causal beliefs of a population. In this perspective the Parliament, for example, has been seen as ‘a product less of intention and design than of blind evolution’ (Dahl, Reference Dahl1998: 21). Actions may come before ideas and purposes. Ends and means may develop simultaneously, and evolving behavioral patterns may post hoc be described, explained, and justified, ‘frozen’ into habits and traditions, and formally codified. Surviving institutions are those that have proved their worth through the test of time.
Inspired by Mill’s sarcastic comment, ‘[I]t is difficult to decide which of these doctrines would be the most absurd, if we could suppose either of them held as an exclusive theory’ (Mill, Reference Mill1962: 3), an institutional perspective conceives political actors neither as engineers with full control nor as fatalists with no range of choice. Institutional developments are neither a direct product of will, planning, and design, nor a mere haphazard by-product of chance events and an ecology of uncoordinated actions. Institutionalism emphasizes the endogenous nature and explanatory power of political institutions. It is assumed that the organization of political life makes a difference and that institutions have dynamics of their own. The theoretical challenge is to understand the shifting mix of deliberate design and adaptive behavior (March, Reference March1981; March and Olsen, Reference March and Olsen1989, Reference March and Olsen2006b; Brunsson and Olsen, Reference Brunsson and Olsen1998).
In the literature, scholars often assumed that institutions survive and flourish because they are well-adapted to their functional (Goodin, Reference Goodin1996; Stinchcombe, Reference Stinchcombe2001) or normative environments (Meyer and Rowan, Reference Meyer and Rowan1977; Scott, Meyer, and Associates, Reference Scott and Meyer1994). There is, however, no reason to believe that processes of adaptation are always perfect in a context of large-scale governmental organizations. Political institutions have historically shown robustness facing comprehensive societal change; institutional stability is possible in situations with high external pressure (Héritier, Reference Héritier2007: 242), and it has even been asked whether government organizations are immortal (Kaufman, Reference Kaufman1976) – suggesting that the external environment has a limited and varying ability to select and eliminate political institutions. Likewise, in spite of accounts of the role of heroic founders and constitutional moments, democracies have limited capacity for institutional design and for achieving the intended effects of reorganizations (March and Olsen, Reference March and Olsen1983, Reference March and Olsen1989; Olsen and Peters, Reference Olsen and Peters1996; Offe, Reference Offe2001).
In contrast with standard equilibrium models, institutionalism holds that history is ‘inefficient’.Footnote 2 ‘Historical inefficiency’ implies that institutions rarely are perfectly adapted to their environments and that the matching of institutions, behaviors, and contexts takes time and has multiple, path-dependent equilibria. The receptivity towards external pressure varies and adaptation is less automatic, continuous, and precise than assumed by equilibrium models (March and Olsen, Reference March and Olsen1989).Footnote 3 Institutions affect the rate of change by the ways in which they adapt their internal structures and processes, by creating actors and providing them with premises of action, and by ignoring or modifying external pressures and influencing environments and thereby future environmental inputs. There is, for example, no guarantee that surviving institutions represent an efficient response to external environments because the rate of external change may outpace the rate of institutional adaptation and because institutions sometimes speed up and sometimes delay or impair external impulses and decisions.
While institutionalism assumes that change and order are generated by comprehensible processes producing recurring modes of action and patterns of change (March, Reference March1981), the possibility of inefficiency in standard processes requires detailed knowledge about the processes through which historical experience is incorporated into institutional structures and processes (DiMaggio and Powell, Reference DiMaggio and Powell1991: 33). The key behavioral mechanisms encoding experience into rules and routines are history-dependent, and in a world of partly autonomous institutions focus has to be moved from a single dominant, coherent, and efficient process to observations of how institutional properties mediate between inputs and outcomes by impacting several ‘imperfect’ and possibly disjointed processes of change. Observation and interpretation of experience, targets and aspiration levels, memories, retrieval of information, capabilities, and responses are all affected by the organization and legacy of institutions (March and Olsen, Reference March and Olsen1989, Reference March and Olsen1995).
Institutionalists therefore need to identify processes and determinants that increase or hamper the ordering effect of political institutions and make history more or less inefficient and to attend to how such processes themselves are stabilized or destabilized (Olsen, Reference Olsen2008a). A well-designed institution ‘is not a stable solution to achieve, but a developmental process to keep active’ (Nystrom and Starbuck, Reference Nystrom and Starbuck1981: xx) and democracies face a grand balancing act between exploitation and exploration. Exploitation involves using rules, routines, and knowledge that are known to work. Exploration involves willingness and the ability to experiment with rules, routines, and knowledge that might, but often do not, provide improvements. Purification of exploitation will make an organization obsolete in a dynamic world. Continuous experimentation will prevent the organization from realizing the potential gains of new discoveries. What is less obvious is the optimal balance between the two (March, Reference March1991).
For an elaboration of this approach, there is a need to specify in more detail what is meant by ‘institution’ and ‘institutionalization’.
Institutionalization, de-institutionalization, and re-institutionalization
Formally organized political institutions, such as the legislature, the executive, public administration, and the judiciary, have for a long time been important research sites for students of politics. Institutionalism, however, is a specific approach that aspires to make sense of how such institutions emerge, function, and change. What makes an approach to government and politics ‘institutional’ (Peters, Reference Peters1999a: 18)? As already argued, the simple answer is that an institutional approach assigns more explanatory power to the organization and legacies of institutions than to properties of individual actors and the broader societal contexts.
A parliament, a ministry, or a court of law, like any formal organization, can be conceived as a rational instrument for a dominant actor that creates, reforms, and eliminates institutions; as an arena for struggle and bargaining among contending groups; as an artifact of environmental forces; or as a transformative institution. Each conception demands different kinds of knowledge. An instrumental perspective and an arena perspective require knowledge about the preferences, beliefs, resources, and strategies of (respectively) the dominant decision-maker(s) and the participants negotiating and re-negotiating the terms of order. An environmental perspective demands knowledge about broad economic, technological and social forces and movements. An institutional perspective requires knowledge about the internal success criteria, structures, procedures, rules, practices, career structures, socialization patterns, styles of thought and interpretative traditions, and resources of the entity to be in focus.Footnote 4 An institutional perspective also requires concepts of ‘institution’ and ‘institutionalization’ beyond everyday language.
Rules, reasons, and resources
Institutionalism conceives an institution as a relatively enduring collection of rules and organized practices, embedded in structures of meaning and resources that are relatively invariant in the face of turnover of individuals and changing external circumstances (March and Olsen, Reference March and Olsen2006b). Constitutive rules and repertoires of standard operating procedures structure institutional behavior and developments by prescribing appropriate behavior for specific actors in specific situations. Structures of meaning, involving standardization, homogenization, and authorization of common purposes, reasons, vocabularies, and accounts, give direction to, describe, explain, justify, and legitimate behavioral rules. Structures of resources create capabilities for acting. Resources are routinely tied to rules and worldviews, empowering and constraining actors differently and making them more or less capable of acting according to behavioral codes.
Institutionalism involves purposeful human agency, reflection, and reason giving as well as rules. Yet, in contrast with models assuming a logic of consequentiality and strategic action where actors maximize their (self-) interest, institutionalism assumes that the basic logic of action is rule-following. Behavior is governed by standardized and accepted codes of behavior, prescriptions based on a logic of appropriateness, and a sense of obligations and rights derived from an identity, role, or membership in a political community and the ethos and practices of its institutions. Actors do not simply please others by acting in accordance with their expectations. Rules are to some extent self-enforcing because actors have internalized the belief that some actions are appropriate, natural, and legitimate (March and Olsen, Reference March and Olsen1989, Reference March and Olsen2006a, Reference March and Olsenb). Members of an institution observe and are the guardians of its constitutive principles and standards. Nevertheless, they have the ability to take purposeful action based on rule interpretation, including the ability to develop and modify normative criteria and identities through collective processes.
The scopes and modes of institutionalized activity vary across political systems, policy areas, and historic time (Eisenstadt, Reference Eisenstadt1965; Berger and Luckmann, Reference Berger and Luckmann1967). Over time political life achieves or loses structure, and the nature of order changes. At some periods in some polities and policy areas, politics is organized around well-defined boundaries, common rules and practices, shared causal and normative understandings, and resources adequate for collective action. At other times and places, politics is relatively anarchic. Boundaries are less well defined. Relations are less orderly, and institutions are less common, less adequately supported, and less involved (March and Olsen, Reference March and Olsen1998: 943–944).
Institutionalization
Institutionalization is both a process and a property of organizational arrangements. Institutionalization as a process implies that an organizational identity is developed and that acceptance and legitimacy in a culture (or sub-culture) is built. There are:
(a) Increasing clarity and agreement about behavioral rules, including allocation of formal authority. Standardization and formalization of practice reduce uncertainty and conflict concerning who does what, when, and how. As some ways of acting are perceived as natural and legitimate, there is less need for using incentives or coercion in order to make people follow prescribed rules.
(b) Increasing consensus concerning how behavioral rules are to be described, explained, and justified, with a common vocabulary, expectations and success criteria. There is a decreasing need to explain and justify why modes of action are appropriate in terms of problem-solving and normative validity.
(c) Increasing shared conceptions of what are legitimate resources in different settings and who should have access to, or control, common resources. The supply of resources required to act in accordance with behavioral prescriptions becomes routinized and ‘taken as given’. It takes less effort to get the resources required for acting in accordance with prescribed rules of appropriate behavior.
Corollary, de-institutionalization implies that existing institutional borders, identities, rules, and practices; descriptions, explanations, and justifications, and resources and powers are becoming more contested and possibly discontinued. New actors are mobilized. Outcomes are more uncertain, and it is necessary to use more incentives or coercion to make people follow prescribed rules and to sanction deviance. Re-institutionalization implies either retrogression or a transformation from one order into another, constituted on different normative and organizational principles.
Since institutions are human products, they exist only because a sufficiently high number of citizens believe they exist (Searle, Reference Searle1995). Institutions require continuously renewed collective confirmation and validation of their constitutive rules, meanings, and resources. Yet all institutions experience challenges, and some turn out to be fragile and unable to reproduce themselves. The basic assumptions on which an institution is constituted and its prescribed behavioral rules are never fully accepted by the entire society (Eisenstadt, Reference Eisenstadt1965: 41; Goodin, Reference Goodin1996: 39). Institutions may recede into oblivion because trust is eroded and rules are not obeyed. There may be rationally motivated dissent and change (Habermas, Reference Habermas1996: 36) and ‘revolutionary violence may contribute as much as peaceful reform to the establishment of a free society’ (Moore, Reference Moore1966: 20).
Institutionalization, therefore, is not an inevitable, irreversible, unidirectional, or monotonic process, and institutionalization, de-institutionalization, and re-institutionalization can follow a variety of patterns (Weaver and Rockman, Reference Weaver and Rockman1993; Rokkan, Reference Rokkan1999; Bartolini, Reference Bartolini2005). Can, then, knowledge about intra- and inter-institutional properties contribute to an improved understanding of how formally organized governmental institutions mirror and maintain a certain kind of order and nonetheless change?
Institutional sources of continuity and change
Organized democracy
Theories of political development have to take into account that modern democracies are ‘organized democracies’ (Olsen, Reference Olsen1983) and that institutions are markers of a polity’s character, history, visions, and identity. Institutions give order to social relations, reduce flexibility and variability in behavior, and restrict the possibilities of a one-sided pursuit of self-interest or drives (Weber, Reference Weber1978). Democratic government, however, consists of a conglomerate of partly autonomous and powerful large-scale formal organizations that operate according to different repertoires of relatively stable rules and standard operating procedures (Selznick, Reference Selznick1957: 1; Allison, Reference Allison1971: 67).
Governmental organizations do, often on their own initiative, what they are trained to do and know how to do, and government can, at least in the short run, deliver only what large-scale organizations (military, police, administrative, health, and educational systems, etc.) are capable and motivated to do. Understanding what feasible alternatives organizational routines and repertoires provide is in particular important in complex situations that require coordinated action by a large number of individuals, organizations, technologies, and resources. Government actions and institutional developments can then be understood by uncovering how organizations enact standard operating procedures. It is necessary to know which organizations government consists of, how tasks and responsibilities are allocated among them, and what standard operating programs and repertoires different organizations have (Cyert and March, Reference Cyert and March1963; Allison, Reference Allison1971; March and Olsen, Reference March and Olsen1995).
Rules and standard operating procedures define satisfactory performance (targets, aspiration levels) and organize attention, interpretation, recruitment, education and socialization of personnel, resource allocation, action capabilities, and conflict resolution. Governmental organizations also avoid uncertainty by stabilizing relations to other significant actors, for example through developing shared understandings about turf and budgets. There is normally limited flexibility in organizational targets and aspiration levels, frames and traditions of interpretation, total budgets and internal allocations, and in external relationships. Resistance to change increases the more organizations are institutionalized, so that structures and processes have value and symbolic meaning beyond their contributions to solving the task at hand, and change is seen as threatening institutional identities, the sense of mission, and emotional attachments (Selznick, Reference Selznick1957: 17; March and Olsen, Reference March and Olsen1983, Reference March and Olsen1989).
Institutions are, nevertheless, not static. Rules and practices are modified as a result of positive and negative experience, organizational learning, and adaptation (March, Schultz, and Zhou, Reference March, Schultz and Zhou2000). Routines, identities, beliefs, and resources can be both instruments of stability and vehicles of change; institutions of government do not always favor continuity over change. In democracies, change is usually incremental, but it can also be path breaking, with a sharp departure from existing practice. The question is when, how, and why routines are challenged and how institutional characteristics affect institutional developments and the likelihood of comprehensive change.
Rules
In institutionalized contexts, foundational rules impact the mix of continuity and change. Constitutions, treaties, laws, and institution- and profession-specific rules are carriers of accumulated knowledge. They define fairly stable rights and duties, regulate how advantages and burdens are allocated, and prescribe procedures for conflict resolution. Institutions may, however, carry the seeds of their own reform. There are rules of constitutional amendment and for who is responsible for initiating and implementing reforms, for example specific departments for planning and organizational development. Change can also be driven by explicit rules institutionalized in specific units or sub-units, prescribing routine shifts within an existing repertoire of rules (March and Simon, Reference March and Simon1958).
For example, constitutional rules protect Rechtsstaat values and limit the legitimacy of sudden, radical change. However, constitutional rules and routines also facilitate and legitimate change such as the transfer of power from one government to another, and the instrumental strand of democratic theory holds that citizens and their representatives should be able to fashion and refashion political institutions at will. Founding assemblies, lawmakers, and governments are assumed to provide a dynamic element, making statecraft through institutional design and reform an important aspect of political agency and an assertion of human will, understanding, and power to shape the world (March and Olsen, Reference March and Olsen1995; Heper, Kazancigil, and Rockman, Reference Heper, Kazancigil and Rockman1997; Goodin, Rein, and Moran, Reference Goodin, Rein and Moran2006: 3). Change is furthermore supported by the institutionalization of critical reflection and debate, legitimate opposition, and the rights for citizens to speak, publish, and organize, including civil disobedience. The mix of rules constraining and facilitating change varies across political systems, and the more heterogeneous a polity, the more likely it is that priority is given to rules protecting individuals and minorities (Weaver and Rockman, Reference Weaver and Rockman1993).
Identification
Institutionalists see identification and the internalization of accepted ways of doing things as a key process for understanding rule-following. Institutions affect individuals, their normative and causal beliefs, and not only their environments. Rules are followed because they are seen as legitimate and not solely because of external incentives, and belief in a democratic order and commitment to democracy’s institutions may be generated through socialization, education, and participation. Humans are born into a world of institutions where normative and causal beliefs are handed down from generation to generation, and the main institutions of the culture are (at least for a period of time) taken for granted. Humans are prepared, and prepare themselves, for different offices and roles. They may be recruited to specific positions on the basis of their normative and causal beliefs, and they are fashioned through on-the-job training and selective exposure to information (Simon, Reference Simon1957; March and Simon, Reference March and Simon1958; March and Olsen, Reference March and Olsen2006a).
People’s habits of mind, including their beliefs in legitimate political organization and rule, may be more difficult to change than formal rules and incentives. However, cultures and sub-cultures may inculcate respect for traditions or emphasize innovation and change, and some institutions, for example the university, are organized around skepticism to existing knowledge, beliefs, and practices. Even a Weberian bureaucracy, designed to be effective, independent of environmental variation and change, and usually seen as rule-bound and inflexible, is founded on beliefs in legitimate change as new knowledge and insights become available, governments shift, and courts of law (re)interpret existing rules.
The ways in which individuals are differently selected and fostered can also be a source of change as well as continuity. Like all organizational processes recruitment, socialization, education, participation, and identification are more or less ‘perfect’ in the sense that they, to different degrees, successfully select or mold people’s mind-sets. There is great variation across institutional settings and over time when it comes to what and who control such processes. Socialization agencies are weak or strong, and institutional cultures are more or less integrated. Participants are ‘social but not entirely socialized’ (Wrong, Reference Wrong1961: 191) and non-conformity is always a possibility. People also often have more than one identity, and change depends on which identity and rules of appropriateness are evoked in different contexts.
Socialization is, for example, affected by organizational growth rates, internal careers, and the length of apprenticeship for top positions, the frequency of promotions and rewards, the turnover of personnel, and the ratio of veterans and newcomers (Lægreid and Olsen, Reference Lægreid and Olsen1978). Institutional identities and memories are enhanced by a permanent civil service, compared to a spoil system such as the United States public administration where identities are weakened, memory is removed, and the ability to learn from experience is reduced because many key actors leave with changes in government (Peters, Reference Peters1996).
Interpretation and search
The impact of rules and identities depends on how they are interpreted. Core assumptions within the tradition of ‘bounded rationality’ in organizational studies are that all humans act on the basis of a simple model of the world and that the office one holds and the organizational setting in which one acts to a large extent provide the premises for action (Simon, Reference Simon1957; March and Simon, Reference March and Simon1958). Existing meaning systems, frames, and traditions of interpretation can be a source of inertia. However, thoughtful and imaginative reasoning about current and historical experience and the meaning of behavioral codes, causal and normative beliefs, and situations can also generate change – even a re-interpretation of an institution’s mission and role in society (March and Olsen, Reference March and Olsen1995). External impulses may also be interpreted in ways that increase or constrain their impacts. For example, global prescriptions of administrative reform have consistently been interpreted and responded to differently depending on national institutional arrangements and historical traditions (Christensen and Lægreid, Reference Christensen and Lægreid2007).
Change can follow from shifting institutional attention. An organization will usually enact the program believed to be most appropriate for the case at hand among the repertoires of options available. Most of the time actors attend to the tasks, targets, and task environments they are responsible for. Bounded rational actors do not constantly attend to institutional issues, if that is not their specific responsibility. Because time, energy, and attention are limited, some challenges are not faced, some opportunities are not realized, and competency traps reduce experimentation and produce ‘lock-ins’ (Arthur, Reference Arthur1989). The organization of attention then affects whether pressure for change accumulates, so that sudden change may follow from an internal re-focusing of attention. The better democratic politics and organizational routines work as feedback mechanisms, ensuring collective learning and continuous adaptation to feedback, the less need there is for comprehensive reform and the less likelihood of sudden breakdowns (Olsen, Reference Olsen1997).
Institutional routines are developed for fairly well structured and recurring problems and situations, and may look inappropriate when applied to ill-structured and non-recurring problems and situations. Searches for alternatives and innovations are initiated when available standard operating procedures are perceived to be unsatisfactory to solve problems, resulting in search in the neighborhood of problems or current alternatives (Cyert and March, Reference Cyert and March1963). Search and innovation can be driven by an internal aspirational pressure caused by enduring gaps between high institutional ideals and actual practices (Broderick, Reference Broderick1970). An example is unattainable democratic ideals that are never completely fulfilled (Dahl, Reference Dahl1998: 31). Institutional ideals can also be deliberately mobilized for change, as illustrated by the development of the European Parliament (EP). While the EP began with few of the functions and competencies usually found in national parliaments, the vision of ‘Parliament’ has been used, in particular in crises situations, to enhance the status and power of the EP (Héritier, Reference Héritier2007).
Search and innovations can, furthermore, follow because people gradually lose faith in institutional arrangements or from sudden performance failure. There can be not only external but also internal disenchantment, discontent, and a loss of faith in the institution and the authoritative interpreters of its mission, history, and future. Typically, taken-for-granted beliefs and arrangements are challenged by new or increased contact between previously separated entities based on different principles. Institutionalized beliefs can then be threatened by realities that are meaningless in terms of the beliefs on which an institution is founded. Unexplainable inconsistencies and incoherence cannot be dealt with by standard operating procedures, and change follows from efforts to reduce inconsistency and generate a more coherent interpretation of existing difficulties (Berger and Luckmann, Reference Berger and Luckmann1967: 103, 107–108). An important aspect of such processes is change in beliefs about what is inevitable and what it is possible to do. For example, for citizens and political leaders to imagine that they could apply reason and will to remake institutions, they had to begin to believe that institutions expressed the will and interests of humans and to discard the medieval belief that institutions reflected the creation and will of God, that institutions had existed since time immemorial, and that they would persist into perpetuity (Lathrop Gilb, Reference Lathrop Gilb1981: 467).
Resources
Under ideal democratic conditions all citizens have equal influence. In practice, the ability to comprehend, implement, and enforce rules, identities, and beliefs and to punish deviance depends on available resources for action. Institutions are defended by insiders and validated by outsiders and cannot be changed arbitrarily. Institutional resources can be mobilized to inhibit externally induced efforts to change as well as to amplify such impulses or initiate change (March and Olsen, Reference March and Olsen1989, Reference March and Olsen1995; Offe, Reference Offe2001).
Institutionalists, therefore, have to attend to the standard procedures through which institutions allocate and re-allocate resources and to how internal redistribution of resources, authority, and power may impact change. How much authority and power are the result of winning a majority in popular elections (Rokkan, Reference Rokkan1966)? How much influence is located in specific positions and roles and, in particular, what are the resources available for those who occupy institutional command posts (Wright Mills, Reference Wright Mills1956)? Resourceful, organized groups in society may initiate change and overwhelm and capture political institutions. Foreign influence, for example externally induced or assisted institutional change, is also well-known from colonialism to today’s reform programs of the World Bank and the International Monetary Fund (Nef, Reference Nef2003: 529). Institutions, however, are to varying degrees vulnerable to external changes in available resources, generating budgetary bonanzas or enduring austerity where expectations and demands are excessive compared to available resources.
Slack institutional resources may work as shock absorbers against environmental change and contribute to continuity. However, slack resources may also create surpluses that generate search, innovation, and change (Cyert and March, Reference Cyert and March1963). Slack resources may, furthermore, support institutional autonomy so that everyday-life inconsistencies and tensions are buffered by specialization, separation, sequential attention, and local rationality. Budgetary starvation or reduced slack are likely to generate demands for joint decisions and coordination, and such demands tend to make conflict and change more likely (Cyert and March, Reference Cyert and March1963).
Arguably, institutional specialization, separation, and autonomy help democracies cope with tensions that create conflicts and stalemates at constitutional moments. Constitutional decisions often generate struggles over the identity of the polity or specific institution. Due to their catch-all character, constitutional decisions easily become ‘garbage cans’ for a variety of ill-structured issues, characterized by competing or ambiguous goals, weak means-end understanding, and fluid participation (Cohen, March, and Olsen, Reference Cohen, March and Olsen1972; Olsen, Reference Olsen2003). Simultaneously, the demands for consistency become stronger. Institutional routines are challenged and it is more difficult to make joint decisions. Therefore, one hypothesis is that democratic systems work comparatively well because their political orders are not well integrated. Rather than subordinating all other institutions to the logic of one dominant center, democracies reconcile institutional autonomy and interdependence. Problem-solving and conflict resolution are disaggregated to different levels of government and institutional spheres, making it easier to live with unresolved conflict (Olsen, Reference Olsen2003 and Reference Olsen2007: Ch. 9).
Unresolved conflict
Institutions are not merely structures of voluntary cooperation and collective problem solving that produces desirable outcomes, and institutional change is not necessarily an apolitical, harmonious process. It cannot be assumed that conflict is solved through social integration and shared values, political consensus, or some prior agreement and ‘governing text’ (constitution, treaty, coalition agreement, or employment contract). Except at the level of non-operational goals, most organizations most of the time exist and thrive with considerable latent conflict (Cyert and March, Reference Cyert and March1963: 28, 117). Change processes assuming a single, unitary designer with well-specified objectives therefore have to be supplemented with processes involving conflict and unequal power (Knight, Reference Knight1992). Tensions and change may follow because those deciding, implementing, and being affected by rules are not identical (Farrell and Héritier, Reference Farrell and Héritier2007; Héritier, Reference Héritier2007), or because the dynamics of rules, beliefs, and resources are not synchronized.
Conflicts over the form of government and how society is to be constituted politically can be destructive as well as a source of innovation and improvement. Key questions are under what conditions democracies are successful in channeling discontent and protest into institutionalized conflict resolution and how different institutions influence how disputes are coped with. For example, political processes produce more or less clear winners and losers, and losers are often supposed to mobilize politically and demand change (Clemens and Cook, Reference Clemens and Cook1999). ‘Winner-take-all’ systems are then more likely to generate institutional oscillation with shifting political majorities while incremental change is more likely in political systems that routinely aim at sharing benefits and costs, including compensation for the losers.
While much of the literature attends to how conflicts between political parties and societal groups are dealt with and how mass mobilization through social movements produce change, institutionalists also have to study how intra- and inter-institutional conflicts within government may drive change. De-institutionalization is seen as creating ‘institutional chaos’ and an ‘institutional vacuum’ (Ágh, Reference Ágh2003: 541). Destroying the ancient régime is perceived as a precondition for clearing the way for a new set of institutions (Moore, Reference Moore1966: 16), and in market economies ‘creative destruction’, generated by entrepreneurs and competition, is seen to guarantee continuous change, as new ways of doing things eliminate outmoded and less efficient and profitable organizational forms and technologies (Schumpeter, Reference Schumpeter1994).Footnote 5
This view, however, has to be supplemented with the possibility that destruction is less complete. Democratic polities are generally uneasy about excessive change and the uneven distribution of gains and losses following from ‘creative destruction’; they usually try to reach compromises that modify the pace of change, compensate losers, and maintain social peace. European processes of transformation also suggest that there may be an asymmetry between institutionalization and de-institutionalization. New institutionalization has taken place at the European level without the predicted de-institutionalization (non-viability, withering, and demise) of the nation state. Rather than ‘creative destruction’, the main pattern has been that new institutions have supplemented rather than replaced national institutional arrangements. The European state has been under strain, but it has endured as a key political institution and contributed to its own transformation (Hurrelmann et al., Reference Hurrelmann, Leibfried, Martens and Mayer2007). Loosely coupled polities with partly autonomous institutions may in particular generate new institutions and keep old ones. Rather than a general ‘creative destruction’, there may be processes of sedimentation, making new and old institutions co-exist even when they are constituted on partly inconsistent principles (Sait, Reference Sait1938; Christensen and Lægreid, Reference Christensen and Lægreid2007; Olsen, Reference Olsen2007).
Tensions within and among institutions, nevertheless, provide a potential challenge to coherence and stability, as institutions organized upon competing principles and rules create problems for each other. While ‘political order’ suggests an integrated and coherent institutional configuration, polities are, as already argued, never perfectly integrated and monolithic. No democracy subscribes to a single set of doctrines and structures, and no grand architect has the power to implement a coherent institutional blueprint. Institutional arrangements are usually a product of situation-specific compromises. They fit more or less into a coherent order and they function through a mix of co-existing organizational and normative principles, behavioral logics, and legitimate resources.
Even a Weberian bureaucracy, the prototype of hierarchical organization, harbors competing claims to authority and logics of appropriate behavior. Bureaucrats are supposed to follow commands rooted in a formal position and public mandates generated through competitive elections. They are expected to be governed by rules, laws, and Rechtsstaat principles, and they are assumed to be dictated by professional knowledge, truth claims, and the democratic doctrine of enlightened government. The three competing claims are also embedded in different institutional contexts, i.e. elected government, courts of law, and institutions of higher education and professions (Olsen, Reference ——2008b). Likewise, diplomats face competing claims because diplomacy as an institution involves a tension between being the carrier of the interests of a specific state and of transnational principles, norms, and rules maintained and enacted by representatives of the states in mutual interaction (Bátora, Reference Bátora2005).
Polities, then, routinely face institutional imbalances and collisions and some of the fiercest societal conflicts have historically been between carriers of competing institutional principles. There are transformative periods characterized by major institutional confrontations and resource mobilization (Weber, Reference Weber1978). An institution may have its raison d’être questioned, and there are radical intrusions and attempts to achieve external control over the institution. There are also stern institutional defenses against invasions of alien norms, combined with a re-examination of the institution’s ethos, codes of behavior, primary allegiances, and pacts with society (Merton, Reference Merton1942; Maassen and Olsen, Reference Maassen and Olsen2007). Sometimes such collisions generate radical change in internal as well as external relationships.
However, while disagreement over inter-institutional organization is a possible source of change, change is unlikely to take the form of an instant shift from a coherent equilibrium to a new one (Olsen, Reference Olsen2004,Reference Olsen2007; Orren and Skowronek, Reference Orren and Skowronek2004; Hurrelmann et al., Reference Hurrelmann, Leibfried, Martens and Mayer2007). For example, strong relationships with other institutions make it difficult to redesign institutions (Peters, Reference Peters1999b), and in tightly coupled systems, change is likely to involve several institutional spheres and levels of government. The more loosely coupled a political order, the more likely are institution-specific processes of change. In fragmented systems, innovation may take place in partly autonomous communities where deviant ideas can be insulated long enough to mature before they are confronted with dominant ideas (March, Reference March2004). Institutions may then be transformed as participants learn from local experience and adjust local linkages rather than as a result of global rationality achieved through some singular grand process, such as deliberate choice, experiential learning, or competitive selection. Adaptation may be myopic, meandering, and ‘inefficient’ (March, Reference March1999) and ‘designs’ local (Goodin, Reference Goodin1996: 28–29).
The long list of relevant mechanism and factors suggests why it has been difficult to build simple models that explain institutional change and continuity. The difficulties are general – institutionalism is not an account that relates solely to institutions of democratic government. It may, nevertheless, be of value to explore in more detail possible implications for how democratic development is conceived. This is, in particular, so in a period when students of formal organizations primarily address business enterprises and economic organization, and innovation literature largely ignores democratic debate, competition, and institutions as a source of innovation and change (Fagerberg, Mowery, and Nelson, Reference Fagerberg, Mowery and Nelson2005).
Democracy as a complex adaptive system
An institutional approach calls attention to one of democracy’s great mysteries: how is it possible that a huge number of potentially chaotic decisions by ‘sovereign’ individuals generate a fairly stable order and a political community capable of making and implementing binding collective decisions? The discussion above suggests that to answer the question it may be useful to see democracies as examples of complex adaptive systems, i.e. self-renewing institutional arrangements that learn from experience. In relatively simple terms, self-organizing local relationships modify existing forms, generate new ones, and create order with properties that each component part does not have. Adaptation based upon a small set of standard rules and procedures generates systems of surprising complexity, so that the system persists while individual components change, and the interaction between the components is responsible for the persistence of the system (Holland, Reference Holland1995, Reference Holland1998; Axelrod and Cohen, Reference Axelrod and Cohen1999).
Unfinished democracy
When ‘democracy’ is used as a fixed category – a specific form of government classified according to stable properties in the taxonomic tradition of Aristotle and Linnaeus – making sense of change implies understanding how non-democratic systems become democratic and vice versa. In a dynamic perspective, however, democracies are being made continuously through processes of institutionalization, de-institutionalization, and re-institutionalization. The nature of democracy is to be unfinished and in the process of becoming rather than static.
Historically democratic institutions have waxed and waned as citizens and their leaders have developed and redefined (often unattainable) normative doctrines and organizational principles of good government towards which rulers and ruled are supposed to orient their behavior. The idea of democracy has been redefined as it has been linked to the city state, the nation state, emerging regional polities such as the European Union, and cosmopolitan democracy. There has been a succession of democratic forms and an open-ended series of institutional origins, transitions, and breakdowns. In many parts of the world there is a gap between citizens and their institutions, and electoral volatility, a reduced number of party members, new political parties and new alliances indicate less strong institutions and less predictability also in developed democracies (Mair, Reference Mair2007).
A customary starting point for making sense of democratic development is to see democracy as government by and for the governed. The people are ‘the only legitimate fountain of power’ (Hamilton, Jay, and Madison, Reference Hamilton, Jay and Madison1964: 11), and political institutions shall reflect the will of the people, understood as equal citizens. Governing, nevertheless, includes responding to citizens’ demands and societal change as well as initiating and driving change through forecasting and planning. Then what kinds of institutions does democracy require (Dahl, Reference Dahl1998)? Historically, different conceptions of ‘democracy’ have been reflected in competing institutional prescriptions. Today there is no stable consensus when it comes to ideas about appropriate ways of organizing government, and there is huge variation among what is in everyday language called ‘democracies’ with regard to how the will of the people is established and implemented in practice.
Until the middle of the nineteenth century, ‘democracy’ was regarded as an unstable and dangerous form of politics and incompatible with personal security and the right of property.Footnote 6 Since then democracy has become more valid. However, its meaning has been lost in ‘the cacophony of competing interpretations’, and contemporary democracies lack meaningful concepts for describing, criticizing, and defending political institutions and practices (Hanson, Reference Hanson1987), including the relationship between democratic norms and other legitimate normative standards. Normative political theory and practice provide incomplete answers – a loose framework rather than clear guidance – and there is an ongoing ‘belief battle’ over democratic ideas (Sartori, Reference Sartori1969: 87) and the desirability of particular political institutions. The democratic belief that political institutions should and can be designed and implemented to achieve preconceived goals also lives side by side with the observation that it is difficult to establish a firm theoretical basis for institutional design (Olsen, Reference Olsen1983: 9).
Have then democracies a unique ability to learn from experience and to adapt to shifting circumstances? If so, through what institutional mechanisms is learning and adaptation taking place? How do democracies adapt to and also adapt environmental circumstances? How do they fit themselves to changing environments and also fit external environments to themselves? Arguably, understanding democracy as a complex adaptive system can supplement accounts that give primacy to learning through a centralized authority and organizer, as well as individual learning through decentralized, voluntary exchange and competition.
Between centralization and decentralization
Democracies are usually seen as organized around making collectively binding rules, executing and applying those rules, sanctioning deviance, and adjudicating rule disputes (Rothstein, Reference Rothstein1996). Democratic organizational thinking, however, has a bias. Primacy is given to the formation of opinion and will in civil society and legislative supremacy. ‘[T]he authorization of rule is derived from elections’ (Przeworski, Reference Przeworski2006: 312), and the contestation of will and power and the selecting and legitimating of government take place in the electoral channel through the competition for people’s votes. That is, demos exercise power through public debates, electoral systems, competition among political parties, votes, and legislatures. Within this frame one center (usually the legislature) acts authoritatively on behalf of society and fashions institutions such as bureaucratic organizations and competitive markets through legal acts (Habermas, Reference Habermas1996: 75, 171).Footnote 7 It is the task of elected leaders to make ‘the state apparatus’ an efficient and smooth-running machine, to exercise control over agencies and agents, and to counteract usurpation and the arbitrary use of public power.
In contrast, Lindblom, relying on economic theories of competitive markets, argues that political science has an ‘overriding disposition’ towards centrality. The ‘intelligence of democracy’ – rationality, coordination, and efficiency – is, however, achieved via dispersion and fragmentation of power and mutual adjustment. Actors ‘have an eye on each other’ and there is coordination without a central organizer, common purpose or identity, or detailed rules (Lindblom, Reference Lindblom1965: 3, 305). New forms evolve and disappear without a deliberate act of design. The resulting system is an artifact, evolving through competition and struggle for advantages and existence among self-interested individuals. When the incentives are right, human actors will follow their selfish interest and at the same time further the common good. Within such decentralized, market-inspired approaches, the primary function of democracy is to aggregate predetermined preferences and resources. Democracies are held together by individual utility calculation and expediency.
An institutional perspective does not deny that the electoral channel is important and that a central authority sometimes has considerable organizing power or that individual autonomous adjustment is a significant process in contemporary democracies. Nevertheless, in contrast with decentralized approaches, institutionalism assumes that institutions can be integrative (March and Olsen, Reference March and Olsen1986). A core task for democratic institutions is to translate a heterogeneous and pluralistic society into a viable political community and to provide long-term systematic arrangements, agreed-upon principles, and procedures that have normative value in themselves. Institutions provide a framework for policy-making that affects but does not determine outcomes, and non-deterministic institutions may even be a precondition for legitimacy and compliance (Pitkin, Reference Pitkin1972; Di Palma, Reference Di Palma1990).
To be a member of a political community, a citoyen rather than a bourgeois, implies that collective life is to some degree governed by socially validated and individually internalized rules, norms, and understandings. Democratic politics is a fundamental process of interaction and reasoning that involves a search for collective purpose, direction, meaning, and belonging. Citizens discuss how they want to live together, what rules to follow, what the common good is, and what resources are legitimate in different institutional spheres – reasoning that may generate a belief in legitimate authority and order and develop empathy and trust (March and Olsen, Reference March and Olsen1986; Viroli, Reference Viroli1992). Democratic governing and politics involve shaping not only history but also citizens’ understanding of it and their willingness to accept it (March and Olsen, Reference March and Olsen1995). Citizens are educable and preferences, measures of success, and identities evolve over time. Democracy’s challenge is ‘to construct institutions and train individuals in such a way that they engage in the pursuit of the public interest […], and at the same time, to remain critical of those institutions and that training, so that they are always open to further interpretation and reform’ (Pitkin, Reference Pitkin1972: 240).Footnote 8 New beginnings are possible when people have a ‘hypothetical attitude’ toward existing institutions and forms of life (Habermas, Reference Habermas1996: 468).
In contrast with centralized approaches giving primacy to the electoral channel, institutionalism assumes that modern democracies are characterized by institutional differentiation. Over time new institutional spheres have split off from older ones and developed their own identities: politics, economics, administration, law, civil society, religion, science, art, and the family. Partly autonomous institutions are constituted on different normative and organizational principles, defining different actors, behavioral logics, arguments, resources, and distributional principles as legitimate. What is appropriate in one institutional sphere is inappropriate in others.
Normatively the idea of centralized, monolithic power in a single branch of government has been attacked as the very definition of majority tyranny and electoral despotism, and demands have been raised for institutions to provide ‘inefficiencies’ in adaptive processes in order to protect individuals and minorities. Limited government, separation of powers, checks, and balances, making decision-makers responsive to different constituencies, constitutions, the rule of law, bills of inalienable rights, and an independent judiciary have been prescribed as instruments for protecting individuals and minorities from misuse of political power and securing representation of all, and not only the majority (Hamilton, Jay, and Madison, Reference Hamilton, Jay and Madison1964: 47, 48). Modern democracies aspire to balance effective problem solving and protection of rights, and shifting conceptions of the appropriate mix are reflected in organizational arrangements.
While centralized approaches link learning and power to the formal decisions of a law-maker and are less interested in how formal-legal decisions are executed and turned into outcomes and effects, there is in practice huge variation across polities and over time when it comes to the role and capabilities of overall coordinating institutions in relation to public entities with legitimacy and resources of their own and allies in society. It is, however, unrealistic to assume à priori that some aspects of governing (e.g. making formal decisions) are ‘political’ while others (e.g. preparing and implementing formal decisions) are ‘apolitical’.
For example, while partially independent courts of law are usually acknowledged, public administration is relegated to a non-political instrument – a rational structure established to achieve coordination and maximize predetermined purposes. This view is not totally without merit, but as a general description it ‘must be rejected as empirically untenable and ethically unwarranted’ (Long, Reference Long1962a: 79, also Selznick, Reference Selznick1957).Footnote 9 Public administration is not a mere instrument for elected leaders. Over the last few years there has been growth in the number of non-majoritarian, regulatory agencies, kept at arms length from politicians (Majone, Reference Majone1996), and in the literature public administration is portrayed as a core institution of modern government, staffed with professionals with their own ethos, standards, and rules of appropriate behavior. Administrators have substantial discretion, control vast resources, and exercise power. They are active participants in the policy process. Public administration is a major point of contact between citizens and the state, a target of citizens’ influence, and is important in creating an image of government in the popular mind. Public administration also has a constitutive dimension: explicating collective interests; protecting values such as universality, equality, and legal security; providing fair implementation of laws and policies; securing predictability, accountability and control, and reducing corruption and favoritism (Peters and Pierre, Reference Peters and Pierre2003; Olsen, Reference Olsen2008a).
A complex institutional ecology
An institutional perspective, then, holds that democracy is a form of ordered rule involving an institutional sphere with the specific task of governing a territory and population. Political institutions have some autonomy from other spheres of society, absorptive and adaptive capabilities, and internal differentiation and coordination of offices and roles with specified authority and responsibility (Huntington, Reference Huntington1968). Political institutions are, nevertheless, embedded in a larger historical-institutional order, differently enabling and constraining individual institutions. Modern democracies form a complex ecology of partly autonomous yet interdependent and interconnected institutions with separate origins, histories, and traditions and different internal and external organization (March and Olsen, Reference March and Olsen1989: 170). There are many, often ‘inefficient’ and not necessarily synchronized and coordinated, institutionalized processes of will formation, decision-making, experiential learning, and adaptation. Therefore, the whole configuration of institutions across levels of government and institutional spheres has to be taken into account (Pierson, Reference Pierson2000; Pierson and Skocpol, Reference Pierson and Skocpol2002; Olsen, Reference Olsen2007). Understanding change requires information about how different types of institutions fit together, their interdependencies and interactions (Powell and DiMaggio, Reference Powell and DiMaggio1991; Scott, Reference Scott1995), and how change in one institution is linked to change in other institutions.
A hypothesis is that in routine and calm periods learning and adaptation largely take place in parallel, fairly autonomous institutional spheres, yet in the shadow of somewhat shared basic understandings or political pacts. Different institutions interpret and respond to external impulses through a set of standard operating procedures and bounded rational, simple models of the world, taking into account only selected parts of the environment. Institutionalized behavioral rules, understandings, and available resources are incrementally modified on the basis of experience, and individual institutions have a reservoir of rules and procedures, and therefore sources of internal variability. However, feedback from the environment is in particular important when large-scale failures and performance crises generate demands for more coordination. Then institutional developments are more likely to be influenced by the interaction, collisions, conflicts, meta-rules, and power struggles between several institutional spheres, adapting to each other, and it becomes less fruitful to study learning and adaptation in each sphere in isolation.
An implication is that there is a need to understand the organizational processes through which compromises and victories in political battles are ‘frozen’ into institutions, sustaining a lasting legacy (Lipset and Rokkan, Reference Lipset and Rokkan1967; Thelen, Reference Thelen1999: 390). An example of such institutionalization is the processes following ‘the birth of an organization’, i.e. after a formal decision is made to establish an entity and provide it with legal competences, offices, staffs, and budgets. Then the organization has to find its place in a larger institutional order, and its identity is shaped as it becomes aware of and adapts structures and practices to opportunities and constraints in the internal and external environments (Simon, Reference Simon1953; Selznick, Reference Selznick1957; Laffan, Reference Laffan1999).
Studies of institutionalization, de-institutionalization, and re-institutionalization, however, require long-term perspectives. For example, the institutionalization of the European state, as the key modern political formation, took centuries as administrative and military capacity was built and these capacities were legitimated through nationalization and culture-building, legalization, democratization and mass mobilization, and the development of the welfare state and social citizenship (Rokkan, Reference Rokkan1999; Bartolini, Reference Bartolini2005). Likewise, the emergence of the European Union (EU), as a multi-level and multi-centered polity, and a possible new stage in the development of the European state and democracy, illustrates the need to take an extended historical perspective. The EU is also an example of a political order with properties that the component part (member state) does not have. The new order has evolved through complex interaction between supra-national, intergovernmental, and trans-national processes producing outcomes that are difficult to predict precisely, even if the post-war trend has been towards more cooperation and integration. Because new institutions have arisen without older ones disappearing, there has been a tendency towards increasing complexity in political life, and the balancing of order and change has been affected by the increased intercourse among member states (Olsen, Reference Olsen2007).
Mea culpa, ideas about complex adaptive systems, have in this paper been used in a loose fashion. The literature on complex adaptive systems has so far not taken much interest in the emergence of democratic political institutions, and there is far to go before (if at all) the mechanisms and factors that influence such phenomena as the emergence of the EU are well understood. Nevertheless, an institutional approach and studies of complex adaptive systems, in particular those looking at organizational implications and conditions for harnessing complexity (Axelrod and Cohen, Reference Axelrod and Cohen1999), share many assumptions and puzzles. Therefore, ideas about complex adaptive systems may be useful for understanding democratic order and change in an institutionally differentiated polity, as a supplement to models assuming a single dominant central learner and organizer and market-inspired models assuming decentralized, individual learning and adaptation.
Much remains …
It is easy to agree that ‘[N]ew institutionalists should specify more rigorously the factors that change institutions and explicate the links between these factors and institutional change’ (Gorges, Reference Gorges2001). There are many unanswered questions. Why are institutions what they are; how do institutions matter and why do some matter more than others (Rothstein, Reference Rothstein1996)? When do routines stop being routine (Immergut, Reference Immergut2006: 241)? How do institutions unleash processes of stability and change simultaneously (Greif and Laitin, Reference Greif and Laitin2004: 636)? Which institutional collisions are likely to be consequential (Thelen, Reference Thelen1999: 397)? Is change in some institutions dependent on continuity in others (March, Schultz, and Zhou, Reference March, Schultz and Zhou2000)? What is the relationship between incremental adaptation and radical change and between the decline of one institutional order and the rise of another (Olsen, Reference Olsen1997 and Reference Olsen2007)? What is the role of intention, reflection, and choice in the development of institutions (Hamilton, Jay, and Madison, Reference Hamilton, Jay and Madison1787–88; Mill, Reference Mill1861)?
It can be difficult to distinguish between competing explanations that give similar predictions and disentangle the impact of institutions (Immergut, Reference Immergut2006: 249; Przeworski, Reference Przeworski2006: 325; Caporaso, Reference Caporaso2007), and processes of change can be linked in complex ways. For example, the emergence of a civil service reform may be part of the efforts to solve a specific problem, but the diffusion of reform may represent a growing legitimacy of these procedures (Tolbert and Zucker, Reference Tolbert and Zucker1983). Patterns also co-evolve and co-dissolve. Party systems, for example, may become de-institutionalized and break down because voters become more critical and less predictable. Yet voters may also lose their identities and shift their predilections because they can no longer make sense of party systems (Mair, Reference Mair2007: 152).
It has been argued that ‘it is a dismal science of politics (or the science of a dismal politics) that passively entrust political change to exogenous and distance social transformations’ (Di Palma, Reference Di Palma1990: 4). Nevertheless, the belief in the explanatory power of political institutions among students of politics has varied over time. The connections between human agency, institutional design, and change have remained ‘bafflingly complex’ (Moran, Reference Moran2006: 158) and today there is no agreed-upon theory explaining how institutions affect change. While scholars of different persuasions tend to agree that change and continuity depend on both institutional and non-institutional properties and that different approaches starting out with different privileged explanatory factors give partial insight, theorizing is frustrated by the need to reconcile the mutual influence of partly autonomous institutions, human agency, and macro-historical forces. All of them, and the relationships between them, matter but there is no agreement on the conditions under which one matters more than the others.
Institutionalism simply claims that relationships between political agency, large-scale societal processes, normative democratic prescriptions, existing institutional arrangements, and institutional development are complex and that knowledge about the functioning of formally organized institutions adds to our understanding of continuity and change in democratic contexts. In contrast with recent reform ideology, democracy’s problem in this paper has been seen as balancing stability and change, institutionalization and de-institutionalization rather than maximizing change. It has been argued that mainstream normative democratic theory is incomplete and has a biased ‘institutional theory’. Institutions and actors have been seen to mutually constitute each other, and contrary to conventional wisdom it has been shown that an institutional approach can deal with institutional dynamics, political agency, conflict, and power differentials.
Attention has been concentrated on how intra- and inter-institutional properties may affect processes through which institutions emerge and change rather than addressing the huge literature on deliberate reform and broad societal processes, including revolutions and wars. Routine processes of rule application, identification, interpretation, attention, search, resource allocation, and conflict resolution have been used to explore possible ‘inefficiencies’ in processes of change and how institutions may enable and constrain human agency and modify external impulses. While ideas about complex adaptive systems have some promise for understanding democratic dynamics, the aspiration of discovering a limited number of principles generating complex systems have certainly not been met in this paper. Rather, in terms of parsimony and clear predictions, the many mechanisms and the probabilistic and context-dependent trajectories of change found relevant may be discouraging. An institutional approach, nevertheless, assumes that institutional developments are better understood by analyzing the underlying processes than by specifying a (long) list of factors for a comparative static analysis of change.
However, ‘institutional change and order’ is probably a too heterogeneous phenomenon to be captured by any simple theory based upon a few grand generalizations and a dominant mechanism of change. An institutional approach invites further exploration of the processes through which institutional structures and processes affect human behavior and change and of how human action is translated into change in governmental institutions. There is a need to specify in more detail the latitude of purposeful institutional reform, environmental effectiveness in eliminating sub-optimal institutions through competitive selection, and the abilities of institutions to adapt spontaneously to deliberate reforms and environmental change in modern democracies (Olsen, Reference Olsen2001,Reference Olsen2008a; March and Olsen, Reference March and Olsen2006a, Reference March and Olsenb). Under what conditions – if any – are environments perfect enough (little friction, perfect knowledge, easy entry, many actors, no externalities) to eliminate non-competitive governmental institutions? For which institutions are there clear, consistent, and stable normative standards and adequate understanding and control so that institutions can be deliberately designed and reformed and actors achieve desired effects? Under what conditions are institutions perfectly adaptive, changing themselves or their environments in ways that create a fairly stable order?
