The dynamics of institutional change include elements of design, competitive selection, and the accidents of external shocks (Goodin 1996, 24–5). Rules, routines, norms, and identities are both instruments of stability and arenas of change. Change is a constant feature of institutions and existing arrangements impact how institutions emerge and how they are reproduced and changed. Institutional arrangements can prescribe and proscribe, speed up and delay change; and a key to understanding the dynamics of change is a clariWcation of the role of institutions within standard processes of change.
Most contemporary theories assume that the mix of rules, routines, norms, and identities that describe institutions change over time in response to historical experience. The changes are neither instantaneous nor reliably desirable in the sense of moving the system closer to some optimum. As a result, assumptions of historical eYciency cannot be sustained (March and Olsen 1989; March 1994). By ‘‘historical eYciency’’ we mean the idea that institutions become in some sense ‘‘better’’ adapted to their environments and quickly achieve a uniquely optimum solution to the problem of surviving and thriving. The matching of institutions, behaviors, and contexts takes time and has multiple, path-dependent equilibria. Adaptation is less automatic, less continuous, and less precise than assumed by standard equilibrium models and it does not necessarily improve eYciency and survival.
The processes of change that have been considered in the literature are primarily processes of single-actor design (in which single individual actors or collectivities that act as single actors specify designs in an eVort to achieve some fairly well- speciWed objectives), conXict design (in which multiple actors pursue conXicting objectives and create designs that reXect the outcomes of political trading and power), learning (in which actors adapt designs as a result of feedback from experience or by borrowing from others), or competitive selection (in which unvarying rules and the other elements of institutions compete for survival and reproduction so that the mix of rules changes over time).
Each of these is better understood theoretically than it is empirically. Institutions have shown considerable robustness even when facing radical social, economic, technical, and cultural change. It has often been assumed that the environment has a limited ability to select and eliminate political institutions and it has, for example, been asked whether governmental institutions are immortal (Kaufman 1976). In democracies political debate and competition has been assigned importance as sources of change. Yet, institutions seem sometimes to encourage and sometimes to obstruct reXection, criticism, and opposition. Even party structures in competitive systems can become ‘‘frozen’’ (Lipset and Rokkan 1967).
The ideal that citizens and their representatives should be able to design political institutions at will, making governing through organizing and reorganizing insti- tutions an important aspect of political agency, has been prominent in both democratic ideology and the literature. Nevertheless, historically the role of deliberate design, and the conditions under which political actors can get beyond existing structures, have been questioned (Hamilton, Jay, and Madison 1787 [1964, 1]; Mill 1861 [1962, 1]). In spite of accounts of the role of heroic founders and constitutional moments, modern democracies also seem to have limited capacity for institutional design and reform and in particular for achieving intended eVects of reorganizations (March and Olsen 1983; Goodin 1996; OVe 2001). Constitutions limit the legitimacy of design. The need for major intervention may be modest because routine processes of learning and adaptation work fairly well and the capability may be constrained by inadequate causal understanding, authority, and power (Olsen 1997).
The standard model of punctuated equilibrium assumes discontinuous change. Long periods of institutional continuity, where institutions are reproduced, are assumed to be interrupted only at critical junctures of radical change, where political agency (re)fashions institutional structures. In this view, institutions are the legacy of path dependencies, including political compromises and victories.5 Massive failure is an important condition for change.
The assumption, that institutional structures persist unless there are external shocks, underestimates both intra- and interinstitutional dynamics and sources of change. Usually, there is an internal aspiration level pressure for change caused by enduring gaps between institutional ideals and institutional practices (Bro- derick 1970). Change can also be rule-governed, institutionalized in speciWc units or sub-units, or be generated by the routine interpretation and implementation of rules. Typically, an institution can be threatened by realities that are mean- ingless in terms of the normative and causal beliefs on which it is founded, and eVorts to reduce inconsistency and generate a coherent interpretation are a possible source of change (Berger and Luckmann 1967, 103). As people gradually get or lose faith in institutional arrangements, there are routine switches between institutional repertoires of standard operating procedures and struc- tures. Reallocation of resources also impacts the capability to follow and enforce diVerent rules and therefore the relative signiWcance of alternative structures (March and Olsen 1995).
Thus, a focus on ‘‘critical junctures’’ may underestimate how incremental steps can produce transformative results (Streeck and Thelen 2005). For example, in the post-Second World War period most Western democracies moved stepwise towards an intervening welfare state and a larger public sector. The Scandinavian countries,
in particular, saw a ‘‘revolution in slow motion’’ (Olsen, Roness, and Sætren 1982). Since the end of the 1970s most Western democracies have moved incrementally in a neoliberal direction, emphasizing voluntary exchange, competitive markets, and private contracts rather than political authority and democratic politics. Suleiman, for example, argues that the reforms add up to a dismantling of the state. There has been a tendency to eliminate political belongings and ties and turn citizens into customers. To be a citizen requires a commitment and a responsibility beyond the self. To be a customer requires no such commitment and a responsibility only to oneself (Suleiman 2003, 52, 56).
Institutions face what is celebrated in theories of adaptation as the problem of balancing exploitation and exploration. Exploitation involves using existing knowledge, rules, and routines that are seen as encoding the lessons of history. Exploration involves exploring knowledge, rules, and routines that might come to be known (March 1991). Rules and routines are the carriers of accumulated knowledge and generally reXect a broader and a longer experience than the experience that informs any individual actor. By virtue of their long-term adaptive character, they yield outcome distributions that are characterized by relatively high means. By virtue of their short-term stability and their shaping of individual actions, they give those distributions relatively high reliability (low variability). In general, following the rules provides a higher average return and a lower variance on returns than does a random draw from a set of deviant actions proposed by individuals. The adaptive character of rules (and thus of institutions) is, however, threatened by their stability and reliability. Although violation of the rules is unlikely to be a good idea, it sometimes is; and without experimentation with that possibility, the eVectiveness of the set of rules decays with time.
It is obvious that any system that engages only in exploitation will become obsolescent in a changing world, and that any system that engages only in explor- ation will never realize the potential gains of its discoveries. What is less obvious, indeed is ordinarily indeterminate, is the optimal balance between the two. The indeterminacy stems from the way in which the balance depends on trade-oVs across time and space that are notoriously diYcult to establish. Adaptation itself tends to be biased against exploration. Since the returns to exploitation are typically more certain, sooner, and more in the immediate neighborhood than are the returns to exploration, adaptive systems often extinguish exploratory options before accumulating suYcient experience with them to assess their value. As a result, one of the primary concerns in studies of institutional change is with the sources of exploration. How is the experimentation necessary to maintain eVectiveness sustained in a system infused with the stability and reliability characteristic of exploitation (March 1991)?
Most theories of institutional change or adaptation, however, seem to be exquisitely simple relative to the reality of institutions that is observed. While the
concept of institution assumes some internal coherence and consistency, conXict is also endemic in institutions. It cannot be assumed that conXict is solved through the terms of some prior agreement (constitution, coalition agreement, or employment contract) and that all participants agree to be bound by institu- tional rules. There are tensions, ‘‘institutional irritants,’’ and antisystems, and the basic assumptions on which an institution is constituted are never fully accepted by the entire society (Eisenstadt 1965, 41; Goodin 1996, 39). There are also competing institutional and group belongings. For instance, diplomacy as an institution involves an inherent tension between being the carrier of the interests and policies of a speciWc state and the carrier of transnational principles, norms, and rules maintained and enacted by the representatives of the states in mutual interaction (Ba´tora 2005).
Institutions, furthermore, operate in an environment populated by other insti- tutions organized according to diVerent principles and logics. No contemporary democracy subscribes to a single set of principles, doctrines, and structures. While the concept ‘‘political system’’ suggests an integrated and coherent institutional conWguration, political orders are never perfectly integrated. They routinely face institutional imbalances and collisions (Pierson and Skocpol 2002; Olsen 2004; Orren and Skowronek 2004) and ‘‘politics is eternally concerned with the achievement of unity from diversity’’ (Wheeler 1975, 4). Therefore, we have to go beyond a focus on how a speciWc institution aVects change and attend to how the dynamics of change can be understood in terms of the organization, interaction, and collisions among competing institutional structures, norms, rules, identities, and practices.
Within a common set of generalized values and beliefs in society, modernity involved a large-scale institutional diVerentiation between institutional spheres with diVerent organizational structures, normative and causal beliefs, vocabularies, resources, histories, and dynamics. Institutional interrelations varied and changed. Institutions came to be specialized, diVerentiated, autonomous, and autopoietic— self-referential and self-produced with closure against inXuence from the environ- ment (Teubner 1993). There are strains and tensions and at transformative points in history institutions can come in direct confrontation. In diVerent time periods the economy, politics, organized religion, science, etc. can all lead or be led and one cannot be completely reduced either to another or to some transcendent spirit (Gerth and Mills 1970, 328–57; Weber 1978).
A distinction, then, has to be made between change within fairly stable institu- tional and normative frameworks and change in the frameworks themselves. For example, there are routine tensions because modern society involves several criteria of truth and truth-Wnding. It makes a diVerence whether an issue is deWned as a technical, economic, legal, moral, or political question and there are clashes between, for instance, legal and scientiWc conceptions of reality, their starting assumptions, and methods of truth-Wnding and interpretation (Nelken 1993, 151).
Likewise, there are tensions between what is accepted as ‘‘rational,’’ ‘‘just,’’ and a ‘‘good argument’’ across institutional contexts. DiVerent institutions are, for instance, based on diVerent conceptions of both procedural fairness and outcome fairness and through their practices they generate diVerent expectations about how interaction will be organized and diVerent actors will be treated (Isaac, Mathieu, and Zajac 1991, 336, 339).
There are also situations where an institution has its raison d’eˆtre, mission, wisdom, integrity, organization, performance, moral foundation, justice, prestige, and resources questioned and it is asked whether the institution contributes to society what it is supposed to contribute. There are radical intrusions and attempts to achieve ideological hegemony and control over other institutional spheres, as well as stern defenses of institutional mandates and traditions against invasion of alien norms. An institution under serious attack is likely to reexamine its ethos, codes of behavior, primary allegiances, and pact with society (Merton 1942). There is rethinking, reorganization, reWnancing, and possibly a new ‘‘constitutional’’ settlement, rebalancing core institutions. Typically, taken-for-granted beliefs and arrangements are challenged by new or increased contact between previously separated polities or institutional spheres based on diVerent principles (Berger and Luckmann 1967, 107–8).
Contemporary systems cope with diversity in a variety of ways. Inconsistencies are buVered by institutional specialization, separation, autonomy, sequential at- tention, local rationality, and conXict avoidance (Cyert and March 1963). Incon- sistencies are also debated in public and a well-functioning public sphere is seen as a prerequisite for coping with diversity (Habermas 1994). Modern citizens have lost some of the naive respect and emotional aVection for traditional authorities and the legitimacy of competing principles and structures have to be based on com- municative rationality and claims of validity. Their relative merits have to be tested and justiWed through collective reasoning, making them vulnerable to arguments, including demands for exceptions and exemptions that can restrict their scope (Kratochwil 1984, 701).
In general, the Enlightenment-inspired belief in institutional design in the name of progress is tempered by limited human capacity for understanding and control. The institutional frames within which political actors act impact their motivations and their capabilities, and reformers are often institutional gardeners more than institutional engineers (March and Olsen 1983, 1989; Olsen 2000). They can reinterpret rules and codes of behavior, impact causal and normative beliefs, foster civic and democratic identities and engagement, develop organized capabilities, and improve adaptability (March and Olsen 1995). Yet, they cannot do so arbitrar- ily and there is modest knowledge about the conditions under which they are likely to produce institutional changes that generate intended and desired substantive eVects.
