This on-line version is the pre-copyedited, preprint version. The published version can be found here:
‘The rise of governance and the risks of failure: the case of economic development’, International Social Science Journal, issue 155, 29-46, 1998.
This article explores the roles of markets, states, and partnerships in economic coordination and considers their respective tendencies to failure. The first section addresses the growing interest in governance and seeks explanations in recent theoretical developments. The second section then asks whether the rise of the governance paradigm might also reflect fundamental shifts in economic, political, and social life, such that governance will remain a key issue for a long time, or is a response to more cyclical shifts in modes of coordination. The third section considers the logic of ‘heterarchic governance’ in contrast to anarchic, ex post coordination through market exchange and imperative ex ante coordination through hierarchical forms of organization. It also offers some preliminary reflections on the nature, forms, and logic of ‘governance failure’. The final section addresses the state’s increasing role in ‘meta-governance’, i.e., in managing the respective roles of these different modes of coordination.
The rise of the ‘governance’ paradigm
Governance has only recently entered the standard anglophone social science lexicon and become a ‘buzzword’ in various lay circles. Even now its social scientific usages are often ‘pre-theoretical’ and eclectic; and lay usages are just as diverse and contrary. Nonetheless, in general terms, two closely related, but nested, meanings can be identified. First, governance can refer to any mode of coordination of interdependent activities. Among these modes, three are relevant here: the anarchy of exchange, organizational hierarchy, and self-organizing ‘heterarchy’. The second, more restricted, meaning is heterarchy (or self-organization) and is the focus of this article. Its forms include self-organizing interpersonal networks, negotiated inter-organizational coordination, and de-centred, context-mediated inter-systemic steering. The latter two cases involve self-organized steering of multiple agencies, institutions, and systems that are operationally autonomous from one another yet structurally coupled due to their mutual interdependence. These two features are especially significant in encouraging reliance on heterarchy. For, whilst their respective operational autonomies exclude primary reliance on a single hierarchy as a mode of coordination, their interdependence makes them ill-suited to simple, blind co-evolution based on the ‘invisible hand’ of mutual, ex post adaptation. Such incrementalism is sub-optimal because it is based on short-run, localized, ad hoc responses and thus takes inadequate account of the complex and continuing interdependence among these autonomous agencies, institutions, and systems.
The ‘self-organization of inter-organizational relations’ is a familiar form of governance in many different contexts. ‘De-centred, context-mediated inter-systemic steering’ is less familiar, however, at least outside the German-speaking world (where it is known as ‘dezentrierte Kontextsteuerung’, Glagow and Willke 1987). Thus an initial account should be useful. This form of governance involves the coordination of differentiated institutional orders or functional systems (such as the economic, political, legal, scientific, or educational systems), each of which has its own complex operational logic such that it is impossible to exercise effective overall control of its development from outside that system. The political and legal systems, for example, cannot control the overall development of the economy through coercion, taxation, legislation, judicial decisions, and so forth. This does not exclude specific external interventions to produce a particular result; it does exclude control over that result’s repercussions on the wider and longer-term development of the whole system. This indicates that there may be better prospects of ‘steering’ systems’ overall development by taking serious account of their own internal codes and logics and modifying the structural and strategic contexts in which these continue to operate; and by co-ordinating these contexts across different systems in the light of their substantive, social, and spatio-temporal interdependencies. Such steering is mediated not only through symbolic media of communication such as money, law, or knowledge but also through direct communication oriented to inter-systemic ‘noise reduction’, negotiation, negative coordination, and cooperation in shared projects (these terms are defined below). Inter-systemic coordination is typically de-centred and pluralistic and depends on specific forms of governance (Glagow and Willke 1987).
Although governance in the sense of heterarchy is found on three different levels (interpersonal, inter-organizational, and inter-systemic), the term itself is often limited to practices on the second level. This is consistent with recent lay usage, in which ‘governance’ refers to the mode of conduct of specific institutions or organizations with multiple stakeholders, the role of public-private partnerships, and other kinds of strategic alliances among autonomous but interdependent organizations. However, insofar as the relevant agencies, stakeholders, or organizations are based in different institutional orders or functional systems, problems relating to inter-systemic steering will also affect the ‘self-organization of inter-organizational relations’ even if they are not explicitly posed in this context.
Etymology, genealogy, and discourse
The anglophone term ‘governance’ can be traced to the classical Latin and ancient Greek words for the ‘steering’ of boats. It originally referred mainly to the action or manner of governing, guiding, or steering conduct and overlapped with ‘government’. For a long time, usage was mainly limited to constitutional and legal issues concerning the conduct of ‘affairs of state’ and/or to the direction of specific institutions or professions with multiple stakeholders. It has enjoyed a remarkable revival over the last 15 years or so in many contexts [as of 1998, BJ], however, becoming a ubiquitous ‘buzzword’ that can mean anything or nothing. The key factor in its revival has probably been the need to distinguish between ‘governance’ and ‘government’. Thus governance would refer to the modes and manner of governing, government to the institutions and agents charged with governing, and governing to the act of governing itself. The analogous German concept of Steuerung (steering, guidance) proved popular in the 1970s and 1980s and for much the same reasons. But it also has a fourth connotation in German through its links to systems theories. Mayntz notes, for example, that, in systems-theoretical terms, governing refers to the deliberate action of bringing an autonomous system as an object of governance from one state into another: whether to stabilize it, re-direct it, or transform it (1993b: 11-12).
This general etymological account does not explain why a relatively dormant concept with limited scope and restricted usage came to be revitalized at a particular moment and has been applied by so many individuals, agencies, and organizations to so many different topics. In theoretical terms, this can be linked to certain paradigmatic crises in the social sciences in the 1970s and 1980s — crises partly due in turn to dissatisfaction with their capacity to describe and explain the ‘real world’. An interest in ‘governance’ as a major theme is rooted in the rejection of several simplistic dichotomies that inform the social sciences. These include: market vs hierarchy in economics; market vs plan in policy studies; private vs public in politics; and anarchy vs sovereignty in international relations. Indeed Scharpf was prompted to write:
Considering the current state of theory, it seems that it is not so much increasing disorder on all sides that needs to be explained as the really existing extent, despite everything, of intra- as well as inter- organizational, intra- as well as inter-sectoral, and intra- as well as international, agreement and expectations regarding mutual security. Clearly, beyond the limits of the pure market, hierarchical state, and domination-free discourses, there are more — and more effective — coordination mechanisms than science has hitherto grasped empirically and conceptualized theoretically’ (Scharpf 1993: 57, my translation).
This dissatisfaction is reflected in recent disciplinary reorientations. Institutional economics, for example, shows growing interest in mechanisms (such as organized markets within group enterprises, clans, networks, trade associations, and strategic alliances) that coordinate economic activities in other ways than exchange and hierarchy. International relations has discovered ‘international regimes’, i.e., forms of international coordination that avoid international anarchy and yet by-pass the nation state — and that have therefore been described as involving ‘governance without government’ (Rosenau and Czempiel 1992). In political science attention has turned to forms of coordination that not only span the conventional public-private divide but also involve ‘tangled hierarchies’, parallel power networks, or other forms of complex interdependence across different tiers of government and functional domains. And the spread of systems theory is linked to growing interest in the self-regulation of complex systems in turbulent environments and the problems posed by attempts to ‘steer’ such systems from outside. In these and other contexts, then, ‘governance’ is seen as a missing ‘third term’ that both critiques and complements more dichotomous approaches.
The real world of governance
I now consider whether the link between conceptual interest in governance and social change involves anything more than transferring old wine into new bottles. Many practices now subsumed under ‘governance’ have been examined under other rubrics. Thus corporatism, public-private partnerships, industrial districts, trade associations, statecraft, diplomacy, interest in ‘police’ (Polizei), policy communities, international regimes, etc., all involve aspects of what is now termed ‘governance’. In this sense, there are clearly pre-cursors of current interest in governance in various disciplines. One could interpret this in at least four ways. First, regardless of the changing importance or otherwise of heterarchy, the significance of governance in lay discourses has changed and this is reflected in social science scholarship. Second, a stable but recently subterranean stream of heterarchic practices has re-surfaced and begun to attract renewed attention. Third, after becoming less significant compared with other modes of coordination, heterarchy has once again become important. And, fourth, an upward trend has continued, is becoming dominant, and is likely to continue to do so. There is a kernel of truth in each interpretation.
The first possibility is suggested by the expansion of governance discourses. These range from ‘global governance’ through ‘multi-level governance’ and the shift from ‘government to governance’ to issues of ‘the stakeholding society’ and ‘corporate governance’. Given the close, mutually constitutive relationship between the social sciences and lay discourses, this suggestion would be worth exploring further.
The second possibility is the persistence of underlying realities beneath the vagaries of intellectual fashion. So-called ‘governance’ mechanisms (as contrasted to markets or hierarchy) have long been widely used in coordinating complex organizations and systems. There have always been issues and problems for which heterarchic governance is, so to speak, the ‘natural’ mode of coordination. Certain forms of interdependence are inappropriate for (or at least resistant to) market and/or imperative coordination. For example, public-private partnership is theoretically well-suited in cases of organized complexity characterized by a loose coupling of agents, complex forms of reciprocal interdependence, and complex spatio-temporal horizons. In addition, different state traditions have given more or less scope for market forces and/or self-regulation to operate in their economies and civil societies. Here one can contrast, for example, the Anglo-American tradition with the differing traditions in France or Germany (Dyson 1980). There are also normative preferences for self-organization in certain contexts. This socially necessary minimum of heterarchic practices makes it all the more curious that they have only recently attracted focused scientific interest. This is almost certainly related to the blindspots associated with particular disciplinary paradigms or prevailing forms of ‘common sense’. Thus, during the post-war period of growth based on a virtuous circle of mass production and mass consumption in North America and Western Europe (hereafter referred to as ‘Atlantic Fordism’), when the ‘mixed economy’ was a dominant paradigm, institutions and practices intermediate between market and state were often neglected. They had not actually disappeared; they were simply marginalized theoretically and politically. Subsequent disenchantment with the state in the 1970s, and with markets in the 1990s, has renewed interest in something that never really disappeared.
A third factor contributing to the rise of governance is the cycle of modes of coordination. All modes are prone to dilemmas, contradictions, paradoxes, and failures but the problems differ with the mode in question. Markets, states, and governance fail in different ways. One practical response to this situation is to combine modes of policy-making and vary their weight over time — thereby shifting the forms in which tendencies to ‘failure’ manifest themselves and creating room for manoeuvre (Offe 1975). The re-discovery of governance could mark a fresh revolution in this process — a simple cyclical response to past state failures (especially those linked to attempts to manage the emerging crisis of Atlantic Fordism from the mid-1970s) and, more recently, market failure (and its associated crisis in corporate governance).
The fourth possibility is that a fundamental secular shift in state-market-society relations has occurred. This implies that important new economic and social conditions and attendant problems have emerged that cannot be managed or resolved readily, if at all, through top-down state planning or market-mediated anarchy. This secular shift reflects the dramatic intensification of societal complexity that flows from growing functional differentiation of institutional orders in an increasingly global society — which leads in turn to greater systemic interdependencies across various social, spatial, and temporal horizons of action. As Scharpf notes:
… the advantages of hierarchical coordination are lost in a world that is characterized by increasingly dense, extended, and rapidly changing patterns of reciprocal interdependence, and by increasingly frequent, but ephemeral, interactions across all types of pre-established boundaries, intra- and interorganizational, intra- and intersectoral, intra- and international’ (Scharpf 1994: 37).
In this sense, the recent expansion of networks at the expense of markets and hierarchies and of governance at the expense of government is not just a pendular swing in some regular succession of dominant modes of policy-making. It reflects a shift in the fundamental structures of the real world and a corresponding shift in the centre of gravity around which policy cycles move.
The rise of governance practices
The rise of governance is partly due to secular shifts in political economy that have made heterarchy more significant than markets or hierarchies for economic, political, and social coordination. I now consider the reasons for this by undertaking three tasks: first, identifying the logic of governance as a distinctive coordination mechanism in contradistinction to markets and imperative coordination; second, distinguishing three main types of heterarchic governance in terms of the sites on which they operate; and, third, considering more fully what societal (or macro-social) changes might have made heterarchy more appropriate as an economic coordination mechanism.
First, the most general case for the rise of heterarchic governance can be made in terms of the evolutionary advantage (in terms of relative capacities to innovate and learn in a changing environment) that it offers in certain types of circumstances. For self-organization is especially useful in cases of loose coupling or operational autonomy, complex reciprocal interdependence, complex spatio-temporal horizons, and shared interests or projects. Mayntz has discussed networks as a form of heterarchic governance in these terms. She suggests that their typical logic is that of negotiation directed to the realization of a joint product, such as ‘a specific technical innovation, a city plan, a strategy of collective action, or a problem solution in public policy’ (Mayntz 1993a:11). I would add that such negotiation typically occurs in the context of more or less complex forms of interpersonal and inter-organizational networking that bring and keep together those involved in negotiation; that the key to successful negotiation is noise reduction, i.e., reducing mutual incomprehension in the communication between different institutional orders in and through attempts to enhance understanding and sensitivity to their distinctive rationalities, identities, and interests; and that, once agreements are reached, they form the basis for negative and positive coordination of activities. In short, if reliance on heterarchy has increased, it is because increasing interdependencies are no longer so easily managed through markets and hierarchies.
Second, I want to distinguish three types of heterarchy in terms of the types of action system being coordinated. These are: interpersonal networking in which individuals represent themselves and/or their functional systems but are not mandated to commit specific agencies or organizations; the self-organization of inter-organizational relations based on negotiation and positive coordination in task-oriented ‘strategic alliances’ based on a (perceived or constructed) coincidence of interests and dispersed control of the interdependent resources needed to produce a joint outcome that is deemed to be mutually beneficial; and more programmatic or mission-oriented inter-systemic steering based on noise reduction (see above) and on negative coordination, i.e., taking account of the possible adverse repercussions of one’s own actions on third parties or other systems and exercising self-restraint as appropriate. These three forms of heterarchic governance are often linked in tangled hierarchies. Thus, interpersonal trust can facilitate inter-organizational negotiation; inter-organizational dialogue facilitates inter-systemic communication; and the resulting noise reduction can promote interpersonal trust by enhancing mutual understanding and by stabilizing expectations.
Third, turning to the macro-social changes that might explain the growth of heterarchy, I focus on the interdependencies in and across the economy and polity. The world economy is being reshaped by a complex dialectic of globalization-regionalization. This has allegedly made it more difficult for (national) states to control economic activities within their borders — let alone global capitalist dynamics. Once the relative coincidence of coherent economic spaces and national territories typical of Atlantic Fordism (see above) was undermined by internationalization of the economy (especially among the advanced capitalist economies), faith in the national state’s capacities to govern the economy was undermined. A corresponding increase in the ‘unstructured complexity’ of the economy on a world scale has triggered attempts on various spatial scales (from local to global) to re-impose some structure and order through resort to heterarchic coordination.
Such attempts are further complicated because capital accumulation has come to depend more heavily on a wide range of extra-economic factors generated through other institutional orders on various spatio-temporal scales. Whilst accumulation has always depended on non-commodified as well as commodified relations, this dependence has become more marked due to a fundamental change in the conditions making for competitiveness. In contrast to the Ricardian approach of maximizing the efficiency with which material factors of production are allocated to different activities, competitiveness is now understood in more Schumpeterian terms as having a ‘structural’ or ‘systemic’ quality (Chesnais 1987; Jessop 1993; Messner 1996). It now depends not only on an extensive range of economic factors but also on the capacity to derive added value from a wide range of extra-economic institutions and relations. This is closely linked to the increased importance of the micro-level and the supply-side in economic performance and the correlative importance of complex interdependence in promoting competitiveness. This has greatly extended the economic and extra-economic interdependencies bearing on competition — whether at the level of the firm, the sector or branch, or specific economic spaces. Such changes have major implications both for the internal and external relations of the organizational units of competition (firms, etc.). These have been well expressed in the following terms:
The traditional models of the large, vertically integrated firm of the 1960s, and of the small autonomous, single-phase firm of the 1970s and part of the 1980s, are replaced by a new type of large networked firm, with strongly centralized strategic functions extending in several directions, and by a new type of small enterprise, integrated into a multi-company local network. Across the network, a system of constantly evolving power relationships governs both the dynamics of innovation and the appropriability of returns to the partners involved. The network firm is attracted towards diversified mass production and the competitive factor of the single firm is the control of complementary assets in the hands of its potential partners (Capello 1996: 490).
These changes also make public-private partnerships and other forms of heterarchy more relevant than conventional legislative, bureaucratic, and administrative techniques. This is seen in a turn from the ‘Keynesian welfare national state’ to a more complex, negotiated system oriented to international competitiveness, innovation, flexibility, and ‘enterprise culture’. The primary coordination instruments in the Keynesian welfare system were the market and the state. They were articulated in a ‘mixed economy’ in which big business, big labour, and the big state often engaged in tripartite concertation at the national or regional level. In the emerging Schumpeterian workfare regime (Jessop 1993) the market, the national state, and the mixed economy have lost significance to inter-firm networks, public-private partnerships, and a multilateral and heterarchic ‘negotiated economy’ (Nielsen and Pedersen 1993). Moreover, in contrast to the primarily national focus of the mixed economy, these new forms of negotiated economy also involve ‘key’ economic players from local and regional as well as national and, increasingly, international economic spaces. This is linked to the partial ‘hollowing out’ of national states through the expansion of supranational government, local governance regimes, and transnationalized local policy networks in an attempt to enhance the ‘de-centred context-mediated steering’ of capitalist economies. And this latter shift poses further coordination problems concerning the management of the inter-scalar as well as inter-systemic dependencies (see below).
Governance success and governance failure
These changes may make anarchic and hierarchic modes of coordination more problematic. But it does not follow that the structural and strategic conditions for effective heterarchy will be always and everywhere sufficiently developed to ensure that it will outperform continued reliance on market forces or top-down control. Large literatures already exist on market and state failure. It is equally important to examine the governance failure and consider what affects its likelihood and the capacities for recuperating or responding to such failure. Noting the problems and risks of governance will help us see through the current rhetoric surrounding ‘public-private partnership’ and the associated tendency to highlight successes and downplay failures (cf. Capello 1996).
The capitalist market has a procedural rationality. This is formal in nature, prioritizing an endless ‘economizing’ pursuit of profit maximization. By contrast, government has a substantive rationality. It is goal-oriented, prioritizing ‘effective’ pursuit of successive policy goals. Market coordination and hierarchy are prey to the problems of bounded rationality, opportunism, and asset specificity (Coulson 1997). Heterarchic governance is based on a third type of rationality: reflexive rationality. The key to its success is continued commitment to dialogue to generate and exchange more information (thereby reducing, without ever eliminating, the problem of bounded rationality); reducing opportunism through locking governance partners into a range of interdependent decisions over a mixture of short-, medium-, and long-term time horizons; and building on the interdependencies and risks associated with ‘asset specificity’ by encouraging solidarity among those involved. It thereby supplements market exchange and government hierarchy with institutionalized negotiations to mobilize consensus and build mutual understanding. Individual economic partners give up part of their autonomy in economic decision-making in exchange for political influence and a better overall functioning of the system; and the state gives up part of its capacity for top-down authoritative decision-making in exchange for influence over economic agent and more effective overall economic performance. In this sense the rationality of governance is dialogic rather than monologic and this requires an investment of time to work effectively.
The conditions for successful heterarchic governance depend on the modes of coordination adopted, the constitution of the objects of governance, and the environments within which relevant actors coordinate their activities to achieve their objectives. Interpersonal networking, inter-organizational negotiation, and inter-systemic steering pose different problems in this regard. Objects of governance will also affect the likelihood of success. For example, governing the global economy, human rights regimes, and transnational social movements clearly involve very different problems. Finally, turbulent environments pose different governance problems from those that are relatively stable. I now refer briefly to some differences among the three types of heterarchic governance and then offer some initial generalizations about governance success and failure.
The simplest form of heterarchy arises from the selective formalization of interpersonal networking. Individual actors build on their past familiarity with others in various interpersonal networks to form a more exclusive, more targeted partnership; partners share an imagined community of interests and orientation to the future and they use selective memories to reinforce trust (Elchedus 1990: 197-8; Luhmann 1979: 16-19). In the first instance, partners represent only themselves; but they may also be regarded as speaking informally on behalf of institutional orders from which they are recruited. If their actions are confined to interpersonal networks, however, partners cannot commit the organizations or institutions from which they may be recruited and/or which they represent symbolically.
The self-organization of inter-organizational relations is usually more complex. It emerges where materially interdependent but formally autonomous organizations, each of which controls important resources, must coordinate their actions to secure a joint outcome that is deemed mutually beneficial. To this end they negotiate to identify common objectives and engage in positive coordination to achieve these aims. The continued pursuit of common long-term objectives typically depends upon the realization of shorter-term objectives and general compliance with established or emerging inter-organizational expectations and rules. Crucial to the success of such partnerships is resource synergy, i.e., the ‘added value’ that comes from partners combining resources rather than acting alone (Hastings 1996). This should be linked in turn to the building of inter-organizational capacities that surpass the powers of any individual member organization.
The most complex form of governance is found in attempts to facilitate the mutual understanding and co-evolution of different institutional orders to secure agreed societal objectives. In addition to inter-personal networking and inter-organizational negotiation and positive coordination, two further mechanisms are crucial here: ‘noise reduction’ and negative coordination (on which, see above). Ideally, noise reduction involves mutual understanding through dialogue rather than the forceful colonization (or penetration) of other systems by the rationale and logic of one, dominant system. Likewise, if negative coordination is not to become a disguised form of imperative coordination, it should also be based on genuine pluralism rather than the sheer dominance of one system, its operational codes, and social dynamic. Together these mechanisms may help to realize an inter-systemic consensus around visions or missions which provide a basis for more specific inter-organizational arrangements oriented to positive coordination of relevant activities around specific objectives. This is the situation that existed in the period of Atlantic Fordism with its commitment to full employment and social welfare as the context for specific inter-organizational projects as well as inter-systemic cooperation. It is something that the neo-liberal hegemonic project also sought to achieve.
Each form of governance has its own distinctive problems. Thus interpersonal networks involve an acute problem of trust as more actors get involved and/or the material stakes increase. Such problems are reinforced on an inter-organizational level by difficulties in securing the internal cohesion and adaptability of individual organizations; and in making their respective operational unities and independence compatible with their de facto material, social, and spatio-temporal interdependence on other organizations. Finally, partnerships that cross institutional boundaries face further problems due to difficulties in mutually coordinating (let alone unilaterally controlling) institutional orders that operate according to their own distinctive logics.
Nonetheless different forms of governance arrangement can also be mutually supportive. Interpersonal trust may ease inter-organizational negotiation and/or help build less personalized, more ‘generalized trust’ as organizations and other collective actors (including inter-organizational partnerships) are seen to sacrifice short-term interests and reject opportunism. In turn, inter-organizational dialogue helps to formulate and represent the identities and/or interests of different institutional orders and so eases inter-systemic communication. It can also generate ‘systemic trust’ (in the integrity of other systems’ codes and operations) by promoting mutual understanding and stabilizing reciprocal expectations around a wider ‘societal project’ as the basis for future self-binding and self-limiting actions. And, as noted above, these different levels of heterarchy are often involved in tangled hierarchies.
Regardless of the level(s) on or across that they operate, attempts to build effective governance mechanisms should include:
(a) Simplifying models and practices that reduce the complexity of the world but are congruent with real world processes and relevant to governance objectives. These models should simplify the world without neglecting significant side effects, interdependencies, and emerging problems. Some bodies may specialize in such model building and/or in monitoring their adequacy.
(b) Developing the capacity for dynamic interactive learning about various causal processes and forms of interdependence, attributions of responsibility and capacity for actions, and possibilities of coordination in a complex, turbulent environment. This is enhanced when actors are able to switch among different modes of governance to facilitate more effective responses to internal and/or external turbulence.
(c) Building methods for coordinating actions across different social forces with different identities, interests, and meaning systems, over different spatio-temporal horizons, and over different domains of action. This depends on the self-reflexive use of self-organization to sustain exchange, negotiation, hierarchy, or solidarity as well as on the specific nature of the coordination problems engendered by operating on different scales and over different time horizons.
(d) Establishing both a common world view for individual action and a system of meta-governance to stabilize key players’ orientations, expectations, and rules of conduct. This allows for a more systematic review and assessment of problems and potentials, of resource availability and requirements, and the framework for conntinued commitment to negative and positive coordination.
Each of these four conditions bears on the problem of establishing secure bases of coordination and giving them a structurally-inscribed strategic selectivity that rewards continued compliance. This does not exclude (and, indeed, may well require) a certain ambivalence and real flexibility in governance mechanisms so that an adequate repertoire of governance routines exists to ensure continued vitality in the face of a turbulent environment. This requisite variety (with its informational, structural, and functional redundancies) plays an important role in the adaptability of intra- and inter-organizational networks and de-centred inter-systemic steering (cf. Grabher 1994; Willke 1992). It promotes the ability to alter strategies and select those that are successful. This may seem inefficient at first sight because it introduces slack or waste into organizations and movements. But it also provides major sources of flexibility in the face of failure. For, if every mode of economic and political coordination is failure-laden, relative success in coordination over time depends on the capacity to switch modes of coordination as the limits of any one mode become evident. This provides in turn the basis of displacing or postponing failures and crises.
Regarding the environment in which heterarchy operates, co-ordination is most likely to succeed where conditions are sufficiently stable and the options sufficiently restricted that reflexive monitoring, interactive learning, and incremental change can occur. Relatively stable, non-turbulent environments facilitate the selection of responses that prove successful and thereby enable governance mechanisms to stabilize them. Conversely turbulence means that any lessons from previous successes or failures may be inapplicable in rapidly changing circumstances. This argument also applies, of course, to the use of imperative coordination. Those who see markets as discovery mechanisms also presuppose some measure of stability in the environment.
The growing fascination with governance mechanisms as a solution to market and/or state failure should not lead us to overlook the risks involved in attempts to substitute governance for markets and/or hierarchies and the resulting likelihood of governance failure. However, whilst there are already extensive literatures on market failure and state failure, there is far less direct, explicit, and focused concern with governance failure. Yet, if both market and state failure are recognized in social sciences, we should also thematize governance failure.
Market failure is usually seen as the failure of markets to provide economically efficient allocations in and through pursuit of monetized private interests (as would, presumably, occur if the market functioned according to the standards of an imaginary perfect market). State failure is a failure to achieve substantive political objectives defined as in the public interest and enforced as necessary against particular interests. Those who believe in the beneficence of market forces, regard state failure as normal and market failure as exceptional; they generally respond to market failure by calling for more market, not less! Conversely, those who believe in the rationality of the state and its embodiment of the public interest, typically consider market failure as inevitable and state failure as something that, if not exceptional, is at least conjunctural. They therefore conclude that it can be solved through improved institutional design, knowledge, or political practice. To recognize the inevitability and centrality of market and state failure alike would surely be a recipe for pessimism (Malpas and Wickham 1995: 39). For those who recognize at least the formal procedural rationality of markets, it might still be possible to adjudge market outcomes as failures in terms of substantive (political) criteria, such as an unjust distribution of life-chances. Likewise, even if one accepts that state elites are motivated by the public interest, political outcomes might still be seen as failures in terms of formal (economic) criteria, such as the oversupply of poor quality, high priced public goods (cf. Mitchell and Simmons 1994).
The criterion for governance failure is not immediately obvious. For there is no pre-given formal maximand or reference point to judge governance success, as there is in the economy with its emphasis on monetized profits and/or the (imaginary) perfect market outcome. Nor is there a pre-given substantive criterion — the realization of specific political objectives — as there is with imperative coordination. For the point of governance is that goals are modified in and through negotiation and reflection. In this sense governance failure would presumably consist in the failure to redefine objectives in the face of continuing disagreement about whether they are still valid for the various partners involved.
But one can also apply procedural and substantive criteria to heterarchy and assess whether it produces more efficient long-term outcomes than the market and more effective long-term outcomes than top-down coordination. This involves shifting perspective somewhat and implies a comparative evaluation of all three modes of coordination in terms of all three rationalities. This can be seen in the increasing interest in heterarchy as a mechanism to reduce transaction costs in the economy in cases of bounded rationality, complex interdependence, and asset specificity. It is also reflected in the increasing interest on the part of the state apparatus in heterarchy as a mechanism to enhance the state’s capacity to achieve political objectives by sharing power with forces beyond the state and/or delegating responsibilities for specific objectives to partnerships (or other heterarchic arrangements) acting in the shadow of hierarchy. In the same way, of course, partnerships (or other heterarchic arrangements) may simplify the pursuit of their own goals by relying on the market or the state to fulfil certain aspects of their jointly-agreed projects. There is an interesting research agenda implied in these reflections but, rather than pursue it further in the present article, I want to consider the limits to heterarchic governance as a mechanism for pursuing economic development. In this sense I will be applying procedural and substantive criteria to heterarchy and thus judging its performance in more instrumental terms.
There are three main sets of factors that limit the success of governance in guiding economic development. The first is inscribed in the very dynamic of capitalism itself and affects all forms of economic and social coordination, including the market mechanism itself. For capitalist growth depends on the market-mediated exploitation of wage-labour — not the inherent efficiency of unfettered markets. Markets mediate the search for added value but cannot produce it. And commodification generates contradictions that cannot be resolved by the market mechanism. This is evident in contradictions inscribed in most basic forms of capitalist market society. Thus the commodity is both an exchange-value and a use-value; the wage is both a cost of production and a source of demand; money is both national money and international currency; productive capital is both abstract value in motion and a concrete stock of time- and place-specific assets in the course of being valorized; and so on. In this sense much of what passes as market failure is actually an expression of the underlying contradictions of capitalism.
This suggests that substituting the state or heterarchy for the market will not eliminate the underlying obstacles to smooth economic performance. For heterarchy does not substitute non-capitalist principles for those of the market or introduce a neutral third term between market and state (let alone between capital and labour). It adds yet another area where the dilemmas, contradictions, and antagonisms of capitalism are expressed — including those often discussed in terms of the conflict between capital accumulation and political legitimacy. This is especially important to grasp because much literature on economic governance focuses on the modalities rather than objects of governance and thereby ignores the distinctive constraints imposed by the self-organizing dynamic and inter-systemic dominance of capitalism.
The second set of constraints concerns the insertion of heterarchy into the broader political system. This particularly concerns the relative primacy of different modes of coordination and their differential access to the institutional support and the material resources necessary to pursue reflexively-agreed objectives. Among crucial issues here are the flanking and supporting measures that are taken by the state; the provision of material and symbolic support; and the extent of any duplication or counteraction by other coordination mechanisms. We can distinguish three aspects of this second set of constraints. First, as both governance and government mechanisms exist on different scales (indeed one of their functions is to bridge scales), success at one scale may well depend on what occurs on other scales. Second, coordination mechanisms may also have different temporal horizons. One function of governance (as of quangos and corporatist arrangements beforehand) is to enable decisions with long-term implications to be divorced from short-term political (notably electoral) calculations. But there may still be disjunctions between the temporalities of different governance and government mechanisms that go beyond issues of sequencing to affect the very viability of heterarchy in the shadow of hierarchy. Third, although governance mechanisms may acquire specific techno-economic, political, and/or ideological functions, the state typically monitors their effects on its own capacity to secure social cohesion in divided societies. The state reserves to itself the right to open, close, juggle, and re-articulate governance not only in terms of particular functions but also from the viewpoint of partisan and global political advantage.
A complicating factor reinforcing the first and second sets of problems is the ‘relativization of scale’ (cf. Collinge 1996). This is a new development and so cannot have the same status as the first two sets. But it is nonetheless important today. Whereas the national scale of economic and political organization was dominant during the postwar economic expansion, the current period sees a crisis of the taken-for-grantedness of national economic and political space. This crisis has not led to the emergence of another spatial scale as the predominant economic level (whether global or local, supra-national or regional) around that the remaining scale levels (however many and however identified) can be organized in order to produce a degree of structured coherence. Instead we witness a proliferation of scales, related in tangled hierarchies rather than simply nested, with different temporalities as well as spatialities. This increases the extent of ‘unstructured complexity’ to which heterarchy is a response; but it also makes it harder for it to succeed as markets and states also face increasing problems. The rediscovery of the local is one manifestation of this but local problems cannot be solved entirely at this level; nor is there some other spatial scale at which meta-governance can be secured so that the local becomes manageable.
The third set is inscribed in the very nature of governance as a process of self-organization. The conditions bearing on governance success also tell us something about those for failure. First, accepting that all efforts at governance are bound to fail because of the incompleteness of the definition of objects of governance and failure to close them off from competing attempts at governance (Malpas and Wickham 1996), one can still usefully distinguish between degrees of success and failure. In this context, governance attempts may fail because of over-simplification of the conditions of action and/or deficient knowledge about the causal relationships that affect the object of governance. This can be especially problematic when the object of governance is an inherently unstructured but complex system such as the global economy. Second, there may be co-ordination problems on one or more of the interpersonal, inter-organizational, and inter-systemic levels. These levels are often related: thus inter-organizational negotiation often depends on interpersonal trust; and de-centred inter-systemic steering involves the representation of system logics through inter-organizational and/or inter-personal communication. Linked to this is the problematic relationship between those engaged in communication (networking, negotiation, etc.) and those whose interests and identities are being represented. Gaps can open between these groups leading to representational and legitimacy crises and/or to problems in securing compliance. Third, there is the ‘governability’ problem, i.e., the question of whether the object of governance could ever be manageable, even with adequate knowledge (Mayntz 1993b; Malpas and Wickham 1996). Fourth, there is a problem of governance without government — the inability to manage the repercussions of many devolved decisions. I address this problem in a later section on meta-governance.
Dilemmas of governance
In addition to these general constraints affecting governance and meta-governance, there are specific dilemmas within individual mechanisms. Here I discuss four such dilemmas:
Cooperation vs competition: capitalist economies operate through an unstable mix of cooperation and competition. One horn of the resulting dilemma is how to maintain interpersonal trust, secure generalized compliance with negotiated understandings, reduce noise through open communication, and engage in negative coordination in the face of the many and varied opportunities that exist for short-term self-interested competitive behaviour — behaviour that could soon destroy the basis for continuing partnership. The other horn is that an excessive commitment to cooperation and consensus could block the emergence of creative tensions, conflicts, or efforts at crisis-resolution that could promote learning and/or learning capacities and thereby enhance adaptability. This horn is especially acute when the environment is turbulent, speedy action is required, incrementalism is inappropriate, and it would take time to build consensus. Such dilemmas have been widely discussed in recent analyses of flexible industrial districts, learning regions, innovative milieux, etc. They also occur politically in the trade-off between partnership and partisanship. For partnerships are typically linked to differential advantages for political parties, tiers of government, and departmental interests as well as to differential economic interests of various kinds. This poses dilemmas both in relation to any given partnership and, even more acutely, in relation to the opportunities that may exist for juggling multiple partnerships to secure partisan advantage.
Openness vs closure: heterarchies operate in complex, often turbulent, environments. They face problems in remaining open to the environment at the same time as securing the closure needed for effective coordination among a limited number of partners. One horn of the resulting dilemma is that closure may lock in members whose exit would be beneficial (e.g., inefficient firms, underemployed workers, sunset sectors) or block recruitment of new social partners (e.g., new firms, marginalized workers, sunrise sectors). The other horn is that openness may discourage partners from entering into long-term commitments and sharing long-term time horizons. This may prompt opportunism in (the potentially self-fulfilling) case that partnerships dissolve or involve high turnover. It is reflected in the choice of maximizing the range of possible actions by expanding relevant bases of membership or favouring the ‘small is beautiful’ principle for the purpose of focused and timely action; and the choice of variable geometries of action versus fixed spatial boundaries for membership of a governance arrangement. An interesting variant of this latter version of the dilemma is whether to permit transnational partnerships or to insist on sovereignty.
Governability vs flexibility: heterarchy is said to permit longer term strategic guidance (lacking in markets) whilst retaining flexibility (lacking in hierarchies with their rule-governed procedures). But this is also the site of a dilemma: that between governability (the capacity for guidance) and flexibility (the capacity to adapt to changed circumstances). This assumes several forms. Reducing complexity through operational rules as a precondition for governing a complex world needs to be balanced against the recognition of complexity to mobilize the ‘requisite variety’ of actors and resources. Avoiding duplication to limit resource costs needs to be balanced against maintaining an adequate repertoire of actions and strategic capacities. A third variant is posed in the choice between exploiting past organizational and inter-organizational learning to standardize around ‘best practice’ and maintaining adaptability in the face of a turbulent environment by avoiding ‘lock-in’ to outmoded routines. This last problem is particularly associated with efforts to impose ‘best practice’ from above rather than encourage diversity and allow for horizontal communication and learning among partnerships.
Accountability vs efficiency: some public-private partnerships are expected to serve the public interest as well as to deliver private benefits. But this blurs the public-private distinction and poses a familiar dilemma in terms of accountability versus efficiency. On the one hand, there are problems about attributing responsibility for decisions and non-decisions (acts of commission or omission) in interdependent networks. These problems are especially acute when partnerships are inter-organizational rather than interpersonal. On the other hand, attempts to establish clear lines of accountability can interfere with the efficient, cooperative pursuit of joint goals. A related dilemma is that public-private arrangements run the risk of allowing the exploitative capture of public resources for private purposes and/or extending the state’s reach into the market economy and civil society to serve the interests of the state or governing party. A third version of this dilemma concerns the relative primacy of economic performance and social inclusion — how far the maximand in public-private partnerships is marketized economic performance as opposed to addressing problems of social cohesion.
These dilemmas can be managed collectively in several ways. Among these are the development of different institutions, apparatuses, or agencies specializing primarily in one or other horn of a dilemma and changing the balance between them through differential allocation of resources, continuing competition for legitimacy in changing circumstances, etc. Likewise, different horns can be handled at different scales. Thus in neo-liberal economies competition is often pursued more vigorously at the national level (privatization, liberalization, de-regulation, etc.) whilst cooperation is pursued more vigorously at the local or regional level (through public-private partnerships) (cf. Gough and Eisenschitz 1996). Different governance arrangements may also be instituted to deal with different temporal horizons. Thus one partnership may have an open structure and long-term horizon, another may be relatively closed and pursue specific tasks or development activities with short-term time horizons.
From governance failure to meta-governance
In discussing ways of handling these dilemmas I have already broached the issue of meta-governance: the ‘organization of self-organization’. This idea should not be confused with a super-ordinate level of government to which all governance arrangements are subordinated. It involves instead the design of institutions and generation of visions that can facilitate not only self-organization in different fields but also the relative coherence of the diverse objectives, spatial and temporal horizons, actions, and outcomes of various self-organizing arrangements. Meta-governance has institutional and strategic dimensions. Institutionally, it provides mechanisms for collective learning about the functional linkages and the material interdependencies among different sites and spheres of action. Strategically, it promotes the development of shared visions that might encourage new institutional arrangements and/or new activities to be pursued to supplement and/or complement existing patterns of governance. In both respects it involves the shaping of the context within which heterarchies can be forged rather than developing specific strategies and initiatives for them. States have a major role here as the primary organizer of the dialogue among policy communities, as an institutional ensemble charged with ensuring some coherence among all subsystems, as the source of a regulatory order in and through which they can pursue their aims, and as the sovereign power responsible ‘in the last resort’ for compensatory action where other subsystems fail (e.g. where markets, unions, or the science policy community have failed). This involves almost permanent institutional and organizational innovation in order to maintain the very possibility (however remote) of sustained economic growth.
Meta-governance does not amount to the installation of a monolithic mode of governance. Rather, it involves the management of complexity and plurality. Thus markets, hierarchies, and heterarchies still exist; but they operate in a context of ‘negotiated decision-making’. Thus, on the one hand, market competition will be balanced by cooperation, the invisible hand will be combined with a visible handshake. On the other hand, the state is no longer the sovereign authority. It becomes but one participant among others in the pluralistic guidance system and contributes its own distinctive resources to the negotiation process. As the range of networks, partnerships, and other models of economic and political governance expand, official apparatuses remain at best primus inter pares. For, although public money and law would still be important in underpinning their operation, other resources (such as private money, knowledge, or expertise) would also be critical to their success. The state’s involvement would become less hierarchical, less centralized, and less dirigiste in character. The exchange of information and moral suasion become key sources of legitimation and the state’s influence depends as much on its role as a prime source and mediator of collective intelligence as on its command over economic resources or legitimate coercion (cf. Willke 1992).
In exercising this meta-governance role, the state provides the ground rules for governance, ensures the compatibility of different governance mechanisms and regimes, deploys a relative monopoly of organizational intelligence and information with which to shape cognitive expectations, acts as a ‘court of appeal’ for disputes arising within and over governance, serves to re-balance power differentials by strengthening weaker parties or systems in the interests of system integration and/or social cohesion, etc.. This emerging meta-governance role means that networking, negotiation, noise reduction, and negative coordination take place ‘in the shadow of hierarchy’ (Scharpf 1994: 40). The need for such a role is especially acute in the light of the wide dispersion of governance mechanisms and the corresponding need to build appropriate macro-organizational capacities to address far-reaching inter-organizational changes without undermining the basic coherence and integrity of the (national) state. This role tends to fall to the state because of its heightened paradoxical position as an institutional subsystem that is simultaneously merely part of a wider, more complex society (and thus unable to control the latter from above) and also a part normatively charged (notably in the last resort) with securing the institutional integration and social cohesion of that society (Jessop 1990).
In conclusion I want to suggest that markets, states, and governance all fail. This is not surprising because failure is a central feature of all social relations. For ‘there is no such thing as complete or total control of an object or set of objects — governance is necessarily incomplete and as a necessary consequence must always fail’ (Malpas and Wickham 1995: 40). Given the growing structural complexity and opacity of the social world, indeed, failure becomes the most likely outcome of most attempts to govern it with reference to multiple objectives over extended spatial and temporal horizons — whether through markets, states, partnerships, or some other mechanism.
This is often recognized. However, whilst failure in the other two modes of coordination is regarded as inevitable, in the preferred mode of coordination it is typically seen as exceptional and corrigible. For example, for liberals, although the state is prone to failure, a turn to the market will solve the problem. If the market fails, however, it can be improved. Conversely, for statists, the response to market failure is government. If government fails, however, then it should be improved. This polarization is reflected both in the succession of governments and in policy cycles within governments in which different modes of policy-making succeed each other as the difficulties of each become more evident.
If we accept the incompleteness of attempts at coordination (whether through the market, the state, or heterarchy) as inevitable, then it is necessary to adopt a satisficing approach to these attempts. This in turn has three key dimensions: a self-reflexive orientation to what will prove satisfactory in the case of failure, a self-reflexive cultivation of a repertoire (requisite variety) of responses so that strategies and tactics can be combined to reduce the likelihood of failure and to alter their balance in the face of failure, and a self-reflexive ‘irony’ in the sense that participants must recognize the likelihood of failure but proceed as if success were possible. Perhaps the supreme irony in this context is that the need for irony holds not only for individual governance mechanisms but also for the commitment to meta-governance itself.
* This paper was initially delivered to the Colloquium on ‘Enjeux des debats sur la gouvernance’, 29-30 November, 1996, at the University of Lausanne. It derives from work for an Economic and Social Research Council research project on local governance (Grant L311253032); it was written while I was Hallsworth Research Fellow at Manchester University.
 For example, the corporate governance of firms, schools, hospitals, etc..
 This is not to deny the claim that modes of governance in part constitute their own objects of governance. Not all constituted objects of governance are amenable to governance as defined in specific governance projects.
 This typology is influenced by the Luhmannian distinction between three levels of social structure (interaction, organization, and functional system or institutional order); and by a correlative distinction between different forms of social embeddedness — the social embeddedness of interpersonal relations, the institutional embeddedness of inter-organizational relations, and the societal embeddedness of inter-systemic relations.
 Asset specificity exists to the extent that assets have limited uses and are immobile.
 In this way they may generate policy synergy, i.e., ‘a process by which new insights or solutions are produced out of the differences between partners’ (Hastings 1996: 259).
 This does not mean, for inter-systemic steering, that systems must abandon their own distinctive codes or undergo de-differentiation. But the individual programmes that they use to specify these codes’ operational implications must be modified at the margins to facilitate the continued negative and/or positive coordination of their respective operations across the different systems involved.
Capello, R. (1996) ‘Industrial Enterprises and Economic Space: the Network Paradigm’, European Planning Studies, 4 (4), 485-498.
Chesnais, F. (1987) ‘Science, Technology and Competitiveness’, STI Review, 1, 85-129.
Collinge, C.J. (1996) Spatial Articulation of the State: Reworking Social Relations and Social Regulation Theory. Birmingham: Centre for Urban and Regional Studies.
Coulson, A. (1997) ‘”Transaction Cost Economics” and its Implications for Local Governance’, Local Government Studies, 23 (1), 107-113.
Dyson, K. (1980) The State Tradition in Western Europe, Oxford: Martin Robertson.
Elechedus, M. (1990) ‘The Temporalities of Exchange: the Case of Self-Organization for Societal Guidance’, in B. Marin, ed., Generalized Political Exchange: Antagonistic Cooperation and Integrated Policy Circuits, Boulder: Westview, 192-216.
Glagow, M. and Willke, H., eds. (1987) Dezentrale Gesellschaftssteuerung: Probleme der Integration polyzentristischer Gesellschaft, Pfaffenweiler: Centaurus.
Gough, J. and Eisenschitz, A. (1996) ‘The Modernization of Britain and Local Economic Policy: Promise and Contradictions’, Environment and Planning D: Society and Space, 14, 203-219.
Grabher, G. (1994) Lob der Verschwendung. Redundanz in der Regionalentwicklung: ein socioökonomisches Plädoyer, Berlin: Edition Sigma.
Hastings, A. (1996) ‘Unravelling the Process of ‘Partnership’ in Urban Regeneration Policy’, Urban Studies, 33 (2), 253-268.
Jessop, B. (1990) State Theory: Putting Capitalist States in their Place, Cambridge: Polity.
Jessop, B. (1993) ‘Towards a Schumpeterian Workfare State? Preliminary Remarks on Post-Fordist Political Economy’, Studies in Political Economy, 40, 7-39.
Jessop, B. (1995) ‘Die Zukunft des Nationalstaats: Erosion oder Reorganization? Grundsätzliche Überlegungen zu Westeuropa’, in F. Deppe and W. Burkhardt, eds., Europäische Integration und politische Regulierung — Aspekte, Dimensionen, Perspektiven, Marburg an der Lahn: FEG Studie Nr. 5, 9-48.
Luhmann, N. (1979) Trust and Power: Two Essays, Chichester: John Wiley.
Luhmann, N. (1992) Social Systems, Stanford: Stanford University Press.
Malpas, J. and Wickham, G. (1995) ‘Governance and Failure: on the Limits of Sociology’, Australian and New Zealand Journal of Sociology, 31 (3), 37-50.
Mayntz, R. (1993a) ‘Modernization and the Logic of Interorganizational Networks’, in J. Child, M. Crozier, R. Mayntz et al., Societal Change between Market and Organization, Aldershot: Avebury, 3-18.
Mayntz, R. (1993b) ‘Governing Failures and the Problem of Governability: some Comments on a Theoretical Paradigm’, in J. Kooiman, ed., Modern Governance: new Government-Society Interactions, London: SAGE, 9-20.
Messner, D. (1996) Die Netzwerkgesellschaft, Köln: Weltforum Verlag.
Mitchell, W.C. and Simmons, R.T. (1994) Beyond Politics: Markets, Welfare, and the Failure of Bureaucracy. Boulder: Westview.
Nielsen, K. and Pedersen, O.K. (1993) ‘The Negotiated Economy: General Features and Theoretical Perspectives’, in J. Hausner, B. Jessop, and K. Nielsen, eds., Institutional Frameworks of Market Economies, Aldershot: Avebury, 89-112.
Offe, C. (1975) ‘The Theory of the Capitalist State and the Problem of Policy Formation,’ in L.N. Lindberg et al., eds., Stress and Contradiction in Modern Capitalism, Lexington: D.C. Heath, 125-144.
Rosenau, J.N. and Czempiel, E.-O. (1992) Governance without Government: Order and Change in World Politics, Cambridge: Cambridge University Press.
Scharpf, F.W. (1993) ‘Positive und negative Koordination in Verhandlungssystemen’, in A. Héritier, ed., Policy-Analyse: Kritik und Neuorientierung, Opladen: Westdeutsche Verlag.
Scharpf, F.W. (1994) ‘Games Real Actors could Play: Positive and Negative Coordination in Embedded Negotiations’, Journal of Theoretical Politics, 6 (1), 27-53.
Willke, H. (1992) Ironie des Staates: Grundlinien einer Staatstheorie polyzentrischer Gesellschaft, Frankfurt: Suhrkamp.