Abstract

Few concepts are as influential or pervasive in contemporary political discourse as that of sustainability. The concept gained currency with the report of the World Commission on Environment and Development in 1987, Our Common Future. Its political relevance was established at the Rio Earth Summit in 1992. It became the basis for national plans, local initiatives, corporate planning and goals, and a UN commission. It has become a central political, social, and ecological discourse of global society. Despite its currency, especially in global settings, the concept has been criticized widely. Among the criticisms are (1) that it is too vague to provide a practical guide for policy makers and (2) that it presumes positive relationships among economic and environmental factors that may not exist. The view taken here is that these are fair criticisms, especially for the commonly-accepted definitions that start with the idea of sustainable development and present the concept more in terms of an aspiration than an operational concept. We focus on sustainability as a concept and propose a more operational approach. The intent of this paper is to set out an exploratory framework for a field and practice of policy analysis based on sustainability. Part One sets out a working definition of sustainability based on a systems framework. The systems include the three standard elements of most definitions (the ecological, economic, and social), but it adds a fourth, the governance system. Sustainability is described as the task of sustaining each of these four systems over time and maintaining a balance among them. Part Two probes further into the basis for and implications of a complex systems approach to sustainability. It argues that taking a complex systems approach leads to different priorities, strategies, and methods compared to conventional analytical tools that are used in environmental policy analysis. The discussion is organized on the basis of three characteristics of complex systems: the tension between stability and 2 instability, resilience in response to stress, and capacities for self-organization. Part Two includes a brief critique of analytical tools that are used routinely in the environmental field. The final section considers the issue of why it matters to develop a field of sustainability policy analysis. Part One: Sustainability as an Analytical Framework There is no shortage of definitions of sustainability. Most commonly-used is that of the World Commission on Environment and Development (1987), of sustainable development as "development which meets the needs of the current generation without compromising the ability of future generations to meet their own needs." Although this captures the two core ideas of the concept-of not foreclosing options for future generations and of reconciling economic and environmental goals-it is hardly the basis for an analytical framework. Sustainable development is defined in the National Environmental Policy Act (NEPA) of 1969 as "to create and maintain conditions, under which humans and nature can exist in productive harmony, that fulfill the requirements of present and future generations." Other definitions stress a balancing the economy and environment goals. Nasrin Systems are self-organizing by being able to search for and maintain equilibrium. Systems survive in their ability to manage and adapt to change and maintain equilibrium. Our premise is that each system is essential for collective survival; no one system should be allowed to it threatens another. 3 There are empirical and normative aspects to these systems, both of which should be incorporated into a field of sustainability policy analysis. The empirical aspect has two parts: (1) all four systems are essential for human well-being and collective survival and (2) they are interconnected and interdependent. Traditional policy analysis focuses on understanding and explaining issues and relationships within systems, with some efforts to make connections among them (such as with cost-benefit analysis and ecosystem valuations). A field of sustainability analysis would aim for a systematic understanding and explanations of issues and relationships among them. The normative aspect is to determine the priority to be accorded to these systems and the value attached to each. The core objective of a field and practice of sustainability policy analysis should be to maximize the opportunities for complementarities and synergies among the four systems. Although trade-offs are inevitable, at least in the short term, the challenge of this field is to be able to identify and evaluate trade-offs in the context of the four systems. Consider the issue of the appropriate "balance" that should be maintained between the economic and ecological systems. Environmental policy conflicts in the US have turned on the perceived conflicts between these two systems. In the past, this has often been framed as a matter of values: Should we allow more pollution, resource development, or ecosystem damage in the interests of growth? An advocate for ecological values would say no; a growth proponent would argue yes. In John Dryzek's (1997) terms, they are advocating competing discourses, based on diverging values, interests, and world views. The empirical aspect of these systems adds another dimension, one that is susceptible to analysis. It is increasingly obvious that economic growth, population increases, and technology development pose stresses that threaten the survival of subsystems within the global ecological system. Climate change, water scarcity, and localized problems like overfishing or desertification are clear examples. As Peter Victor (2008) argues in Managing without Growth, the economic system is dynamic and may expand indefinitely while the ecological system is more static and subject to fixed limits (its carrying capacity). The choice thus turns from one of competing values to one of collective survival. Whatever one's views on the inherent value of nature, at some point ecological degradation threatens economic success while undermining capacities for governance and social progress. For this reason, At the same time, certain degrees and forms of economic growth are not only consistent with but may enhance ecological protection. More affluent societies are more likely to invest in pollution control, seek cleaner production methods, and use energy efficiently. They also exhibit slower rates of population growth, better education and health care, better status for women, and improved governance capacities. Likewise, the social system supports norms of behavior and networks of trust that contribute to the vitality of the other systems. In particular, both business transactions within the economic system and community initiatives to manage natural resources within the ecological system are more likely to succeed where social connectedness is higher. To the extent that the social system is threatened by distortions in the economic or environmental spheres, virtuous cycles of social capital creation and reinforcement may deteriorate. Because these systems are crucial, yet interconnected, they define three long-term imperatives for any society. The ecological imperative is "to remain within planetary biophysical capacity." This includes most environmental issues. The economic imperative is "to ensure and maintain adequate standards of living for all people." The focus here is on material sell-being and security. The social imperative is "to provide social structures, including governance systems, which effectively propagate the values people wish to live by" (77). The larger imperative may be seen as the need to sustain each of these systems while maintaining an appropriate balance among them. Just what this "appropriate balance" means is an issue that has to be worked out through the political process. A frequent criticism of the sustainability concept is that it does not provide clear criteria for making such choices. There is no reason that it should. 5 It does, however, offer an analytical framework for making choices. The definition and application of that framework constitutes the field of sustainability policy analysis and is the basis for the arguments here. Still, there are limits to any system's ability to adapt. Systems fail, whether they are classified as ecosystems (think of the Aral Sea or New England fisheries), political systems (consider the literature on failed states), or economic systems (the Great Depression, Weimar Germany). Systems may fail as a result of internal pressures or stresses from other systems. Unabated economic growth without attention to limits of the biosphere, for example, leads to long-term, irreversible climate change and affects the survival of ecosystems, economies, and states. Persistently high unemployment and rampant inflation cause political instability. The environmental security literature argues that resource scarcity and ecosystem degradation will be a major source of political conflict and instability (Matthew 2009). This paper proposes an addition to the conceptual, systems-based model and to writing on sustainability generally-that politics and governance be separated from the social system. In the literature, the social component (the social system) has been defined vaguely, with a strong normative content. The goals of equity, fairness, participation, transparency, access to health care and education, and gender rights define the principles on which the social system is based. The vital tasks of governance as an essential underpinning for the survival of the other systems are typically given short shrift in the sustainability literature. This is a significant omission, given the role of stable and effective governance to managing the other three systems. The governance system describes the societal capacity to make and carry out collective decisions that maintain the values people wish to live by. In our approach, governance is a precondition for successfully meeting the imperatives of the other three systems. Simply stated, political instability or corrupt and ineffective governments make it nearly impossible to sustain the other systems, let alone to maintain an appropriate balance among them (Fiorino 2010). In sum, we propose a framework for sustainability policy analysis consisting of: 1. four overlapping, interdependent systems: economic, ecological, social, and governance (see Levels of Sustainability Analysis We propose three levels of analysis as a starting point. The macro focuses on a national or global level (e.g., the European Union, China, or Germany). A meso level could describe a geographical region (the Great Lakes, Chesapeake Bay) or economic sector (steel, chemicals, autos, power generation, or semiconductors). A micro level refers to applications of the sustainability concept in communities or organizations (e.g., Boulder and Austin or Intel and Honda). Indeed, an advantage of the sustainability concept is that it may be applied usefully at multiple levels of analysis. We illustrate this with examples from macro and micro levels, where more of what we term sustainability analysis has been done. In our framework, sustainability policy analysis involves an explicit focus on the relationships among the four systems. Although "within-system" analysis will contribute to the understanding of intersystem relationships, our primary concern is with relationships among rather than within systems. To 7 illustrate the possible macro-level applications, this section briefly considers research that aims to explain differences in national environmental performance. This research is only a subset of a much larger body of work that studies the links among governance and economic development or other inter-system relationships. For example, political science research has studied many of the relationships among governance and economic growth indicators 8 This research also confirms the existence of positive relationships among the other systems. Democratic governance and economic growth appear to be positively linked. Similarly, higher incomes have been linked to improved status for women, better health care and education, and other aspects of the social system. Findings on any specific relationships between specific institutional characteristics (e.g., presidential compared to parliamentary systems) are inconclusive, although there is evidence that systems with a greater capacity to integrate policies across the systems demonstrate better environmental performance on some measures. This research illustrates the value of sustainability analysis and research. The micro level of sustainability policy analysis focuses on communities and organizations. This is the level at which policy debates, planning, and indicators development has been most active. Although applications of sustainability at this level are uneven and far from universal, there is research on both the business case and the community development (local government) case for sustainability. This discussion focuses on the corporate setting, where the sustainability concept has been used widely. In the United States, the early applications of the sustainability concept occurred in a corporate context. The literature on corporate sustainability depicts leading firms as moving in stages from reluctant compliance, to acceptance and compliance, to proactive pollution prevention, to a strategic focus on sustainability. The 1992 Rio Earth Summit served as a catalyst for many in business, leading among other initiatives to the 14001 series of ISO (International Standards Organization) environmental management systems and formation of the World Business Council for Sustainable Development (Schmidheiny 1997). In practice, firms like Johnson & Johnson, IBM, Intel, and 3M regularly engage in sustainability analysis. An application of an inter-systems approach is a standard return on investment (ROI) analysis of energy, water, or materials efficiency measures. Others are analysis of the business value of increased market share for green products or actions to manage regulators and competitors Researchers also have investigated the links between strong environmental performance and financial success. Studies have found positive relationships, in which firms ranking high in environmental performance and policies are more profitable (King and Lenox 2001). They conclude that strong environmental performance is not inconsistent with financial success. However, the direction of causality is difficult to determine. Are the more environmentally progressive firms in a better position to succeed financially? Or do greater profits allow firms to invest in more progressive and strategic policies? Or are innovative and proactive policies a sign of strategic leadership and thus an overall indicator of likely business success? It is difficult to answer these questions (although the authors lean toward the third explanation), in light of the research challenges. What matters is that research on corporate sustainability has explored relationships among the economic, ecological, and social systems in business settings. Our vision of sustainability policy analysis thus is based on a reconceptualization of the concept around four interconnected, interdependent systems. The challenge for contemporary governance is to sustain each while maintaining a balance among them. The concept of sustainability cannot determine those choices, but it may be used to define a framework for making them. The analytical foundation for making those choices is what is proposed here as a field of sustainability policy analysis. Given the centrality of complex systems in our approach to sustainability, the discussion turns now to that topic. 10 Part Two: Complex Systems as the Foundation for Sustainability Policy Analysis The first part of the discussion set out a framework for sustainability that emphasizes the existence of four systems and the need to sustain each and to balance the relationships among them. This second part explores the characteristics of complex systems and their implications for policy analysis, specifically as this would influence the conceptual foundations for a field of sustainability policy analysis. Conventional methods of policy analysis and program evaluation rest on the assumption that cause-and-effect relationships in the systems of interest are reasonably stable, linear, and knowable. Costbenefit analysis, for example, begins with the assumption that the program or project under study will generate the impacts for which the costs and benefits are being calculated. Risk assessment assumes that the contribution of a given level of chemical exposure to the probability of contracting a particular disease, for example, can be separated out from other causative factors and their interacting effects. Environmental impact statements, economic impact analyses, and any other "impact" analysis depend upon a basic input-output model of causation, which in turn reflects the idea of public policy as purposive activity by government to transform inputs (the resources marshaled by a program) to generate outputs (the goods or services produced by the program) to achieve desired ends (impacts our outcomes). Dave Snowden (2003, p 462) refers to this idea as "the assumption of order … [which] implies that an understanding of the causal links in past behavior allows us to define 'best practice' for future behavior." In many instances, this type of order is probably quite a reasonable assumption to make, but there are other instances in which omitted variables, unseen interactions among variables, and non-linear dynamics (such as tipping points) are likely to be important explanatory factors. Awareness of this problem has driven a steady stream of methodological advances in statistical modeling and econometrics for decades. Where highly complex systems are concerned, it seems unlikely that the models will ever catch up with the complexity of the phenomena being modeled. If they did, they could cease to be useful as simplified representations of reality. 11 If sustainability policy analysis simply amounted to a bigger, better input-output model with a larger number of variables, more connections among variables, and greater interactivity, it would make a contribution to the field. But it probably would not warrant its own subfield niche. As this section will attempt to show, sustainability analysis represents a more radical departure from conventional policy analysis than any expansion of existing models. It begins with an effort to work with complexity rather than "taming" it With the relatively clear cause-and-effect problems somewhat under control, what is left are the "wicked problems" (Rittel and Webber 1973). 1 1 Rittel and Webber were talking largely about social problems, but the label can be applied to other spheres of policy and planning, all of which touch the social systems at some point. We are now sensitized to the waves of repercussions generated by a problem-solving action directed to any one node in the network, and we are no longer surprised to find it inducing problems of greater severity at some other node. And so we have been forced to expand the boundaries of the systems we deal with, trying to internalize those externalities." Thus, wicked problems arising from complex systems share the defining characteristic of emergence: 1. Any whole complex system is greater than, and often different from, the sum of its parts. This is because agents within the system (human or not) do not act independently all of the time, and their interactions can generate surprising behaviors-everything from a standing ovation to political revolution-that are impossible to predict in advance. This nonlinear, selfreinforcing phenomenon is known as "emergence." 2. Although higher-level, emergent behaviors may be difficult to predict, they often follow identifiable patterns characterized by "retrospective coherence," meaning that "[e]mergent patterns can be perceived [in retrospect] but not predicted [in advance]" (Snowden 2003, 469). Thus, complexity should not be conflated with chaos, which is unpatterned even in hindsight. But neither should analysts be less than humble about their ability to detect these patterns at all, let alone anticipate them. Systems continually change and evolve as agents interact and new patterns emerge. It is risky to assume that cause-and-effect relationships will persist over time, which makes it difficult to assess policy interventions based on expected impacts. The emergent view of causation is not particularly radical; nor is it new. Many practicing policy analysts would probably endorse the concept, at least in principle, as a good description of many 13 phenomena. The problem is that emergence simply cannot be incorporated into the usual techniques of policy analysis. Indeed, scholars seeking to model emergence have had to invent their own computeraided methods, under the banner of computational agent-based models, which resemble the simulations used in computer games (see Because systems-based thinking is a defining characteristic of our view of sustainability analysis, the complex systems/emergence paradigm offers criteria for judging any proposal as being "sustainable." These criteria are organized according to the three characteristics of complex systems noted earlier: stability, resilience, and self-organization; all contain pairs of contrasting forces that must be balanced. The Tension Between Stability and Instability Merriam-Webster's on-line dictionary defines stability as 1a "the strength to stand or endure," and 1b "the property of a body that causes it when disturbed from a condition of equilibrium or steady motion to develop forces or moments that restore the original condition." Kenneth Boulding (1956) noted that most physical and human systems tend toward equilibrium, with one type of force inducing a counter-force to balance it. For example, a rise in average temperature induces species to migrate to higher altitudes and latitudes where temperatures are cooler; a rise in oil prices induces increased demand for fuel-efficient cars. Most systems contain components with this cybernetic or thermostatic control. Complex systems also may exhibit dramatic instability, in the form of riots, stock market crashes, political revolutions, bank failures, forest fires, epidemics, and sudden declines in species. 2 2 It is precisely these types of phenomena that inspired much of the inquiry that we now call complex adaptive systems research Even at times of relative stability, new patterns of emergence are constantly in process, transforming the system bit by bit or laying the foundation for a sudden, discontinuous change in the future. Thus, it can be said that 14 complex systems are characterized by both stability and instability. In other words, most systems tend toward equilibrium and possess a capacity to restore equilibrium when it is disturbed (stability), but they also hold the potential for abrupt change at almost any time if reinforcing behaviors are triggered and balancing forces suppressed (instability). Often, the abrupt system changes qualify as corrections in a system that had become unbalanced, such as an over-valued stock market in the case of a crash. In other situations, a systemic correction may go too far, leading to a wildly swinging pendulum. The ability of systems to self-manage and adapt through small and large corrections is discussed below under resilience. Baumgartner and Jones (1993) borrowed the term "punctuated equilibrium" from evolutionary biology to describe the interactive relationship between stability and instability in policymaking. According to their theory, forces of stability (entrenched interests, institutional inertia, and limited horizons of policymakers) hold policies in place and resist anything more than incremental changes to those policies until the policies become so out of date, and the need for change so obvious, that the forces of reform finally break through (punctuation). The oscillations between stability and sudden bursts of reform occur largely because key players tend to focus disproportionate attention on particular stimuli that cross their path at particular times (Jones and Baumgartner 2005). The habit of locking into some ideas and not others causes policymakers to neglect the need for changes in some areas, leading to periods of stability. As these information processing errors accumulate, they become harder to ignore. As a result, "Decisions are always catching up to reality; generals are always fighting the last war" (Jones and Baumgartner 2005, p. 334). The concept of punctuated equilibrium can apply to phenomena across all systems and even overlaps to some degree with the recently popularized idea of tipping points. It poses a formidable challenge to policy analysts to find ways of challenging conventional wisdom when policy gets stuck in an error-induced equilibrium. Cross-systems analysis may provide some leverage. The tensions between stability and instability also apply to the implementation phase of policy. Policy stability contributes to the phenomenon of regulatory capture, in which representatives of a regulated industry (such as oil rig operators in the Gulf of Mexico) form inappropriately cozy relationships with regulators (such as the Department of the Interior's Minerals Management Service). 15 The result is a dangerous equilibrium in which lax monitoring, assessment, and enforcement contribute to a culture of negligence and risk-taking in a self-reinforcing, vicious cycle. One month after the Deepwater Horizon oil rig exploded, the Minerals Management Service was restructured to clarify roles and separate conflicting functions. It remains to be seen whether this deliberate destabilizing is sufficient to break the old patterns of capture and nurture more desirable attitudes and behaviors among regulators and industry. Those who seek to design policies that are more resistant to regulatory capture face another formidable challenge at the institutional intersection between the economic and governance systems. The lesson from all of this for sustainability policy analysts is clear but challenging: Policy needs to be designed to strike a balance between reasonable stability and predictability of institutions and programs, on one hand, and adaptive mechanisms that can respond effectively to inevitable instability in the social, economic, and ecological spheres on the other. Striking such a balance requires deep understanding of the forces that generate stability and instability within complex systems, which in turn should suggest interventions that have the potential to nurture productive forms of stability while disrupting unproductive forms. However, the uncertain nature of causation in these systems means that the forces of stability and instability, and the emergent patterns that they create, are not apparent using conventional analytical techniques. An example is risk assessment, which focuses on harms that have already been identified; it does not tell the analyst how to search for other risks, emergent or fully developed, that may have gone unnoticed. Cost-benefit analysis and impact assessments also tend to focus on identifying and measuring the likely efficiency and effectiveness of proposed actions, rather than devising new ideas for action that work with the systems in question. Even basic research into system dynamics can promise only so much, due to the lack of reliable connections between past experience and future expectations. For these reasons, understanding systems often requires methods that do not generally fall into the category of "analysis" at all. The "adaptive management" approach to natural resources, for example, qualifies more as a blend of action research and responsive decision-making than policy analysis. The basic idea is to design policies that can function both as experiments and as evolving solutions, with the 16 goal of testing and comparing the effects of different management approaches by putting them in place, monitoring results, and adjusting practices as learning occurs (Williams et al 2007, Walters 1986, Holling 1978. It has been used for decades to manage forests, wildlife populations, river systems, watersheds, fisheries, and other resources. The lessons learned from adaptive management experiments may gradually expand understanding of how policy intervention affects complex social and ecological systems. Adaptive management works with rather than dispels uncertainty. showing that government can play a constructive, non-controlling role in self-organization processes. Resilience: Fostering Heterogeneity in Policy Analysis Stability and resilience stand in tension 17 C.S. Ecologists know that heterogeneity is basic to resilience. The same story can be told about virtually any form of command-and-control policy. As long as regulatees maintain a larger repertoire of activities than the regulations can cover they maintain the upper hand within the system. As policy makers scramble to update laws and regulations to cover as many undesirable activities as possible, the list of rules can grow so lengthy that no one understands them and the regulators struggle to enforce them (Bardach and Kagan 1982). Meanwhile, many regulatees will have focused on meeting the bare expectations of the law and avoiding detection, rather than meeting the larger goals of the law, such as safety, sustainability, or fairness. For these reasons, Snowden (2005, p 25) concludes that "[c]ore habits and sound ethical training beat rules and rule compliance any day," because good habits and internalized norms, once developed, can be applied to a variety of situations, including ones that no one could have anticipated. The truth in Snowden's conclusion stems from the greater variety inherent in habits and ethical norms, compared to rules, and the benefits of variety for system resilience. Running through the rapidly expanding literature on complexity theory's relevance to public policy and administration is a general preference for decentralized decision making, empowerment of agents to organize their own solutions, and light-handed interventions aimed at facilitating desirable 19 rather than prohibiting undesirable behavior. This is a manifestation of what Snowden (2009, p. 3) calls "distributed cognition," or the capacity of networks to influence policy design and decision making based on a deep understanding of local context and the power of existing social relationships. The paradigm of adaptive co-management, described in the previous section, follows in this general school of thought. The strong pro-decentralization view is well represented in Emery Roe's 1998 book, entitled Taking Complexity Seriously: Policy Analysis, Triangulation and Sustainable Development, which favors local decision-making processes over global ones for their ability to differentiate more finely among the facts of each local case and their propensity to develop diverse, customized management strategies that respect local people's perspectives. Global approaches, by contrast, tend to reduce the richness of local wisdom to a generic set of unrealistic demands for reducing population and consumption and elevating ecosystem over other values via the precautionary principle and the principle of intergenerational equity. We should therefore look forward to an enlightened period when all top-down, command-and- Acknowledgment of complexity, therefore, often requires a similarly complex portfolio approach to policy, one with both horizontal and vertical dimensions. On the horizontal plane, multiple different instruments or mechanisms may be used to influence a system. On the vertical plane, those instruments 20 may be designed and implemented at different levels-macro, meso, or micro. The question is how to decide which levels to include in a policy and in a policy analysis. That question relates to the substantive question of where to draw the boundaries of the systems of interest. Rittel and Webber neatly summarize the trade-offs involved (1973, p. 165), "There is nothing like a natural level of a wicked problem. Of course, the higher the level of a problem's formulation, the broader and more general it becomes: and the more difficult it becomes to do something about it. On the other hand, one should not try to cure symptoms: and therefore one should try to settle the problem on as high a level as possible." Similarly, Kenneth Boulding (1956, p. 197) warned that "we always pay for generality by sacrificing content, and all we can say about practically everything is almost nothing." In other words, as we draw