mercredi 19 mars 2008


2 Ethics and sustainable development: an adaptive approach to environmental choice* Bryan G. Norton:

1. Introduction
Most writing on environmental ethics concerns the dichotomy between humans and non-humans, and much of the work in the field has been motivated by the effort to escape ‘anthropocentrism’ with respect to environmental values. Resulting debates about whether to extend ‘moral consider ability’ to various elements of non-human nature have been, to say the least, inconclusive, and writings in this vein have had no discernible impact on the development of sustainability theory or on public policy more generally (Good paster, 1978). In this contribution, a new approach to re-conceptualizing our responsibilities toward nature is proposed, an approach that begins with a re-examination of spatiotemporal scaling in the conceptualization of environmental problems and human responses to them. Before turning in the following sections to a description of this emerging approach to management – sometimes called ‘adaptive management’ – I will in this introductory section briefly summarize the current situation in environmental ethics.
Discussions in the field of environmental ethics, which emerged as a separate subfield of ethics in the early 1970s, have, as just noted, turned on defining and explaining key dichotomies (Norton, 2005). This trend originated in the publication, by the historian Lynn White, Jr, of an influential essay (1967), ‘The historical roots of our ecologic crisis’, in which he declared that Christianity ‘is the most anthropocentric religion the world has seen’, setting the stage for a spate of responses by ethicists who questioned the longstanding ethical divide between humans and non-humans. Environmental ethicists have, accordingly, focused on the dualisms of modernism: humans vs non-humans, moral exclusivism – the view that all and only humans have intrinsic value, and the underlying dichotomy between matter and spirit. From 1970 until the early 1990s, these dichotomous formulations dominated environmental ethics as the question of where to draw the crucial line between those beings that are morally considerable and those that are morally irrelevant seemed so seminal a question that the field could not proceed without some resolution of it, and yet discussions of ‘intrinsic’ or ‘inherent’ value shed little light on practical questions about what to do. Worse, emphasis on these dichotomies created an irresolvable conflict with environmental economists, blocking any integration of philosophical and economic discourse (Norton and Minteer, 2002). Because economists insist that all values are values of human beings (consumers), they are in ontological disagreement with environmental ethicists, who wish to shift the line of moral consideration to include non-humans and their interests.
The debate over intrinsic value could of course be brought to bear upon questions of sustainable development, as it seems reasonable for a non- anthropocentrist, who attributes intrinsic value to some non-humans, to advocate sustainable use of ‘resources’ for all intrinsically valuable beings.
As the debates have actually evolved, however, this has not been a nexus of active discussion – the debate about sustainable development has been staged at the edges of mainstream, environmental economics and of the emerging competitor, ecological economics, both of which count human values only. Environmental ethicists, rejecting this exclusivism, have argued indiscriminately against all attempts to assess the economic and instrumental uses of the material world only for the satisfaction of human needs and demands. Thus, by objecting to the economic framework of analysis (because it is anthropocentric), environmental ethicists have been at cross purposes with both sides in the debate about how to define sustainable development. By the 1990s, a few philosophers began to see that this unfortunate stalemate between economic approaches and environmental philosophy rested mainly on ideological commitments and a priori theories, theories that for non-empirical reasons attempt to force all environmental value into a single valuational currency. No empirical evidence can be brought to bear upon whether nature has intrinsic value, and commitments to valuing objects as consumable items with a price are likewise based on a priori assumptions.
Worse, the categorical nature of the debate has encouraged all-or-nothing answers to complex management problems, and a conceptual polarization that leads to direct oppositions and an inability to frame questions as open to compromise.
If one instead adopts pluralism, accepting the fact that humans value nature in many ways, and considers these values to range along a continuum from purely selfish uses to spiritual and less instrumental uses, it is unclear – and not really very important – where to ‘separate’ one kind of value from another (Stone, 1987; Norton, 2005). If we think of natural objects as having many kinds of value, arguments about why we should protect nature slide into the background and the focus moves to protecting as many of the values of nature as possible, for the longest time that is foreseeable. Of course there will be disagreements about priorities and immediate objectives, but if policies are devised to protect as much of nature as possible for the use and enjoyment of humans for as long into the future as possible, then it is perhaps not crucial whether those values preserved are counted in one theoretical framework or another.
The viewpoint advanced here is referred to as environmental pragmatism, which is advanced as a philosophy of environmental action that begins with real-world problems, not with abstract, theory-dependent questions regarding what kind of value nature has (Light and Katz, 1996; Norton, 2005). Environmental pragmatism can be seen as a third way in environ-mental ethics: it bypasses the theoretically grounded questions of environ-mental ethics and focuses on learning our way out of uncertainty in particular situations. If the ‘true’ value of natural systems is unknown to day, this is all the more reason to save them for the future, where their full and true value may be learned.
Further, pragmatism complements the search for sustainable development because it is a forward-looking philosophy, defining truth as that which will prevail, within the community of inquirers, in the long run. This feature makes it a natural complement to the theory of sustainable development and acts as the unifying thread in the justification of preservation efforts at all scales: this forward-looking sense of responsibility and commitment to learning our way to sustainability can be thought of as pragmatism’s contribution to the theory of sustainable development (Lee, 1993;Norton, 1999; Norton, 2005).
In the remainder of this chapter, I will propose one approach to a new environmental philosophy, a philosophy that is more geared to learning to be sustainable than in defining what kind of good nature has. This philosophy emphasizes social learning and community adaptation, and it derives its method more from the epistemology of pragmatism than from theoretical ethics.
2. Adaptive management
To introduce the adaptive management approach, I will briefly explain how it rests on three intellectual pillars, and then propose a more explicit definition of adaptive management before undertaking to elaborate the theory by discussing each of these pillars in more detail.
I. A Commitment to a Unified Method: Naturalism. Attempts to separate factual from value content in the process of deliberation are rejected; there is only one method for evaluating human assertions, including assertions with all kinds of mixes of descriptive and prescriptive content, and that is the method of experience – active experimentation when possible, and careful observation otherwise. The scientific method is embraced as the best approach to evaluating hypotheses about cause and effect, but also about what is valuable to individuals and cultures.
II. A Relationship between Values and Boundaries. The values of people who care about the environment are expressed in the ways they (a)‘bound’ the natural system associated with a given problem, and (b) the choices they make in focusing on physical dynamics they use to ‘model’ those problems.
III. A New Approach to Scaling and Environmental Problems. Building on this idea, scalar choices in modeling environmental problems, if made a topic for open public discussion, might provide insight into the temporal and spatial ‘horizons’ over which impacts will be measured, and processes of change monitored. In policy, they direct the formation of effective administrative strategies for addressing problems; scientifically, careful attention to the dimensions and models developed in response to environmental problems might clarify problem formulation and illuminate public discourse.
This approach to management is often referred to as ‘adaptive management’ in North America, but it is practiced elsewhere in varied forms and with different names. Adaptive management, which can be understood as a search for a locally anchored conception of sustainability and sustainable management, sets out to use science and social learning as tools to achieve cooperation in the pursuit of management goals (Walters, 1986;Lee, 1993; Gunderson et al., 1995; Gunderson and Holling, 2002;Norton, 2005). In the United States, the ideas were first articulated by the scientific and philosophical forester, Aldo Leopold, who emphasized the importance of multi-scalar adaptation in his essay, ‘Thinking like a mountain’, and who advocated scientific management throughout his career.
Three characteristics can be taken to define a process of adaptive management:
_ Experimentalism: adaptive managers respond to uncertainty by undertaking reversible actions and studying outcomes to reduce uncertainty at the next decision point.
_ Multi-Scalar Modeling: adaptive managers model environmental problems within multi-scaled (‘hierarchical’) space–time systems.
_ Place-Orientation: adaptive managers address environmental problems from a ‘place’ which means problems are embedded in a local context of natural systems but also of political forces.
By profession, most adaptive managers are ecologists and most discussions to date have emphasized learning our way out of scientific uncertainty; these ecologists have paid less attention to developing appropriate processes for evaluating environmental change and for setting intelligent goals for environmental management. Here, we will incorporate the ideas of these ecologists and expand them to include learning about social values as an integral part of the adaptive management process.
3. Naturalism: the method of experience
As noted above, much discussion in environmental ethics has centered on the debate between anthropocentrists and non-anthropocentrists, between those who limit moral consider ability to humans and those who extend human consider ability into the non-human world. Unfortunately, environmental ethicists have not paid as much attention to another controversial dichotomy, that between ‘facts’ and ‘values’ – between descriptive and prescriptive language. Analytic philosophers have been very cautious about mixing facts and values in argumentation, a trend initiated by David Hume, who promulgated ‘Hume’s Law’, which is usually taken to deny the possibility of deducing an ‘ought’ proposition from any body of ‘is’ propositions. Recently, two prominent environmental ethicists have argued, adopting arguments reminiscent of Hume, for forsaking science and descriptive studies and concentrating on ‘intrinsic values’ in the effort to protect natural systems, processes and elements.
In particular, J. Baird Callicott (2002) and Mark Sago (2004) have both argued that environmentalists should play down instrumental arguments for saving species and biodiversity, basing their main arguments on the ‘intrinsic value’ of nature. Sagoff says: ‘indeed environmental policy is most characterized by the opposition between instrumental values and aesthetic and moral judgments and convictions.’ (2004, p. 20). He goes on to argue that ‘Environmental controversies . . . turn on the discovery and acceptance of moral and aesthetic judgments as facts.’ (p. 39). Unfortunately, he describes no means to separate fact from fiction in assertions that this or that has intrinsic value and explicitly claims that scientific arguments have no bearing on defending environmental values or goals. Callicott (2002) joins Sagoff in sharply separating science from ethics and instrumental uses from non-instrumental appreciation: ‘We subjects value objects in one or both of at least two ways – instrumentally or intrinsically
– between which there is no middle term.’ (p. 16). Callicott goes on to emphasize the subjective source of these intrinsic values:
All value, in short, is of subjective provenance. And I hold that intrinsic value should be defined negatively, in contradistinction to instrumental value, as the value of something that is left over when all its instrumental value has been subtracted (‘intrinsic value’ and ‘noninstrumental value’ are two names for one and the same thing).
Emphasizing the personal and the subjective nature of intrinsic valuings,he says: ‘Indeed, it is logically possible to value intrinsically anything underthe sun – an old worn-out shoe, for example.’ (Callicott, 2004, p. 10).Callicott and Sagoff, then, have called for a strategy of emphasizing intrinsic values over instrumental uses of nature in arguing for the protection of nature. In doing so, they rely on a sharp dichotomy between descriptive andprescriptive discourse, and on sharply separating instrumental reasons for protecting nature from non-instrumental reasons. These non-natural qualities are, apparently, apprehended through intuition or created by emotional affects, and they seem ill-suited to provide inter-subjectively valid or convincing reasons for environmental action.
A more realistic – and less theory-driven – view of the relation between factual and evaluative discourse is advocated by B.A.O. Williams (1985), who argued persuasively that, in ordinary discourse, fact-discourse and value-discourse are inseparable; when philosophers separate them, they do so on the basis of a specialized theory, such as logical positivism. In the ordinary discourse in which citizens discuss and evaluate their environment, these discourses are inseparable; to insist on partitioning policy discourse into fact-discourse (positivistic science) and value-discourse is to artificialize that discourse. There is an alternative, of course. Following pragmatists such as C.S. Peirce and John Dewey, one can advocate a pragmatic epistemology for environmental science and policy discourse, a discourse conducted so as to maximize social learning among participants (Dewey, 1927; 1966; Lee, 1993). This epistemology insists upon a single method – the method of experience – and this method applies equally to factual claims and evaluative ones. Following Dewey, assertions that something or some process is valued are taken as a hypothesis that that thing or process is valuable. Pursuing that value, and acting upon associated values, provides communities with experience that can support or undermine the claim that the thing or process is indeed valuable. Non-naturalism thus construes environmental values in ways that are not easily related to scientifically measurable indicators. If the public and policy makers are going to support environmental actions, it will be necessary to cite values and to explain and justify environmentally motivated actions, but it is difficult to see how one would link ‘non-natural’ qualities of nature with empirically measurable indicators. Insistence upon a sharp separation of facts from values, means from ends, and instrumental from non-instrumental values makes connections between ecological change and social values more abstract, theoretical and tenuous. It makes the integration of the discourses of environmental science and environmental value virtually impossible. Worse, it estranges values from management; science, creating a situation in which managers must look outside the adaptive process for indications of social value; they must either turn to economists’ measurements of consumers’ unconsidered preferences, or they can ask environmental ethicists to divine the nature of nature’s non-natural qualities.
So, rejecting non-naturalism, the first pillar of my proposed approach is a form of methodological naturalism. This method, while not expecting deductions from facts to values, relies on the open-ended, public process of challenging beliefs and values with contrary experience. From these challenges, we expect attitudes, values and beliefs to change – but the changes cannot be justified by deductive arguments flowing one way from facts to values. The changes needed to support a new conservation consciousness are usually reorganizations and re-conceptualizations of facts, not deductions from value-neutral facts. The specific means by which assertions of value are connected will be through the development and refinement of measurable indicators that reflect values articulated by the stakeholders who represent multiple positions within the community. Pluralism is operationalized in process as communities participate in choosing multiple indicators, as will be discussed in the next two sections.
Values and bounding While one need not challenge Hume’s Law to see a non-deductive connection between factual information and values, two assumptions that Hume made in formulating his law should be challenged. By stating the law as a prohibition against deriving ‘ought’ sentences from ‘is’ sentences, Hume implied that fact-discourse and evaluative discourse could be sharply separated, and that the difference would announce itself syntactically via the evident copula. In real discourse, they are all mixed together in ordinary speech; to separate them artificializes normal discourse in important ways. This argument, however, raises an inevitable question: How, exactly, do values manifest themselves in scientific, descriptive literature, which claims to be ‘value-free’ and is apparently ‘scrubbed’ of evaluative language before publication in scientific journals? In order to answer this question, it is useful to follow Funtowicz and Ravetz (1990; 1995) in distinguishing between ‘curiosity-driven’ (discipline-driven) science and ‘mission oriented’ (problem-driven) science. Authors who place their research in disciplinary journals succeed, to varying degrees, in purging evidence of values from their scientific papers. Adaptive management, however, is an active, mission oriented science and, as Funtowicz and Ravetz argue, it often takes place in contexts where stakeholders have different perspectives and interests. In these contexts, scientific models and reports that are taken to bear on management decisions will, in effect, be ‘peer reviewed’ not only by appropriate disciplinary scientists, but also by scientists in different fields, and by interested laypersons. This places a transparency requirement on scientific discourse: if science is to be advanced as a guide to controversial policies, then that science must be explainable – and explained – in ordinary speech that requires no scientific credentials to understand.
When attention shifts from disciplinary science to mission-oriented science, values slip back into the discourse, because participants are proposing and evaluating policies from their own perspective, given their own models of the problems. So, if we want to find values implicit in scientific work, we should look closely at the discourse of management science. The values and interests of participants are coded into the choices they make to ‘model’ the problem – to bound the problem spatially, to form a temporal horizon, and to describe a function of the system that is considered problematic. These values are often embedded in the choices individuals and groups make when they choose/develop a ‘mental model’of the problem they are addressing.
A historical example may help to illustrate what is claimed here. Chesapeake Bay, on the East Coast of the US, is among the most productive – and loved – bodies of water in the world. The Bay is the mouth of the Susqhehanna River, and many other tributaries that drain a huge portion of the Northeastern United States. By the 1970s there were multiple danger signals that the Bay was becoming polluted, even if it was unclear what was driving the widespread changes in Bay functioning, especially the increasing turbidity and consequent die-back of the vast underwater grass flats that formed the base of the Bay’s food-web. Until the 1970s, when the US Environmental Protection Agency (EPA) undertook a detailed scientific study, pollution issues had mostly centered around toxic and point source pollution problems, including polluting industries and inadequate sewage treatment in a densely packed area of residences, agriculture and industry. It was learned that, while environmental monitors were paying attention to small-scale, local variables, a large-scale variable associated with a larger- scale dynamic – one driven by the total input of nutrients into the bay from its tributaries – posed a slower-moving, but more profound threat to bay health. Agricultural and residential run-off of nitrogen and phosphorous was causing increased turbidity, reducing submerged aquatic vegetation beds, and causing algal blooms and anoxia in deep waters. The rich farm-lands of Pennsylvania, the Piedmont, and the coastal plain all drain into the Chesapeake. To save the Chesapeake, it would be necessary to gain the cooperation of countless upstream users of the waters that eventually enter the bay, a monumental task, since Pennsylvania and the District of Columbia, situated upstream on tributaries, have no coastline on the Bay and no direct stake in its protection. Nevertheless, against all odds, the larger Bay community – enabled by the EPA study and countless private research efforts – succeeded in transforming the public consciousness to think of the Bay as an organic, connected watershed. Tom Horton, an environmental journalist and activist said it best when, at the height of this period of intense social learning, he wrote:
‘We are throwing out our old maps of the bay. They are outdated not because of shoaling or erosion or political boundary shifts, but because the public needs a radically new perception of North America’s greatest estuary’ (Horton, 1987, pp. 7–8). He pointed out that, as the problem with bay water quality expanded beyond point-source pollution, to include non- point sources, residents of the area had to change their mental model of the processes of pollution; and they had to address activities throughout the watershed, adopting a model that includes all the lands contributing run-off to the bay. What is important to learn from this analysis is that the ‘transformation’ of the Bay from an estuary into a watershed occurred in a context of mission-oriented science and it was as much a process of transformation of public consciousness as it was a change in scientific understanding. It was a dramatic change in perspective that was driven by values – an outpouring of love and commitments not to let the Bay become more unhealthy. In order to address the problem of Bay water quality, it was necessary to create a new ‘model’ of what was going wrong. The shift in models led to a public campaign, driven by the deep and varied values residents felt toward the Bay, which was marked, for example, by the outstanding success of the Chesapeake Bay Foundation, a private foundation that advocates, educates and supports science to guide Bay management. So, we have here an example of a value-driven re-mapping of a complex natural system, how it works, and how pollution is being delivered into it. We can say that a new ‘cultural model’ was formed (Kempton et al., 1995; Kempton and Falk, 2000; Paolisso, 2002), and Chesapeake Bay management, while not perfect, of course, has been a model of cross-state cooperation as serious steps have been taken throughout the watershed to reduce non-point-source as well as point-source pollution.
How should we interpret this transformation? A scientific finding that the Bay was threatened by processes outside its currently conceived boundaries, interacting with the strength of the love for the Bay as a ‘place’, created a new model that more accurately represented the problems of the Bay, and also expanded the sense of responsibility of residents and users of the Bay. The public understanding embraced the larger system, and they shifted their attention to addressing non-point-source pollution problems throughout the watershed. One could correctly argue that it was values –the love felt for the Bay – that was driving the acceptance of these models; it would be just as correct, however, to say that it was the scientific studies that analyzed the problem as watershed-sized that enveloped and embodied that love in a new ecophysical model of the Bay and its problems. Residents and officials of the Bay area, upon being convinced that the Bay’s health was threatened, and that a large part of the problem came from the larger-scale watershed system, shifted to a larger perspective on Bay health, a perspective that is more aligned with a scientific understanding of the problem faced. This shift in perspective, however, is not just scientific: it expresses a deep and varied set of social values that residents and stake-holders feel toward the Bay. And, when Horton describes the change in hydrological and cartographic terms, the underlying truth is that the shift to a watershed-sized model was the expression of an implicit value, a sense of caring for the health of the Bay as a part of one’s way of life. The love and respect residents had for the Bay, once the nature of the threat was better understood, expressed itself in a ready embrace of the Bay as a watershed. Their local valuing came to express a community consensus in goals and values, transforming a local consciousness into a regional consciousness and sense of responsibility. Through social learning, the residents of the area discovered how to ‘think like a watershed’, and began living in a larger ‘place’ than before. Social values are imputed to environmental and ecological systems implicitly in the process of developing ‘models’ – either cultural or scientific – of the problem that needs addressing. These models, if they are similar across all participants in public deliberations, can be very helpful in developing common understandings and in undertaking experimental actions. If they are very different, communication may be difficult, and environmental problems remain recalcitrant, dividing communities and undermining cooperative and experimental action. Inmany cases, communities are paralyzed because they have not had the kind of social learning experience that took place in the Chesapeake region, and co-operative action to address pressing and perceived problems are gridlocked. Differing values and interests – according to the hypothesis of this Part – thus inform and shape the models that participants use to understand environmental problems in their areas.
Diversity of perspective and differences about value are thus key aspects of difficulties in deciding what, exactly, is the problem to be addressed.
5. Scaling and environmental problem formulation Environmental disputes are so difficult, among other reasons, because it is so difficult to provide a definitive problem formulation. This feature was well explained by Rittel and Webber (1973), who distinguished ‘benign’ and ‘wicked’ problems. Benign problems, they said, have determinate answers and when the solution is found, the problem is uncontroversially ‘solved’. Mathematics and some areas of science exemplify benign problems. Wicked problems, on the other hand, resist unified problem formulation; there is controversy regarding what models to use and what data are important. Rittel and Webber suggest that wicked problems, because they are perceived differently by different interest groups with different values and goals, have no determinate solution because there is no agreement on the problem formulation. They can be ‘resolved’ by finding a temporary balance among competing interests and social goals, but as the situation changes, the problem changes and becomes more open-ended. Rittel and Webber explicitly mention that wicked problems have a way of coming back in new forms; as society addresses one symptom or set of symptoms, new symptoms appear, sometimes as unintended effects of treatments of the original problem.
Most environmental problems are wicked problems; they affect multiple values, and they impact different elements of the community differently, encouraging the development of multiple models of understanding and remedy. While resistance to unified problem formulation is endemic to wicked problems, and requires iterative negotiations to find even temporary resolutions and agreements on actions, one aspect of wicked problems – the temporal open-endedness which often attends wicked problems and brings them back in more virulent form as larger and larger systems are affected –may be susceptible to clarification through modeling. Ecologists have introduced ‘hierarchy theory’ (HT), as a set of conventions to clarify space–time relations in complex systems (Allen and Starr, 1982; Holling, 1992; Norton, 2005). HT can be characterized by two axioms (which happen to coincide with the second and third key characteristics of Adaptive Management listed in the ‘Introduction’). HT encompasses a set of models of ecological systems that are characterized by two constraints on observer and system behavior:
(i) The system is conceived as composed of nested subsystems, such that any subsystem is smaller (by at least one order of magnitude) than the system of which it is a component, and (ii) all observations of the system are taken from a particular perspective within the physical hierarchy. A major addition, encouraged by environmental pragmatism, is to expand
(ii)to (ii): All observations and evaluations orient from a particular perspective within the physical hierarchy. An effect of this innovation is to make environmental values, evaluation and social learning about values endogenous to the broader, adaptive management process.
This conceptual apparatus allows us to see human decision-makers as located within layered subsystems and supersystems, with the smallest subsystems being the fastest-changing, and the larger systems changing more slowly. These larger, slower-changing systems provide the environment for adaptation by subsystems (including organisms and places –composed of individuals and cultures). This convention allows us to associate temporal ‘horizons’ with changing features of landscapes as is illustrated in the famous metaphor used by Aldo Leopold, a forester and wildlife manager. Leopold set out to remove predators from the Forest
Service ranges he managed in the Southwestern US.When the deer starved for lack of browse, he regretted his decision to extirpate wolves, chiding himself for not yet having learned to ‘think like a mountain’ (Leopold,1949). He had not yet, that is, understood the role of the targeted species in the broader system. When he came to understand that role, he accepted responsibilities for the long-term consequences of his decisions, and advocated wolf protection in wilderness areas. Leopold’s account parallels the above case of Chesapeake Bay. In both cases, human activities – intended to improve the lot of human consumers of nature’s bounty – threatened larger-scale dynamics. Thinking like a mountain – or a watershed – requires accepting responsibility for the impacts one’s decisions will have on subsequent generations. Accepting this responsibility is inseparable from adopting a larger ecophysical model of the system under management. At this point in time, armed with some knowledge of changing systems and how to model them, we begin to accept moral responsibility for actions that were once thought to be morally neutral. In both cases, accepting moral responsibility – and a sense of caring – were inseparable from adopting a changing causal model of what has happened to deer populations on Leopold’s metaphoric mountain, and to submerged aquatic vegetation in the Chesapeake. Chesapeake Watershed residents, busily plying their trades and tending their lawns, discovered that the ways in which they were pursuing their economic wellbeing could turn the Chesapeake into an anaerobic slime pond. In both cases the total impacts of individual actions to improve individual well-being turn out to reduce the ratio of opportunities to constraints faced by subsequent generations.
Using this framework of actions embedded within nested, hierarchical systems, it is possible to articulate a new approach to evaluating changes in human-dominated systems. Human management of the environment takes place within environmental systems as they are embedded in larger and larger – and progressively slower-changing – supersystems. Each generation is concerned for its short-term well-being (personal survival), but also must be concerned to leave a viable range of choices for subsequent generations. Given our expanding knowledge of our impacts on the larger and normally slower-changing systems that form our environment, it seems reasonable also to accept responsibility for activities that can change the range of choices that will be open to posterity.
A concept of sustainability nicely ‘falls out’ from this conception of adaptive management, in that a ‘schematic definition’ of sustainability can be constructed on the axioms of adaptive management, provided only that prior generations accept responsibility for their impacts on the choice sets of subsequent generations. Given this rather sparse set of assumptions and hypothetical premises, it is possible to provide a simple and elegant definition of sustainability, or rather what might better be called a definitional schema for sustainability definitions (Norton, 2005). Because of the place-based emphasis of adaptive management and the recognition of pervasive uncertainty, there is only so much that one can say about what is sustainable at the very general level of a universal definition. Speaking at this level of general theory, sustainability is best thought of as a cluster of variables; local communities can fell in the blanks, so to speak, to form a set of criteria and goals that reflect their needs and values. While local determination must play a key role in the details, adaptive management, and its associated definitional schema, makes evident the structure and internal relationships that are essential to more specific, locally applicable definitions of sustainable policies.
The two principles of hierarchy theory, when embodied in models, place individual actors in a world that is encountered as a mixture of opportunities and constraints; some of the chooser’s choices result in survival: the chooser lives to choose again. If the chooser survives and has offspring, the offspring will also choose in the face of similar but changing environmental conditions. Some choices of others lead to death with no offspring. Other choices lead to continuation and to offspring who will face a similar, but possibly a changing array of possibilities and limitations. This is the basic structure of an evolution-through-selection model that interprets the environment of a chooser as a mixture of opportunities and constraints; it contextualizes the ‘game’ of adaptation and survival and can be represented as in Figure 2.1A.
Community-level success, in other words, requires success on two levels: at least some individuals from each generation must be sufficiently adapted to he environment to survive and reproduce and, for the population to survive over many generations, the collective actions of the population must beappropriate for (adaptive to) its environment. Since humans are necessarilysocial animals (because of the long period of helpless infancy of individuals), individual survival depends also on reasonable levels of stability in the‘ecological background’ and in the cultural context, the stage on which individuals act. Successful cultures develop specific adaptations appropriate totheir place, adaptation to the cycles and constancies of background systemsthat usually change more slowly than individual behaviors. This simple From this simple framework, a schematic definition of sustainabilityemerges: individuals in earlier generations alter their environment, using upsome resources, leaving others. If all individuals in the earlier generationsover-consume, and if they do not create new opportunities, then they willhave changed the environment that subsequent generations encounter,making survival more dificult.A set of behaviors is thus understood as sustainable if and only if its practice in generation m will not reduce the ratioof opportunities to constraints that will be encountered by individuals insubsequent generations n, o, p.
Although the model has a ‘flat’, schematic character, it could also be given a richer, normative-moral interpretation, as is hinted at by use of theterms opportunities and constraints.Ifwe stipulate that the actors arehuman individuals, then the simple model provides a representation ofintergenerational impacts of decisions regarding resources; our little modelcan thus be enriched to allow a normative interpretation or analogue. If weaccept that having a range of choices is good for free human individuals,we can see the structure, in skeletal form, of the normative theory of sus-tainability. An action or a policy is notsustainable if it will reduce the ratioof opportunities to constraints in the future.
Each generation stands in this asymmetric relationship to subsequentones: choices made today could, in principle, reduce the range of freechoices available to subsequent generations. Thus it makes sense to recognize impacts that play out on multiple, distinct scales. If it is agreed thatmaintaining a constant or expanding set of choices for the future is good,and that imposing crushing constraints on future people is bad, our littlemodel has the potential to represent, and relate to each other, the short- and long-term impacts of choices and to allow either a physical, descriptiveinterpretation or a normative one.
This schematic definition, understood within the general model of adaptive management, captures two of our most important basic intuitionsabout sustainability: (1) that sustainability, incorporating a multi-scalar and multi-criteria analysis, refers to a relationship between generations existing at different times – a relationship having to do with the physical existence of important opportunities – and (2) that this relationship has an important normative dimension, a dimension that cannot be captured by economic measures alone, but one that involves important questions of intergenerational equity. Thus we can tentatively put adaptive management – complete with a schematic definition of sustainability – forward as a useful and comprehensive approach to environmental science and management. Adaptive management, in this context, encompasses the experimental search for better understanding, better goals, and better decisions.
It has been claimed that, provided a community accepts responsibility for its impacts on the future and the set of choices (adaptations) available to future people, a plausible definition of sustainability results. Next it is necessary to show how multi-scalar evaluation of impacts of actions can be correlated with a pluralistic approach to environmental values. If it is recognized that some actions – or aggregations of actions – of individuals threaten a valued aspect of the environment on a multi-generational time scale, there arises a competition between the ‘good’ of current individuals (consumption and increased individual welfare) and the ‘good’ of future people (whowe can expect to want to face a broad array of opportunities to adapt totheir environment as they see ?t). Further, if we accept that (followingHierarchy Theory) these goods are associated with different social and ecological dynamics,which unfold at different scales, it may be possible to identify public policies that protect both kinds of goods; or, it may be possibleto find an acceptable balance among the values if they turn out to be competing (Norton, 2005).
In a pluralistic value system – if it is embedded in a multi-scalar system
– Some human values can be associated with faster (‘economic’) processesof production and consumption. Protection of native vegetation andimproving bay water quality, on the other hand, are associated with alarge-scaled system and with values that, because they unfold at differentscales and are supported by different processes, need not compete witheconomic values in real multi-scaled systems. It becomes conceivable tofind win–win policies that provide adequate increments of individualwelfare, but which do so in a way that does not destroy options open forfuture choosers.Multi-scalar thinking, an emphasis on experience, and a forward-looking, pragmatic, problem-oriented attitude have been argued to be ade-quate to adaptive management processes, even though the goal of‘sustainable development’ is not yet clearly defined. By recognizing that wecan learn from experience, and by developing multiple criteria associatedwith different scales, it is possible for a community – much as theChesapeake community did – to learn itself into a new set of indicators, anew set of concerns, and a whole new understanding of their place and thespace around that place. If environmental ethics is to contribute to pursuitof sustainable development, that contribution seems more likely to comefrom the pragmatic line of analysis, functioning as a ‘philosophy’ of adaptive management, than from sterile discussions of which elements of naturehave intrinsic value and moral considerability.

3 The capital approach to sustainability Giovanni Ruta and Kirk Hamilton
1. Introduction
It is a matter of fact that sustainability has been adopted by many scientists, prime ministers and citizens alike as a goal for the world we would liketo live in, and yet that its measurement is largely non-existent. The purposeof the chapter is to approach the measurement challenge the way an economist would: if sustainability means leaving future generations with at leastas many opportunities as we have today, then the way to achieve this is bypassing on to future generations a level of capital that is at least as high asours today.
The measurement of sustainability can then be likened to an accountingexercise in that the object being measured is capital, very much the sameway a firm would report the value of buildings, machinery and trademarksin its books at the end of each year. But when we start thinking about a country’s capital, produced assets – such as buildings and machines – are not enough to describe the complex set of elements which form the base forthe production of well-being. The chapter starts by establishing the conceptual link between sustainability and wealth. Next, the methods and tools underpinning the wealth estimates are explained followed by a presentation of the main highlights from recent findings on wealth estimation.
This discussion draws on the results published by World Bank (2006a)which presents estimates of ‘total’ capital, or wealth, for nearly 120 countries. A further section is devoted to the components of intangible capital: a major determinant of wealth. Finally, the policy implications of the capital approach to sustainability are presented.
2. Sustainability, wealth and well-being Most people will agree that sustainable development is something that is desirable, like happiness, yet few will be able to pinpoint its practical implications. A myriad of definitions have been proposed but it has not been easy to find one that simultaneously satisfies economists, ecologists, sociologists, philosophers and policy makers. The problem in part relates to uncertainty about the object of sustainability, rather than the idea itself.
What is it that ought to be sustained?
Natural scientists and ecologists will typically respond to the question above by stating that it is the capacity of the ecosystem that needs to be sustained. Concepts such as diversity and resilience become then useful in addressing the complex measurement issues. An ecologically based measure of sustainability is especially important in those cases in which the natural resource is critical to survival. The ozone layer and the oceans truly provide services that can hardly be thought of as replaceable. A world economy that depletes the ozone layer cannot be considered sustainable.
More generally, however, identifying sustainable development with a halton all ecosystem transformation would probably come at prohibitive costs for the economy.
A more comprehensive approach would identify sustainable development with the maintenance of a non-declining level of a number of ecological, social and economic indicators. While appealing, a problem withthis approach is that it is dificult to make claims about sustainability whensome indicators increase while others decrease.Would a society be sustainable if equity is enhanced while natural resources are lost? In this chapter,we argue that what needs to be sustained should be a comprehensive object.
In particular, we argue that the concept of social well-being should be thestarting point. One may even emphasize that well-being, or utility, is simplythe result of the different elements of what constitutes development,including a clean environment, income and social relations.
The question of ‘what’ should be sustained will automatically lead to concerns about measurement. And measuring well-being is indeed a non-trivialmatter. Yet, this is where economics makes a crucial contribution. It turnsout that, if properly measured, capital or wealth constitute an appropriatemeasure of social welfare. Following the lead of the Brundtland Commission,the issue was clearly put by Pearce et al. (1989) who argued that sustainablewell-being is possible if the next generation inherits ‘a stock of wealth . . . noless than the stock inherited by the previous generation’ (p. 34). Wealth, orcapital assets, becomes the object of the sustainable development paradigm.
From well-being to wealthA myopic approach to sustainability will typically consider well-being asapproximated by income. To have sustained well-being, the quantity ofgoods and services produced in an economy should not decline from oneyear to the next. A defendant of this proposal might point to the fact that,by and large, higher income leads to higher well-being. Moreover, growthof income is important to address social goals such as poverty alleviation.Income measures, however, do not say much about sustainability. Higherincome does not necessarily mean higher sustainability, in the same way asa higher fishery catch does not necessarily mean a bigger fish stock.
The fact that income, or for thatmatter consumption, does not have a directwelfare connotationwas highlighted in a seminal paper by Samuelson (1961).Assume you observe two countries,AandB.Both countries produce the samelevel of income but while A consumes it all, B saves a part of its income andinvests it into productive capital. Citizen A is consumingmore than citizen B butgiven country B’s saving effort, B will soon be able to generate a higherlevel of income and increase its consumption possibilities. In order tocompare well-being between the two countries, current income provides amisleading signal: while starting from the same level of income, B will soonbe able to produce more, owing to its saving effort. Current consumptionsimilarly provides amisleading signal.The choice has to bemade ‘in the spaceof all present and future consumption...the only valid approximation to ameasure of welfare comes from computing wealth-like magnitudes notincome magnitudes’ (Samuelson, 1961, pp. 50, 57).
Irving Fisher (1906) provided the original insight that current wealthequals the present value of future consumption. For the relationshipbetween current and future consumption and wealth to hold, one shouldhowever make sure that, in the latter, all assets that are needed for the generation of well-being are included. Fisher (1906) identified three types ofassets: immovable wealth, comprising of land and the fixed structures uponit,movable assets, or commodities, and human beings. As we shall see, theseassets remain of interest although terminology has changed and more categories have been added to this list.
From wealth to sustainability If wealth is the correct measure of well-being, sustainability can be expressed in terms of changes in wealth.A major strength of the capital approachto sustainability is the fact that it provides a simple and forward-looking guide to policy makers.
Consider the following definition of sustainability: a development path is sustainable if social well-being, that is, the present value of current and future consumption, does not decline at any point along the path:
Given that social welfare equals wealth, a simple sustainability test requires that wealth does not decline over time. In other words, the level of net saving, adjusted to take into account the net changes in natural and human capital, should be positive for the economy to be sustainable.
The strength of this definition of sustainability is that it provides a forward-looking guide to policy. Decision makers at time t do not usually know what utility or well-being will look like far in the future. But theydon’t need to. To achieve sustainability, the only thing the committed policymaker should worry about is that current net saving be positive.
In making this claim, we implicitly adopt a paradigm which allows forthe possibility of replacing natural capital with produced capital. This approach has the weakness of not being able to account for irreplaceable assets such as biodiversity hot-spots and the oceans’ regulating function over the global climate. Low substitutability critically hinders sustainability. Substitutability refers to the extent to which an asset, for example natural resources, can be replaced by another asset, for example man-madecapital, in the production process. If substitutability is low, that is, the elasticity of substitution between man-made capital and exhaustible natural resources is less than one, sustainability is not possible in the absence of technical progress (Dasgupta and Heal, 1979).
Pearce and Atkinson (1993) and Pearce et al. (1996) have highlighted the advantages and limits of the so-called ‘weak sustainability’ rule. While undermined by the existence of irreplaceable and unique assets, weak sustainability has the non-trivial advantage of being easy to apply and still provide a strong signal: ‘even on a weak sustainability rule many countriesare unlikely to pass a sustainability test’ (Pearce and Atkinson, 1993,p. 105). Hamilton and Clemens (1999) calculated the first country-widegenuine saving rates for developing countries, showing that the greatestwealth dissipation is taking place in many of the poorest countries in theworld. Chapter 18 deals explicitly with the theory and practice of genuinesaving. For present purposes, it suffices to say that genuine saving measuresthe true rate of saving of an economy, after accounting for the depletionof natural resources, investments in human capital and damages from(certain) pollutants.
The advantage of measurability
The capital approach to sustainability provides an answer to the measurability dilemma. Measurement requires that our computation of wealth (a)be comprehensive and (b) use the right prices. Comprehensiveness meansthat not only should produced capital be counted as wealth but also naturalresources, human capital and social capital should be accounted for. Thenext section describes the estimation issues. While substantial progress hasbeen made in the measurement of natural capital, many assets are leftoutside due to the lack of data.Groundwater and fishery stocks, for example,are not included in the measures of natural wealth presented in this chapter.
Human and social capital is still very hard to measure. The approach here isto compute it as the difference between total wealth and the sum of the tangible components of wealth (produced capital and natural capital).Proper accounting prices are required to measure the individualcomponents of wealth. This is not difficult for marketed, produced goods.
It is, however, a challenge when it comes to non-marketed items, for whichprices are not directly observable. Asset prices are intimately related to thescarcity of the asset. If an economy is running out of clean water, citizenswill usually have to pay higher prices for potable water. As many environmental and natural resources are provided at no charge, the market price isusually a bad signal of their scarcity and modelled accounting prices needto be estimated.Knowing the composition of wealth helps inform policy making
The wealth estimates not only provide a measure of well-being, they alsoprovide useful insight into the composition of capital assets in an econ-omy. Policies to foster sustainability depend on the relative endowmentsof resources a country has available for the generation of well-being.Economic management for sustainability can be equated to a process ofportfolio management, in which economic decisions entail the transformation of one resource into another.Forested areas can be transformed into cropland; oil rents can be invested in school facilities. Sustainability is not about keeping this or that asset intact, but rather about keeping the system’s ability to produce well-being. Sustainable development in an oil country, such as Venezuela, will
4 mean investing resource rents in human or physical capital. Development need not only entail the transformation of natural capital in other assets.
In a resource-poor, rural based economy such as Ethiopia, sustainable development means keeping, and possibly increasing, the land’s capacity to produce an economic surplus, which only then can be invested in other assets. In biodiversity-rich countries, such as Peru, sustainability will entailmanaging pristine areas so as to maximize revenues from sustainableforestry, tourism and bioprospecting research.
Knowing the basis of a society’s welfare is a desirable objective. The nexttask is to understand how concrete estimates of total wealth can be obtained.
3. The architecture of the wealth estimatesBroadly speaking, total wealth is composed of produced capital, naturalcapital and intangible capital, where the latter is an aggregate includinghuman, social and institutional capital. Rather than summing up thesethree components, the estimation proceeds by first estimating total wealth,then produced capital and natural capital and finally calculating intangiblecapital as the difference between total wealth and the sum of produced andnatural capital.
Estimating total wealthTo measure total wealth, and in line with Fisher (1906), Hamiltonand Hartwick (2005) show that the current value of wealth, composed ofThere are a number of estimation methods available for the calculationof physical capital stocks. Some of them, such as the derivation of capitalstocks from insurance values or accounting values or from direct surveys,entail enormous expenditures and face problems of limited availability andadequacy of the data. Other estimation procedures, such as the perpetualinventory method (PIM) are cheaper and more easily implementable sincethey only require investment data and information on the assets service life and depreciation pattern. Here, the following PIM formula was used to compute the value of machinery, equipment and structures:
i where I is the value of investment in constant prices and 0.05 is a geo-
5 metric depreciation rate.Urban land was valued as a fixed proportion of the value of physical capital. This is a fallback for the more palatable and data intensive option of using country-specific proportions. A constant proportion equal to
24 per cent is then assumed.
Natural capital
Natural capital is the sum of non-renewable resources (including energy resources such as oil, natural gas and coal, and mineral resources), cropland; pasture land, forested areas (including areas used for timber extraction and non-timber forest products) and protected areas.
The PIM is not useful in valuing natural capital, given that most naturalresources are accumulated over a very long time span. The present valuemethod is used in most cases. This method consists of computing thepresent value of a given natural resource net rents over the life span of theresource. When data on rents (or benefits) is not available, the opportunity cost method is used instead.
_ Sub-soil assets. Estimating future rents for sub-soil assets is subjectto a high level of uncertainty. Here the simplifying assumption thatrents grow at a constant rate is used. Moreover, an average life of a mine is assumed to be 20 years (this may vary from country to country though and from one resource to the other).
_ Timber resources. The predominant economic use of forests has beenas a source of timber. Timber wealth is calculated as the net presentvalue of rents from roundwood production. The estimation thenrequires data on roundwood production, unit rents and the time toexhaustion of the forest (if unsustainably managed). Notice that theuse of rents to value capital implicitly assumes that the timber valueof a forest is given by the currently exploitable timber, rather than thevolume of the resource itself.
_ Non-timber forest products. Average world values (from Lampietti and Dixon, 1995) are applied to a share of the country’s forest.
_ Cropland.Given the lack of data on land prices, land values are computed on the basis of the present value of land rents, assuming thatthe products of the land are sold at world prices. The return to land iscomputed as the difference between the market value of output cropsand crop-specific production costs. Nine representative crops areselected based mainly on their production significance in terms ofsowing area, production volume and revenue. The ninerepresentativecrops considered are: maize, rice, wheat, banana, grapes, apples,oranges, soybean and coffee. A country’s overall land rent is calculated as a weighted average (weighted by sowing areas) of rents fromthe crop categories. A projected growth in production (land areas areassumed to stay constant) is assumed based on Rosengrant et al.
_ Pasture land. The returns to pasture land are assumed to be a fixedproportion of the value of output. On average, costs of productionare 55 per cent of revenues, and therefore returns to pasture land areassumed to be 45 per cent of output value. Value of output is basedon the production of beef, lamb, milk and wool valued at inter-national prices.As is the case for cropland, this rental share of outputvalues is applied to country-specific outputs of pasture land valuedat world prices. A projected growth in production is assumed also inthis case (Rosengrant et al., 1995).
_ Protected areas.Values are obtained using, as a proxy, the lower ofthe unit values of cropland and pasture land; an imperfect and conservative measure of the opportunity cost of protecting land areas.
Precise estimations are very difficult to undertake and country-specific data are sparse.
Intangible capital Even after accounting for produced capital and a large set of naturalresource assets, the wealth estimates show that most countries’ wealth iscaptured by what we call ‘intangible capital’. By definition, intangiblecapital captures all those assets that are unaccounted for in the wealthestimates. It includes assets such as the skills and know-how embodied inthe labour force: human capital. It also encompasses social capital; that is,the amount of trust among people in a society and their ability to worktogether for common purposes. Finally, it includes those elements ofgovernance that boost the productivity of the economy. For example, if aneconomy has a very effcient judicial system, clear property rights and aneffective government, the effects will be picked up in the form of highertotal wealth and thus will increase the ‘intangible capital’ residual.
The intangible capital residual also includes other assets which, for lackof data coverage, could not be accounted for in the wealth estimates. Themain omissions include coastal and marine resources, such as fisheries, andthe net depletion of renewable natural resources such as underground waterand environmental services.
4. The highlights of the capital estimates
Country-specific estimates of total capital are presented in World Bank
(2006a). Table 3.2 summarizes the results by region, income group and forthe world as a whole. High energy and mineral exporters are treated as aseparate group. The relative distribution of assets in these countries is suchthat the aggregates would tend to overestimate the role of natural capital particularly sub-soil – in the groups such countries are in.
A quick glance at Table 3.2 reveals the following.
Firstly, the average world citizen ‘owns’ a total wealth of nearlyUS$96 000. The number becomes US$90 000 if oil exporters are included.
This level of wealth is comparable to the one for Brazil (US$90 000), Libya (US$89 000) or Croatia (US$91 000).
Second, total wealth in high income countries is several times higher than in low income countries (column 2). This fact is only partially due to the use of nominal exchange rates as opposed to purchasing power parity (PPP) exchange rates typically used to compare welfare between highincome and developing countries.
Third, natural capital is higher in value in high income countries than in low income countries (column 3 and Figure 3.1). This evidence contradicts a common perception that high income countries have ‘usedup’ their natural resources.
Fourth, the share of natural capital in total wealth decreases with income(column 6 and Figure 3.2). The world’s poorest countries – particularly in
South and East Asia – depend heavily on natural resources. Developmentcannot be pursued without maintaining an ever watchful eye on how natural resources are managed.
Lastly, intangible capital – an aggregate including human capital, the quality of institutions and governance – constitutes the preponderant form of wealth, an insight that goes back to the very origins of modern economic thinking (columns 5 and 8).
Points three and four above are particularly relevant from the perspective of sustainability. Natural resource abundance is also a characteristic of wealthy economies. What we are observing is better management of resources such as agricultural land (resulting in higher yields) and forests (resulting in timber rents that are sustained over time).
Yet, low income countries are more dependent (in terms of relative share) than high income countries on natural resources. This provides useful information. What we are observing is low levels of diversification and low levels of intangible assets such as education and efficient institutions. Given the importance of natural capital on the wealth of poor countries, one should look at the individual sub-components (Table 3.3).
If one excludes large resource exporting countries, which constitute a group by themselves, land resources (columns 5–7 in Table 3.3) are very important in low income countries, with a value of 75 per cent, followed by sub-soil assets (column 2), with 17 per cent. In middle income countries land resources account for 61 per cent of natural capital, while sub-soil assets account for 31 per cent of the total.
The importance of land resources (that is, cropland, pasture land and protected areas) decreases with the level of income. This fact is partly the effect of using international prices for agricultural products, a procedure that overestimates the value of land in countries with subsistence agricultural production. However, the results also suggest a potential poverty–land dependence trap in low income countries. Countries in which land resources account for more than one third of total wealth – such as Niger, Burundi, Moldova to name a few – all belong to the low income country group.
By contrast, high dependence on sub-soil assets is not necessarily a char-act eristic of low income countries. Countries which are rich in mineral and energy resources may be found in each of the income groups. Rents from sub-soil assets can be key in raising countries out of poverty, but do not represent a sufficient condition: high rents require efficient management in order to achieve poverty reduction (see Chapters 13 and 14).
5. Understanding intangible capital
Given its role in the wealth numbers, one should look more closely at the intangible capital component. Regression analysis can help us pinpoint its major determinants. Three factors – average years of schooling per capita, rule of law, and remittances received per capita – explain 89 per cent of the 7total variation in the residual across countries. Figure 3.3 shows the relative importance of each factor, with rule of law accounting for 57 per cent and schooling accounting for 36 per cent of intangible capital.
Table 3.4 reports the marginal returns, measured at the mean, to unit increases in the three factors for each level of income. Increasing the average stock of schooling by one year per person, increases total wealth per capita 8 by nearly $840 in poor countries, nearly $2000 in middle income countries and over $16 000 in high income countries. A one-point increase in the rule of law index (on a 100 point scale) boosts total wealth by over $100 in poor countries, over $400 in middle income countries, and nearly $3000 in high income countries. Larger stocks of produced capital – usually at higher income levels – will also boost the returns to education and governance. This helps to explain the wide ranges in marginal returns as countries get richer.
The analysis of intangible capital provides useful insight for policy makers. Education expenditure can obviously play a role, but these expenditures have to be effective in actually creating human capital. Investing in rule of law is clearly complex – an efficient judicial system for example calls not only for competitive salaries but also for competent institutions that can be trusted by citizens and entrepreneurs alike. The returns to doing so, however, are potentially very large.
6. The capital approach to sustainability: implications
A key contribution of the economic debate of sustainability is that it sets the ground for measurement. Hart wick (1977) demonstrated that under some stringent conditions, non-declining real wealth implies non-declining consumption. More in general, non-declining real wealth is associated with non-declining social welfare. The bottom line is that comprehensive measures of wealth and its changes appear as meaningful indicators to track sustainable development. Saving, in particular, constitutes a significant measure of sustainability and one that provides useful insight for policymaking. Chapter 18 analyzes thoroughly the theory and evidence related to genuine saving.
By looking at comprehensive wealth, the objective is to understand the potential for the creation of well-being in a country. This approach revives the ideas of the classical economists, who identified not only man-made capital but also labour and natural resources as determinants of production. From the numbers, it is evident that the components of wealth vary widely across regions and according to level of income. Managing each component of the portfolio and transforming efficiently one type of asset into another is germane to a country’s development policy.
Implications for policy makers
Economic decisions are usually the domain of finance and economy ministers and seldom take into account environmental concerns. The capital approach to sustainability expands the responsibilities of economic management to include the management of natural resources, human capital and institutions. The wealth estimates indicate that the development process entails a diminishing dependence on natural resources while increasing the reliance on human skills and the country’s social and institutional infrastructure. Notice that this need not occur at the expense of environmental degradation. While less important in relative terms, natural resources are larger in absolute terms in richer countries.
Managing development in the poorest countries requires the recognition of the role of natural resources as a source of subsistence. In aggregate, natural capital represents a quarter of wealth in low income countries.
Throughout the world, many rural households depend on the services of forest ecosystems, fisheries and agricultural land for subsistence. These resources are typically renewable, and the management challenge essentially entails sustainability in use. Institutions and social arrangements that foster conservation include the clear definition of property rights and the control of corruption and poaching. On a positive note, there must be policies geared toward increasing the productivity of assets, so as to allow growing income and consequently higher savings to finance investment.
In resource-rich countries, natural resources are a fundamental source of development finance. Fiscal policies should be geared toward capturing resource rents. Examples include energy royalties, taxes on tourism revenues, underground water tariffs. Public expenditure should give priority to high return investments, as opposed to the more commonly observed excessive public consumption expenditure (see Chapters 13 and 14). This may prove difficult with fiscal shocks, typical of oil countries, and low absorptive capacity. In the short term, investment in financial assets may be a better option compared to an unsustainable increase in current expenditures. Botswana, for example, has been able to manage diamond revenues successfully through a strict budget balance rule.
Investment in man-made, human capital and reliable institutions is crucial. Governments should invest in education, an efficient judicial system and rule of law and policies aiming at attracting remittances.
Implications for economists and statisticians
Good decision-making requires good information. The wealth estimates discussed in this chapter constitute a contribution to this work on ‘greening’ the national accounts. Including monetary estimation of natural capital in a country’s macroeconomic balance sheet is important in representing the actual sources of welfare for the country. The economic valuation of environment and natural resources is the basic building block of a comprehensive accounting system. Valuation can usefully inform monitoring and enforcement, decision-making through cost–benefit analysis and fiscal policies.
Asset prices have to reflect the social worth of capital, which in turn reflects its social scarcity. Moreover, achieving sustainability critically depends on the substitutability of man-made capital and natural resources.
The substitutability issue is also a measurement problem. Valuing total wealth as the sum of produced, natural and human capital relies on the assumption that assets are substitutable. It must be possible to deplete one resource and substitute it with other assets, for our assumed ‘weakly’ sustainable world to hold. If assets are irreplaceable, while being essential for the production of well-being, physical measures must complement monetary measures of capital.
7. Summing up
The discussion in this chapter was motivated by the need to adopt a pragmatic measure of current generations’ bequest of opportunities to future generations. A narrow definition of ‘opportunities’ – associated with capital – was identified. Wealth per capita, measured as the sum of all assets that allow the production of well-being, was thus our measure of ‘opportunities’. To the extent that future generations are left with a level of total wealth per capita at least as high as today’s wealth, then we are on a sustainable path, at least from a weak sustainability perspective.
Where is the wealth of nations? The estimates of comprehensive wealth and its components go beyond a simple sustainability test and provide insights about what constitutes a country’s base for producing well-being. By and large, wealth is about intangible assets. Intangible does not mean indefinable. In fact, a very strong association between education attainment, governance and institutions on one side and intangible capital on the other is found. A society investing in skilled workers, trusted institutions and efficient government is building the very basis of welfare creation.
These sorts of intangible assets explain the high level of wealth of countries in Europe, North America and East Asia.
How about natural capital? The wealth estimates suggest the importance of natural resources management in maintaining wealth in the poorest countries in the world. For the average citizen of Ethiopia, natural capital –particularly crop land – constitutes more than 40 per cent of available assets. Depleting forest resources and degrading agricultural soils will impair the prospect for poverty alleviation. Sustainability here not only requires investing resource rents into some form of capital. Within a subsistence economy, it means managing natural assets so as to provide the basis for income. Mineral deposits, once discovered, can only be depleted.
Sustainability here means investing resource rents in some form of capital.
The Hartwick rule for sustainability has, however, been neglected by many resource-rich countries, leading to consumption levels that are unsustainable, explaining economic downturns.
Finally, to measure sustainability truly, the focus has to be on changes in wealth. The wealth estimates provide an insightful vision of the world and its prospects for generating welfare. Any sensible sustainability test should, however, look at the change in capital rather than at the stock.
Chapter 18 introduces genuine saving – that is, the annual change in totalreal wealth – as a measure of sustainability. Breaking down total wealthinto its components is a major step forward in the analysis of country-level endowments and welfare generation possibilities. The estimatesmade available here contribute to this work even if data constraints limitour ability to measure some assets in a comprehensive way. This work has just begun.
4 Sustainable developments in ecological economics Jeroen C.J.M. van den Bergh
1. Introduction
The notions of ‘sustainable development’ and ‘sustainability’ are inter-preted in various ways. This has become most clear perhaps in the field of ecological economics, where different disciplines have offered particularperspectives on these notions. Ecological economics (EE) was founded atthe end of the 1980s. It integrates elements of economics and ecology, aswell as of thermodynamics, ethics, and a number of other natural andsocial sciences to provide for an integrated and biophysical perspectiveon environment–economy interactions. EE expresses the view that theeconomy is a subsystem of a larger local and global ecosystem that limitsphysical growth of the economy. At the same time, it is critical of the dominant paradigm of (environmental and resource) economics, characterizedby rational agents and equilibrium thinking. Instead, EE is characterizedby the use of physical (material, energy, chemical, biological) indicatorsand comprehensive, multidisciplinary systems analysis. Both features areconsistent with the fact that (un)sustainable development, generally seen asan important dimension of performance of the overall systems level, occu-pies a central position in the study of EE.
All intellectual founders and antecedents of EE have written extensivelyabout sustainable development, even if not using this particular terminology. For example,H.E.Daly proposed the idea of a ‘steady state economy’,associated with the objective to minimize the use of materials and energy‘throughput’ in the economy (Daly, 1991). In addition, he has suggested theIndex of Sustainable Economic Welfare (ISEW: see also Chapter 19) as asustainable welfare indicator (Daly and Cobb, 1989). K.E. Boulding pro-posed the opposition between the ‘cowboy economy’ and the ‘spaceshipeconomy’ (Boulding, 1966). The spaceship metaphor can be seen as aprecursor to the modern view on sustainability from a global environmental perspective. Finally, C.S. Holling (1973, 1986) has originated the notion of resilience (Chapter 5), which has proven to be a fruitful and distinctiveway of thinking about sustainable development.
This chapter tries to provide a broad sketch of ideas, approaches and policyangles that ecological economics has offered in the study of sustainable development. The result is the following structure. Section 2 discusses the distinctive character of ecological economics approaches to sustainable development as compared with mainstream economics. Section 3 then examines the well-known opposition between strong and weak sustainability. Section
4addresses the sustainability of open systems, involving issues like spatial sustainability and sustainable trade. Section 5 deals with measurement of, and models for, sustainable development. Section 6 discusses policies specifically oriented towards sustainability. Section 7 concludes.
2. Ecological versus environmental economics
An important distinction between ecological economics (EE) and environmental and resource economics (ERE) relates to scale versus allocation. ERE studies optimal allocation orefficiency of using scarce resources. Consistent with this idea is the objective to optimize social welfare and thusstrive towards an optimal level of external costs. Daly (for example, 1992) argues that ERE has, however, neglected the issue of optimal physical scale or size of the economy. Consistent with this neglect, ERE tends to regard sustainable development as identical to sustainable growth. EE, on the other hand, sees sustainable development more in line with the oldernotions of development and structural change. Not surprisingly, history,institutional context and poverty receive much more attention in EEdiscussions and analyses of the concept. Somewhat related is the fact thatERE, or at least many of its proponents, does not seem to take physicallimits to growth as seriously as supporters of EE. This might have to dowith optimism about both the inventiveness of humans (technical progressand problem-solving in general) as well as about the stability of nature andenvironmental systems to withstand pressure caused by humans. Possibly,
EE generally assumes a longer time horizon than ERE. In this sense, thedifferent approaches to sustainable development – optimistic versus precautionary – bear a strong relationship with the different positions in the growth debate (van den Bergh and de Mooij, 1999).
The main goals and criteria for evaluating developments, policies and projects differ between EE and ERE. The dominant criterion of ERE is efficiency (or sometimes a more limited version, such as costs-effectiveness). EE is best characterized by a ‘precautionary principle’ linked to environmental sustainability, with much attention for ‘small-probability–large-impact’ combinations. This precautionary principle is closely related to a concern for instability of ecosystems, loss of biodiversity, and environmental ethical considerations (‘biocentric ethics’).
Efficiency is in EE of secondary concern. Furthermore, whereas in EREdistribution and equity are secondary criteria, ‘distribution’ is often in EE considered a more important criterion. In line with this, EE emphasizes (basic) needs, North–South welfare differences, and the complex linkbetween poverty and environment. In addition, a recent emphasis in the literature is that it is impossible to analyse distribution and efficiency perfectlyseparately, as the latter depends on the former (Martinez-Alier andO’Connor, 1999).One argument here is that preferences are interdependent and income distribution a?ects individual well-being. Subjective welfare studies show that relative rather than absolute income is an important factor of happiness (Tversky and Simonson, 2000; Brekke and Howarth, 2002; van Praag and Ferrer-i-Carbonell, 2004).
3. Strong versus weak sustainability Sustainability and sustainable development have been defined, interpreted and analysed in various ways (see Pezzey, 1989, 1993; Toman et al., 1995). Beckerman (1994) has argued that these notions serve no purpose as they are already captured in the concept of intergenerational welfare optimization.
Responses by Common, Daly, El Serafy and Jacobs in Environmental Values
vol. 4 (1995, issues 1 and 2) and vol. 5 (1996, issue 1) oppose this view. In par-
ticular, the opposition between strong and weak sustainability has received much attention in the literature (Ayres et al., 2001).
Weak sustainability
Weak sustainability has been defined using notions like ‘economic capital’ and ‘natural capital’ (Cabeza-Gutés, 1996). Economic capital comprises machines, labour and knowledge. Natural capital covers resources, envir-
onment and nature. Weak sustainability is defined as maintaining ‘total capital’, defined as the ‘sum’ of the two types of capital. Evidently, under this goal the substitution of natural capital by economic capital is allowed for. The methodological aspects of this approach are most clearly expressed in economic growth theory (Solow, 1974, 1986; Hartwick, 1977). This theory translates weak sustainability into intergenerational equity (Toman et al., 1995). Sustainability is usually interpreted as a constraint on eco-
nomic growth, namely non-decreasing welfare. This is quite a strict crite-
rion, as any temporary decrease in welfare implies an unsustainable development. Pezzey (1989) has referred to ‘sustainedness’ in this respect, since such a pattern can be assessed only after the fact. As a weaker alter- native criterion, Pezzey (1993) proposes ‘survivability’, according to which a reduction in welfare is allowed as long as the level of consumption exceeds some subsistence level.
In the general economic case, social welfare is a function of utility, which is di?cult to operationalize. In practice, simple models often equate utility to (aggregate) consumption, defined as gross output less investment. This gives rise to ‘Hicksian sustainability’, or non-decreasing consumption, which is equivalent to ‘Hartwick–Solow sustainability’ defined in terms of maintaining the total capital stock of society.
Strong sustainability Strong sustainability, on the other hand, requires that every type of capital – economic and natural – is maintained separately, or that even, at a lower level of disaggregation, capital stocks are maintained. Various motivations for strong sustainability exist: ? Natural resources are considered as essential inputs in economic production, consumption or welfare that cannot be substituted for by manufactured or human capital. Life support functions of nature and environment are often mentioned here.
_Acknowledgement of environmental integrity and ‘rights of nature’ (bioethics).
_ Risk aversion in combination with irreversible changes in natural capital. In this context the terms stability, resilience, (bio)diversity and ecosystem health (Costanza et al., 1992) are often mentioned.
Within EE frequently a particular type of (un)sustainability is pointed out, namely the stability and resilience of ecosystems. Stability is defined at the level of biological populations. This means that variables return to equilibrium values after perturbation. Resilience (resistance to change, or robustness) is defined at the system level and refers to the maintenance of organization or structure and functions of a system in the face of stress (see Chapter 12). Perrings (1998) mentions two alternative approaches to resilience: one is directed at the time necessary for a disturbed system to return to its original state (Pimm, 1984); the other is directed at the inten-
sity of disturbance that a system can absorb before moving to another state
(Holling, 1973). In line with the latter interpretation resilience has been phrased ‘Holling sustainability’, as opposed to weak ‘Solow–Hartwick sustainability’ (Common and Perrings, 1992). The comparison shows that
EE studies pay much attention to the sensitivity of ecosystems at a micro level, often in applied studies, whereas ERE extends economic growth theory with environmental variables, emphasizing determinism and coarse long-term trends in a macro approach that lacks micro detail. From this perspective EE and ERE approaches to sustainability can give rise to complementary as well as contradictory insights.
‘Very strong’ sustainability, as supported by the Deep Ecology movement and those who believe in the ‘right-to-life’ of other species, would then imply that every component or subsystem of the natural environment, every species, and every physical stock must be preserved. A compromise version of strong sustainability focuses on preserving ecosystems and envir-
onmental assets that are critical for life-support or unique and irreplace- able. The ozone layer is an example of the first; songbirds or coral reefs might be an example of the second. Another way of formulating such a compromise is that a minimum amount of certain environmental assets should be maintained, based on the idea that these assets are partly complementary to economic assets and partly substitutable by the latter.
How to judge or resolve the opposition?
The opposition between strong and weak sustainability is ultimately a question about the substitutability between the products and services of the market economy and the environment, or the substitution of natural by produced capital (including human capital or knowledge). This has often been discussed in the context of production processes (see the special issue of Ecological Economics, vol. 22(3) (1997) on the contributions of Nicholas
Georgescu-Roegen to ecological economics). However, the distinction also applies to consumption and individual welfare. This is most clearly expressed in the notion of lexicographic preference orderings, which is con-
sistent with the Maslow pyramid (Stern, 1997). It denies universal substi-
tutability. This is consistent with findings in experiments and stated preference valuation (Spash and Hanley, 1995; Gowdy, 1997).
A problem with the weak sustainability approach as formalized in growth theory with environment or resources is that this was formulated explicitly for non-renewable resources, not for complex biological systems.
Moreover, the tools of growth theory – deterministic dynamic optimization models with one dynamic equation describing the environment – are too rough to incorporate scientific facts of complex evolutionary (irreversible) living systems. Therefore, growth theory cannot offer a complete, and perhaps not even a relevant, perspective on sustainability.
Resilience can be considered as a global, structural stability concept, based on the idea that multiple locally stable ecosystem equilibria can exist.
Sustainability can thus be directly related to resilience. In line with this, weak sustainability can cause extreme sensitivity to either natural disturbances (for example, diseases in the case of agriculture focusing on only a few crops: see Chapter 22) or economic disturbances (international financial markets as in the case of the small Paci?c island nation of Nauru: Gowdy and McDaniel, 1999). Such extreme sensitivity or lack of resilience of regional systems in the face of external factors is a telling argument against weak sus-
tainability. Traditional economic models with environment and resources do, however, not address resilience, fluctuations and cycles. Business cycle theories might be useful in this respect (Young, 1996). Indeed, one may wonder why other types of dynamic macroeconomics – apart from growth theory – have seen so little application in environmental economics, for example, to address questions related to the interaction between sustain- ability and unemployment. Finally, it is very likely that the truth is in between weak and strong sustainability. Perfect substitutability is not real-
istic, but neither is maintenance of all individual environmental stocks and biological populations.
4. Spatial sustainability and sustainable trade
When talking about sustainability, scale and openness of a system are important. Openness means that the system may affect other systems and be affected from outside, either by other regions or by the global system. A relevant question about sustainability in an open (regional/national) system context is whether trade can substitute for nature at the local level.
The international dimension of environmental problems and policy has received much attention over the last decade. Nevertheless, this has pre- dominantly concerned attention for international trade with traditional economic welfare- or externality-based models. Dynamic issues of regional sustainability and its counterpart sustainable trade have hardly received attention. As a result, much is known about the efficiency of trade but not about its sustainability. This would require some merger of dynamic theories (including possibly growth theories), trade theories, resources and externalities. The result is a very complex system.
Countries with a history of resource depletion and ecosystem damage may look sustainable. Indeed, numerical results in Pearce and Atkinson
(1995) show that this is the case for the Netherlands and Japan, both of which have hardly any forest land. This hints at the problem of sustain- ability of open regions or countries, which evidently can surpass local sustainability limits by engaging in international trade.
Daly and Cobb (1989) have expressed the opinion that insights from traditional comparative advantage theory have less relevance these days as the assumption of immobile capital flows no longer holds. They conclude, referring to statements by J .M. Keynes, that production of products should, whenever feasible, take place in the home country. An additional argument for this view is that sustainability at a regional scale can be better controlled in an autarchic than an open region.
In order to ‘measure’ regional unsustainability, Wackernagel and Rees
(1996) have formulated the ‘ecological footprint’ (EF: see also Chapter 20) and applied it to countries (as well as other spatial units). They conclude that many countries, in particular small ones, use directly and indirectly more surface area than is available inside their national boundaries.
Evidently, this is compensated by international trade. Wackernagel and
Rees try to argue on the basis of the EF that autarchy is to be preferred to a trading region. Van den Bergh and Verbruggen (1999) criticize the EF indicator and applications:
_ The EF is an example of ‘false concreteness’: the resulting land area is hypothetical and too crude a measure of various types of environ- mental pressure.
_ The EF method does not distinguish between sustainable and unsus-
tainable land use, notably in agriculture.
_ Aggregation of di?erent environmental problems occurs through an implicit weighting that lacks any motivation.
_CO emissions due to burning fossil fuels are translated, on the basis
2 of an arbitrary ‘sustainability scenario’ (forestation to capture CO ), 2 into hypothetical seizure of land. Comparing the EF of countries with their available land area implies that national consumption should remain within boundaries defined by national production opportunities, which represents a normative and arbitrary ex ante anti-trade bias. Relatively small or densely populated countries (in terms of available land area) need, for evident reasons, to trade a large part of their national income. Spatial scales indeed correlate strongly with the proportion of trade in consumption. For illustration: cities trade 100 per cent of their consumption, and the world as a whole is autarchic. Use of the EF thus seems to suggest that we should get rid of cities, but this neglects agglomeration effects and comparative advantages.
An adequate approach to assess spatial sustainability and sustainable trade should not start from any biases but instead allow the question to be addressed of whether concentration of people in space is desirable from a global sustainability perspective. Positive externalities of concentration (for example, agglomeration effects) and of trade (comparative advantages) should be taken into account and traded off against negative environmen-
tal externalities (Grazi et al. 2007). In addition, the various negative impacts of trade in social and political dimensions, such as weakening com-
munity structures and preventing individual human perception of ecologi- cal impacts of consumptive decisions, should be taken into account.On the other hand, attention needs to be given to the negative consequences of reducing international trade, such as destabilizing international agree-
ments, trade wars and less diffusion of knowledge and technology.
5. Measurement and models
Many studies have developed indicators for sustainable development. As a result, di?erent approaches are available. These can be classi?ed as follows: ? Ecological (for example, biodiversity) versus physical (material or energy) indicators.
_Stock (capital) versus flow indicators.
_Source versus effect indicators.
_ Monetary versus other indicators.
_ Sustainability versus progress indicators (green and sustainable GDP measures, ISEW, GPI).
Indicators suffer from two main problems. First, often they aggregate information in a way that does not give rise to useful indicators from either a social welfare or environmental sustainability perspective (Ebert and
Welsch, 2004). Secondly, they often represent a supply side perspective, suggesting value theories much in the spirit of the Marxian labour value theory.
EE has produced several of these, such as energy indicators (energy value theory), ecological footprints (land value theory), and MIPS (material value theory). Economists are critical of such theories, as since Marshall it is widely agreed that values represent relative scarcity, which is the result of an interaction of demand and supply. This is not to say that one market dimension cannot sometimes dominate. For example, basic needs may become unsatisfied once absolute supply limits have been reached.
Models of sustainable development come in various types. Simple models from population biology (ecology) have been incorporated in eco-
nomic models of renewable resources, which perhaps can be seen as the most simple approaches to the sustainability problem. Specific models have been developed for the analysis of fisheries, forestry and water manage-
ment. EE has tried to move beyond such models by including advanced insights from ecology (see Folke, 1999). Resulting studies deal with one or more of four levels: biological populations (multispecies), ecosystems, bio- physical processes (for example, hydrology, climate change), and coevolu-
tion of economic and environmental systems.
A particular model of interest here is the ‘four box model’ for terrestrial ecosystems as proposed by Holling (1986). It depicts ecosystems and their changes in a two-dimensional diagram with the axes ‘stored capital’ (biomass) and ‘connectedness’ (complexity of the food web). Ecosystems can repeatedly move through four phases: ‘exploitation’, ‘conservation’, ‘release’ and ‘reorganization’. The ‘release’ phase can be initiated by forest fires, storms and outbreak of diseases. Such dynamics of ecosystems have given rise to questions about their stability and resilience. In the above- mentioned ‘four-box model’, management aimed at artificially prolonging a certain phase, notably ‘conservation’, can in fact reduce the resilience of the system. For example, checking small forest ?res, which leave seeds intact, tends to result in an accumulation of forest biomass. This in turn will increase the probability of the occurrence of a large forest fire, going along with very high temperatures, which can destroy plant seeds and thus prevent the ‘reorganization’ phase from occurring successfully.
A range of other economic–ecological models exists, focusing on ecosys- tem management and integrated systems ranging from regions to the globe
(Costanza et al., 1993; Rotmans and de Vries, 1997; van den Bergh et al.,
2004). Integrated ecological–economic modelling has been practised since
at least the early 1970s. One can be modestly optimistic about the feasibil-
ity of formal linking of economic and ecological models, but it requires significant financial and human resource investments. Such investments have been undertaken in some areas of application, notably in the area of climate change and policy, but less so in the area of ecosystem management modelling.
Costanza et al. (1997, p. xxii) state that the integration of economics and ecology is hampered by the lack of space in economic theories and models.
Although it is true that mainstream economics has largely assumed away space and spatial externalities between economic agents, the statement neglects the large area of spatial economics. This covers regional, urban and transport economics as well as spatial informatics – mainly the application of geographical information systems (GIS). GIS applications are nowadays often considered an essential input to integrated spatial models, because they allow the capturing of interactions between economic and ecological phenomena at a detailed spatial scale. It is not beforehand clear, however, that using a high spatial resolution will always be fruitful. Whereas many ecological and hydrological processes are amenable to a grid-based descrip-
tion, most economic processes operate at higher scales. This explains, for instance, why a method like ‘cellular automata’ has been more popular in landscape ecology than in spatial economics (Engelen et al., 1995).
Simultaneous changes in the economy and the environment are sometimes referred to as coevolution. Strictly, this notion means that variation in either subsystem depends on the other subsystem (Norgaard, 1984; Faber and
Proops, 1990; van den Bergh, 2004a). Coevolution thus re?ects mutual selec-
tion of economic and environmental systems that creates a unique historical development. In this sense EE is close in spirit to evolutionary economics, which is characterized by concepts like diversity, selection, innovation, path dependence, and lock-in (Mulder and van den Bergh, 2001). The evolution-
ary perspective suggests that systems are adaptive and coincidental rather than optimal. Some of these notions can and have been translated into evolutionary, notably multi-agent models (van den Bergh, 2004b; Janssen,
2002). Such models depend on boundedly rational agents, which in fact can be seen as a response to the critique of EE on the rational-agent assumption that underlies much of traditional environmental economics.
Finally, within EE modelling of sustainable development attention is given to describing structural change. In this context ‘industrial ecology’ and ‘industrial metabolism’ are relevant areas of research (Graedel and Allenby, 2003; van den Bergh and Janssen, 2005). They combine environ- mental science, economics and analysis of technologies to realize a minimal environmental pressure caused by substance and material flows.
Important strategies studied include ‘dematerialization’, recycling and reuse, waste management and enhancing durability of products. This is what Herman Daly would associate with keeping constant or reducing resource throughput.
6. Sustainability policy
Can one distinguish between sustainability policies and other environmental policies? One view is that the former include all environmental regulation since this will affect the degree of (un)sustainability. Another view is that certain policies or instruments are speciffically focused on long term sustainability issues. A few examples are as follows. First, if it is recognized that a transition from the current unsustainable system to a sustainable one is prevented by the lock-in of certain technologies, notably fossil fuel- based, then un-locking policy is needed. Price corrections are clearly Insufficient as increasing returns to scale play a dominant role. Stimulating diversity, for example, through subsidies, support of niche markets and public R&D are important elements of un-locking policy (Unruh, 2002).
Second, policies for sustainable development can include theoretical insights such as investment rules that stimulate constant total capital (Hartwick, 1977) and intergenerational transfers to compensate for environmental changes (Howarth and Norgaard, 1995). Both fit the weak sustainability approach, as substitution of natural capital is allowed for. Costanza (1994) in addition mentions three instruments. First, a natural capital depreciation tax would stimulate consumption in a more sustain- able direction. The result would be a shift from use of (and investment in) non-renewable to renewable resources. Second, a ‘precautionary polluter pays principle’ could stimulate caution in making decisions with much uncertainty about the occurrence and size of environmental damage.
Third, a system of ecological tariffs as countervailing duties would allow countries or trading blocs to apply strict policies (including the previous suggestions) so as to make sure that producers would not be stimulated to move overseas. The result would be that ecological costs would be reflected in prices of both domestically produced and imported products.
A number of instruments have been proposed to address the uncertainty and complexity surrounding ecosystems and sustainability. The notion of ‘safe minimum standards’ (Ciriacy Wantrup 1952) points to the fact that efficiency means exploring the borders, whereas in many circumstances – characterized by a large degree of uncertainty – it would be better to take account of safety margins (see Chapter 6). A flexible instrument to do this is an ‘environmental bond’ (Perrings, 1989; Costanza and Perrings, 1990).
An investment or project that is surrounded by much uncertainty concern-
ing environmental consequences is complemented by an insurance bond with a value equal to the maximum expected environmental damage. This bond functions as a deposit that is completely or partly refunded (with interest) depending on the amount of environmental damage that has resulted from the respective investment project. If environmental damages are nil the entire deposit is returned; in cases of actual or threatening negative environmental effects the deposit serves to compensate or prevent damage. This instrument can, among others, be applied to land reclamation, investment in infrastructure, transport and treatment of hazardous (toxic, nuclear) substances, and location of agriculture and industrial activities near sensitive nature areas. As a consequence of environmental bonds, the (expected) private costs of such activities will increase, causing investors to decide more conservatively, and so take account of environmental risks associated with their activities and investment projects.
Economists traditionally analyse uncertainty by defining ‘states of the world’ with associated probabilities, and maximizing an expected benefit function. Fundamental or complete uncertainty, that is surprises, implies a different approach, namely ‘adaptive management’ (see Chapter 2). This is based on the idea that management of complex and uncontrollable systems requires an interaction between experimental research, monitoring, learn-
ing processes and policy choices, with the objective to learn from disturb-
ances. This recipe has been applied to problems of fisheries, agriculture (ecological alternatives for pesticides) and forestry. Adaptive management also covers an interaction between various disciplines, experts and ‘stake- holders’ (Holling, 1978;Walters, 1986; and Gunderson et al., 1995). Similar advice follows from an evolutionary perspective (Rammel and van den Bergh, 2003).
A number of studies in the field of EE have examined the environmental policy implications of alternative theories of economic behaviour, which stress bounded rationality of economic agents, both consumers and producers (van den Bergh et al., 2000; Brekke and Howarth, 2002). Alternative theories or elements thereof include ‘satisficing’, lexicographic preferences, relative welfare, habits and routines, imitation, reciprocity, myopia, changing and endogenous preferences, and various models of behaviour under uncertainty. Some insights relevant to sustainability policy are as follows.
First, policies aimed at changing consumer preferences make sense when sovereign preferences are inconsistent with long-run goals of sustainability (Norton et al., 1998). Second, a ‘hierarchy of needs’ perspective relates to the notion of strong sustainability in that it emphasizes uniqueness and non-substitutability of goods and services provided by nature (Stern, 1997; Blamey and Common, 1999). It suggests that individuals may be unwilling to make a trade-off between economic and environmental goods or ser- vices. Finally, policy under uncertainty should reckon with strategies like imitation and pursuit of wealth, and aim at increasing or maintaining diversity of knowledge, technology and behaviour (Roe, 1996).
7. Conclusions and future research
This chapter has covered a broad spectrum of issues related to sustainability and sustainable development. Ecological economics offers a distinctive approach to sustainability, which includes much attention for ecosystem resilience. The opposition between weak and strong sustainability is some- what artificial, as the realistic or inevitable approach lies somewhere in between. Ecological economists nevertheless often tend to move in the direction of strong sustainability. Whereas global sustainability and sustainable development have received an enormous amount of attention, spatial sustainability and sustainable trade are grossly neglected issues. The large and growing amount of literature on international trade and environment adopts essentially a static perspective. The analysis of spatial sustainability requires an integration of insights and approaches from growth theory, international trade theory, resource economics and ecology. No one has yet succeeded in doing this and it seems likely that analytical approaches will fall short. In the area of sustainability policy various concrete suggestions offered by ecological economics were discussed. More theoretical and empirical research seems needed into which sustainability policies match the various types of bounded rationality that characterize the behaviour of economic agents.

5 Ecological and social resilience W. Neil Adger
1. Introduction
The world needs to be resilient to change. Sustaining life, sustaining well- being and sustaining the environment into the future increasingly means adapting to new circumstances and potentially unpredictable perturbations and challenges. New technologies for example have unforeseen consequences while demographic and cultural changes bring about new challenges for sustainable living. Setting single goals and universal prescriptions for sustainable development across the world seems increasingly unrealistic and potentially counter-productive. In these circumstances, a new emphasis on building resilience, and recognition of the linkages between elements of society and the ecosystems on which they depend, seems a sensible contribution to sustainable development. But understanding what the resilience of a social–ecological system might be, and the identification of the mechanisms which link the wider environment with human well-being, are far from trivial.
Resilience is a property of a system. In ecological sciences, resilience relates to the properties of ecosystems at different scales, rather than populations. There has been a significant evolution of the concept of resilience in ecology over the past decade in terms of its measurement and in terms of understanding how resilience interacts with other system properties such as diversity and stability. It has been demonstrated empirically that resilience is an essential factor underlying the sustainability of natural resources and ecosystem services (Gunderson and Holling, 2002). Resilience therefore is defined in relation to changes in ecosystems which are in turn related to human use and pressure on the natural world. To link resilience with sustainable development, it is therefore necessary to define the resilience of the actual interaction between humans and nature: the resilience of social-ecological systems is a central objective of sustainability. A social-ecological system in this context is, for example, a natural resource and its resource users. Examples of social-ecological systems are a fishery, a managed forest ecosystem, and the interaction of the carbon economy with global atmospheric sinks and climate (Gunderson and Pritchard, 2002).
Social elements of these coupled systems include the well-being and the governance of access and regulation to the resources in question. The resilience of a social-ecological system is made up of a number of elements: the amount of disturbance a system can absorb and still retain the same characteristics and controls on function and structure; the degree to which a system is capable of self-organization; and the ability to build and increase the capacity for learning and adaptation (Carpenter et al., 2001; Berkes et al., 2003).
The ultimate goal of sustainable development is to promote use of the environment and resources to meet the needs of present society without compromising the future. What then does knowledge of resilience con tribute to meeting such goals? First, resilient social-ecological systems have within them the ability to absorb shocks and hence maintain ecosystems and governance structures maintaining options for future users. Resilient systems can, in other words, cope, adapt or reorganize without sacrificing the provision of ecosystem services. Second, a loss of resilience in social- ecological systems is often associated with irreversible change, the creation of vulnerabilities for marginalized elements of society, and the reduction of flows of ecosystem services. Even actions and strategies which are apparently rational in the short run can reduce resilience. Hence building resilience is compatible with sustainable development and indeed provides a superior framework for analysing sustainability in the context of irreversibility, surprise and non-marginal change. The chapter outlines examples of where management of resources for resilience brings about benefits for sustainability, including adapting to climate change and managing the consequences of disasters. It proceeds by examining how resilience is currently understood across the natural and social sciences, explains elements of social resilience, and discusses hypotheses concerning how they interact with ecological resilience thereby explaining how resilience is a component of sustainable development.
2. Ecological resilience
The resilience of an ecological system relates to the functioning of the system, rather than the stability of its component populations, or even the ability to maintain a steady ecological state. Ecosystems have diverse properties which ecologists have sought to measure – these form the basis of normative statements about sustainability and sustainable utilization of ecosystems (Holling and Mefie, 1996). Many tropical terrestrial ecosystems, for example, have stable and diverse populations but are relatively low in resilience. Similar ecosystems in temperate regions with apparently low diversity can exhibit greater resilience.
Different ecosystem types, from terrestrial and marine environments, display a number of common features (following Holling et al., 1995; Gunderson, 2000; Gunderson and Holling, 2002). First, change in most ecosystems is not gradual but rather is triggered by external perturbations, and is episodic. Second, spatial attributes in ecosystems are not uniform but are skewed in their distribution and patchy at different scales ‘from the leaf, to the landscape, to the planet’ (Holling et al., 1995, p. 49), with the implication that scaling up of management solutions cannot simply be aggregated across scales. What works for a single location will not work for a whole eco-region. Finally, ecosystems often have more than one equilibrium: the functions which control ecosystems promote stability, but other destabilizing influences, such as physiological reaction to pathogens create diversity and resilience. These attributes lead to a range of implications for understanding resilience and for management.
From declining fish stocks in the Pacific, through to land use change in the Sahel, ecosystems have been shown to be subject to periodic shifts into states which are often less desirable for, but are often triggered by, human use (Scheffer et al., 2001). Figure 5.1 documents examples of shifts in human used ecosystems from one stable state to another across a number of ecosystem types. These shifts are often triggered by single events such as a tropical storm impacting on coral reefs or through fires and their impact on forest ecosystems. Sometimes they are caused by longer-term events such as the removal of one predator from an ecological system.
In Figure 5.1, the initial state is in column 1 and shows that, in relation to the two major state variables for each ecosystem (x and y axis), there may be more than one equilibrium position. For the ecosystems highlighted, from coral reefs to lake ecosystems, human action has reduced the capacity of ecosystems to cope with perturbations. The causes may be the over- exploitation of an important species (for example over-grazing of grasses, over-harvesting of fishes) or chronic stress such as pollution and nutrient loading. Over time the probability increases that the ecosystem will flip into the states represented in column 4 of Figure 5.1, which tend to be simplified, ‘weedy’ ecosystems characterized by lower levels of ecosystem services (Folke et al., 2004). The undesirable states in column 4, such as algae- dominated reefs, also tend to be difficult to reverse, because they tend to be caused by changes in so-called ‘slow’ variables such as land use, nutrient stocks and reduction in long-lived organisms (Folke et al., 2004).
Within the ecological sciences there is a continued focus on the relation- ship between diversity (the common focus of conservation practice) and resilience. The links between diversity of species and the stability of ecosystems now appear to be more widely accepted (Folke et al., 2004). An emerging new area is that of the diversity of response within ecosystems to external perturbations – this is the observation that different species providing the same function within ecosystems have different mechanisms for retaining the resilience of the system (Elmqvist et al., 2003). This raises the possibility that response diversity increases the likelihood for renew and reorganization to the desired states in column 1 in Figure 5.1.Respons diversity is an inherent characteristic of ecological populations, how’ve and cannot easily be managed by human action.
The case of coral reefs provides a good example of the nature of resilienc of ecosystems and interactions with human use. Periodic natural disturbance has been shown to be an important element promoting the divers it and resilience of coral reef ecosystems (Nyström et al., 2000). But coral resilience is reduced through chronic stress as a result of human activities on land: for example through agricultural pollution or poorly treate sewage, and through over-fishing (Jackson et al., 2001). Observation throughout the tropics, and particularly in the Caribbean, demonstrate that many sites only have half the live coral cover of three decades previously (for example Gardner et al., 2003). Resilience is being reduced through inappropriate fisheries management, as well as through indirect mechanisms such as land development or clearance, as well as through natural events such as hurricane damage or freshwater sediment inputs.
Nyström and colleagues (2000) outline the ecological pathways of these changes highlighted in Figure 5.1. Coral reefs once dominated by hard corals, attractive to reef fishes and as nurseries for many commercial species as well as for tourism, have changed state in a number of locations in the Caribbean to systems dominated by fleshy algae. The triggers for these changes are often natural, but the chronic stresses are human. Over-fishing of key reef species and nutrient loading into coastal areas from agriculture and sewage present one set of stresses – algae can multiply and smother coral growth. The coral reefs of the Caribbean in some cases persisted since the role of fish species in keeping algae at bay was taken over by sea urchins.
But ultimately the chronic stress on coral reefs resulted in a change in state when 99 per cent of sea urchins in particular locations were wiped out by a novel pathogen. These phase shifts in coral reefs have been observed in other areas, for example as a result of persistent or high El Niño events which increase sea surface temperatures beyond the thermal stress limits of corals. In all instances of phase shifts, ecological theories are not good predictors of whether systems will return to previous states (Hughes et al., 2005).
Phase shifts and stresses to environmental systems are also apparent in the arena of climate change (human-induced as well as natural). The present global ‘experiment’, of perturbing the world’s climate system by increasing global concentrations of carbon dioxide and other greenhouse gases, could bring about many unknowable and irreversible phase shifts in ecological, physical and ultimately human systems. Such phase shifts and threshold effects in climate change are increasingly referred to as abrupt or rapid climate change. Examples include significant warming (that is more than 6ºC) of the earth’s atmosphere because of positive feedbacks in the carbon cycle; melting of the West Antarctic Ice Sheet leading to 5–7 metres of sea level rise; or collapse of the thermohaline circulation of the Atlantic
Ocean (Alley et al., 2003). But, as Hulme (2003) points out, these possible abrupt changes in climate are different in their characteristics – they may be abrupt in the sense of being an unexpected change in the direction of a trend, abrupt because of the rate of change, or abrupt because some threshold has been exceeded.
There are, of course, precedents for localized abrupt climatic changes in human history (Diamond, 2004).Hulme (2003) argues that the Sahelian dry period from the 1960s to the 1980s, when precipitation fell by 30 per cent in most areas, represented a directional change from the previous decades which were steadily wetter. Clearly the anticipated phase shifts in climate are difficult for societies to adapt to and represent a major perturbation to social-ecological resilience. This is particularly so when social resilience is dependent on decisions that lock the technologies and societies into inflexible patterns of resource use. If decisions on building irrigation schemes and dams are based on the mean river flows from a wet period, as was the case in East African river systems (Conway, 2005), this leads to loss of resilience when a phase shift occurs.
In summary, the resilience of ecological (and physical) systems is increasingly understood to be reliant on mechanisms associated with diversity and with slowly changing environmental variables. Resilience promotes both the production of socially useful ecosystem services and provides a stable environment for human use of these services. Loss of resilience is, from a human perspective, undesirable.
3. Social elements of resilience
A key component of the emergent resilience analysis in ecology is the recognition that ecosystems do not exist in isolation from the human world.
The stability and resilience, as well as the value and cultural significance, of most of the world’s ecosystems are therefore intimately bound up in their human use. As the examples of environmental change above show, human use of natural systems reduces resilience at many scales. But from the traditional societies of hunters and gatherers, to the subsistence and commercial use of the world’s farmlands, human use has the potential to be both sustainable and resilient. This section examines the economic arguments for resilience and the determinants of social dimensions of resilience.
But many processes of economic development are not sustainable or resilient, including the reliance on fossil fuels and the fetishism of consumption. Economic growth, involving unsustainable resource use or use of the environment causing chronic stress on ecosystems, creates vulnerabilities and makes society more sensitive to shocks. In economic terms, ecological resilience itself is therefore important for human well-being for three reasons (Arrow et al., 1995). First, as outlined above, discontinuous change in ecosystem functions is associated with a loss of productivity and of ecosystem services. Second, the irreversible (or reversible only at significant resource costs: Mäler, 2000) impacts of a loss of resilience affect the portfolio of options for future use. Hence losing resilience reduces positive option values attached to the environment. Third, Arrow et al. (1995) argue that loss of resilience and more to unfamiliar states (column 4 in Figure 5.1) increases the uncertainties associated with environmental interactions.
In other words, dealing with unfamiliar and undesirable states has added costs, and hence entails a loss of welfare.
These economic reasons for preserving ecological resilience are, however, only part of the story. Sustainable development brings a normative domain to the relationship between ecological resilience and society. Sustainable development necessarily relates to human values: what is desirable, what is undesirable, and for whom. Thus the stable ecological states in column 4 in
Figure 5.1 may be ecologically poor and unproductive from a human-use perspective and hence unsustainable (see Norton, 1995). As Levin et al. (1998) point out, ‘resilience makes no distinctions, preserving ecologically or socially undesirable situations as well as desirable ones’ (p. 225). A social-ecological resilience compatible with sustainability needs to consider societal demands for ecosystem services, equity, vulnerability in the distribution of resources, and the governance of resources. Resilience in social-ecological systems includes the ability for positive adaptation despite adversity and hence involves human agency. The social elements of resilience are therefore bound up with the ability of groups or communities to adapt in the face of external social, political or environmental stresses and disturbances (Adger, 2000) and highlight the necessity of collective action. If formal and informal institutions themselves are resilient, they can promote wider resilience. Institutions (including modes of socialized behaviour as well as more formal structures of governance or law) can be persistent, sustainable and resilient depending on a range of parameters. The persistence of institutions of governance depends, for example, on legitimacy and on selecting environmental risks which resonate with the institutions’ agenda. Thus the resilience of institutions is based on their historical evolution and their inclusivity or exclusivity and how effective they are in ‘oiling the wheels’ of society. Resilient communities are promoted through integrating features of social organization such as trust, norms and networks. The cultural context of institutional adaptation, and indeed the differing conceptions of human environment interactions within different knowledge systems, is central to the resilience of institutions. These cultural contexts and local technical knowledge tend to be overlooked in considering equity and economic efficiency aspects of the sustainable use of natural resources (Gadgil et al., 2003). Hence the resilience of communities is not simply a matter of the economic relations between them, but is determined, as with social capital, by their inclusivity and degree of trust. The nature of social resilience can be inferred from perturbations and coping with change. Adger (2000) hypothesizes that social resilience is a function of resource dependency. The more resource-dependent a society, the more tightly coupled it is to the ecosystem functions and services on which it depends. Fishing communities depend on the abundance and migration patterns of fish stocks, as well as the integrity of habitats, the regularity of ocean currents, and the competition for fish from other fishing communities as well as natural predators. Hence fishing communities are resource-dependent. But they can maintain and build resilience through promoting diversity in livelihoods or even migrating with fish stocks (Adger et al., 2002). Resource dependency is the reliance on a narrow range of resources leading to social and economic stresses within economic and eco- logical systems. So, for example, the dependence of economies on mineral or renewable resources depends on how much of the economy is reliant on their mineral production; how volatile the world markets are in these commodities; and how much boom and bust there is in these commodities. Auty (1998 and Chapter 13) argues that resource endowments of minerals and high dependency ratios partly explain trajectories of development and the ultimate destiny of resource-dependent societies. The preoccupation with capturing the benefits of resource endowments during boom times in oil-rich or forest-rich countries impedes the creation of economic linkages, land reform and diversification of the economies (see discussion in Vincent, 1992; Neumayer, 2005). Dependency, whether on sub-soil or on living resources, brings its own set of problems and does not necessarily promote resilience.
The direct dependence of communities on ecosystems is an influence on their social resilience and ability to cope with shocks, particularly for food security and coping with hazards. Resilience can be undermined by high variability and exploitative relationships in the market system or natural or induced disturbance in the environmental system. Resilience therefore depends on the diversity of the ecosystem as well as the institutional rules that govern social-ecological systems.
4. Sustainability, resilience and adaptive management
Can resilience be enhanced to promote sustainable development? Action to promote resilience implies management based on the recognition of the dynamics and patchiness outlined above, and on the recognition of values and dynamics of institutions that create and constrain human use.
Promoting resilience is therefore directly dependent on the recognition of community engagement in resource management – particularly in areas where communities rely on ecosystem health for their own well-being or livelihoods – as a means of preserving ecosystem integrity. It is also dependent on the recognition of different worldviews and knowledge systems that can, without reference to standard science, formulate successful knowledge of functions of the environment and successful institutions to manage these functions (Berkes, 1999). Integrated conservation and development approaches that include collaborative resource management would appear to be central to reducing vulnerability and increasing resilience to improve the well-being of those societies and ecosystems dependent on natural resources. In many situations, where full knowledge about a system does not exist and optimum productivity is not an obtainable goal, an iterative management process that is informed and evolves through an ongoing learning process is about the best that can be achieved. Adaptive management (see also Chapter 2) not only pursues the goal of greater ecological stability, but also that of more flexible institutions for resource management (Olsson et al., 2004).
Promoting resilience requires flexibility and adaptation in decision- making on resource use and conservation. Hence it is argued that adaptive management of resources can improve the resilience of people and the environment and reduce vulnerability (Olsson et al., 2004). Under such an approach, an evolving management process for social as well as ecological systems is developed through iterative and learning processes. So can adaptive management ensure the resilience of social systems over time in the face of external stresses and perturbations? Clearly individuals and communities have been adapting to change throughout history. Societies have coped with climate variability through adopting new technologies, adapting their locations or moving their settlements (Diamond, 2004).
Not all adaptations are sustainable and there is recent historical evidence that large-scale, systematic changes in regional climate have had profoundly negative consequences for many societies in the past. But collective response and institutional resilience remain the dominant factor in sustaining adaptation. When faced with contemporary climatic perturbations in the Canadian Arctic, the Inuvialuit people of Sachs Harbour have been making short-term adjustments to their resource management (Berkes and Jolly, 2001). Their adaptations include switching hunted species and changing the timing and methods of hunting. Flexibility within cultural traditions and networks makes other forms of adaptation possible for this community, such as food-sharing networks and intercommunity trade.
Newly evolving co-management institutions are creating linkages across scales (local, regional, national and international) and hence transmitting local concerns to a wider audience and also being able to draw on the same wider community for assistance and advice. In a globalizing world, networks and learning opportunities cross traditional scales –engagement and exchange are both local and global processes at the sometime (Berkes, 2002). The autonomy that allows recognition of different forms of knowledge’s important. Olsson and Folke (2001) examine the local knowledge of ecosystem processes for a coastal crayfish fishery in Sweden and argue that the collective management of this resource involves institutions at diverse scales. They find that local-level institutions for direct management (harvesting strategies and seasonal patterns, for example) have been self-organizing, have created spaces for evolutionary reorganization, and give precedence to knowledgeable individuals. These institutional characteristics, they argue, provide evidence both of the importance of local knowledge at the ecosystems scale, and that evolution of institutions takes place through strategies of adaptive management as they move to higher and deeper levels of knowledge.
Adaptive management requires, at its core, retaining flexibility in the relationship between social resilience, changing property rights and institutional evolution. Coastal districts in Vietnam, for example, are impacted seasonally by landfall typhoons and coastal storms. Although fishing, farming and other activities have evolved to cope with this risk over the millennia, the radical redirection of the economy during the 1990s towards individual responsibility and private property and away from central planning diminished the resilience of many systems and resources, from upland forests to coastal communities reliant on aquaculture (Adger et al., 2001). Social-ecological resilience is important in the context of vulnerability to disasters. Changing resilience over time directly affects the ability to copewith perturbations, to recover and to adapt. Following the 2004 Asian tsunami, there is emerging evidence that those areas in South and South East Asia where ecosystems such as mangroves had previously been lostwere those that suffered the greatest impact. Importantly, traditional resource management institutions have played an important part in post-disaster recovery and rebuilding the resilience of communities (Adger et al., 2005). Coping with extreme weather events such as hurricanes also testssocial and ecological resilience. The Cayman Islands, for example, has implemented adaptation actions at national and community levels but suffered significant impacts from Hurricane Ivan in 2004. Tompkins (2005) found that social learning, a diversity of adaptations, and the promotion of strong local social cohesion and mechanisms for collective action have all enhanced resilience and continue to guide planning for future climate change. In Trinidad and Tobago, networks associated with present day coral reef management also play a key role in disaster preparedness and in building resilience (Tompkins and Adger, 2004).
There is growing evidence and experience of adaptive management building resilience, from traditional environmental management systems through to government-led collective action and experimentation with new institutional arrangements. A key lesson for adaptive management is that the nature of relationships between community members is critical, as is access to, and participation in, the wider decision-making process.
5. Conclusions
Resilience constitutes a radical critique of the traditional objectives of resource management. It is required because of the failure of institutions, ecological science, or economic policies to reverse the unsustainable management of resources or to reduce the large-scale environmental consequences of resource use. Resilience involves recognizing the dynamics of systems and functions that ecosystems play in protecting and facilitating human society and in promoting the robustness or resilience of ecological systems. But at the same time, flexibility and resilience are important characteristics of societies where environmental and societal risks permeate decision-making. The promotion of resilience of social-ecological systems is therefore a normative and ethical issue, not simply a descriptive theory of a natural state of the world. Global economic interests, property rights abuses, and asymmetric access to power and information combine to create conditions where environments become critical, and populations become vulnerable. As vulnerability is lowered and criticality reduced, so resilience increases. But in an ecological sense, resilience relates to the functioning of the system, rather than the stability of the component populations.
Resilience is the key to sustainability in the wider sense. Resilience, in both its social and ecological manifestations, is an important criterion for the sustainability of development and resource use, since all human welfare is ultimately dependent on the biosphere and its sometimes surprising nature.

6 Benefit–cost analysis and a safe minimum standard of conservation Alan Randall
1. Introduction
The Brundtland Commission definition of sustainability – meet(ing) the needs of the present without compromising the ability of future generations to meet their own needs (World Commission on Environment and Development, 1987) – would be satisfied by any arrangement that succeeds in maintaining welfare for the indefinite future. The goal of sustaining welfare can be met, in principle, by arrangements that allow great scope for substitution in production and consumption and rely, as time unfolds, on continuing technological progress and accumulation of capital to compensate for population growth and depletion of natural resources (Solow, 1974). Life may well be different in the future, just as life today is different from just a few generations ago, but it will be at least as satisfying. That is the promise of approaches that seek to sustain welfare – weak sustainability, the Hart wick rule, and green accounting (see Chapters 3, 17 and 18).
The idea that welfare is what should be sustained accords well with post-industrial-revolution human experience in the well-off countries. Our production systems and consumption bundles keep changing and the old ways of doing things disappear apace, but it all seems to be making us better-off.
Those concerned with sustainability could hardly take seriously a weaker form of sustainability. After all, weak sustainability places a lot of faith in technology, substitutability of capital for natural resources, and the ability of markets to transmit the right incentives. Many economists agree that sustaining welfare is the appropriate goal, but tend to assume that well-functioning markets will attain it automatically. That is, they agree with the weak sustainability goal, but question the need for explicit weak sustainability policies.
Among the environmental community and the public at large, more demanding commitments to sustainability have their dedicated promoters.
Strong sustainability – roughly, the commitment to compensate for depletion of exhaustible resources by augmenting economically-equivalent capital and/or renewable resources, and to limit the use of renewable resources to a sustainable level (‘cut a tree, plant a tree’) – offers an alternative to weak sustainability, one that assumes much less about the substitutability of capital for natural resources (see Chapter 4). There are also sustainability concepts that are less global, and more particular and local. The goal may be to sustain particular natural resources for reasons that are prudential (they might be essential for human welfare), or aesthetic (they are much appreciated for their contribution to human satisfaction, or perhaps for their own sake). Respect for, and attachment to, place may motivate local sustainability concepts that are related only remotely to worries about the world running out of something essential for human welfare.
Here, I do not propose to argue for or against any particular concept of sustainability. Instead, I simply assert that there is a certain commonsense appeal to the notion that sustaining welfare is a reasonable business-as-usual goal, but that attention to particular resources makes sense when there are plausible threats of resource crises. I argue below that people who find this a commonsense sort of approach will find much to like about a policy framework that, for business-as-usual resource allocation decisions, relies on markets supported by public actions that pass a benefit–cost filter, but invokes a safe minimum standard (SMS) of conservation principle for guidance when crises loom regarding particular natural resources.
In what follows, I summarize the moral arguments for attending to benefits and costs for business-as-usual decisions, and argue for explicit morally-justified constraints to deal with exceptional threats. The SMS is proposed as one such constraint to deal with threatened resource crises, and it is shown that this conception of the SMS has clear implications for SMS design, providing an internally consistent specification of the intolerable cost clause and endorsing early warning and implementation of SMS policies. Then, some
2key implications for doing benefit–cost analysis (BCA) and implementing the SMS are highlighted. Finally, I discuss ways of embedding the SMS in policy processes, and offer some concluding comments.
2. The search for ethical justifications
Benefits and costs are morally considerable we begin with a search for convincing reasons why the public decision process ought to be concerned with benefits and costs. One way to frame the question is: are there good reasons to believe that a benign and conscientious public decision-maker has a duty to consult an account of benefits and costs (Copp, 1985; Randall, 1999)? The traditional epistemological approach to ethics suggests that good reasons should be founded in a theory of right action, allowing us to conclude that benefits and costs are serious considerations in the search for right action.
When called upon to defend the systematic use of BCA in public decision processes, economists are likely to start talking about the need to impose a market-like efficiency on the activities of government (for example, Arrow et al., 1996). BCA can be defended as an instrument for accomplishing just that, but the fundamental question remains: why impose a market-like efficiency on the activities of government? We need convincing arguments why market efficiency is good in its own domain, and why it should be emulated in the government domain. As I argued in 1999 (Randall, 1999, pp. 251–2), the efficiency approach to right action is problematic, even if we concede the considerable instrumental virtues of efficiency. A more promising avenue (I believe) is to argue that BCA provides an acceptable account of preference satisfaction, and preference satisfaction matters ethically. In the extreme, consider welfarism: the goodness of an individual life is exactly the level of satisfaction of the individual’s preferences, and the goodness of a society is a matter only of the level of satisfaction of its members. From these premises, economists have developed, invoking various assumptions and restrictions as necessary and convenient, the whole apparatus of welfare change measurement, of which BCA is the direct practical implementation.
Welfarism is a particular kind of axiology, the theory that goodness is a matter of value (Vallentyne, 1987): particular in that it confines considerations of value to consequences alone, and considers only welfare when valuing consequences. And axiology is particular among moral theories, being just one of the foundational ethics in the western tradition. The others are Kantianism, which defines right action as that which is obedient to moral duties derived ultimately from a set of universal moral principles; and co Tractarianism, in which right action respects the rights of individuals. Both of these theories are deontological, because the justification of Kantian moral imperatives and of individual rights requires appeal ultimately to some asserted principle. It is now generally conceded (Williams, 1985) that the epistemological moral theories, axiological and deontological, all are wrong (or at least seriously incomplete) about some things that matter morally. By casting welfarism as a particular kind of axiology, we give it legitimacy as a moral theory, but at the cost of conceding that it too is wrong (or at least seriously incomplete) about some things.
Benefits and costs cannot count for everything Hubin (1994) asks us to consider benefit cost moral theory (BCMT): the theory that right action is what-ever maximizes the excess of benefits over costs, as economists understand the terms benefit and costs. Note that BCMT is founded in welfarism, but implemented according to rules of welfare measurement that weight individual preferences by endowments thus emulating the market, but introducing the morally-unsettling property that the preferences of the well-off count for more. It is hard to imagine a single supporter of such a moral theory, among philosophers or the public at large. Instead, we would find unanimity that such a moral theory is inadequate, and an enormous diversity of reasons as to exactly why.
Value pluralism Given the inadequacy of the epistemological moral theories, it seems unlikely that any one will defeat the others decisively (Williams, 1985). This existential value pluralism suggests that the task of the thoughtful moral agent in the policy arena is, then, to find principles that can command broad agreement and serve to guide society toward consensus on particular real world policy resolutions. Taylor (1989) point’s out that value pluralism is not just morally-inarticulate relativism; it is a search for principles that provide moral guidance for action. Benefits and costs must count for something the failure of BCMT is hardly an argument that BC considerations are morally irrelevant. Hubin offers the analogy of democratic moral theory: right action is whatever commands a plurality of the eligible votes. This too is a thoroughly unacceptable moral theory. Nevertheless, democratic institutions flourish in a wide variety of circumstances, and good reasons can be found for a society taking seriously the wishes of its citizens expressed through the ballot. So, the gross inadequacy of democratic moral theory serves to justify not abandoning democratic procedures but nesting them within a framework of constitutional restraints, and all of this embedded in a public life where moral and ethical issues are discussed openly and vigorously.
It turns out that one cannot imagine a plausible moral theory in which the level of satisfaction of individual preferences counts for nothing at all 5 (Hubin, 1994). Examining a broad array of contending moral theories, it turns out that preference satisfaction counts for something, in each of them.
Clearly, benefits and costs, among other concerns, are morally considerable.
Public roles for benefit and cost information to this point, we have concluded that a society of thoughtful moral agents would agree to take seriously an account of benefits and costs, within some more complete set of principles. At this point, the interesting questions are about what else, beyond preference satisfaction, might one want to consider, and in what manner might one want to take account of those things. One approach treats benefit and cost information as simply one kind of decision-relevant information.
Benefit–cost analysis to inform decisions, rather than to decide issues
Suppose that respect for benefits and a cost is one of a set of principles that together provide a framework for public decisions. The notion that benefits and costs cannot always be decisive in public policy, but should nevertheless play some role, is congenial to many economists (for example, Arrow et al.,
1996, p. 221). But there are at least two kinds of problems with this approach. First, it leaves unanswered the question of exactly what role. Are there particular situations and circumstances in which an account of preference satisfaction should be ignored entirely, and others in which it should be decisive? How should an account of preference satisfaction be weighted relative to other kinds of information? Can the answers to these questions be principled, or must they always be circumstantial? Second, it opens the door to ‘flexing’ BCA – if other considerations matter, that must be because
BCA gets it wrong in some systematic ways, so why not try to fox these6 problems. If the one true moral theory is ever-elusive, then it follows that the perfect decision criterion is impossible, which renders foolish the project of perfecting BCA.
A benefit–cost decision rule subject to constraints an alternative approach would be to endorse a benefit–cost decision rule for those issues where no overriding moral concerns are threatened. Benefits and costs could then be decisive within some broad domain, while that domain is itself bounded by constraints reflecting rights that ought to be respected and moral principles that ought to be taken seriously. This would implement the commonsense notion that preference satisfaction is perfectly fine so long as it doesn’t threaten any concerns that are more important.
The general form of such constraints might be: don’t do anything disgusting. The basic idea is that a pluralistic society would agree to be bound by a general-form constraint to eschew actions that violate obvious limit son decent public policy. This kind of constraint is in principle broad enough to take seriously the objections to unrestrained pursuit of preference satisfaction that might be made from a wide range of philosophical perspectives. Examples of such constraints might include: don’t violate the rights that other people and perhaps other entities might reasonably be believed to hold; be obedient to the duties that arise from universal moral principles, or that could reasonably be derived there from; don’t impose inordinate risks upon the future, in pursuit of immediate but modest benefit; and, don’t sacrifice important intrinsic values in the service of mere instrumental ends. In each of these cases, the domain within which pursuit of preference satisfaction is permitted would be bounded by non-utilitarian constraints; and these constraints themselves would be determined by serious moral agents in pluralistic processes.
A safe minimum standard of conservation is a commonsense precaution
The safe minimum standard of conservation was proposed by Ciriacy-Wantrup (1968) and defended by Bishop (1978) as a rational response to uncertainty about the workings of environmental systems. Given the intuitive plausibility of carelessly exploiting a resource beyond the limits of its resilience, society should pre-commit to preserving a sufficient stock of the renewable resource to ensure its survival.
Economists raised two kinds of objections to the SMS as a utilitarian response to uncertainty. First, in order to adopt an SMS constraint voluntarily, a rational utilitarian would need to have sharply discontinuous preferences. Second, Bishop’s (1978) attempt to show that a risk-averse utilitarian would rationally adopt an SMS constraint – formally, the SMS is the maxim in solution – failed. Writing with Ready, Bishop (Ready and
Bishop, 1991) conceded that game theory did not support his earlier attempt at a utilitarian justification of a discrete interruption of business-as-usual when the SMS constraint was reached. The quest for an internally
10 -consistent utilitarian justification of the SMS remains elusive.
Farmer and Randall (1998) take a very different approach. Rather than attempting to derive the SMS constraint from any particular epistemological moral theory, they argue from existential moral pluralism that the SMS is best framed as a decision heuristic adopted for good reason: a sharp break from business-as-usual that – given the fear of possible disastrous consequences from anthropogenic modification of environmental systems about which we know so little – could earn the allegiance of moral agents operating from a variety of principles. Three principled intuitions that we would expect to be honoured widely – the existence of future humans disvalued; the welfare of future humans is valued; and moral agents should resolve these intergenerational concerns in the context of their intergenerational obligations to each other – provide substantial justification for this kind of SMS.
3. Implications for implementation
I have offered justifications for adopting a policy framework that, for business-as-usual resource allocation decisions, relies on markets supported by public actions that pass a benefit–cost filter, but invokes a safe minimum standard of conservation principle for guidance when crises loom regarding particular natural resources. It follows that the practical implementation of these decision tools should serve effectively the purposes that justify them.
Implications for doing BCA
The reasons for agreeing to take benefits and costs seriously in the policy process are reasons why preference satisfaction matters morally. It follows that BCA should provide an acceptable account of preference satisfaction.
In the Appendix, a stylized BCA framework is provided that enables us to identify the essential characteristics of the benefit–cost criterion. The underlying value system is homocentric, instrumentalist and welfares. The environment is regarded as a resource, an instrument for serving human purposes. Humans do the valuing, and value at the household level derives exclusively from the satisfaction of human preferences. Value is aggregated across households according to the potential Pareto-improvement (PPI) criterion, which is consistent with Bent Amite utilitarianism. Since voluntary exchange and contract Arian political processes honour the actual Pareto-improvement (PI) criterion, the PPI can be interpreted, albeit with important caveats, in market and contract Arian terms.
Proposals are evaluated according to the ‘with and without’ principle, which requires that both baseline and with-project conditions be projected into the distant future. Benefits and costs are discounted to reflect the opportunity cost of capital, and expressed in present value terms. While the BCA model is presented in deterministic terms, uncertainty about future conditions can be recognized by expressing the valuations in ex ante expected value terms.
Hubin argues that BCA does in fact provide an acceptable account of preference satisfaction. Its main weakness in this respect, the endowment-weighting of preferences, stems directly from its reaching out to market institutions, efficiency logic, and contract Arian epistemological ethics; and it can be argued that these accommodations gain, as well as lose, legitimacy for BCA. If BC analysts wish to claim, based on the justifications provided here, that the public has a duty to take BCA seriously, and then the analyst’s them-selves have a duty to implement the PPI valuation framework rigorously and carefully. The result would be BCAs that depart from customary practice – to the extent that customary practice retains some remnants of BCA’s roots in financial feasibility analysis – in several ways. Less attention would-be paid to market prices and demands, while more attention would be paid to public preferences for public goods and the non-market values those preferences imply, and to willingness-to-sell as the appropriate measure of costs. We found, much earlier in this essay, that a claimed need to impose a market-like efficiency on the activities of government provides an implausible justification for taking benefits and costs seriously. Now, we find that a sounder justification for BCA entails an obligation on the part of the analyst to pay more than customary attention to preferences and less than customary attention to market outcomes.
Implications for implementing the SMS
Farmer and Randall (1998) argue that the SMS constraint makes most sense when cast transparently as a discrete interruption of business-as-usual, imposed to act upon firm, and often non-utilitarian, intuitions that to permit threatened destruction of a unique renewable resource would be foolish and (perhaps) morally wrong. The justification for this discrete switch has implications for the construction and implementation of the Farmer–Randall
(FR) SMS. For illustrative purposes, we assume a renewable natural resource
11 with a logistic regeneration function (Figure 6.1). With deterministic regeneration, S represents the minimum resource carried forward in order to min avoid resource exhaustion. The Ciriacy–Wantrup SMS addresses the stochastic nature of regeneration – it is safe in the sense that it carries forward a sustainable stock of the resource even in the worst-case regeneration scenario. The FR SMS – designed to respect the heuristics that moral agents value future humans and their welfare, but resolve these intergenerational concerns in the context of their intragenerational obligations to each other –is set at SMS*, which provides for an essential harvest, D . Min The essential harvest concept is most powerful in the case of an essential resource, where it has moral and practical implications for public choice.
Moral theories encounter serious difficulties in dealing with intergenerational problems, but one thing seems clear: no serious moral theory demands that a generation decimate itself for the benefit of future generations. The SMS, in the multigenerational context, can be effective only if each succeeding generation reaffirms the SMS commitment. Not only that, but each current generation in its turn would abide by the SMS only if it confidently expected succeeding generations to do the same – otherwise, in the end, little is gained by current sacrifice. Moral and practical reasoning lead to the same conclusion – in the case of an essential resource, the SMS must be set at SMS* to allow for essential harvest by each succeeding generation. The FR SMS emphasizes early warning and early implementation of conservation policies that require relatively modest sacrifices on the part of society. Since unilateral withdrawal from any intertemporal obligation is always a possibility, conservationists have a strong interest in keeping the costs of conservation tolerably low.
Many SMS proponents envision using an SMS to ensure preservation of unique and valued natural resources (often biotic), whether or not they are strictly essential to human welfare. For the case of an inessential renewable resource, practical reasoning reaffirms the logic of the essential resource case. Imagine that some minimal harvest or use of the resource enjoys strong political support (in the extreme, is politically essential). Then an
SMS* policy is recommended for practical reasons – again, conservationists have a strong interest in keeping the costs of conservation tolerably
12 low. Moral reasoning is murkier in this case, because moral theories differ as to what obligations humans may have toward unique and much- appreciated entities that are ultimately inessential to welfare.
Defining the intolerable cost The standard rendition of the SMS policy prescription contains an escape clause: the SMS should be maintained unless the costs of so doing are intolerably high (Bishop, 1978). At the outset, the ‘intolerable cost’ clause was tacked on to the SMS, ad hoc. More recent authors have offered quite different analyses aimed at bringing the intolerable cost inside the SMS framework. Rolfe (1995) proposes an SMS for risk-averse utilitarians, in which the limits of tolerable cost are defined
13 by willingness to pay for risk reduction. Randall and Farmer (1995) call on the concept of essential harvest, D, to define both SMS* and the min intolerable cost – any SMS obligation requiring that a generation forgo the essential harvest is ipso facto intolerable.
4. Embedding SMS in policy and management – what is needed?
It has become commonplace to characterize support for the SMS among environmental economists as wide but shallow. Yet Berrens (2001) argues that the SMS is attracting much more than cursory attention in the literature, in resource/environmental economics textbooks, in laws with rather sweeping application (the US Endangered Species Act, ESA), and in limited local policy applications. A very broad-brush review suggests that the ESA has evolved, via amendments and conventions adopted to guide, application, much along the lines of the SMS. To relieve ‘excessive’ economic burdens, land can be excluded from the designated critical habitat, or a species may be exempted from protection, provisions that parallel the intolerable cost escape clause in SMS. Nevertheless, the ESA fails to capture an essential feature of the FR SMS, the early-warning trigger designed to keep the costs of conservation tolerably low – and it might be argued that the much lamented ‘train wreck’ collisions of interests that make ESA so controversial are the inevitable result of this omission.
There is a modest amount of literature on local implementation of SMS procedures. Berrens reports, favorably and with only modest reservations, on several local applications of ESA. Woodward and Bishop (1997) argue that procedures drawing on the SMS and precautionary principle traditions make sense when policy makers face a wide divergence of beliefs among the experts they consult. Farmer (2001) reports a case where stake- holder convention processes were much improved by restructuring them around SMS concepts. Woodward and Bishop (2003) develop criteria for sustainability-constrained sector-level planning.
Implementation of a serious SMS-based policy requires that society monitor the landscape for indicators that warn of a particular risk of a resource crisis and, when the alarm is sounded, take seriously the call for avoidance/mitigation measures beyond those justified by ordinary welfare considerations. That much is agreed by most SMS proponents. But what comes next? The answer depends on what status we accord the SMS. It could be argued that the SMS, to be effective, must be codified into statute law (as happened, roughly, with ESA) or even constitutional law, or at least incorporated in administrative rules. An alternative view (Michael Farmer, personal communication) is captured in the idea of ‘principles that guide’. On-the-ground policy practitioners should be bound (by law or regulation) to certain broad-brush principles and encouraged to interpret these principles in practice via some kind of serious policy dialogue. This stands in contrast to formal technocratic planning procedures on the one hand, and abdication to stakeholder-consensus processes on the other. Concluding comments This chapter has elucidated the moral foundations of benefit–cost analysis and argued that it provides commonsense guidance for business-as-usual policy. While some economics textbooks argue that, in an ideal economy, resource crises are impossible, a mainstream economics literature has arisen that takes sustainability issues seriously indeed (Pezzey and Toman, 2002).
However, BCA (even the extended BCA that includes non-market and passive use values, and incorporates risk-aversion into the value estimation procedures) – by conflating uncertainty and gross ignorance of how natural systems work with ordinary risk – provides an unconvincing response to sustainability threats. The safe minimum standard of conservation was proposed by Ciriacy-Wantrup to address this perceived deficiency in business-as-usual economic thinking.
Some commentators have expressed concern that the SMS is fundamentally inconsistent: the SMS exception, as a break from business-as-usual cannot be justified by whatever justifies business-as-usual. But this insistence on internal consistency seems out of step with recent developments in philosophy. The search for the one epistemological moral theory that defeats all others seems hopeless, and much current thinking in ethics is aimed at finding robust principled ways to translate diverse moral sentiments among ethically inclined persons so that a rule deemed moral is at least possible.
Many economists have assumed unquestioningly that a credible SMS must be a utilitarian SMS. Thus, Rolfe proposes an SMS that is little more than extended BCA – at best a warning flag raised in information-poor situations to remind the analysts to bend over backwards to give uncertainty and non-use values their due. Others (Bishop, Ready and Bishop) invoke extreme risk-aversion in the quest for a utilitarian SMS.
The Farmer–Randall SMS proposed and defended here is a substantive SMS that calls for an explicit policy switch made for good reasons. It is motivated not just by uncertainty in the real world, but also by ambiguity concerning what we as a society care about, especially when the distant future is at issue. This substantive SMS is guided by principles adopted by serious moral agents in the absence of a complete and convincing epistemological moral theory. From this perspective, the economists’ impulse to retreat into more familiar moral territory (for example, front-loading a lot of risk aversion into a BCA) should be resisted – it simply does not take principles very seriously.
The BCA subject to SMS framework proposed and defended here would honor weak sustainability for business-as-usual circumstances, but reserve a strong sustainability instrument targeted to particular, credible threats of resource exhaustion. As such, it respects the modern experience of technical progress and increasing welfare even as substitution in production and consumption proceeds apace, and the reasonable instinct for caution as we continue to push at the frontiers of what can be known about our planet’s capacity to support future welfare.

Aucun commentaire: