Analyzing sustainability transitions as a shift between socio-metabolic regimes.

This essay seeks to specify the theoretical choices and assumptions involved in studying sociometabolic transitions, such as sustainability transitions, in a way that distinguishes them from mere "changes". These generalizations draw on experiences with the empirical analysis of historical transitions on various scale levels. This perspective is illustrated by using material and energy flow data to demonstrate global sociometabolic regime transitions during the 20th century.

Distinguishing transitions from other changes

A scientific treatment of sustainability transitions – or, for that matter, any transition as distinct from just “change” – requires a certain conceptual clarity and self-discipline: clarity in defining the unit of analysis, in how to distinguish different phases or stages, and in how to conceptualize the directionality of time.

What is an adequate unit of analysis? The focus ought to be on a theoretically and operationally identifiable system. The system should be self organizing and sufficiently complex to maintain itself under changing conditions. For such a system, there would be environmental boundary conditions: If they are transgressed, major features of the systems functioning will change. In the extreme case, the system may collapse (if it is an organism, die), or else it may resume its self organization in a new “state”. Both would then be called a transition. Can such a system be or become “sustainable”? Which criteria could be used to evaluate sustainability? In our opinion, the key answer is the following: the system is unsustainable if it behaves in ways that actively bring about those very boundary conditions in the system's environment. So the unit of analysis cannot just be the system itself, but the system in interaction with relevant other systems in its environment. A system as such cannot be judged as sustainable or unsustainable. In the case of social systems, as well as with organisms, the system is unsustainable if it triggers changes in the environment that threaten its basic metabolic requirements.

The second consideration, again very abstract, is the distinction of stages or phases. The typical model of alternating phases is the S-curve (Rotmans et al., 2001), although other models have also been considered, such as the so-called “lazy eight” (Berkes and Folke, 1998), lock-in situations or system collapse (Tainter, 1988), or “tipping points” in earth systems (Lenton et al., 2008). From the notion of transition, there follows an understanding that no linear, incremental path leads from one state or phase to the other, but rather a possibly chaotic and dynamic intermediate process, or a discrete “jump” from one state of the system into another. One has to be aware, though, that these distinctions are extremely sensitive to the observer's choice of scale. From a wider perspective something may appear as a continuous process, progressing steadily. But from a closer perspective the same process may appear as whimsical, sharply fluctuating. For example, the process of walking from a certain distance looks like a linear movement; from close distance, one sees muscles contracting and relaxing again, weight shifting from one leg to the other, so the process appears as cyclical. From a still closer and shorter perspective it would appear as transition. Thus descriptions of processes as transitions or as gradual change do not necessarily exclude each other. One type of process may well be nested into the other. Nevertheless, the idea of a system gradually behaving ever more sustainably (as suggested in theories of ecological modernization sometimes, see Mol and Spaargaren, 1998) does not comply with the term “sustainability transition”.

A third consideration relates to the order of phases or stages, in other words, the understanding of directionality of time. This directionality can either imply consecutive stages of a developmental type (like Herbert Spencer's notion of evolution, or Marxist historical materialism, or Rostow's stages of economic growth), or it may follow a Darwinian type of evolutionary theory by assuming the future to be contingent upon the past but an open process into the future: you know the mechanisms driving it but not where it will lead to. In the first case, when a developmental model is employed, each consecutive stage follows with a certain necessity from the previous stage, and it is, as a rule, considered superior, more mature. The progress to this more mature stage can be accelerated or delayed. In the second, “Darwinian” case, the direction of change is principally unknown (Gould, 2002). Many people believe earlier transitions (such as the industrial revolution) to have been of a developmental type, simply human progress. Can we think of sustainability transitions as a kind of inevitable, logical step beyond the past, leading to a more mature state of the system?

The socio-metabolic approach makes certain choices with regard to these distinctions. It says the appropriate unit of analysis is society, interpreted as a socio-metabolic system (Fischer-Kowalski and Weisz, 1999) that interacts with systems in the natural environment. It claims that a transition to a (more) sustainable state implies a major transformation, on a par with the great transformations in history such as the Neolithic or the Industrial Revolution (Haberl et al., 2011). Sieferle (2001) goes even as far as stating that industrial society as such is but a transitory stage from agrarian society to a very different as yet unknown type of social organization. And finally, there is the presumption that a sustainability transition is both inevitable and improbable. It is inevitable, because the present sociometabolic dynamics cannot continue for very long any more, and it is improbable because the changes need to depart from known historical dynamics rather than being a logical step from the past into a more mature future state.

Does it make sense to regard transitions as a shift between socio-metabolic regimes?

A regime, according to the socio-metabolic approach, is rooted in the energy system a society depends upon, that is the sources and dominant conversion technologies of energy.1 The theory of socio-metabolic regimes has been developed by Sieferle, 1982, Sieferle, 2001 and elaborated by Fischer-Kowalski and Haberl (2007). Depending on the reasons for and the speed of an energy transition, parts of the system may at a certain point in time be under different energy regimes: urban industrialized centers, for instance, may coexist with traditional agricultural communities, or industrialized countries with agrarian colonies. Such a “synchronicity of the asynchronic” (Füllsack, 2011) influences the overall course of transitions. How these processes evolve is contingent upon specific conditions. The socio-metabolic approach shares with complex systems theory the notion of emergence: neither can one state be deliberately transformed into the other, nor can the process be fully controlled. One is confronted with self-organizing dynamics (Maturana and Varela, 1975) to which orderly governance or steering cannot be applied. The sustainability transition with regard to energy needs to be a change away from fossil fuels, and probably back to solar energy again, thus somehow reversing the historical transition from the agrarian to the industrial society and ongoing contemporary “development” that was and is a shift from solar energy to large scale fossil fuel use.2

What drives socio-metabolic regime transitions? On such a broad and long term scale one cannot easily talk about actors and their deliberate efforts. What one can mainly analyze is structural change of interlinked social and natural systems, across a broad range of variables. Among these, the socio-metabolic approach focuses on a relatively narrow set describing the society-nature interface for which quantitative measurements can be reliably obtained in very different contexts. The advantage of this self-restraint is that it is possible to demonstrate the interconnectedness of socio-economic changes and changes in natural systems (between population growth, diets, land use and species extinction, for example) and to generate models for important biophysical requirements and boundary conditions for system perpetuation. When an energy regime changes, society and its metabolism alter, and also the natural systems it interacts with. A regime can be characterized by the socio-metabolic profile of the society involved, and the associated modifications in natural systems that occur either as an unintended consequence (such as resource exhaustion or pollution) or as intentional change induced by society (such as land cover change).3

Based upon the socio-metabolic approach, research has been undertaken into historical cases of transitions. Winiwarter (2003) and McNeill and Winiwarter (2004) have approached the issue of soils as the main resource base of agriculture, and made a modeling effort to determine when for example a historical agricultural village would have to be given up (Winiwarter and Sonnlechner, 2000). Krausmann et al. (2008) have shown the role of resource and land scarcity for European history on various scale levels. There was a modeling effort to determine the limits to city growth under agrarian conditions, given certain yields and the constraints to land transport, while considering metabolic needs for food, construction material and firewood, as well as varying rates of appropriation of agricultural surplus by cities (Fischer-Kowalski et al., 2004). These are no more than examples of first efforts to understand what happens when social systems challenge the boundary conditions of their environment and transgress their own coping capacity, and attempts at explaining under which conditions transitions (collapses sometimes) occur. Under the agrarian regime, this is easier to determine, as its resource base is much narrower and local constraints play a key role.4

Table 1 makes an attempt at using the insights gained from such historical studies. It spells out a set of hypotheses on a more general level, so that they might also be applicable to a sustainability transition.

Table 1

Framework conditions favourable or unfavourable to unleashing socio-metabolic transitions.

	New resources/opportunities
	Not perceivable	Perceivable, appear promising
Previous resources/opportunities
Still intact	Status quo maintained	Status quo defended + eventual expansion
Threatened or exhausted	System collapse	Transition dynamics triggered

It is claimed that what drives a transition is the structural exhaustion of opportunities, and at the same time the opening of new opportunities (see Fig. 1). If only previous opportunities are exhausted, and no substantial new opportunities open up, one may rather expect system collapse (Diamond, 2005). If previous opportunities are not exhausted when new resources/opportunities offer themselves, vested interests in the status quo will often be strong enough to prevent change. This case seems an inherently unstable situation, though: As long as the interest groups benefiting from the use of the “old” resources are very strong (such as the landed aristocracy at the beginning of coal utilization, or the oil industry at present5), they may delay the use of the new opportunities for a long time, maintaining the status quo.6 They also may give in gradually and allow for the additional utilization of the new resources that are connected to different interest groups. This may result in expansion (building one resource use upon the other), but in the longer run would give rise to a transition.

Fig. 1

Global rates of energy and materials use across the 20th century per capita and year. Energy use is measured as Domestic Energy Consumption (DEC) which equals Total Primary Energy Supply (TPES) as reported by IEA, plus nutritional energy of humans and livestock, in Gigajoules. Domestic Material Consumption (DMC) is measured in tonnes according to MFA methodology as standardized by Eurostat (2007). On the global level, as international trade equals out, DMC is equivalent to global annual (raw) material extraction. Materials comprise the main groups: biomass (for human and animal nutrition, combustion and other uses), fossil fuels, metal ores and industrial minerals, and minerals used in construction (such as sand, gravel and limestone).

Both dimensions in Fig. 1 involve objective criteria (such as changes at the interface of society and nature), and subjective elements of human perception and learning. New energy sources, superior in energy density and cost, have of course been the prototype of new opportunities, and such a grand new opportunity at present is not in sight. But probably also other new opportunities, such as sophisticated solar technologies in combination with low-energy IT, could play the same role.

The dynamics of global socio-metabolic change in the 20th century

The environmental historian McNeill (2000) summarized the 20th century using the ironical title “Something New Under the Sun”. According to the statistics he assembled, there is rarely any dimension of human social life and interference with the environment that has not undergone a rapid expansion worldwide. During this one century, many indicators of human activity and human use of the environment exceeded the fivefold growth of the human population, substantial in itself, sometimes by an order of magnitude. Taking McNeill's reconstruction of the 20th century seriously, humans are driving a biophysical explosion in limited space. This explosion, as is to be briefly demonstrated further down, derives from multiple transitions. The countries that had already undergone a transition from the agrarian to the industrial socio-metabolic regime during the 19th century made a transition from the coal-based to the oil-based mode. During the same period, many other countries started their transition from an agrarian regime to the coal based and oil based mode on a faster track. Some countries have not as yet really started this transition (Lankao et al., 2008).

Haberl et al. (2011) have interpreted, after analyzing a large number of country metabolic profiles, the S-curve in Fig. 1 as a global result of such transitions. Fig. 1 shows metabolic rates of global energy and materials use. Energy is measured as Domestic Energy Consumption (DEC)7 in Gigajoules per capita, and material use as Domestic Material Consumption (DMC) in tonnes per capita. The interpretation, in short, is as follows: The dynamics of global energy and materials use was, for most of the 20th century, marked by the highly developed industrial countries. The apparent global patterns may be distinguished into a coal-dominated phase of industrial development that raised overall energy and materials use but at the same time population—thus while there was huge growth in capital, per capita consumption stagnated. This phase started before 1900 and ended globally around 1930, a date also marked by a world economic crisis. The next phase, dominated by the use of petroleum, realized what many dreamed of: a sharp increase in consumption opportunities for each person, in other words the American Way of Life.8 This pattern dominated the global metabolic profile from around 1930 to the early 1970s, as can be gathered from the steep increase in per capita metabolic rates in Fig. 1 during that period.

In the early 1970s, a new pattern emerged: there was again a relative stagnation of metabolic rates, but this time in combination with low population growth. Global GDP/capita continued to rise. After the year 2000, in the industrial countries, per capita energy and material consumption continued to stagnate, but global metabolic rates again turned sharply upwards (Fig. 1).

Interpreting changing growth patterns of material and energy use as socio-metabolic transitions seems only justified if confirmed by a broader analysis. Such an analysis is ongoing and requires giving attention to demographic variables, technological variables, political variables and last but not least economic variables, on various scale levels. It extends the scope of this paper to go into any detail. Nevertheless, we found it surprising that such a simple quantification as global metabolic rates seems to reflect major other well-known changes.9 Demographically, the coal phase was linked to massive population growth in the industrializing countries; and one can register a substantial drop in population growth around 1930 (and again after 1970). Technologically, the breakthrough of coal use was associated with the steam engine and railroads, while the breakthrough of petroleum was linked to automobiles, electricity and the industrialization of agriculture, the “green revolution”. The phase after 1970 was marked by the rise of information and communication technologies. Politically and culturally, the transition to coal was connected to the rise of the British Empire, and the transition to petroleum to American hegemony that possibly started to decline around the time of the US oil peak in 1973 (Hubbert, 1971). From then on, there is a continuous rise of “emerging economies” mainly emulating the patterns of the previous industrial transformation. This rise is also responsible for the sharp turn upwards in metabolic rates after the year 2000. We found it very surprising how many events (such as global crises) and changes in structural variables fitted neatly to apparent turning points in a timeline of per capita energy and materials use. This encourages us to talk of transitions in this context.

At the time of writing, the economic crisis years 2008–2010 are not documented in the data yet. They surely caused a substantial downturn of the curves. Will the curves later resume their course upwards? According to our guess, they will not: a period of strong fluctuations will ensue, driven by a rising and volatile oil price (global oil peak?), supply shortages and a new trend of price increases for important raw materials.

From the perspective of a potential sustainability transition, the phase of stagnation of metabolic rates (but not of global income) between the early 1970s and the late 1990s is very interesting. Quite obviously, there had occurred structural change terminating the rapid increase in per capita material and energy consumption, in industrial countries and globally.10 This may have signaled a historical window of opportunity for physical degrowth in industrial countries, in alliance with the rise of new information and communication technologies. But politically, this window of opportunity was ignored, or even forcefully shut. Instead of utilizing the structural opportunity of heading for a less material and energy intensive lifestyle in the OECD world and more calmly accepting the emerging economies to catch up, the “limits to growth” message (Meadows et al., 1972) was discarded.11 Most political efforts were invested in returning to the higher growth rates of the previous period. This strategy was not fully successful in re-establishing the previous high rates of economic growth. It also missed out on the chances to stabilize employment and mitigate rebound effects by reductions in working time, to stabilize climate change by taxing fossil fuels, or to tackle social inequality. Policy measures were directed at maintaining business as usual, and despite major accidents (Three Mile Islands, Chernobyl, now topped by the meta-GAU in Fukushima), huge public investment continued to be spent for nuclear energy in denial of a possible lower energy future based on renewable sources.

Can the socio-metabolic approach provide conceptual and empirical guidance to sustainability transitions?

What distinguishes a socio-metabolic analysis conceptually and methodologically from other approaches to study historical transitions? Conceptually, it differs from storylines on technological change or from storylines on economic cycles by systematically bringing in empirical information about biophysical variables and attributing an important role to nature. In the socio-metabolic approach, nature matters in terms of providing easily accessible, high density energy from limited geological sources as a key ingredient of the “economic growth engine” (Ayres and Warr, 2009). Forces of nature matter as stochastic events in demonstrating the limitations of human technology, such as winds over Chernobyl distributing radioactive particles all across Europe, hurricanes over the Mexican Gulf destroying offshore drilling platforms, or sea level rise threatening the survival of island states and influencing climate negotiations. But more systematically, natural systems matter as they coevolve with human interventions and exert pressure upon societies to keep on changing.

Methodologically, this approach to transition analysis is but in an early stage. Much effort has been devoted to creating consistent long term databases with variables across many domains and across several scale levels, from local communities (or small pixels) to countries, regions and the global level. Statistical analysis has gone into identifying phase-specific interrelations between variables and trends. A few modeling exercises have helped to generate missing data and perform consistency checks. But adequate modeling techniques allowing the reconstruction and simulation of structural change are still pending (see www.matisse-project.net). It may be expected that the declining belief in 19th century type of progress continuing forever will, in combination with unexpected and possibly catastrophic events, trigger new approaches for analyzing and modeling major transformations of society-nature relations.