2nd Order of Sustainability

8/14/2019 2nd Order of Sustainability

1/19

ARTICLE IN PRESS

ANALYSIS

Second-order sustainabilityconditions for the development of

sustainable innovations in a dynamic environment

Christian Sartorius *

Fraunhofer Institute for Systems and Innovation Research, Breslauer Str. 48, D-76139 Karlsruhe, Germany

Received 5 September 2003; received in revised form 6 July 2005; accepted 6 July 2005

Abstract

In particular radical innovations can be important means to achieving improved sustainability. Due to the existence of path

dependency and lock-in, however, the transition from one technological trajectory to another, more sustainable one is often

impeded by significant barriers. Fortunately, these barriers are by their nature subject to substantial changes in time; so, it makes

sense to carefully distinguish between periods of stability (showing high barriers) in which the given trajectory can hardly be

left and periods of instability (characterized by low barriers) where a new trajectory can be reached more easily. The latter

distinction matters since sustainable innovations often rely on governmental regulation and the economic burden arising from

this regulation will be much lower in periods of instability. Moreover, due to the complexity and dynamics of change in theirrespective environments, innovations are generally associated with fundamental uncertainty such that it becomes impossible to

predict the degree of sustainability yielded by specific innovations in the longer run. Under these circumstances, it is essential to

facilitate the change between trajectories and to allow for the possibility to select between a variety of alternative trajectories

within a process of trial and error. Sustainability as viewed from this evolutionary perspective is therefore better understood as

the general capability to adapt, that is, to readily change from less to more sustainable technological trajectories. Since the latter

kind of sustainability determines the conditions under which the former kind (i.e. sustainability related to a specific technology)

can be achieved, the two kinds are respectively called second-order and first-order sustainability.

Finally, a series of determinants (and corresponding indicators) from the techno-economic, political, and socio-cultural

sphere is identified which, after proper measurement and weighting, allow for making an assessment whether and when the

incumbent industry is sufficiently destabilized and the political system rendered sufficiently favorable to the new, more

sustainable technology such that a transition to the preferred trajectory is possible without too much effort.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Sustainability; Innovation; Path dependence; Lock-in; Uncertainty; Indicators

1. Introduction

Innovations play a crucial role not only as the

basis of the persistent economic growth prevailing

0921-8009/$ - see front matterD 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.ecolecon.2005.07.010

* Tel.: +49 721 6809 118.

E-mail address: [email protected].

Ecological Economics xx (2005) xxxxxx

www.elsevier.com/locate/ecolecon

ECOLEC-02304; No of Pages 19

DTD 5


2/19

ARTICLE IN PRESS

especially in developed countries since the beginning

of the industrial revolution (see Schumpeter, 1934;

Nelson and Winter, 1982, for evolutionary; Romer,

1986, for neoclassical perspectives on innovation-driven growth); they are also an important, if not

the only, means for maintaining the sustainability of

this development, that is, for avoiding destruction of

the natural environment and exhaustion of natural

resources that may be needed by all our descendents

in order to maintain at least the current level of

wealth (see Rennings, 2000 for an overview). How-

ever, innovations towards sustainability are often

associated with substantial costs. From the point of

view of environmental economics this is due to the

fact that environmental innovations internalize exter-nal costs for which the innovator does not receive a

compensating benefit. By contrast, Porter and Van

der Linde (1995) claim that these costs can be sub-

stantially reduced, if, rather than merely redressing

the consequences of existing technologies (e.g. by

end-of-pipe solutions), innovation is understood as

an integrated process avoiding environmental extern-

alities right from the beginning. The remaining costs

can even be turned into a benefit, if, due to its more

fundamental character, an innovation avoids both

external and internal costs.

Despite the basic attractiveness of this kind of

innovation, employing them is far from rendering

the path towards sustainability self-sustaining for

two reasons. On the one hand, all environmentally

more benign substitutes after a while tend to give rise

to unforeseen environmentally hazardous side effects

such that, in the longer run, new technological

(including organizational) substitutes have to be gen-

erated again and again. Moreover, the development

and becoming effective of new substitutes takes time,

allowing related technology branches to exhibit envir-

onmental externalities in their turn. Due to the higheruncertainty associated with fundamental innovations,

they will show this tendency even more markedly

than incremental innovations succeeding within one

paradigm. As a consequence, sustainability will gen-

erally remain temporary and more or less incomplete.

On the other hand, fundamental technological

change requires the transition from one technology

paradigm to another and, therefore, is not only less

likely to occur and but also associated with higher

uncertainty and risk than innovation along a given

trajectory (Dosi, 1982, 1988). Accordingly, the fre-

quency of environmentally sound and economically

profitable fundamental innovations will remain low

unless they are supported by policy instruments spe-cifically referring to the causes of paradigm formation

and the related lock-ins. Klemmer et al. (1999) to

some extent point in this direction when they

acknowledge that a mix of regulative measures is

needed to properly account for the complexity of

circumstances in which innovation arise.

In this paper, both time and uncertainty will be

accounted for more thoroughly as crucial conditions

of technological development in general and espe-

cially with regard to sustainability. In particular, it is

assumed that, along with the change in circumstances,periods of stability of a given technological trajectory

(where establishing a new paradigm requires much

effort) alternate with periods of instability (where such

a shift is more easily achieved). It is further assumed

that it is possible to identify and even strategically use

the latter phases of instability in the search for the

lowest possible cost of achieving a higher degree of

sustainability. In order to justify this claim, Section 2

starts with a discussion of the relevance of innovation

in the context of sustainability from both the neoclas-

sical and ecological economics perspective. In Sec-

tion 3, an evolutionary framework is used to show

how potential progress towards greater sustainability

by means of innovations may be hampered by com-

plexity, uncertainty, path dependency and lock-in.

While identifying the strategic elements for overcom-

ing these shortcomings, Section 4 specifies the con-

ditions for the more ready identification and

implementation of sustainable innovationsa prop-

erty we call second-order sustainability because it

refers to the dynamic interrelation between innova-

tions rather than the innovations themselves. In order

to make use of this dynamic concept of sustainability,Section 5 identifies a variety of its potential determi-

nants and indicates how they may be applied. Finally,

conclusions are drawn in Section 6.

2. Innovation and sustainability

A very wide-spread understanding of innovation is

reflected in the definition used by the OECD (1997),

which distinguishes (1) process innovations allowing

C. Sartorius / Ecological Economics xx (2005) xxxxxx2


3/19

ARTICLE IN PRESS

to produce a given quantity of output (i.e. goods or

services) with less input, (2) product innovations

characterized by the improvement of existing, or the

development of new, goods or services and (3) orga-nizational innovations including new forms of man-

agement. While the exact meaning of even this

relatively simple definition of innovation depends on

the methodological and normative background

assumed by its respective applicator, it is even more

difficult to relate to each other the concepts of innova-

tion and sustainability. In this section, I will try to

elucidate this relationship first from the neoclassical

and then from the ecological economics perspective.

2.1. Environmental innovation in the neoclassicalcontext

While, in the above-quoted OECD definition, pro-

cess innovation is explicitly efficiency-oriented, the

meaning of improvement or novelty in the definition

of product innovations is not further specified. If, as is

often done in economic contexts (see e.g. Rennings,

2000), the notion of innovation is further qualified by

its distinction from the term invention, the explicit

reference to the marketability of an innovation implies

that also product innovation is considered in terms of

efficiency, providing more benefit (as revealed by the

customers willingness to pay) at the same cost or the

same benefit at the lower cost. So, in the (neoclassi-

cal) economic context, the complete monetary com-

mensurability of costs (for inputs) and prices (for

outputs) renders it fairly easy to identify innovations

in a given set of new processes or products.1 By

contrast, maintaining this commensurability is essen-

tially impossible, if the above concept of innovation is

to be extended beyond the realm of human prefer-

encesfor instance into ecological sustainability

(compare Munda, 1997).In particular since the beginning of the industrial

revolution growing human production and consump-

tion led to ever more frequent, more persistent and

more severe adverse impacts on the natural environ-

ment, representing an overall increase in economic

sustainability (i.e. persistent growth) at the expense

of ecological sustainability.2 Neoclassical environ-

mental economics tried to resolve this trade-off by

determining (shadow) prices for those uses of nature

giving rise to negative external effects and imposingthem by means of Pigou taxes and allowance trading

(Pigou, 1920; Coase, 1960). Although functional in

the neoclassical setting, both approaches often proved

to be ineffective in practice for two reasons. First, due

to asymmetries in the stakeholders endowment with

knowledge and other resources and the public good

character of most parts of the environment, it is

impossible to determine the true willingness to pay

(i.e. the price) for the services of nature with sufficient

accuracy. And, second, even if this price could be

determined sufficiently exactly, it may be doubtedwhether the two targets of satisfying the needs and

wants of humans and meeting the requirements for the

natural environment to sustain could be brought to

coherence. Reasons for this are, among other things,

the difference in time horizon between myopic

humans and long-term processes in nature and, most

of all, the basic ignorance of most individuals con-

cerning the wide variety of causeeffect relationships

in nature (I will deal with this point more extensively

in Section 3). From the neoclassical perspective, this

implies that the economy within or, respectively, with-

out its natural environment would not converge to the

same equilibrium and, therefore, possible disturbances

1 In this context, organizational innovations are not mentioned

explicitly because, with regard to their inputoutput relation, they

can be treated like either process or product innovations.

2 Since the Brundtland report (WCED, 1987) sustainability is

usually discussed as a state or, better, a development in which three

kinds of interests are met simultaneously: (1) the interest of the

present generation to generally improve their actual living condi-

tions (i.e. economic sustainability), (2) the search for an equalization

of the living conditions between rich and poor (i.e. social sustain-

ability), and (3) the interest in an intact natural environment that is

capable of supporting the needs of future generations (i.e. ecological

sustainability). Since social sustainability including the (re)distribu-

tion of natural resources and the benefits drawn from their use aresubject to intense political discussion and continued negotiations

especially between developed and developing countries, the norma-

tive character of this issue is readily accepted as an argument to

exclude it from the scientific discourse. Although balancing the

interests of succeeding generations is a normative issue as well,

the lacking possibility of the future generations to participate in the

corresponding political discussion is in this case taken as a justifica-

tion and as a potential for science to make fruitful contributions.

Consequently, the discussion of sustainability particularly among

economists essentially focuses on the question how to allow for the

strongest possible growth now without compromising the potential

for growth to persist in the future.

C. Sartorius / Ecological Economics xx (2005) xxxxxx 3


4/19

ARTICLE IN PRESS

of the economyecology relationship cannot be cor-

rected by merely relying on the market mechanism

(Rennings, 2000).

2.2. Strong sustainability and innovation

As a consequence of the failure of market competi-

tion alone to induce the innovations necessary to bring

about ecological sustainability, a different approach

has to be used. In search for such an approach, it is

worthwhile to look more closely at the distinction

between weak and strong sustainability and at the

sustainability indicators developed in relation to the

latter concept.

From the anthropocentric point of view, sustainabledevelopment implies the preservation of a pool of

natural resources and man-made capital that provides

each generation with the opportunity to have its activ-

ities based on equivalent sets of man-made and natural

capital. This conceptualization of sustainable devel-

opment as bnon-declining wealthQ (Pearce et al., 1989)

finds two basically different expressions. On the one

hand, economists in the tradition of Hartwick (1978)

and Solow (1974, 1986) argue that a society using an

exhaustible stock of resources could enjoy a constant

stream of consumption over time if it invested all the

rents from tapping on those resources, that is, if it held

the overall capital stock constant. Evidently, this weak

approach to sustainability is based on the implicit

assumption that both natural and man-made capitals

are complete (i.e. reversible) substitutes. While this

assumption may be met in some cases, it does not

hold in general because, first, for many types of

natural assets (e.g. an endangered species, a habitat

or the ozone layer) technical substitutes do not exist

and once brought about many changes turn out to be

irreversible (Munda, 1997). Moreover, the argument

developed above clearly indicates that the mechan-isms for specifically identifying and implementing

suitable technologies or inducing necessary innova-

tions do not exist in the neoclassical framework

underlying weak sustainability.

On the other hand, concepts of strong sustainabil-

ity, which are characteristic for ecological economics,

specify the natural capital in terms of its physical

function rather than the costs of actual damage caused

to it. The logic of this approach is based on the

assumption that, in order to continue to rely on certain

essential functions of the environment (e.g. assimila-

tion of waste or supply with resources), the ecosystem

or at least certain parts of it have to be kept intact.

Accordingly, substitutability has to be proven in eachspecific case rather than simply being assumed.

Although this approach does not exclude monetiza-

tion in principle (e.g. in terms of the opportunity costs

of the avoided or restricted use of the environment),

the (how ever aggregated) monetary figure does not

suffice to eventually specify the state of sustainability.

Instead, it is necessary to follow the following three-

step procedure and to (1) identify those elements of

the natural capital that are essential for the mainte-

nance of the ecosystems stability or, better, its ability

to recover from distressing impacts (i.e., resilience),(2) select those elements that are related to, and

possibly endangered by, economic activities, and (3)

derive a set of indicators each of which reflects the

actual condition of a specific aspect of the environ-

ment and puts it into relation to the sustainable state as

determined by any suitable management rule (see

Opschoor and Reijnders, 1991). Typical examples of

the latter approach are Pressure-State-Response (PSR)

indicators like the one employed by the OECD. Here,

the causes of environmental problems (bpressureQ),

the actual state of the environment (bstateQ), and

efforts to solve the problem (bresponseQ) are moni-

tored and quantified in separate modules (OECD,

1993).

The role of innovation in the latter framework

consists in modifying existing, or implementing

new, technologies in such a way that identified pres-

sures are relaxed and problematic environmental

states are improved. So far, however, this concept is

still quite limited such that it needs further qualifica-

tion and extension.

2.3. Critical loads and non-linearity

An important qualification of PSR-like schemes

refers to their implication that it is generally possible

to quantify the effect of an innovation in terms of

reduction of those processes or their side-effects that

caused the corresponding pressure in the first place. In

reality, however, many counter-measures later turn out

to be themselves not without side-effects such that the

relaxation of pressure in their target field may go

along with the increase of pressures in other fields.



5/19

ARTICLE IN PRESS

Rennings and Wiggering (1997) explain how this lack

of innovative efficiency arises. The logic underlying

the PSR approach implies a correlation according to

which stronger (weaker) efforts to counteract an envir-onmental problem by means of the best-available

technology generally lead to the alleviation (enhance-

ment) of the pressure and, thus, to the improvement

(deterioration) of the condition of the environment.

Unfortunately, in the natural environment, such a

blinearQ relation between causes and effects is not

the rule. In contrast, effects like the following are

frequently observed. Although in an agriculturally

dominated region the intense use of mineral fertilizers

was common practice for quite a while, contamination

of the ground-water with nitrate could be observedonly recentlywith a strongly increasing rate. In this

case, the existence (and transgression) of carrying

capacities or buffer capacities gives rise to non-linear

processes which typically show sudden changes or

even jumps. Returning to a sustainable state then

not only requires the reduction of emissions below

the respective critical load or critical level. Since the

latter may itself be adversely affected by the harm, it

additionally requires the repair of the damages that

had so far been caused by the excess emissions.

With regard to the role of innovation as a consti-

tuent of response in the PSR scheme, the non-linearity

basically implies a substantial element of uncertainty.

However, uncertainty is not limited to the adverse

effects that innovation is supposed to reverse. The

innovation itself is a source of uncertainty in so far

as it can be the source of lacking sustainability unfore-

seen at the moment of its implementation. More about

the causes of uncertainty and approaches to deal with

it will be said in the following.

3. Sustainable innovations and the evolutionaryperspective

Section 2 has elaborated on the possible impact of

innovation on ecological sustainability and on the

dependence of this impact on the underlying concepts

of sustainability and the economic paradigms related

to them. It could be shown that the preconditions for

innovations effectively responding to emerging eco-

logical challenges are rather demanding. This is not

only due to the basic structural complexity of inter-

action between a wide variety of elements in the

economy as well as in the natural environment; it is

even more due to the specific temporal interrelated-

ness of these elements. In order to deal with thesedifficulties, a closer look will be taken at concepts like

uncertainty, irreversibility, path dependency and coe-

volution which are closely related to innovation in the

sustainability context and in general and extensively

discussed in the evolutionary branch of economics.

3.1. Fundamental uncertainty and the trial-and-error

approach of evolution

It is the wide variety and high complexity of

interactions between human actors and between thelatter and their natural environment that renders

human (economic) activities as well as their environ-

mental effects highly unpredictable particularly in the

long run (see also Section 2.3). However, the uncer-

tainty accruing in this context is not just a matter of

probability distributions within a known or assumed

set of possibilities and therefore cannot be accounted

for by the concept of risk. Instead uncertainty is better

characterized as ignorance in the face of novel, fun-

damentally unpredictable, events. So the question

arises how to deal with this fundamental uncertainty.

If complete knowledge about the set of available

alternatives is lacking, actors cannot maximize the

expected utility of alternative choices and, thus,

rational decisions cannot be made. One approach to

the solution of this problem was made by Simon

(1957) who proposed that human decision-making

in situations of incomplete knowledge may better be

described as being based on bounded rationality.

However, the boundedly rational decision-makers

striving for an acceptable (i.e. dsatisficingT) rather

than a maximum level of utility still requires some

knowledge as to which goals are attainable in principle. Additionally, even bounded rationality assumes

fixed sets of individual preferences that basically

include all possible alternativesan assumption that

simply turns out to be underdetermined in the face of

real novelty.

Therefore, it may be advisable to look at the solu-

tion of (long-run) problems related to fundamental

uncertainty and endogenously changing preferences

from a completely different perspective: Darwins

approach to evolution in nature. Like society, nature



6/19

ARTICLE IN PRESS

is characterized by the complex interaction between

its constituents, the living organisms and their physi-

cal environment, and thus by the existence of funda-

mental uncertainty and non-linearity which togethercan give rise to the formation of new species or the

sudden extinction of major parts of the existing bio-

sphere as well as for the persistence of existing spe-

cies over prolonged periods of time. In order to

bmanageQ such unpredictable processes, nature relies

on the principles of heritance, random variation and

natural selectionwith diversity created by random

mutation and recombination within the existing

genetic pool and selection resulting from continuous

competition between species with inherited properties

for a limited set of resources.A further step toward an increased problem solving

capability in nature and, ultimately, in man is based on

the capability of an organism to undergo specific or

individual adaptation to varying circumstances and to

transmit the acquired knowledge to other organisms

that is to learn and communicate. While evolution on

this level is based on social norms, individual values,

and ideas rather than material genes, the basic princi-

ples nevertheless remain essentially unchanged (Sar-

torius, 2003, especially chap. 4). Initially, the

perception of a problem leads to the assessment of a

variety of alternative approaches to its solution. Those

approaches giving rise to the solution of the problem

are selected; those that fail are rejected. The solutions

with the better performance are further modified and

tested in subsequent rounds of selection.3 The wider

the variety of alternative approaches the higher is the

probability that at least one of them may perform

better than in the status quo. With respect to human

behavior, special use of evolutionary principles has

been made by many proponents of evolutionary eco-

nomics: in his search for new business opportunities,

for instance, Schumpeters (1934) entrepreneurassumes significant risk but, at the same time, gives

rise to novelty; Hayek (1978) interprets market com-

petition as a process of selection (and detection) of

innovations by means of the willingness-to-pay on the

demand side; and Nelson and Winter (1982) show

how profit may serve as the selecting force thatleads to the persistence of some innovations and to

the vanishing of most others. A particular case of

evolution leading to the solution of unprecedented

problems is the selection of cooperation rules on the

group level, a task that could never be fulfilled by

individuals on the basis of their mere rationality

(Hayek, 1988; Sartorius, 2002). In this context, envir-

onmental and social sustainability can respectively be

interpreted as cooperation (i.e. fair behavior with the

potential of winwin situations) between succeeding

generations and different parts of the same generation.The relevance of fundamental uncertainty and the

corresponding problem-solving capability for sustain-

ability is quite evident. Human activities frequently

generate adverse environmental side-effects which,

due to the complexity of their interaction with the

environment, are often unforeseen (see Section 2.3).

In the search for (long-term) sustainability, it therefore

makes little sense to exclusively rely on the causes of,

and solutions to, specific environmental problems

since they may be subject to considerable variation

over time. This does not at all imply that the determi-

nation of critical substances and the application of

critical thresholds do not make sense. Especially in

the short run they are even indispensable. However, in

the long run, that is, in the time perspective in which

the sustainability concept is usefully applied, the pro-

cess leading to sustainability also has to account for

the conditions under which the identification of pro-

blems as well as the search for the corresponding

solutions and their translation into the appropriate

measures takes place. Rather than referring to specific

innovations whose characterization as being sustain-

able can only be temporary, sustainability should beviewed as a property of the system and determined

with reference to the systems general capability to

bring about a variety of potentially useful innovations

and, should the occasion arise, to allow for the ready

implementation of the most promising alternative. In

short, sustainability also, and from the evolutionary

perspective predominantly, includes the flexibility and

versatility of the entire system to allow for a quick and

effective response to whichever environmental pro-

blem arises (see Erdmann, 2000).

3 Note that selection of the best alternative would only be possible

in a static environment with very low complexity. In reality, the

higher degree of complexity leads to the emergence of local rather

than global optima and, due to the dynamics of the system, the

successive choice of better alternatives influences the actual speci-

fication of the respectively best alternative. Therefore, selection in

the evolutionary approach adopted here refers to the better, but not

to the best (see also Rammel, 2003).



7/19

ARTICLE IN PRESS

3.2. Coevolution and the extension of innovation

beyond the technological sphere

The concept of coevolution basically refers to thefact that the development of an organism does not

simply follow the conditions set by its environment,

but that, in the process of adaptation, the organism

itself is a source of change for this environment. In

biology, coevolution usually describes the mutual

adaptation of ecologically related species such as

butterflies and plants (e.g. Ehrlich and Raven, 1964)

which, under certain circumstances, can give rise to an

arms race known as the Red Queen effect. In ecolo-

gical economics, coevolution refers to the socio-eco-

nomic development as a process of adaptation to achanging environment while being itself a source of

this change ( Norgaard, 1994; Gowdy, 1994). While

evolution in the socio-economic sphere was shown to

have the potential of giving rise to better adaptation to

the natural environment (Cavalli-Sforza and Feldman,

1981; Boyd and Richerson, 1985), coevolution in this

context typically accounts for the mutual interference

between socio-economic and natural developments

which, depending on their specific characteristics,

can facilitate or hinder innovation processes leading

to lesser or greater sustainability. In the case of spray-

ing pesticides in agriculture, for instance, the forma-

tion of resistance is a clear indication for a decrease in

sustainability.

With regard to sustainability-related innovations,

coevolution has several crucial implications. First,

the mutual interaction between several social spheres

increases complexity giving rise to a higher degree of

uncertainty that needs to be managed by human actors

trying to pursue a sustainable development (see Sec-

tion 3.1).

Second, coevolution involving the social or cul-

tural sphere has the potential of giving rise to a highdegree of diversity with regard to flexibility and

adaptability to temporally or spatially varying condi-

tions (Munda, 1997). This potential is however con-

trasted by the possibility of the Red Queen effect

which is likely to give rise to maladaptation and the

reduction of diversity. These two opposing effects are

the basis for the trade-off between diversity and flex-

ibility on the one hand and economic efficiency on the

other (Rammel, 2003), which will be further discussed

in Section 4.

Third, the coevolution (especially the Red Queen

effect) is an intriguing example of path dependence

where the developmental path can not easily be shifted

from one trajectory to another (see Section 3.3).Forth, and most evidently, coevolution implies that

successful innovation in general, and successful sus-

tainable innovation in particular, has to acknowledge

the involvement of, and mutual interaction between,

more than the mere technical and economic spheres.

The current human way of life being coherent with the

existing institutions (including codified rules and

social value and belief systems) and giving rise to

technology-caused transgressions of the sustainability

boundary in many and profound ways and, contra-

riwise, the support of this lifestyle by just thesetechno-economic conditions are an evident instance

of coevolution. Accordingly, efficiency changes under

the proviso of sustainability may be achieved more

readily through an integrated approach employing

institutional or social in addition to technical innova-

tions. With regard to the aim for increased sustain-

ability, it is therefore necessary to broaden the view

from the merely technical towards the social and

political aspects of innovations. In accord with these

thoughts, Klemmer et al. (1999, see also Rennings,

2000) broadly define the term denvironmental

innovationT as all measures of relevant actors that

lead to the development and application of new

ideas, behavior, products and processes and, thereby,

contribute to a reduction of environmental burdens or

to ecologically specified sustainability targets. This

may include process and product innovations, organi-

zational changes in the management of firms, and, on

the social and political level, changes in environmen-

tally counter-productive regulation and legislature,

consumer behavior, or lifestyle in general. This

emphasis on social innovations is all the more impor-

tant because unsustainable development itself is oftenthe result of btechnology outpacing changes in social

organizationQ (Norgaard, 1994, p. 16). Moreover, after

an intense and extended discussion in environmental

economics about the brightQ instruments towards an

environmentally sound, sustainable development, it

more and more turns out that there is not a single

suitable instrument. Instead, it seems to depend on the

respective circumstances (e.g. type of competition or

existence of information asymmetries), whether Pigou

taxes, markets for pollution rights, the setting of stan-



8/19

ARTICLE IN PRESS

dards, or even temporary subsidization of promising

innovations is the more effective instrument

(Rennings, 2000). Jaenicke (1999) even goes one

step further by claiming that the relevance of instru-ments for environmental policy has generally been

overemphasized. Instead, the discussion should

focus on other elements of a successful environmental

policy such as long-term goals, mixes of instruments,

policy styles, and constellations of actors.

Altogether, the above emphasis on social and

political aspects makes clear that the success of

sustainable innovations depends on more than their

mere technical (or even economic) superiority. This

is all the more evident when, following the sugges-

tion of Section 3.1, sustainability is considered as the property of an entire system rather than a specific

innovation.

3.3. Irreversibility and path dependence

Beside fundamental uncertainty and the need for

diversity following from the preceding sections, the

complexity of multiple-interaction systems has

another at least equally important consequence for

the sustainability discussion. If the sequence of events

within a complex system was described by means of

several independent parameters, careful analysis

would reveal non-ergodicity. That is, of all basically

possible states only some are likely to occur in any

single moment. Whether or not a given state is likely

to arise accordingly depends on the past or, more

exactly, on the succession of states preceding the

actual statea phenomenon called path dependence.

In biology, this issue is discussed among evolutionary

biologists and ecologists in the context of the phylo-

genetic development of organisms and successions in

ecosystems. Gould and Eldredge (1977), for instance,

emphasize that the complex architecture of all moreadvanced organisms strongly limits the potential for

successful further mutations and, thus, better adapta-

tion. The reason for this is that most changes that may

be advantageous from an isolated perspective may not

be so in a complex context in which advantageousness

requires the meeting of many strict preconditions. As

a consequence, a variety of mutations would have to

come up simultaneously which is very unlikely to

occur. So, once a certain amount of genetic informa-

tion has accumulated within an organism and hap-

pened to be arranged in a sufficiently complex system

of mutual interaction, the entire system is stabilized

against further change (Waddington, 1969). Interest-

ingly, the parameter decisive for the stability of thesystem is the average number of interaction from one

element to others and not so much the number of

genes (Kaufman, 1995).

With regard to sustainability, path dependence

plays a particularly important role in three respects.

First, as shown above, the wide variety of life forms in

nature represents a large source of solutions for pro-

blems not only in the natural environment but also in

the human spherefor the assimilation of wastes, the

production of food, and the design of pharmaceuticals,

to mention just a few examples. Every species evi-dently represents a piece of knowledge that could

potentially be useful for present or future generations.

Against the backdrop of path dependence, however, it

is also clear that the loss of any species leads to a loss

of such knowledge that is irreversible. For every

species is the outcome of a succession of phylogenetic

stages in which the formation of every single stage is

based on the existence of its respective predecessor

a fact that renders it impossible to reconstruct a spe-

cies once it has been lost.

Second, even when knowledge is not directly

acquired from models in nature, but derived through

trial and error in the scientific process, this does not

imply that all knowledge is equally accessible.

Instead, technical knowledge generation is character-

ized by technological paradigms (Dosi, 1982, 1988).

Within such paradigms, knowledge acquisition occurs

gradually along the respective trajectoriesby the

systematic variation of single parameters and the

selection of those variants showing the desired effect

most markedly. Incremental innovations proceeding

along such a path are to some extent predictable but

the marginal cost-to-effect ratio is subject to increasesuch that maintaining the profitability (in economic

terms) of innovations becomes increasingly more dif-

ficult. With regard to sustainability, end-of-pipe solu-

tions fit into this category because they add

environmental soundness to an existing technology.

According to economic wisdom, they do this at

increasing marginal costs.

An alternative route is the search for radical

innovations leading to a transition between trajec-

tories in different paradigms. While this approach



9/19

ARTICLE IN PRESS

has the potential of achieving much better profitabil-

ity in economic contexts, it is characterized by a high

degree of uncertainty representing a substantial

threshold for typically risk-averse people. In the sus-tainability context, Porter and van der Linde (1995)

emphasize that integrated environmental innovations

(where harm is avoided from the beginning rather

than redressed after its generation) can be so efficient

that the environmentally beneficial effect induced by

suitable regulatory measures is achieved at no addi-

tional cost.

The third aspect of path dependence to be

addressed here refers to the induced resistance-to-

change and, thereby, to some extent relates to the

second. It plays an important role in the discussionabout technology development and is of central

importance for the objective of this paper: the search

of determinants for a sustainable technology develop-

ment. Innovations and the introduction of new tech-

nologies often are the key instruments to the

(temporary) avoidance or redressing of adverse envir-

onmental effects. However, even if negative external

effects were completely internalized and the new

technology turned out to be technologically and envir-

onmentally superior to the existing one, successful

commercialization and diffusion into the market can-

not be taken for granted. A frequently quoted example

for this kind of failure of a superior technology to

prevail refers to the design of typewriter and computer

keyboards (David, 1985). Although the totality of

users could benefit from the use of a better design

that allows for a significantly higher writing speed, the

traditional QWERTY keyboard is maintained because

just for the first users of any new alternative, a devia-

tion from the dominant design would cause costs that

are much higher than the expected benefits (Arthur,

1988). While network externalities are the relevant

factor in the latter case, a variety of other effectswill be identified in Section 5 that lead to the lock-

in of a conventional technology and, accordingly, to

the lock-out of its superior challenger.

4. Second-order sustainability

While traditional approaches to achieving (strong)

sustainability typically start with the identification of

the technical or social causes of a current lack of

sustainability and then point to possible alternatives,

the implications of fundamental uncertainty, coevolu-

tion and path dependency go far beyond such an

assessment of specific innovations. In order to accord-ingly develop a more comprehensive conception of

sustainability, it is necessary to return once again to

the shortcomings of the traditional approach of attain-

ing sustainability.

4.1. Knowledge gain through trial-and-error

First, in most societies, and all the more in all

advanced economies, the common wealth is yielded

by the complex interaction of numerous individuals in

the context of a variety of technologies and social andpolitical institutions. Due to the uncertainty prevalent

in such complex systems (see Section 3.1), it is

impossible to predict all the specific effects of any

human intervention. This argument explains the some-

times low success in fitting a new technology or

institution into a given setting in general. And it

particularly explains the frequent failure of technical

or institutional innovations in terms of sustainability.

As a consequence, the search for increased sustain-

ability, while in principle being the result of human

action, will usually not be the well-specified outcome

of a man-made plan. Hayek (1973) referred to this

phenomenon as the bfailure of constructivist

rationalismQ and identified mans constitutive lack of

knowledge as its main cause. Instead of looking for

the one and only perfect substitute, Hayek suggests, it

therefore appears much more promising to engage

into a trial-and-error process based on a variety of

potential substitutes.

4.2. Diversity as precondition for trial-and-error

The second argument in disfavor of the specificreplacement of a non-sustainable technology (or insti-

tution) also refers to the uncertainty aspect, but it

focuses on the effect of dynamic change rather than

the mere lack of knowledge. Since, in a complex

system, the change in one component always gives

rise to a change of the restrictive conditions for all

others (compare the discussion of coevolution in

Section 3.2), it is little surprising that sustainability

as achieved by the employment of whichever technol-

ogy or institution (e.g. property rights or social pre-



10/19

ARTICLE IN PRESS

ferences) can only be a temporary state of a system.

Even those interventions that successfully redress

instances of lacking sustainability at first will them-

selves change the entire system in such a way thatnew losses of sustainability are likely incurred in the

futureeither due to their interaction with other com-

ponents or directly by themselves. In order to main-

tain sustainability over longer periods of time, it is

therefore not sufficient to simply solve a given pro-

blem; rather the problem-solving capacity must keep

pace with the rise of new problems. So, the Darwinian

process of trial-and-error has to cope with timea

scarce resource especially in dynamic systems; and in

order to do so, two preconditions need to be met

which at first appear to be given quite naturally, butin fact do not come for free. Variation, the first pre-

condition, implies the existence of a wide variety of

potential alternatives on which selection can act. For

socioeconomic systems in general, Matutinovic

(2001) shows that diversity is a systematic and resi-

lient property the lack of which could provoke

instability and eventually lead to the collapse of the

system. Since in present-day economies the selective

effect of market competition is rather strong, self-

sustained maintenance of a high degree of diversity

cannot be taken for granted. Especially with regard to

the uncertainty associated with long-term sustainabil-

ity problems, it may therefore even be necessary to

actively keep competition in a more early stage

against the self-enforcing advantages of productiv-

ity-increasing specialization (see Kemp, 1997).

Accordingly, it is evident that this diversity is costly

since, (1) for the supply of promising technologies

society needs to promote learning, that is, to invest

into human capital. More specifically, incentives for

an engagement into R&D have to be provided for the

respective firms. (2) Prior to eventually reaching mar-

ket diffusion and successful commercialization, parti-cularly the more radical inventions may additionally

need governmental support (e.g. through subsidization

or the creation of niche markets). (3) Finally, the

partial suspension of market forces needed to maintain

a certain degree of diversity and keep competition in

an early stage not only leads to the less efficient

adaptation of technologies to the existing uses; (4) it

also prevents part of the cost-saving potentials of

economies of scale and scope or learning effects

from being realized. The trade-off we face here is

one between (short-term) economic and (long-term)

sustainability-related efficiency.

4.3. Lock-in resolution as precondition fortrial-and-error

In contrast to the preceding arguments, the third

argument against the possibility of an easy substitu-

tion of more sustainable technologiesand respec-

tively the second precondition for the successful

employment of trial-and-error processesrelates to

the systemic integration of established technologies

and institutions rather than their potential substitutes.

For even in the presence of a variety of alternative

solutions, selection and further development of themost suitable technology by means of the market

forces will remain ineffective so long as the estab-

lished technology is subject to strong stabilization and

withstands its displacement by even strongly superior

alternatives.4 In Section 3.3, this resistance to change

of the established technology known as lock-in was

shown to be caused by a wide variety of effects of

which a more complete account will be given in

Section 5. Sustainability is particularly affected by

such a lock-in because the more radicaland thereby

often more effectiveinnovations (Rennings, 2000;

Ashford, 2002) face more opposition than the less

effective incremental ones because they belong to a

new paradigm. To the extent that lock-in effects are to

be undermined, economic actors again have to bear

the cost of refraining from the realization of the

corresponding economies of scale, learning and net-

work effects, etc. (see Section 4.2). Although, this

time, the trade-off between short-term and long-term

efficiency is basically a purely economic one, it has

important consequences for sustainability.

4 The careful reader will have recognized that, on the one hand,the proposed approach relies on the selective capacity of (market)

competition to identify and gradually improve more sustainable

technologies while, on the other hand, competition has to be par-

tially suspended in order to create the diversity on which selection

can act. This apparent contradiction is resolved by temporal, spatial

or functional disjunction of the two functions. While temporal

disjunction can be achieved by the mere alternation of phases of

variation and selection, spatial and functional disjunction, respec-

tively, imply implementation and testing of new technologies on a

locally or application-specific markets. The latter two approaches

also constitute the basis for strategic niche management (Kemp et

al., 1998).



11/19

ARTICLE IN PRESS

4.4. Second-order sustainability as adaptive flexibility

While the specific problem-solving capacity of cer-

tain innovations gives rise to sustainability in specificcircumstances and for limited periods of time, it is the

total number of such solutions or, more concisely, the

context-dependent trial-and-error process giving rise to

their implementation that brings about sustainability in

more general termsin the long run and in dynamic

contexts. The latter idea conforms well with the view of

Kemp (1997), Van den Bergh and Gowdy (2000),

Rammel and Van den Bergh (2003) and Rammel

(2003) that sustainability is the result of a strategic

process (rather than a certain state) trying to deal

with uncertainty and unpredictable emerging propertiesby means ofbadaptive flexibilityQ. This emphasis of the

process character does, of course, not render specific

sustainable innovations dispensable in the search for

sustainability in general, but due to their conditional

effectiveness they represent sufficient (rather than

necessary) conditions of sustainability whereas the

conditions for the effective working of the basic trial-

and-error process are necessary (but not sufficient)

ones. Since, from the functional perspective, bgeneralQ

sustainability determines the conditions under which

bspecificQ sustainability can be achieved, the two kinds

of sustainability describe the function of a system on

two different levels with general sustainability repre-

senting the more basic level. Since sustainability, and

more so its lack, is immediately perceived as specific

instance of resource or environmental problems

whereas the working of its general problem-solving

capacity (albeit more fundamental) is less immediately

evident, I refer to the two aspects as first-order and

second-order sustainability, respectively.

In the preceding parts of this section, a variety of

measures was mentioned that would increase diversity,

improve selection and, thus, support second-order sus-tainability, but most of these measures would come

with significant (opportunity) costs only. Conversely,

the lack of second-order sustainability caused by the

unwillingness to pay this price leads itself to the incap-

ability to adapt to changing circumstances and, thus, to

a loss of welfare that arises from the high cost of

redressing or functionally replacing a damaged envir-

onment. In this trade-off between the costs and benefits

of second-order sustainability, the optimum degree of

diversity is not easily determined ex ante. However,

also this optimum can be approached in a trial-and-

error mannerby the gradual change and subsequent

assessment of the conditions for second-order sustain-

ability particularly in those industries and economicsectors where the most severe violations of first-order

sustainability are encountered.

5. Determinants of second-order sustainability

In Section 3.3, it was suggested that certain struc-

tural properties of a given technology can severely

restrict the probability with which new innovations

may become effective. The way in which these states

of rigidity are sometimes discussed (David, 1985) ormodeled (Arthur, 1988) in the literature could imply

that such states of stability are omnipresent and, once

they turn up, tend to persist for prolonged periods of

time. Not surprisingly, many economists (e.g. Lie-

bowitz and Margolis, 1994) are convinced that latter

position crossly overstates the relevance of network

externalities, as this would allow them to become the

cause of almost ubiquitous market failure. In the latter

debate, an intermediate position is taken by Witt

(1997) who, while principally acknowledging the

relevance of network effects, limits their general

importance for the function of the market to certain

restricted periods of time. So periods of stability tend

to alternate with periods of instability where new

networks can be formed. Such a period in which the

direction of technological progress is flexible is

referred to as a bwindow of opportunityQ (Witt,

1997). Disregarding these windows could severely

hamper, if not completely inhibit, the introduction of

any useful innovation. And even when, in the pursuit

of sustainability, a new (sustainable) technology was

successfully pushed by governmental regulation with

no regard at the specific circumstances, the differencebetween stable and unstable phases would be worth a

lot of money. It will therefore be the main objective of

this section to identify those factors that allow poli-

tical and other decision makers to make a well-

founded judgement as to whether the preference for

a potentially sustainable innovation is based on eco-

nomic, social and political feasibility.

The first set of factors will be economic ones. It

will become evident in the following that the variety

of relevant effects is wider and their respective time



12/19

ARTICLE IN PRESS

pattern more diverse than may have been implied by

the repeated reference to network externalities in pre-

vious parts of this paper. Additionally, it is a special

characteristic of many sustainable technologies that,beyond the competitive disadvantage frequently aris-

ing from their failure to internalize reduced external

costs, the government typically plays a crucial role in

overcoming existing barriers to competitiveness in the

relevant markets. In doing so the government inevi-

tably faces opposition from those whose interests are

negatively affected: the incumbent industry and other

groups paying the price for the measures taken. Typi-

cally, a government or policy makers in general are

not inclined to neglect such an opposition unless the

promoting forces from other parts of the society aresufficiently strong. More so, major techno-economic

changes require a general openness or even a readi-

ness to change (i.e. a phase of instability) on the part

of the political system. For these reasons, the techno-

economic factors will have to be supplemented by

both, political and social factors. The selection of

these criteria occurred on the basis of a priori theore-

tical plausibility considerations and ex post after the

screening of relevant case studies (Sartorius and Zun-

del, 2005). Due to the large number of relevant fac-

tors, it is not possible to present them here at length;

for a more detailed discussion, the reader is therefore

referred to Sartorius and Zundel (2005, ch.2).

5.1. Determinants of (in)stability in the

techno-economic system

5.1.1. Economies of scale

Economies of scale account for the greater effi-

ciency of larger manufacturing devices. They are typi-

cally measured on the firm level in terms of average

unit cost as a function of output rate. As these average

costs decrease with increasing scale, they give rise tostrong competitive disadvantage for new technologies

which, at the beginning of their life cycle, cannot

immediately engage into large-scale production.

5.1.2. Economies of scope

Economies of scope account for synergies between

different production lines from the common use of

certain resources, intermediate products, or production

facilities. While economies of scope lead to important

cost decreases for the established industry, the mutual

dependencies between existing processes renders it

more difficult for a radically new technology to

become competitive.

5.1.3. Learning by doing

Unlike the cases of economies of scale and econo-

mies of scope, the cost decreasing effect of growing

experience in designing, constructing (dlearning by

doingT), and using production facilities (dlearning by

usingT) is usually expressed as the percentage of cost/

price reduction per doubling of the cumulative pro-

duction output in the respective branch. While learn-

ing effects give rise to a large potential for further cost

reductions for any new technology, they confront it

with a high cost disadvantage in the beginning.

5.1.4. Network externalities

Network externalities refer to the fact that the

utility derived from the use of a given technology is

positively correlated with the number of its users.

Alternatively, a technology can be subject to network

externalities if it relies on another technology that

forms a network in its turn. The weaker the depen-

dence on the established network or the better the

compatibility, the smaller is the entry barrier for the

new technology.

5.1.5. Sunk cost

Investment into a new technology can cause sig-

nificant sunk costs if it renders useless an old technol-

ogy prior to its complete depreciation. While sunk

costs represent opportunity costs of any new technol-

ogy, they do not come to bear in competitive markets.

Instead, they are relevant whenever market access is

restricted by other causes. The rate of capitalization in

the relevant industry and data about the investment

cycle can be used to assess sunk costs; but this

analysis needs to be supplemented by the competitivestructure of the industry in question (see below).

5.1.6. Market structure

Although natural or regulated oligopolies or mono-

polies do not exclude competition in principle, such

market structures will provide the corresponding firms

with strong incentives to maintain the existing market

barriers, engage in strong activities to preserve these

monopoly rents and neglect innovative activities.

While the innovation-related forces of market competi-



13/19

ARTICLE IN PRESS

tion may be characterized as biased in favor of the

established technology by (above-mentioned) increas-

ing returns to adoption, any non-competitive market

structure will stabilize the technological status quoeven more because it does not give rise to innovation

in the first place.

5.1.7. Potential versus risk

In order to replenish their earned innovation rent

and, thus, maintain their current profit margins within

a competitive market environment, entrepreneurs

occasionally have to complement their technological

portfolios with more radical innovations. Since the

latter are associated with higher risk, an (expected)

strong potential (including its regulatory conditions)will be decisive for the success or failure of this

technology being adopted.

5.1.8. Demand

To be considered an economic substitute for an

existing technology, a new technology at first has to

fulfill certain functions of the former. In order to

attract the attention and raise the specific demand of

consumers and investors that would prefer the more

familiar, established technology over its otherwise

quite dissimilar counterpart, a new technology has to

fulfill certain extra-functions to overcome this inertia.

5.1.9. Niche markets

If the entry barrier for a new technology is high, it

may need a long period of subsidization until general

competitiveness is achieved. At the same time, partial

competitiveness may be achieved much sooner under

certain, geographically or culturally specified, favor-

able conditionsoften called a niche market. Since

the existence and extent of niche markets can be

decisive for reaching competitiveness of a new tech-

nology in general, the strategic use or the artificial(regulatory) creation of such niche markets can be an

important approach to the successful implementation

of a new technology as suggested by the strategic-

niche-management approach of Kemp et al. (1998).

5.2. Determinants of in-/stability in the political system

The basic characteristics of the political system

generally play an important role in allowing a new,

more sustainable technology to prevail. As a precon-

dition for this to happen, the political system either

must be in favor of the new technology from the

beginning or it needs to be destabilized itself in the

first place. While in the former case, structural char-acteristics of the political system play the most impor-

tant role, both structural and procedural aspects are

important in the latter. The following enumeration will

begin with the structural factors.

5.2.1. Institutional embeddedness

Many technologies, particularly those related to

environmental protection, are subject to substantial

political regulation determining which external effects

a technology is allowed to exert and which (and how)

others must be avoided. In this context, the closemutual relationship between the established technol-

ogy and its regulatory environment tends to adversely

influence the competitive position of any (radically)

new competitor. An example for the self-stabilizing

effect that needs to be overcome by a new technology

is the reference many regulations make to the state of

the art (related to the established technology) for

solving an environmental problem.

5.2.2. Interest groups

While it is a matter of political culture how influ-

ential corporate bodies or individual actors can be in

principle, it depends on the specific circumstances

which effects they actually give rise to. Basically,

the power of an interest group is known to be crucially

dependent on the size of the group, the homogeneity

of its interests, its organization, and the resources it

controls (Olson, 1965). Other important factors are the

economic relevance of the industry or its history and

cultural integration. Particularly in mature industries

with strong market power, lobbying may pay even for

single firms as investing in a useful regulatory envir-

onment is more profitable than investments in tech-nological innovations (Berg, 1995) with the

corresponding stabilizing effect for the established

technology.

5.2.3. Asymmetry of knowledge

For the solution of environmental problems, gov-

ernments and political administrations need external

advice. As long as the problem has not attracted too

much public attention, the necessary information is

most convenient obtained from the industry that



14/19

ARTICLE IN PRESS

Table 1

Factors determining the stability or instability in each of the three subsystems and the indicators used for their operationalization

Effect Indicators Operationalization

Economies of scale cost (or price) development as a function of actual outputaverage capitalization of industry statistical data

identification of investment cycles recurrent phase-shifted cycling of prices and investment

Sunk costs

political regulation cost of retro-fitting after regulation,

delayed investment due to expectation of uncertain measures

Economies of scope pattern of interactions between

production lines

number and relevance of interactions between the old (new)

technology and the entire production network

Learning by doing cost (or price) development as a function of cumulative output

direct competition with (an)other

network(s)

market share(s) of the competitor(s),

availability of gateway technologies

Network externalities

need for compatibility with comple-

menting infrastructure or periphery:

existence of public standards availability of an adapter which requirements are met?cost of the adapter, legal admission possible, payable royalties

Market structure degree of competition as a function of

market concentration

market share of the biggest firm(s), Herfindahl index, legal

regulations

riskiness availability of capital marginal interest rate, capital share of venture capitalistsPotential / risk

problem solving capacity

realization of an innovation rent

technical properties (benchmarks), associated costs

readiness to pay for extra-functions market research

existence of natural niche markets higher prices, non-applicability of the established technology

Techno-economicsubsystem

Extra-demand

creation of artificial niche markets by

means of regulation

(eco-)taxes, tradable certificates, cost of retro-fitting the old

technology

Institutional

embeddedness

subsidies

protection

norms and standards

financial support, tax breaks

duties, other barriers to trade

specificity of specification

Interest groups resources under control (power)

structure; degree of homogeneityinfluence; earlier success

number and economic importance of represented firms/sector

market shares, concentration index(qualitative)

Asymmetry of

knowledge

influence of industry in hearings

number of industry-independent

research institutions/projects

(qualitative)

number, financial support, number and size of commissioned

projects

Parliamentary

majorities

stability of majorities size of majority, stability of constituting coalition (number and

relation of parties)

Election cycle distance to the next election ditto.

political scandals deception by possible interest holdersSingular constraints

catastrophes accidents, unexpected discoveries

probabili ty of legislative initiatives number and relevance of potential init iators, number of cases

legislative vs. administrative

regulation

number of laws referring to ordinances, actual number of

ordinances

reassessment and resubmission cycles deadlines, frequency, possible consequences

corporate structure number, size, and frequency of political involvement of

corporate organizations

participation frequency and extent of incorporation of political outsiders

(e.g. NGOs) into the decision process

Politicalsubsystem

Decision-making

procedures

supranational structures share of regulation that is not subject to national legislation

relevant publications in scientific

literature, contributions to conferences

number of relevant articles (keyword search) in journals etc.;

identification of seminal articles and quotation circles

Scientific confirm-

ation of threat to

sustainability independence of research sources and quantity of research sponsoring

Public concern about

lack of sustainability

Socio-cultural

subsystem

Public acceptance of

possible solutions

relevant articles in newspapers, reports

in broadcast,

formation of major protest campaigns

number of articles/reports over time

number and size of campaigns

Source: own compilation.



15/19

ARTICLE IN PRESS

caused the problem. According to the life cycle theory

of bureaucracies, initially independent (regulatory)

authorities will thus successively merge theirinterests

with those of the established industry (Martimort,1999). This bregulatory capture of bureaucraciesQ

often leads to quick and at most half-hearted solutions

related to the dominant technology. By contrast, more

radical changes can only be expected, if the necessary

knowledge comes from more independent sources

notably state-financed scientific research.

5.2.4. Parliamentary majorities

Especially more radical changes are often not

unanimously supported since the associated improve-

ments go at the expense of the established regime.Even if its basic attitude would tend to render a

government supportive of the corresponding change,

its actual realization will ultimately depend on the

strength and stability of the majority on which it can

rely. From this perspective, a large, stable majority

basically opens the potential for more radical changes

than does a minute or unstable one.

5.2.5. Election cycle

One of the most prominent stylized facts in poli-

tical science states that more radical political changes

usually occur at the beginning of an election period

while incremental changes, if not political standstill,

follow at the end (Troja, 1998). With regard to envir-

onmental innovations this implies a potential for a

political window of opportunity in the post-election

period. Unfortunately, empirical tests so far failed to

confirm this effect of the election cycle (Horbach,

1992). A special popularity of environmental regula-

tion, an eminent problem pressure or, like in Germany,

the temporal alternation between state and federal

elections could be reasons for this.

5.2.6. Singular constraints

The costs and, thus, the scope of each regulatory

measure is subject to a budget constraint. While the

power of the interest groups behind technologies gen-

erally influences the allocation of governmental

resources, it depends on the social appreciation of

environmental protection or the reputation of the

involved parties whether the incumbent industry can

defend its subsidies or has to share it with its more

sustainable competitors. In this respect, singular (i.e.

exogenous) events like political scandals and envir-

onmental or other disaster can bring about sudden

changes.

5.2.7. Decision-making procedures

Since it is not possible here to extensively analyze

the entire political decision-making process, just a few

criteria will be presented that may allow for a basic

characterization of the procedural aspects of a political

system with regard to the stabilization or destabiliza-

tion of a specific technology.

(1) It is an important aspect of political culture

whether the initiatives for regulatory acts typi-

cally come from single actors (e.g. president,members of parliament) or major bodies (gov-

ernment, parties, or the parliament). In general,

the former tends to give rise to more radical

(i.e. destabilizing) changes than those of the

latter.

(2) The relation between legislative bodies and

executive administration determines whether

a regulation is enacted by means of a law that

has to pass a lengthy parliamentary approval

procedure or whether this can be done by refer-

ring to an ordinance that is quickly adopted by

the administration alone.

(3) Obligatory reassessment and resubmission

cycles ensure that any existing regulation does

not lead to the stabilization of the respectively

benefiting technology.

(4) Participation of larger parts of the society (e.g.

NGOs, public research institutes) in the search

for more sustainable solutions will not only

facilitate the search for knowledge but also

increase and widen the support for (often more

radical) solutions.

(5) Finally, it is important how a country is incor-porated into supranational structures (e.g. EU,

WTO). While this limits a countrys possibility

to implement innovations in an idiosyncratic

manner, it broadens the scope and efficacy of

many sustainable innovations.

5.3. Factors of change in the socio-cultural system

Public attention to a (perceived) problem and

subsequent worry about its potential consequences



16/19

ARTICLE IN PRESS

play a key role in provoking political reactions direc-

ted to solving the problem or, at least, alleviating its

consequences. This is all the more true in the context

of environmental protection since due to their long-term relevance and public-good nature, environmental

problems and their solutions are rarely issues that

allow a politician to derive major benefits for himself.

While awareness and concern by a considerable part

of the population is neither sufficient nor necessary

for political action to be initiated, their lack will

usually lead to a failure or, at least, major delay in

acting accordingly.

Mass media play an important role not only as

transmitters for the corresponding information but

also for the assignment of meaning and valuationto the underlying problem. The relation between the

media and their readers, listeners, or watchers is

characterized by mutual interaction giving rise to

positive and negative reinforcement The scientific

verification of an environmental problem, which

often stays at the beginning of such an dissue atten-

tion cycleT (Downs, 1972), is identified through

scanning the scientific literature for relevant key-

words and trying to identify seminal publications

through the tracing back of references. On the

other hand, public concern about these problems

can be measured to some extent by counting relevant

articles in newspapers and reports in other mass

media. Additionally, it may be necessary to account

for the more qualitative aspects of concern and

valuation, as the authors of relevant articles often

differ in their basic attitude towards a given environ-

mental problem. It is also important to realize that

the attention of mass media to any given problem

usually tends to decline more rapidly than the atten-

tion of the public in general.

Table 1 summarizes the comprehensive list of

determinants of periods of instability elaboratedabove including the corresponding indicators and

their potential operationalization.

5.4. Windows of opportunity as periods of higher

second-order sustainability

In order to identify periods of greater or lesser

second-order sustainability by means of these indi-

cators, it needs to be pointed out first that sustain-

ability correlates strongly with the instability

(=flexibility) of the established technological regime

and the political and social conditions supporting it.

So, second-order sustainability will be strongest

when the window of opportunity is most widelyopen and it will be weak when the window is closed.

In order to identify a window of opportunity, an

aggregation of its determinants is necessary. Since

the direct comparison of all these indicators on the

basis of a common denominator (e.g. monetary

value) is not possible, however, any comparison

can in the end only be of qualitative nature. There-

fore, the following scheme of aggregation is used to

arrive at least at a relative measure of second-order

sustainability.

In the techno-economic sphere, all factors essen-tially work in parallel. High sunk costs add to the

stability of the incumbent technology as well as does

extended learning. Niche markets for the new tech-

nology on the other hand destabilize the incumbent

technology. None of these factors relies on another

one to become effective. So, even if one effect became

zero, the other factors would remain unaffected. Their

mode of aggregation is additive.

By contrast, in the socio-cultural system, (scienti-

fic) verification of an environmental problem is a

necessary (but not sufficient) prerequisite for the for-

mation of public concern. Conversely, public concern

alone sometimes is little effective until the exact

causes for an environmental problem are scientifically

verified and unless an acceptable solution exists. So,

all factors work in sequence with the combined effect

yielded by multiplying the constituents.

In the political system, both effects are found.

While structural and procedural factors in general

appear to complement each other in a multiplicative

way, the specific structural (or procedural) factors tend

to work in parallel.

With regard to the relationship between the entiresystems, the political system not surprisingly is of

central importance because in the end, it brings

about the regulation necessary to achieve greater sus-

tainability. However, the political system hardly

works on its own; it needs impulses from the other

systems: destabilizing impulses (for the existing

regime) come from the society disapproving the lack

of sustainability and from the new, more sustainable

technological or institutional alternatives; opposite

stabilizing impulses come from the incumbent indus-



17/19

ARTICLE IN PRESS

try that caused the environmental problem and the

loss of sustainability in the first place. Fig. 1

summarizes how the composite indicator of sustain-

able technology development is constructed from itsconstituents.

6. Conclusion

In particular radical innovations can be important

means to the achievement of improved sustainability.

Due to the existence of path dependencies, however,

the transition from one technological trajectory to

another, more sustainable one is often impeded by

significant barriers. Fortunately, these barriers are bytheir nature subject to substantial changes; so, it

makes sense to carefully distinguish between periods

of stability (with high barriers) in which the given

trajectory can hardly be left and periods of instabil-

ity (characterized by low barriers) where a new

trajectory can be reached more easily. With respect

to sustainability, the latter distinction is particularly

important for two reasons. First, more sustainable

innovations often rely on governmental regulation.

In periods of instability, the economic burden arising

from this regulation will be much lower than in

periods of stability; so, a given budget will yield a

much better sustainability effect in the former case

than in the latter. Second, due to the complexity and

changes in their respective environments, innova-

tions are generally associated with fundamental

uncertainty such that it becomes impossible to pre-dict the degree of sustainability resulting from spe-

cific innovations in the long run. Under these

circumstances, it is essential to ensure flexibility

including the possibility to select between a variety

of different trajectories in a process of trial and error.

Sustainability as viewed from this evolutionary

perspective may therefore better be understood as

the general capability to readily change between

different technological trajectories. Since the latter

kind of sustainability determines the conditions

under which the former kind can be achieved,we call the two kinds first-order and second-order

sustainability.

In order to undergo successful diffusion, most

sustainable innovations rely on regulatory measures

especially in the beginning of their (economic) life-

cycles. When looking for the factors determining

periods of (in-)stability, the political system enacting

this regulation therefore is of central interest. How-

ever, while basically allowing for the convergence

of both technological progress and sustainability, the

political system itself can neither give rise to the

search for sustainability nor bring about the appro-

priate innovations in the first place. This is where

the socio-cultural and, of course, the techno-eco-

nomic sphere itself enter the focus of attention as

emitters of positive impulses. Additionally, negative

impulses like those coming from the incumbent

industry need to be taken into account. After all, a

series of factors (and corresponding indicators)

could be identified which, after proper weighting

and prioritization, allow to make an estimation

whether, and possibly when, the incumbent industry

is sufficiently destabilized and the political systemrendered sufficiently favorable to the new, more

sustainable technology such that a transition to the

preferred trajectory is possible without the lowest

effort possible.

Acknowledgement

Funding of this research by the German Federal

Ministry for Education and Research (grant

Fig. 1. Reconstruction of a measure of second-order sustainability

from its constituent factors in the techno-economic, political, and

socio-cultural sphere.



18/19

ARTICLE IN PRESS

07RIW5C) is gratefully acknowledged. I thank Guido

Bunstorf, Jan Nill, Stefan Zundel and an anonymous

referee for valuable comments on earlier versions of

this paper.

References

Arthur, W.B., 1988. Self-reinforcing mechanisms in economics. In:

Anderson, P.W., Arrow, K.J., Pines, D. (Eds.), The Economy as

an Evolving Complex System. Addison-Wesley, Redwood City,

pp. 9 31.

Ashford, N.A., 2002. Government and environmental innovation in

Europe and North America. American Behavioral Scientist 45

(9), 1417 1434.

Berg, C., 1995. Technologischer Fortschritt und okonomische Reg-

ulierung. Lang Publ., Frankfurt/Main.Boyd, R., Richerson, P.J., 1985. Culture and the Evolutionary

Process. The University of Chicago Press, Chicago.

Cavalli-Sforza, L.L., Feldman, M.W., 1981. Cultural Transmission

and Evolution: a Quantitative Approach. Princeton University

Press, Princeton, NJ.

Coase, R.H., 1960. The problem of social cost. Journal of Law and

Economics 3, 1 44.

David, P.A., 1985. Clio and the economics of QWERTY.

American Economic Review, Papers and Proceedings 75 (2),

332337.

Dosi, G., 1982. Technological paradigms and technological trajec-

tories. Research Policy 11, 147162.

Dosi, G., 1988. Sources, procedures and microeconomic effects

of innovation. Journal of Economic Literature 26 (Sept.),11201171.

Downs, A., 1972. Up and down with ecology. The issue-attention

cycle. The Public Interest 28, 3850 (Summer).

Ehrlich, P., Raven, P., 1964. Butterflies and plants: a study in

coevolution. Evolution 18, 586 608.

Erdmann, G., 2000. Innovation, time, and sustainability. Paper

Presente

Documents

2nd Order of Sustainability