IN SEARCH OF INDICATORS OF SUSTAINABLE DEVELOPMENT

Environment & Management

VOLUME 1
 IN SEARCH OF
INDICATORS
OF SUSTAINABLE
DEVELOPMENT

Edited by

Onno Kuik and Harmen Verbruggen

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

ISBN 978-94-010-5431-7 ISBN 978-94-011-3246-6 (eBook)
DOI 10.1007/978-94-011-3246-6

coverphoto:
© Bram de Hollander

Printed on acid-free paper

AU Rights Reserved
© 1991 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1991
Softcover reprint of the hardcover 1st edition 1991
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
 CONTENTS

Preface vii

1. Indicators of sustainable development: an overview 1

Harmen Verbruggen and Onno Kuik

2. Towards sustainable development indicators 7

Hans Opschoor and Lucas Reijnders

3. Note on the correction of national income for environmental losses 29

Roefie Hueting and Peter Bosch

4. GNP and sustainable income measures: some problems and a way out 39
Hans Opschoor

5. Natural Resource Accounting: State of the art and perspectives 45

for the assessment of trends in sustainable development
Jaap Arntzen and Alison Gilbert

6. The predictive meaning of sustainability indicators 57

Leon Braat

7. The AMOEBA approach as a useful tool for establishing sustainable 71

development?
Ben ten Brink

8. Towards sustainability: indicators of environmental quality 89

Helias Udo de Haes, Maarten Nip and Frans Klijn

9. Contours of an integrated environmental index for application in 107

land-use zoning
Joop de Boer, Harry Aiking, Ella Lammers, Vera Sol and Jan Feenstra

Notes on the contributors 121

List of workshop participants 125

 vii

Preface

In 1989 the Dutch government published a National Environmental Policy Plan (Dutch
abbreviation NMP). This NMP is based on the book Concern for Tomorrow. a
national environmental survey by RIVM (the National Institute of Public Health and
Environmental Protection). A major conclusion of the RIVM study was that emissions
of many pollutants had to be cut by 70 - 90 % in order to reach environmental
quality goals. The government accepted the RIVM analysis and consequently ClUTent
Dutch environmental policy aims at large reduction of pollutants. Another conclusion
of the RIVM study was that such high reduction goals would not be easy to achieve
by technological means alone, and that thus structural changes would be required.
These changes could eventually lead to sustainable development, which now forms the
major focus of Dutch government national environmental policy.

This being so, the Dutch government requested that RIVM in subsequent issues of
Concern for Tomorrow should investigate the options for sustainable development.
One first step has been to request the Institute for Environmental Studies of the Free
University, Amsterdam, to organize two workshops that would bring together scientists
from different disciplines to outline the options for "measuring" sustainable
development. The papers resulting from the workshops have been put together in the
volume now before you. I am convinced that the results of this research are
interesting beyond the borders of the Netherlands and I am therefore pleased that the
organizers of tlle workshops have taken the initiative to translate and publish the
papers. Comments from colleagues from various comers of the world are most
welcome and may be sent to the editors.

Leen Hordijk
Head Environmental Forecasting Office
National Institute of Public Health and Environmental Protection
P.O. Box 1
3720 BA BILTHOVEN
The Netherlands
 1

Indicators of sustainable development: an overview

Harmen Verbruggen and Onno Kuik

"Is this country's or that region's economic performance more sustainable in 1991
than it was in 1981?". Finding measuring rods to answer questions like this one, was,
according to Opschoor and Reijnders (Chapter 2), the loosely formulated objective of
the workshops that were organized in the fall of 1989 and early 1990 by the Institute
for Environmental Studies of the Free University of Amsterdam at the request of the
Netherlands' National Institute of Public Health and Environmental Protection (RIVM).
The papers presented at these workshops, which were attended by both scientists and
policy-makers, form the core of this publication. In this introductory chapter we try
to put the different contributions into perspective and highlight some of the topics that
were discussed at the workshops.

The workshops were organized because, although 'sustainable development' is

becoming a key concept and even a goal in Dutch and international environmental
policy, there are no measuring rods or yardsticks to measure practical policy initiatives
against this goal. Unless there is some clear measure or at least some indicator of
sustainable development, the effectiveness of environmental or other policy towards
this goal can not be assessed. As ten Brink (Chapter 7) points out, it is not so much
that environmental information, on which a policy of sustainable development must
be based, is missing; it is the fragmentary, often qualitative and very detailed nature
of the information that hampers its direct usefulness in policy making. What we need
is adequate information that is tailored to quantitative environmental objectives.
"Adequate" means information:
- which gives a clear indication whether objectives will be met,
- on the system as a whole,
- of a quantitative character,
- understandable for non-scientists,
- containing parameters which can be used for longer time periods.
The search for indicators of sustainable development means the search for policy
relevant and coherent environmental information which adheres to these criteria. But
what exactly are indicators? In measurement theory the term indicator is used for
the empirical specification of concepts that cannot be (fully) operationalized on the
basis of generally accepted rules (Vos et ai, 1985). Their primary function lies in
simplification: indicators are a compromise between scientific accuracy and the
 2

demand for concise information. More specifically, indicators may be used for two
intertwined purposes (ibid.):
1. planning: problem identification, allocation of socio-economic resources and
policy assessment; and
2. communication: notification (warning), mobilization and legitimation of policy
measures.
There are other ways to bring the threats to sustainable development to the attention
of economic planners. The best way would be to reach a full integration of economic
and natural resource accounts. However, this is not readily achievable, due to lack of
data and as yet unsolved methodological problems. Given this state of affairs, it is
preferable to monitor sustainable development with a set of "quick and dirty"
indicators. As we have seen above, indicators are a compromise between scientific
accuracy and the demand for information.

Have the workshops been successful? Yes, inasmuch as they contributed to the mutual
understanding of policy-makers and scientists of each other's needs and possibilities
regarding sustainability indicators. Important questions regarding the type, dimensions,
scope and construction of indicators were raised and discussed. The quality of the
individual papers and the urgency of the subject led to the decision to present the
results of the workshops to an international audience. TIle workshops were not
successful inasmuch as we did not find the measuring rod to answer the sustainability
question; but this could hardly be expected at this stage where there is not even
consensus on the exact operational meaning of the concept of sustainable development.
This does not mean, however, that we think that indicator development should wait
until the last questions about this concept have been answered: we see indicator
development and the operationalization of sustainable development as a two-way,
cross-fertilizing activity.

What should be measured by indicators of sustainable development? To answer the

question whether the development of a region or a nation is sustainable or not, we
must first decide which developments are important for the sustainability issue. As
Opschoor and Reijnders (Chapter 2) point out, this question has scientific as well as
ethical angles. Do we just look at the sustainability conditions for the human species,
or do we include other species as well? Although this point was not completely
settled, most workshop participants agreed that sustainability should not be defined
from a purely functionalist or 'narrow' economic perspective. The ecological
sustainability or viability of economic development was stressed. In this interpretation,
the emphasis is on the preservation of prudently and stringently defined environmental
capital to be passed on 'intact' to future generations, as a development potential.
Environmental capital refers to the quantity and quality of the natural resource base
of a region. Indicators of sustainable development should take account of the
'integrity' of natural elements and structures, and of the 'diversity' of species and
systems.
 3

To be more precise, a measure of sustainable development should include indicators

for the pressure from society on the environment (pollution, resource use), and
indicators for the state of the environment (ecological integrity or bio-diversity). Both
(sets of) indicators should confront actual flows or states with sustainable flows or
states. Opschoor and Reijnders (Chapter 2) and Hueting and Bosch (Chapter 3) take
these sustainable flows and states as exogenous to the indicator development process:
sustainable flows and states at some point in time must be defined by scientists or
by the policy process. This sustainable 'reference' is, of course, critical to the validity
of the indicators. Opschoor and Reijnders make many valuable suggestions on this
subject. The requirement that this 'reference' should be 'beyond doubt and dispute'
(Braat, Chapter 6) seems to be aimed too high to have much practical meaning.
National Income Accounts, for instance, have been disputed as long as they exist;
nevertheless, they have served many practical purposes.

Braat (Chapter 6) advocates a somewhat different approach in which actual

developments in the man-environment system are modelled in a dynamic simulation
model to generate future values for selected socio-economic and environmental
variables. The future pattern of development of these variables can then be assessed
to be sustainable or not. Braat points out that there is not one sustainable future, but
many different ones with different levels of key-variables (population, income,
environmental quality). From the set of possible sustainable futures, a set of
acceptable sustainable futures must be selected. This selection must be left to the
political system. This is a challenging programme, indeed. However, given Braat's
general requirements that pertain to indicators -they must be attractive and
representative, have a scientific basis, and be quantifiable- it is this last requirement
(quantifiable) which, at least in the short term, seems difficult to meet.

The same applies to an approach which is based on Natural Resource Accounting

(Arntzen and Gilbert, Chapter 5). Natural Resource Accounting seeks to correct or
supplement the system of national (economic) accounts with information on the
quantity and quality of natural resources, on the basis of the correct view that a
nation's welfare is not only dependent upon the quantity and quality of man-made
capital and human resources, but also on its stock of natural resources. Arntzen and
Gilbert give a very readable account of the attempts that have been made in various
countries (e.g., Norway, Indonesia) to integrate natural resources in the prevailing
system of national accounts. However stimulating their discussion, their conclusion
points to the fact that it cannot be expected that Natural Resource Accounting will
in the short term succeed in fully integrating the environment into the economic
accounts.

Not all is said and done on this particular approach, however. Hueting and Bosch
(Chapter 3) quote an Indonesian minister who, being confronted with the theoretical
problems in constructing a money-measure for environmental losses, remarked: "If a
 4

theoretically sound indicator is not possible, then find one that is rather less
theoretically sound." The approach advocated by Hueting and Bosch is, as they call
it, a practical solution for a theoretical dilemma. The theoretical dilemma -the
problems of a monetary valuation of losses of environmental functions- is explained
both by Hueting and Bosch and by Opschoor (Chapter 4), but their conclusions differ.
Opschoor rejects the monetary approach and focuses on dimensionless indicators
(fractions, percentages). Hueting and Bosch stick to a (somewhat simplified) monetary
approach. Their disagreement is not so much theoretical, but has a practical origin.
They tend to disagree on the expressiveness or attractiveness of the different formats
(monetary versus non-monetary) for policy-makers, and especially for economic
planners. This is an important issue; it would take more social research to settle the
question whether and to what extent economic planners are more impressed by
monetary or non-monetary information about the environment.

These fundamental questions and possible approaches are dealt with in the first five
chapters of this Volume. The next three chapters are of a more practical nature. In
Chapter 7, ten Brink describes an ecological indicator which was developed for, and
is actually used by, Dutch water authorities, for the purpose of formulating and
evaluating goals for the biological component of water systems. Ten Brink argues that
this indicator -the so-called AMOEBA- could well be used in a wider ecological
context to provide a useful indicator for sustainable development. Starting point of the
AMOEBA approach is the formulation of a reference situation for a specific
ecosystem. This is a situation in which the system has not at all, or only slightly,
been influenced by human activities. The assumption is made that the closer one
comes to the point of reference, the larger the guarantee for ecological sustainability.
The search for ecological objectives -or sustainable development- can be reduced to
the question: what is the maximum acceptable distance to the point of reference? The
reference is expressed in numbers, distribution and/or health of a number of selected
species. AMOEBA depicts reference, objective, and present-day situation in an
attractive graphical form.

In Chapter 8 Udo de Haes et al. discuss the possibilities of quantifying ecological

objectives -sustainability standards- for terrestrial ecosystems. The situation for
terrestrial ecosystems is more complicated than that for aquatic ecosystems because
for man-made ecosystems, e.g., agricultural land, the 'reference' cannot be found in
undisturbed nature. An environmental quality target must be constructed. Of course,
this target has to relate to the present or potential functions of the area considered.
On the highest level, a distinction is made between cultural and natural areas. Further
distinctions are made on the basis of a hierarchic classification of ecosystems. Targets
must be formulated for 26 ecodistrict types in the Netherlands. Targets must relate to
physical, chemical and biotic variables. These variables must have relevance for
environmental policy, be predictable and measurable, and, preferably, have some
 5

appeal to the general public and/or policy-makers. Udo de Haes et at. present a
practical example of their approach for a lowland peat area.

In the final chapter, de Boer et ai. discuss the development of an integrated

environmental index, combining air pollution, noise, odour and risk of calamities, for
use in land-use zoning. The problems that were faced in developing this index may
provide lessons for the development of sustainability indicators. Firstly, de Boer et
ai. have combined empirical data and value judgments in the construction of their
index. In the preceding chapters of this Volume it was made clear that this
combination is basic to the development of sustainability indicators. Secondly, the
chapter shows how the development of an index (or indicator) is shaped by the
particular (policy) context for which it is intended. The authors distinguish between
two methods for developing the index: a scientific method, based on health effects
and a policy-oriented method, based on target values of quality standards. The
difference in approach resembles the difference in approach of Braat versus Opschoor
and Reijnders. It is, therefore, interesting to see how de Boer et ai. evaluate both
approaches. Firstly, they conclude that it would be premature to choose a method
without additional research. Secondly, they point to the fact that, although preferred
from a scientific point of view, the scientific approach can easily become deceptive
because of all kinds of (necessary) simplifications. From a policy-oriented point of
view, an index based on target levels of legal standards would be preferable. The
authors doubt, however, whether such an index would give a meaningful representa-
tion of the combined influences on the quality of living conditions. Of course, these
conclusions can be easily transferred to the area of sustainability indicators. It seems
that we have to keep navigating between Scylla and Charybdis: on the one hand
indicators which are conceptually sound, but which can be deceptive because of the
simplifications that are forced by our limited knowledge of man-environment inter-
actions; on the other hand indicators that are more or less, in Hueting's words,
practical solutions to theoretical dilemmas.

Nonetheless, all participants to the workshops agreed that the development of

indicators for sustainability should be vigorously pursued, preferably along the
following lines:

1. Indicator development needs to be a multidisciplinary effort. Integration of

contributions from natural and social sciences on this subject is essential.

2. It would be ideal to integrate natural resources into national economic

accounts. Monetary valuation of natural resources is desirable, in the first
instance to try to internalize natural resource use in economic decisions. It
seems, however, that this ideal cannot comprehensibly be achieved in the
foreseeable future.
 6

3. A complete set of Satellite Accounts of natural resources in physical

dimensions would also be very desirable. Although efforts will and should
continue to develop these accounts, it is unlikely that these efforts will show
results in the near future either.

4. In the short term, therefore, (physical) indicators seem the most promismg
option. Indicator development and the development of natural resource accounts
(in monetary and/or physical terms) must not be viewed as contradictory, but
should be seen as cross-fertilizing activities.

5. Indicators should include both causes and effects of environmental degradation

and resource depletion. These indicators could also be used in economic-
ecological modelling exercises. The workshop participants welcomed the
conceptual framework provided by Opschoor and Reijnders (Chapter 2).
Several partial approaches, for instance the AMOEBA approach, may well be
integrated into this approach.

We would like to conclude this introductory Chapter by recalling the old truth which
says that scientific progress is for only 1% brought about by genius, and for 99% by
sweat and hard labour. We hope that there is some genius hidden in the following
pages; but we are certain that it will take much sweat and hard labour to arrive at
a set of indicators which will more or less meet the objectives sketched above.

Finally, we would like to thank Mrs Dita Smit for the hard labour and genius that
went into the careful wordprocessing that made this publication possible.

References

Vos, J.B., J.F. Feenstra, J. de Boer, L.c. Braat, J. van Baalen (1985). Indicators for
the State of the Environment. R-85/l, Institute for Environmental Studies, Free
University, Amsterdam.
 2

Towards sustainable development indicators

Hans Opschoor and Lucas Reijnders

1. Introduction

The economy and the natural environment interact. The condition of one is of
importance to the other. On the one hand, economic activity is based on the continued
availability of sufficient material and energy resources and an environment that is
sufficiently clean and attractive. Insofar as the economy is based on renewable
resources, the proper functioning of natural processes and systems may become an
essential precondition for society's continuity. On the other hand, by discharging
pollution and by other features associated with human activities, society is interfering
with these environmental processes and systems. In this paper a first attempt is made
to arrive at a system of indicators of the condition of the environment in terms of its
capacity to sustain economic activity. Sustainability indicators reflect the
reproducibility of the way a given society utilizes its environment. Hence, they differ
from classical environmental indicators: they do not simply reflect environmental
conditions or the pressures on the environment, but they indicate to what degree
certain pressures or environmental impacts the earth can deal with in a long-term
perspective, without being affected in its basic structures and processes. We refer to
this capacity of the environment as 'ecological viability'. In a sense, therefore,
sustainability indicators are normative indicators: they relate actual, 'objective'
developments to a desirable condition or goal.

Loosely formulated, the objective of this exercise is to find measuring rods that can
assist researchers and policy evaluators in answering questions such as: "is this
country's or that region's performance more sustainable in 1991 than it was in
1981 ?". Hence we need to start from some presupposition of what these measuring
rods ought to look like (Section 2), and we need to consider the implications of
applying them to specific countries or regions (the imputation problem of Section 3).
We will come face to face with the need to give substance to the notion of ecological
viability (Section 4) and how this is to be defined when taking into account
possibilities to replace natural assets by man-made ones (the substitution problem of
Section 5). Sections 2 through 5 relate societal developments and environmental
change to various notions of sustainability. We shall subsequently touch upon two
 8

more technical issues (the aggregation and indicator construction problems, Section
6) and finally present some examples of actual or possible sustainability indicators
(Section 7).

2. Environmental Indicators and Sustainability Indicators

For the purposes of this Chapter environmental indicators can be defined as quantita-
tive descriptors of changes in either (anthropogenic) environmental pressure or in the
state of the environment. The former type of environmental indicators will be referred
to below as 'pressure indicators' and the latter as 'environmental effect indicators'.

Environmental pressure indicators express (changes in) the amounts/levels of

emissions, discharges, depositions, interventions, etc. in a predetermined region. The
pressures exerted by society on the environment are commonly categorized as follows:
a) pollution, b) overexploitation of resources, and c) landscape and ecosystem(s)
and/or organisms modification. Pollution entails the introduction into the environment
of substances or energy residuals that (may) have a negative impact. Overexploitation
refers to ways and levels of 'cropping' or 'harvesting' natural resources, so that their
future supply is at risk. Modification of ecosystems and landscapes can take the form
of changes in physical structures in such a way or at such levels that the systems'
integrity is in jeopardy by, for instance, too large a reduction in size of groundwater
table changes. Even within small countries such as the Netherlands the number of
different types of pressures on the various environmental compartments (air, water
soil), resources, landscapes and ecosystems tends to become very large. Environmental
pressures can be regarded as structural or incidental shocks that are transformed and
tra?sported in a variety of natural processes (biological, chemical, hydrological,
atmospheric) manifesting themselves into changes in conditions in the environments
of various receptors. These receptors include human beings, populations of plants and
animals, resources, ecosystems, landscapes, and artefacts. The relevant environmental
conditions can be regarded as so many dimensions of the concept of 'environmental
quality': the potential(s) of the environment to satisfy demands by the various
categories of receptors.

Environmental effect indicators express the consequences of environmental quality

changes in terms of their effects on certain (predetermined) receptors, as enumerated
above. For human beings, for instance, effect indicators could include repercussions
on the pattern of welfare over time, and health indicators. For other species one could
monitor environmental effects by looking at qualities and sizes of populations, niche
size or biotopes. At the ecosystems' level, effect indicators could include integrity,
biological diversity and buffering capacities.
 9

When attempting to develop sustainability indicators, one has to make choices as to

the relevant types of environmental change. We opt for a broad approach, whereby
a wide range of receptors as listed above is taken as the starting point. This implies
an elaboration of the notion of sustainability beyond that of purely anthropocentric
'functionality' or utility (see Section 4). If, furthermore, one wishes to develop a
manageable set of indicators in terms of their numbers, one has to address the issue
of how to aggregate, or select from, the large number of actual environmental
(pressure or effect) changes (see Section 6). Starting from our initial summing up of
types of environmental pressure, we suggest that there are at least three areas for
which sustainability indicators are required: (i) pollution, (ii) resources, (iii) biological
diversity.

As was noticed in Section I, sustainability indicators are not simple 'state indicators'
but rather indicators of states vis-a-vis some reference situation; the latter could either
be some past environmental state l , or a future one that is regarded as more desirable
than the present. Sustainability indicators are thus more than mere state descriptors;
they are normative measures of the 'distance(s), between current states and the
reference situation. We are not merely interested in the amounts of acid produced over
a certain area or deposited in some region, but we wish to relate such indicators to,
e.g., policy-determined emission maxima or critical loads. These reference conditions
operate as 'Plimsoll-lines' (the metaphor is Herbert Daly's) and the sustainability
indicators measure the distance between the actual water level and the Plimsoll-line.
As such, information on these distances may help in answering questions such as: is
there scope for further economic development in a region? what is the urgency of
taking pressure alleviating measures? is society moving towards an unsustainable
pattern of economic activity? It goes without saying that there will be no single
Plimsoll-line in this case: it will be a set of standards and conditions, and
wlfortunately these various elements may in fact be interrelated, thus providing
analysts with a rather diffuse description of the reference situation. One way of
visualizing this multidimensionality is by using the 'AMOEBA' model (Chapter 7, this
Volume), where a circle's circumference depicts the reference situation and each
radius is a dimension of sustainability; actual situations are points on these rays that
will normally not lie on the circumference.

3. Environment and economy as open systems

An economy may be defined as a set of productive and consumptive activities (and

the actors involved therein) within certain territorial boundaries, but the environmental

J A good example of indicators using a past situation as a reference point is provided by ten Brink in his
notion of 'AMOEBA' (Chapter 7, this Volume).
 10

pressure and hence the question of an economy's sustainability draws attention to

environmental changes beyond these boundaries.

Environments are open systems linked by various processes of transportation,

migration, etcetera. The environmental impacts of economic activities may thus be
transferred through environmental processes from one place to another, even beyond
national boundaries. Acid deposition in Sweden as a consequence of combustion in
the UK is one example; a higher incidence of flooding in BangIa Desh due to
deforestation on the Himalayan slopes is another. Effects such as holes in the ozone
layer and the greenhouse effect are examples on a global scale of the often long
paths of pollutants through space, and the distances between sources and points where
effects manifest themselves. Harvesting activities by one economy on a shared
resource will affect the quality of the resource for all that use it. This phenomenon
of spatial interdependency has several consequences. Firstly, as the environmental
consequences of an activity may extend to other countries, the full sustainability
impact of that activity involves adding the environmental impacts in all affected
countries. Secondly, as a country's environmental quality may be burdened by influxes
of pollution from abroad, there may be a divergence between an economy's direct
burden on the environment on the one hand, and the change in its own environmental
capital on the other.

Disregarding recharge and regeneration, country A's environmental capital E(A) is

decreased by: (i) the influx M(B) of degradation from other countries B, and (ii) A's
own domestic environmental pressure P(A). A's total Environmental Pressure EP(A)
(i.e. the total of all environmental impacts of all relevant economic activities) equals
P(A) plus X(A), which represents all of A's impacts elsewhere. These variables are
related as follows:

dE(A) = P(A) +M(B) = EP(A) - X(A) + M(B)

This leads to a first moment of choice. Monitoring the development of an economy's

environmental capital E (i.e. the total of all resource stocks, including environmental
quality, within the territory), and monitoring the development of a country's EP, are
exercises leading to different outcomes. One has to choose between monitoring either
one, or monitoring both. We opt for the latter: monitoring both.

Furthermore, economies are normally open systems with levels of production and
consumption linked by international trade flows. In such cases, a part of the
environmental burden of a certain product or a certain activity in country A may
occur in country B. An example is the deforestation and soil exhaustion in Thailand,
associated with the cultivation of the cassava which, in the form of tapioca, is
exported as pig feed to Dutch intensive pig farmers (van Amstel et al. 1986).
Similarly, by exporting certain products, country A may itself face only part of the
 11

environmental problems associated with that product, namely the production related
ones; it will not suffer the use related ones in the country of destination, C (Venne
et al. 1989). An example is provided by drins or other pesticides that used to be
produced in Europe (with pollution consequences there) but were mainly used in third
world countries (with ecosystems damage and waste related problems in the
consuming part of the world).

This raises two further questions: (l) should these environmental consequences of
forwardly or backwardly linked foreign activities be aggregated to obtain a measure
of overall environmental pressure; and (2) if so, in country B or C where they occur,
or in country A where the activity is located that ultimately gave rise to these effects?
The first question arises when one considers the environmental impacts of a certain
activity to be the total of the impact of that activity itself and the impacts of all
backwardly and/or forwardly linked activities. In terms of a country's sustainability
this is no problem if all activities take place within the boundaries of that COWltry,
but this may not be the case. For reasons of putting responsibility for environmental
degradation where it intuitively belongs, one may in certain cases wish to add the
impacts in countries Band/or C to those of A. This might be the case where one
country deliberately uses the environments of other countries as additional resources
to its own, without the other country being in a position to dissociate from such
trade relationships (Opschoor 1989b). Examples of such asymmetric relationships can
be found in South to North trade in primary products. However, for practical reasons
it appears better to abstain from attempting to allocate the cumulative environmental
pressure (cumulated, that is, over the various stages in the product life cycle) to one
single country where one specific stage occurs. Various attempts at empirically doing
so (either by applying input-output techniques -even in single-cowltry situations- or
simply statistically analyzing trade flows in terms of countries of origin and the likely
environmental consequences of the production stage in those countries) have remained
relatively unsuccessful (James, Jansen and Opschoor 1978; Vos 1982, Venne et al.
1989). Ignoring the environmental consequences in country B implied in its exports
to A of intermediate or final products might result in too favourable an assessment
of the environmental pressure related to A's economic process, but on the other hand,
these consequences are the result of economic activities in B and will therefore show
up in that country's environmental account. And if reasonably symmetric relationships
in terms of market power prevail, then the country suffering the environmental costs
of a trade link may be asswned to have wilfully accepted it, as a price to be paid for
the overall gain through trade.

Based on practical considerations it is recommended to incorporate international trade

related transboundary redistributions of accwnulated environmental pressure only if
sufficient empirical data is available to incorporate these interrelationships.
 12
4. Environmental Viability

TIle requirement of environmental viability is an ambivalent one in at least two

respects: a) one may give it substance from an anthropocentric or an ecocentric point
of departure; and b) one may regard it either in a static or in a dynamic context.

To begin with the first ambivalence: recently a plea was made for ensuring 'sustain-
able development' (WeED 1987) defined as: a pattern of development that meets the
needs of the present generation without jeopardizing the ability of future generations
to meet their own needs. This definition is clearly anthropocentric. The notion of
sustainable development, however, is again not unambiguous (ct. Opschoor 1987 and
1989a). One can stress several aspects, notably the socioeconomic one and the
environmental one. Given a sufficiently long time horizon and certainty about all
relevant interactions between the economy and the ecology, the two would coincide;
but for all practical purposes, the two emphases reflect different policies. As time
horizons become shorter and as one is prepared to take more risks as to the
environmental repercussions (or their reversibility) of current activities, the two
interpretations may diverge:
(i) In "sustained economic growth" the emphasis is on economic growth within
some (often rather relaxed and imprudently defined) side conditions related to
environmental quality and resource utilization. Put in formal terms this interpretation
implies: positive growth rates of consumption per capita, either for all future periods
('strong SD', Pearce et al. 1988) or such that the net present value exceeds zero
('weak SD', ibid.);
(ii) Alternatively, the "environmental sustainability" or "viability" of development
could be stressed. In this interpretation, the emphasis is on the preservation of
prudently and more stringently defined environmental capital (natural resource base
plus environmental quality) to be passed on 'intact' to future generations, as a
development potential.

Here, the latter interpretation is followed (and expanded below). This implies the
continued presence of healthy and productive resource regeneration systems and
adequate inputs (qualitatively and quantitatively) to allow these systems to function
from at least an anthropocentric perspective. In order to render the notion of
sustainability more operational, one has to be explicit about the exact nature of
environmental quality, E for short. E can be regarded as a set of individual resource
stocks and environmental quality levels, and we can say that 'sustainability' means
that the time derivative of each element of the vector E must be positive or zero.
Alternatively, we can seek some sort of aggregate A (or vector R with fewer elements
than E) in physical terms (e.g., aggregating energy resources by expressing them in
oil equivalents or Joules) and require that dNdt (or dR/dt) be positive or zero.
 13

From the point of view developed so far, constraints on environmental exploitation

can only be relevant if they follow from a logic based on an intertemporally extended
societal self-interest. However, sustainability constraints may also result from other
perspectives. For instance, ethical views may lead to restrictions on environmental
exploitation based on the 'rights' to existence and development of nonhuman species
and natural systems. These rights would then in fact curb human use rights.
Effectively this might imply placing a value on the 'integrity' of natural elements and
structures, and on the 'diversity' in terms of species and systems (so-called 'biological
diversity'). Integrity or diversity can to some degree be regarded as functional from
an economic, anthropocentric perspective (e.g., the potential direct economic value of
species the qualities of which have not yet been investigated or discovered, for e.g.,
nutritional or medicinal purposes). Secondly, diversity and integrity can to a large
extent be regarded as a precondition for sustainability, as has been argued above. But
these two functionalist lines of argument might not be sufficient to protect all possible
species and ecosystems. For example, it could well be argued that the sustainability
of the biosphere (or, for that matter, of China) would not be at stake if the Panda
bear were to become extinct. A case could be made (on noneconomic grounds) that
integrity and diversity are of a significance beyond the domain of a functionalist
approach; as noted earlier, we accept that case. That means that we prefer to define
'sustainability' broadly, i.e. from a position with due regard for the 'interests' of other
species. This gives additional support for one proposal made earlier, namely that a set
of indicators for (or 'reflectors' of) environmental quality change ought to include one
or more indicators for biological diversity or ecosystem integrity, related to some
reference situation.

The second point to deal with is the choice between a static and a dynamic
interpretation, especially when attempting to establish operational definitions of
biological diversity. From a geological and evolutionary point of view, one cannot
really defend taking any past situation as a reference point. Rather, one could think
in terms of a steady state for resource stocks or resource stock potentials, and for
natural cycles and ecological ('life support') processes, providing an ever preserved
base line situation from which evolutionary development can take place. There still
are choices to be made then, especially as to the choice-specific species and systems
for which conditions are to be ensured. Based on notions of prudence (or precaution)
there is, moreover, a case for defining reference conditions safely beyond the minima
as defined by the current state of knowledge. A further point to decide relates to
spatial distribution. Species and ecosystems development does not necessarily have to
be guaranteed at every point in space where they happen to be present today. Local
extinction is not necessarily incompatible with sustainability (broadly defined).
National or regional authorities will have to define to what degree it is desirable to
maintain current levels of biological diversity. In the Dutch situation, for example, we
would argue that a further reduction of ecological integrity is unacceptable, and hence,
 14

that a base line of environmental/ecological conditions is established such that this

integrity is safely ensured.

5. Substitution

In a functionalist perspective it is conceivable that one natural resource replaces the

other (e.g., sugar cane as a substitute for fossil energy). When resources can thus be
replaced, the unsustained use of a particular resource may not pose a problem in
terms of economic survival of the activities using that resource. Substitution
possibilities may also exist between natural resources and nonnatural ones, such as
produced capital, knowledge and know-how. Hence, technological development or
innovation may lead to an expanded range of options for substitution of one resource
for another, and hence may ease the problematic nature of the sustainability issue.
This fact implies that it may in actual practice prove complex to give concrete
substance to the notion of 'sustainability'. It may be more advantageous to sell and
burn up a given fossil energy resource, and invest the revenues in the development
of alternative natural resources or even of artificial ones, than to preserve the fossil
resource.

Economists, in substantiating the notion of sustainability, often feel inclined to

transform all natural resource into one aggregate economic value V (expressed in
monetary terms) by using existing or calculated resource prices. This entails the ideas
(0 that correct values for each resource exist or can be determined, and (ii) that
depletion plays no role or can be neutralized by substitution of one natural resource
for another. This would then lead to the condition for sustainable development that
dV/dt>O. Some economists (e.g., Solow 1986) have gone to the extreme of requiring
no more than the nonnegativity of some value aggregate of all forms of capital,
natural or man-made. This would imply the possibility of substitution, without
constraints, of produced capital or even knowledge for natural assets. This may be
considered the economist's equivalent of a perpetuum mobile. Pearce et al. modify
the latter approach: they add a set of (physical) constraints on the use over time of
certain essential stocks ER (these constraints may imply a 'critical minimum stock'
approach). At first sight, this is an attractive' modification; however, environmental
considerations might in fact turn ER into a set with a rather substantial number of
elements, which makes it less convincing or effective. Moreover, the set of prices
required to carry out the transformation into value is notoriously lacking due to
market imperfections. In Section 4, we advocated an approach in which the elements
of ER would be selected from a noneconomic perspective, which in many cases
implies that reference to market prices becomes irrelevant.

Still, an operational definition of sustainability requires an answer to two related

questions: a) over which span of time do we wish to ensure sustainability, and b)
 15

how does one deal with proposals that allow for the substitution of man-made assets
for environmental resources? The two questions are related to the extent that future
research and development may broaden the scope for such substitution considerably.

The first question may be answered by the requirement that a set of physical stocks
and conditions is handed over that at least ensures economic and evolutionary
development potentials at their current levels. This means maintaining (or enhancing)
the quality of the present environmental infrastructure and biological diversity, or (at
least) a steady state in terms of all essential environmental structures and processes.
Examples are: putting no more fertilizer on agricultural land than is taken up by
crops and livestock; depositing no more acid than present ecosystems can safely
absorb or buffer; cropping trees at or below maximum sustainable yield.

The second question in our view leads to an approach of prudence or risk aversion
when it comes to assessing the possibilities for substitution of one resource for
another and for science's and technology's capacities to continuously render new
substitution options. This can be incorporated into the Pearce approach. Stocks of
resources will then only include proven stocks, and only proven new technologies
using resources sustainably will be accepted as modifying the dependency of natural
resources and hence the sustainability situation. For essential renewable resources this
approach entails that: (i) the stock levels to be maintained must be high enough to
safely ensure optimal sustainable offtake, and (ii) the quality of the regenerative
systems instrumental in regrowth processes be maintained beyond safe minimum levels
of environmental standards.

6. Selecting and Building Indicators

The notion of indicators is an old one. OECD began work on Social Indicators in
1970 (Fox 1987), which resulted in a list of 33 specific indicators grouped under
eight headings (OECD 1982). One of these is: "Physical Environment". The indicators
listed there that relate to the natural environment are: (i) Exposure to Air Pollutants,
and (ii) Exposure to Noise. It is clear that this does not adequately cover the
environment-economy interactions we are concerned with here. A new set of
Indicators for Sustainability needs to be developed. In the remainder of this Section,
the focus will be on indicators more directly related to environmental quality and
stocks. We will comment briefly on the following points: (i) the areas for which
indicators would be needed; (ii) the indicators' scope, (iii) formal features, and (iv)
the process of developing indicators.
 16

Indicator Areas

Indicaturs would have to be derived from the specific characteristics of the economy-
environment system. Some have argued in favour of one single overall indicator of
environmental capital but it is felt that at least 3 are needed:
* pollution;
* resources: renewable, nonrenewable (and semi-renewable); and
* biological diversity.
We base this on the following considerations.

In a functionalist perspective one can distinguish between several types of sources

of environmental services:
(i) nonrenewable resources such as oil reserves, iron ore deposits; no
substantial natural augmentation or renewal takes place;
(ii) renewable resources such as forest stands, fish populations, agricultural
crops, where regeneration of the resource is a function of initial stock
levels and the quality of the regenerative systems;
(iii) semi-renewable resources such as soil fertility, solar influx, rainfall and
groundwater levels: natural processes provide -at a given moment in time
and at a given point in space- a recurrent but limited -and often uncertain-
supply, where this supply is not a function of initial (stock) levels but
sometimes of other environmental factors.
Ideally, sustainability indicators reflect this variety in circumstances and resource
types. In what follows we develop some resource indicators especially for renewable
and nonrenewable resources.

The absorptive capacities of an ecosystem for pollution and disturbance fall under
category (iii) of resources, but are normally analyzed separately from the perspective
of environmental pressure and environmental effects. Below, we shall follow this
practice and develop some pollution related sustainability indicators.

A third category of sustainability indicators we propose to use has no necessary

relationship with economic functions of the environment vis-a-vis society, but has to
do with the need to monitor ecological integrity or the 'naturalness' of landscapes and
ecosystems by reflecting (changes in) biological diversity (i.e. both species diversity
and ecosystems diversity). The rationale for including this aspect of environmental
change has been given in earlier Sections.

Indicator Scope

The literature on indicators of sustainability has produced a wealth of dimensions

that would have to be incorporated in them (e.g., Liverman et al. 1988), including
 17

(apart from 'sustainability') efficiency and equity. It can be argued that 'integrity'
and 'manageability' are important dimensions as well. It is suggested here that future
work will focus on the notions of sustainability, integrity and manageability. Thus,
indicators would be required with a scope wide enough to reflect:

a) the factual developments in the use of environmental resources; here, one may
be interested in establishing macro indicators of:
a.1) the overall environmental pressure EP, i.e. the environmental impacts of
economic activities and environmental management (for instance,
aggregates of pollution, resource inputs, spatial claims);
a.2) the change of environmental capital E, or the 'state of the environment'
(aggregate indicators for stocks of environmental assets, ambient
environmental qualities, biological diversity).

b) the potentials for management towards sustainability; this would, for instance,
require indicators for:
b.1) current or anticipated development in science and technology in terms of
environmentally relevant products, processes, inputs;
b.2) the development of managerial tools, such as: appropriate institutions for
environmental resource management, policy instruments, budgets, public
support.

In what follows we shall concentrate on sustainability and integrity (biological

diversity), and bypass the subject of indicators of manageability, important though it
is. In other words, a.l and a.2 are elaborated here.

Indicator Features

In order to facilitate international comparison, indicators would have to be formated

identically or analogously as much as possible. It is imperative to stress the need to
severely limit the number of indicators, if they are to play a part in public decision
making.

Liverman et ai. (1988) have considered criteria that could be used in selecting
indicators:
- sensitivity to change in time
- sensitivity to change across space
- sensitivity to change over social distribution
- sensitivity to reversibility
- sensitivity to controllability
- predictive ability
- integrative ability
 18
- relative ease of data collection
- relative ease of application.
It may not be possible to incorporate all these criteria adequately, or to do so while
at the same time observing the recommendation that only a limited number of
indicators is built.

Indicators: Aggregates or Selections

An important choice regarding formal features of indicators has to do with the nature
of the indicators: are they to be true aggregates or transformations of underlying,
more specific indicators, or are they to be selected ('typical' or 'representative' or
'critical') from larger sets? It is felt that the former is preferable but not always
feasible, due to the vast amount of information necessary to duly reflect all relevant
environmental processes and due to difficulties in formal aggregation.

Data reduction is a necessary step in indicator development. Sometimes one can

relatively easily aggregate (e.g., when adding energy resources in terms of caloric
value or oil equivalents, or in using acidification properties in combining S02' NOx
and NH) emissions or depositions). Where such transformations using physical or
chemical properties can no longer be made, economic weighing practices may be
considered. Use could be made of market prices or stated preferences to add otherwise
incomparable phenomena, but we have doubts as to the stability and acceptability of
such procedures (see Chapter 4).

Sustainability indicators ideally provide insight into factual developments in the

environment vis-a-vis certain reference values reflecting objectives or past values
considered to be more desirable. This provides another possibility of aggregating the
large number of individual indicators: determining the ratio of current environmental
conditions and the corresponding reference values, and using mathematical techniques
to aggregate these dimensionless figures in some way. Multicriteria analysis might be
of great value here.

Indicator Development

The development of appropriate sets of such physical indicators is a laborious

undertaking and is likely to involve many 'arbitrary' decisions (often based on
pragmatic grounds such as data availability) on which variables to select and how
to aggregate them (see previous paragraph). Logical steps in a process of deriving
indicators would be:
 19

1. identification of the main natural elements of environmental capital and their

interactions: ecosystems, life support systems, biogeochemical and hydraulic
cycles, biological diversity, habitats, and the levels of integrity (completeness,
'naturalness') and purity (degree of pollution);

2. identification of the economically relevant features within these elements and

their relationships to specific economic activities (either as inputs into, or
receptors of outputs of, these activities);

3. selection of those elements that are quantitatively and/or qualitatively at risk,

and a further analysis of these elements in terms of: a) their significance in
regenerative and resource support systems, and b) substitution options for these
resources in economic activities;

4. setting of standards/targets/critical levels with respect to the elements selected

in 3., in relation to the notions of sustainability and minimum biological
diversity to be maintained;

5. construction of indicators reflecting the development of environmental capital

from the elements selected in 3., either by building aggregate variables or by
picking specific items from that set of elements.

In step 5, one could opt for several possibilities:

a) express the selected items as rates or flows;
b) express them as rates-to-stocks;
c) if step 4 has been completed: express them as rates-to-goals or stocks-to-goals.
Options b) and c) would enable further data reduction by turning the indicators into
dimensionless figures. The procedure outlined here would preferably be fed into an
information system linked with multicriteria methods, so that alternative choices in
the various steps and alternative weights could be followed through in terms of their
impact on the indicator values.

7. Examples of Sustainability Indicators

In this part of our paper we would like to present some examples of sustainability
indicators. We will restrict ourselves essentially to indicators reflecting pressure on
the environment. Examples of indicators relating to environmental effects will be
given in the papers of Udo de Haes and ten Brink in this Volume. Our examples
will mainly relate to deviations from a steady state.
However, we will also consider the case that an initial state cannot be considered
acceptable in view of sustainability, in which case a steady state continuation of the
initial state is not sustainable.
 20
Maintenance of a steady state is one of the operational definitions of sustainable
development. A steady state is a dynamic state in which changes tend to cancel each
other out. An example of a steady state is a constant atmospheric concentration of
carbon dioxyde (C0 2). Such a constant concentration is the net result of a sizable
emission and an equally sizable sequestration of CO2•
A steady state operationalization of sustainable development is appropriate (1) if the
initial state is acceptable in view of sustainability; (2) if there is no time lag between
changes in environmental parameters and effects (as occurs in the case of atmospheric
COJ; (3) if sustainable development refers to many generations, and, (4) if at the end
of the day there is no substitution of physical resources (like groundwater or ores) by
nonphysical resources (like money or inventions).

Maintenance of a steady state in terms of resources, species and pollution would

imply the following:
- use of (conditionally) renewable resources should - within a specified area and
time span - not exceed the formation of new stocks. Thus, for instance, yearly
extraction of groundwater should not exceed the yearly addition to groundwater
reserves coming from rain and surface water.
- use of relatively rare nonrenewable resources, such as fossil carbon or rare
metals, should be close to zero, unless future generations are compensated for
current use by making available for future use an equivalent amount of
renewable resources.
Thus, for instance, the use of rare metals like lead, indium or copper should be
subject to virtually complete reuse. Dispersive use of lead in petrol would, for
instance, violate this criterion.

Also, use of fossil carbon would be acceptable, provided that, for instance, an
equivalent amount of biomass or other capturing devices for solar energy are put
aside for use by future generations, and future generations are compensated for shifts
in exploitation of fossil carbon following from exhaustion of convenient fossil carbon
sources by this generation.
- Significant, though limited, use of relatively abundant nonrenewable resources
such as iron or aluminum meets the steady state criterion, provided that there
is compensation for an increase in exploitation efforts following from
exhaustion of easily accessible and minable resources by this generation.
- Pollution that gives rise to accumulation of pollutants in one or more
environmental compartments (e.g., atmosphere, sea, soil) in a first
approximation violates a steady state operationalization of sustainability. The
same holds for long-lasting pollution (for instance, groundwater pollution,
radioactive pollution around Czemobyl), the safety of which is not established.
Exposure to man-made mutagens affecting the germ line (involved in
reproduction) should be close to zero. Violations of these first approximations
 21
to a steady state may be acceptable if future generations are fully compensated
for associated damages.
- As to natural species in a first approximation the rate of extinction of species
should not exceed the rate of origin. Additional steady state requirements may
relate to diversity of ecosystems, integrity of ecosystems and the conditions
for development of ecosystems.

A steady state operationalization of sustainability is rather strict. It is possible to use

less stringent criteria for sustainability, for instance, depending on the perceived
absence of unacceptable harm from pollution or extinction of species if it remains
below specified levels.

Thus, for instance, Krause et al. (1989) have suggested a nonsteady state criterion
for sustainable global warming. Whereas a steady state approach would require
constant temperatures, they propose an upper level to global warming of O.l°C per
decade and an overall man-induced equilibrium warming of 2.5°C, because such a
warming would be compatible with adaptive possibilities of species and would remain
within past natural fluctuations in the presence of homo sapiens.

Indicators for conformity to or deviation from a steady state

Indicators for sustainability should first and foremost indicate whether or to what
extent a criterion for sustainability is met. The extent to which a criterion for
sustainability is met may, however, also show temporal change, and more specifically
a trend. Deviation from a steady state, for instance, may increase, decrease or remain
roughly constant with time. Such a trend may also be reflected in an indicator. So,
starting from a steady state operationalization of sustainability, one may define two
indicators. First: a sustainability indicator indicating whether or to what extent the
steady state criterion is met at a specified point in time or over a specified time span.
Second: an indicator that reflects the temporal trend with respect to a steady state.
Later we will introduce a third kind of indicator, which is appropriate in situations
in which the initial state cannot be considered acceptable in view of sustainability.
This we will call the sanitation indicator.

An indicator reflecting conformity to maintenance of a steady state is defined to be

positive or zero when the steady state criterion is met. A positive value is given
when there is improvement; for instance, when use of renewable resources is smaller
than addition to stocks or when pollution levels decrease. An indicator reflecting a
temporal trend with respect to a steady state is held to be zero or positive when the
trend leads to conformity to a steady state either consistently or in due course. An
indicator indicating conformity to a steady state will be negative when the steady state
criterion is not met; for instance, because pollution levels increase. An indicator
 22

reflecting a temporal trend will be negative when developments do not lead to

conformity to a steady state.

Indicators reflecting pressure on the environment may differ in spatial scope. For
instance, the scope may be worldwide, continental, national or regional, or refer to
a river basin. Values, including signs, of indicators may be different dependent on
their geographical scope. Thus, for instance, although worldwide there is a loss of
forests, violating the steady state criterion and thus presumably giving rise to a
negative value for the sustainability indicator involved, particular countries may
expand their forests, and this may be reflected in positive values for national
sustainability indicators.

Applications of the steady state criterion may give rise to a large variety of
environmental pressure indicators. Here we would like to give some examples of
such indicators, and the signs these may have.

- The rate at which natural species die out is currently worldwide roughly 106
times the rate of origin of species (May, 1988). Thus an indicator reflecting
conformity to a steady state development should in this case be strongly
negative. (It would be zero if extinction rates equalled rates of origin.)
Estimates on the impact of business as usual up to the end of this century
suggest that the rate of extinction (as a percentage of all remaining species)
may remain roughly unchanged or even increase (Myers, 1979; May, 1988).
Thus a sustainability indicator reflecting a temporal trend relative to a steady
state development should also be strongly negative.
- The rate of increase in atmospheric concentrations of fully halogenated
chlorofluorocarbons (CFC's) is estimated to have varied between 4 and 16%
by the end of the 1980s (Prather and Watson, 1990; Reijnders and Kroeze,
1990). Because these CFC's currently cause significant and increasing
deterioration of the ozone layer, an indicator reflecting current conformity to
a steady state development will be negative. However, a 1990 London
agreement on protection of the ozone layer aims at a phase-out of CFC
production in industrialized countries by the year 2000 and in developing
countries by 2010. Although actual deterioration of the ozone layer is also
dependent on other halogenated compounds, a phase-out of CFC production
may contribute to stabilization of damage to the ozone layer and even to a
final recovery of the ozonesphere. Thus a sustainability indicator reflecting the
relevant temporal trend in conformity to a steady state development may be
positive.
- Forested areas in the Netherlands are expanding. Currently there is a net
increase in the amount of recoverable wood produced (addition to stock minus
exploitation and die-back). Thus the indicator reflecting conformity to a steady
state development may be considered positive. Long-term perspectives for
 23

forests in the Netherlands, however, are poor. Increasing acidification of soil

threatens 'S,)% of current forests with die-back (Nationaal Milieubeleidsplan,
1989), whereas a rise in atmospheric temperature may also negatively affect
Dutch forests. Thus the relevant indicator reflecting the temporal trend is
probably negative.
- Use of fossil carbon in industrialized countries is far from zero. Of total fossil
carbon used probably less than 1-2% is recycled, whereas the remainder is
transformed into wastes, including pollutants. There is some compensation for
future generations through the planting of additional forest and the
development of solar and wind energy, but overall there is no conformity to
a steady state operationalization of sustainability. Thus the indicator reflecting
compliance with sustainability of fossil carbon use should be negative. For the
near future efforts have been announced by several countries to improve
energy efficiency and increase recycling and the use of renewable sources of
energy. Although this will not lead to compliance with a steady state use of
carbon in the near future, the indicator reflecting the trend in compliance may
be less negative than the indicator reflecting current conformity to a steady
state development.

Dimension and size of indicators reflecting conformity to a steady state

If one sticks to an environmental capital approach to sustainability, it will not be

possible to use the same dimension for all indicators. There is no sensible physical
transformation that transforms, for instance, the dimension species (with which a
steady state of living nature may be partially defined) into dimensions referring to use
of fossil carbon or acidification of soil. However, there is the possibility of using the
same dimension for groups of environmental variables. For instance, the use of
nonrenewable resources may be related to presumable reserves, and may be measured
with similar units (for instance, yearly use as a percentage of presumed total reserves).
Similarly, greenhouse gases (N 20, tropospheric ozone, methane, carbon dioxyde and
a number of halocarbons) may be lumped together on the basis of their global
warming potential (for instance, in W/m or °C). Similarly SO.. NO x and NH 3 may
be brought together on the basis of their acidifying effect on soil and surface water.
Table 1 gives a number of examples of possible environmental pressure indicators,
their aggregation levels, dimensions and sizes.
Table 1. Dimension and size of possible environmental pressure indicators reflecting compliance with a steady state

object dimension size

use of renewable resources % of total stock added or lost addition to stock - use - loss
in a specified time span and area total stock

use of nonrenewable % of presumable reserves lost in minus use

resources a specified time span and area presumable reserves

species number or percentage of species species originated minus species become extinct
lost in a specified time span and area in a specified time span and area
(may be divided by total of species)

acidification of soil acid equivalents in a specified time span neutralization by soil minus acid deposition ~
and area in a specified area and time span

global warming W/m' or °C added in a specified time span combined amount of greenhouse gases lost
in sinks minus emission of greenhouse gases
multiplied by global warming potential

depletion ozone layer ozone depletion (in % or absolute) in a combined amount of ozone layer depleting
specified time span substance lost in sinks minus emission of
ozone depleting substances multiplied by
ozone depleting potential

soil pollution amount quantity of pollutant in a specified time amount of a pollutant eliminated from soil
span plus made inactive minus amount added to soil
 25
Sanitation indicators

In a number of cases the situation (initial state) is such that a steady state with
respect to this situation cannot be considered sustainable. Examples of current
situations that are unacceptable from a sustainability point of view are chemical waste
dumps and the hole in the ozone layer over the Antarctic. In such cases an indicator
may reflect the extent to which sanitation is necessary before the situation may be
considered sustainable. In the case of soil polluted by chemical waste a sanitation
indicator may reflect the amount of polluted soil which is unfit for multifunctional
use. Such a soil sanitation indicator will remain negative until a sanitation programme
generating multifunctional soil from polluted soil is completed. Increasing the
deviation from a sustainable state will increase the size of the sanitation indicator. So,
for instance, further loss of ozone from the ozone hole over the Antarctic will make
the relevant sanitation indicator more negative.

8. Discussion and Recommendations

The development over time of 'environmental capital' or even of 'environmental

pressure' is not captured by traditional 'success indicators' such as GDP or National
Income. Several proposals have been formulated to amend this by correcting GDP,
but this approach threatens to be incomplete, or to lack transparency and credibility
(see Chapter 4, this Volume).

A notion of 'sustainable income' could be defined and perhaps even quantified, only
if so-called 'defensive expenditure', environmental stock depreciation, and the value
of remaining environmental degradation could be assessed. Placing monetary values
on them must be expected to continue to cause difficulties for the decades to come;
some barriers are of an ethical nature. Given this situation, but also due to a need to
explicitly know the quality of the environment, it is believed that there will always
be a need for indicators expressing the development over time of environmental
quality in physical terms.

Sets of environmental indicators should be developed at the level of resources or

activities within a given country. In order to playa part in 'merging environment and
economics in decision making' (WCED 1987) and policy development, the number
of indicators proposed must be small. These indicators will have to cover the areas
of: (i) resources (of all kinds: renewable, nonrenewable, semi-renewable ones), (ii)
pollution, and (iii) the biological diversity or integrity of ecosystems.

Indicators are to reflect developments vis-a.-vis net environmental pressure (EP) and/or
environmental capital (E). Preferably, they are of the types of rates-to-stocks, or rates-
to-goals.
 26

The scope of these indicators will have to be broad, ideally including observed
sustainability and integrity impacts, but also the potentials for managing economic
behaviour towards these objectives. The paper has not developed the notion of
indicators for managing capacities, but focused on the dimensions of sustainability
and integrity.

References

Amstel, A.R van et al. (1986). Tapioca for the Dutch Livestock Industry. IES Report
R-86(7, Free University Press Amsterdam.
Fox, K.A. (1987). 'Environmental Quality in a New System of Social Accounts".
Archibugi, F. and P. Nijkamp (eds.): Economy and Ecolo&y: Towards Sustainable
Development. Dordrecht/London: Kluwer Ac. Publ. 189-203.
James, D.E., H.M.A. Jansen and J.B. Opschoor (1978). Economic Approaches to
Environmental Problems. Amsterdam: Elsevier Scientific Publ.
Krause, E., W. Bach, J. Koomey (1989). Ener&y Policy in the Greenhouse. Er Cerrito.
Liverman, D.M., M.E. Hanson, BJ. Brown and R.W. Meredith, Jr. (1988). Global
Sustainability: toward measurement. Environmental Management Vol. 12, no. 2,
pp. 133-143.
May, RM. (1988). Science 241: 1441 - 1449.
Myers, N. (1979). The Sinkin& Ark. Pergamon Press, Oxford.
Nationaal Milieubeleidsplan (1989). Kiezen of verliezen. Staatsuitgeverij, Den Haag.
OECD (1982). The OECD List of Social Indicators. Paris: OECD.
Opschoor, J.B. (1987). Sustainability and Chan&e (in Dutch: inaugur.al address).
Amsterdam: Free University Press.
Opschoor, J.B. (1989a). No Delu&e After Us: Preconditions for Sustainable Use of
the Environment (in Dutch). Kampen: Kok Agora.
Opschoor, J.B. (1989b)"North South Trade, Resource Degradation and Economic
Security". Bull of Peace Proposals Vol 20 (2): 135-142.
Pearce, D.W., E.B. Barbier and A. Markandya (1988). Sustainable Development and
Cost Benefit Analysis. London:IIED/UCL.
Prather, M.J., RT. Watson (1990). Nature 344: 729 - 734.
Reijnders, L., C. Kroeze (1990). Prevention of climate chan&e. Stichting Natuur en
Milieu.
Solow, R.M. (1986). "On the Intergenerational Allocation of Natural Resources".Scand.
J. Ecs. 88(1)141-149.
Velme, H.M (ed) , E.E.M. Baars, J.F. Feenstra J. et al. (1989). Environment and
International Trade. Publicatiereeks Milieubeheer 1989/1. Min. of Envir.
Management, The Hague.
Vas, J.B. (1982). "Consumptieve Activiteiten, Milieuverontreiniging en Energieverbruik
in Nederland". In: Aiking H. et al. (eds) (1982), Mozaiek van de
Milieuproblematiek. Amsterdam: Free University Press.
 27
World Commission on Environment and Development (1987). Our Common Future.
Oxford: Oxford University Press.
 3

Note on the correction of national income

for environmental losses

Roefle Hueting and Peter Bosch l

1. Shadow prices necessary for correction

One of the objections to using national income as an indicator of welfare is that it

does not include environmental losses and losses of other resources. The study "New
Scarcity and Economic Growth" (R. Hueting, 1974) developed the following train of
thought with respect to a correction for these losses. The environment can be
described as a collection of possible uses, environmental functions or simply functions.
When use of a function is at the expense of another function, or threatens to be so
in the future, the environment acquires an economic aspect: loss of fWlction then
occurs. Losses of function are costs. We thus have a problem of choice: which use
gets priority and which has to be forgone? As information on the basis of which these
choices can be made shadow prices for environmental functions have been sought. For
this purpose it has been endeavoured to construct a supply and demand curve.

The supply curve consists of an elimination costs curve. Elimination is defined as

doing away with the burden on the environment, either by technical measures or by
reducing (where necessary to zero level) the activity causing the burden. Both the
expenditure on the measures and the drop in volume caused by reduction of activities
are interpreted as costs. The construction of the supply curve as defined here may
entail technical difficulties but not problems of theoretical economics.

The demand curve usually cannot be calculated because in most cases it is impossible
to quantify the intensity of the (immaterial) needs. Sometimes it is possible to derive
the demand curve partially from the fact that people compensate for loss in functions
by spending. Other methods have also been developed, such as willingness to payor
willingness to accept, but these are not suitable to arrive at the construction of a
complete demand curve. Insofar as people are directly affected by environmental
losses, the approach might be justified. Many environmental losses, however,

I This contribution has also been published in StaJislkal Journal of the United NaJions Economic
Commission for Europe. VoL 7, no. 2, 1990, pp. 75-83. lOS Press.
 30

constitute part of a process which may lead to the disruption of the life support
functions of our planet and endanger the living conditions of generations to come, and
therefore cannot be considered separately. In all these cases the "willingness to pay"
or to accept approach is pointless (R. Hueting, 1989).

As a result, shadow prices cannot be found and correction of national income for
environmental losses is impossible. Only one minor correction can be carried into
effect, viz. for that part of the elimination, recovery and compensation costs, that is
entered as final delivery in the System of National Accounts (SNA). Such a correction
for double counting certainly yields valuable information, but also conjures up the
danger of "pars pro toto"l if it goes no further, for, as is well known, most
environmental losses are not restored or compensated (R. Hueting, 1989).

2. A practical solution for a theoretical dilemma

The call for national income to be corrected to include all environmental losses has
been steadily growing since the publication of "New Scarcity". This is clearly
expressed in the following remark made by the Indonesian minister for Population and
Environment: "In my policy-making I need an indicator in money terms for losses in
environment and resources, as a counterweight to the indicator for production, viz.
national income. If a theoretically sound indicator is not possible, then think up one
that is rather less theoretically sound"z.

TIle answer to this is obvious: an estimate based on standards. The setting of

standards was also discussed in "New Scarcity", but the point was not elaborated then
because the question "What standards are to be set and by whom?" could not be
answered. This situation has now changed. In the past decade society worldwide has
increasingly declared itself in favour of sustainable economic development. This can
be conceived as a preference voiced by society on which standards can be based for
a sustainable use of environmental functions, instead of on (unknown) individual
preferences. In technical terms this means that in the familiar diagram of the supply
and demand curve for environmental functions we have to determine a point on the
abscissa which represents the standard for sustainability (Figure 1). A perpendicular
on this point intersects the supply curve; the perpendicular replaces the (unknown)
demand curve. The point of intersection indicates the volume of activities, measured
in terms of money, involved in attaining sustainable use of the function. The volume
will often be a mix of necessary technical measures and the necessary reduction in

To take a part for the whok.

2 This suggestion was made during a working visit of Roejie Hueting to Indonesia in 1986.
 31
activities which. even after application of the measures. will be required to attain
sustainable use of the function.

&ul1ders/
year

T \
\
\
demand

correction
\
amount

standard for .vallablllty of ----+

susta1nability envlronmental functions
(in physical parameters)

Note: H,e supply curve, or elimination cost curve, consists of the costs of ~asures for eliminating
various degrees of loss of function. The (theoretical) demand curve should be derived from
individual preferences. The practical 'demand curve I is based on a standard for sustainable use of
envirorrnental tlllCtions.

Figure 1. Supply and "demand" curves for environmental functions.

The standards can be related to environmental functions. Thus it is possible to

formulate. for example. the way in which a forest should be exploited in order to
attain a sustainable use of its functions. Sustainability then means that all present and
future uses remain available. As for renewable resources such as forests. water. soil
and air, it holds that as long as the regenerative capacity remains intact the functions
remain intact. e.g.• the function 'supplier of wood' of forests. the function 'drinking
water' of water. the function 'soil for raising crops' of soil and the function 'air for
physiological functioning' of air. Practically this means that, for instance. emissions
of cumulating matters such as PCB's, heavy metals, nitrates and carbon dioxide may
not exceed the natural buffering capacity of the environment and that the erosion rate
may not exceed the regenerative power of the soil. As for nonrenewable resources.
such as oil and copper. 'regeneration' takes the form of research and bringing into
practice flow resources such as energy derived from the sun (wind. tidal. collectors,
photo-voltaic-cells), the recycling of materials and the development of substitutes for
these.

The measures required to meet the standards may range from selective cutting of
trees. reforestation, building terraces, draining roads, maintaining buffers in the
landscape, selective use of pesticides and fertilizers. to building treatment plants.
recirculation of materials. introducing flow energy. altering industrial processes.
 32

making more use of public transport and bicycles instead of private cars and making
use of space that leaves sufficient room for the survival of plant and animal species.
Of course, no measures can be formulated for irreversible losses. If plant and animal
species become extinct, no restoration measures are possible. The same probably holds
for the total loss of the topsoil of a mountainous area. An arbitrary value then has
to be assigned to these losses, of which only one thing can be said for certain: the
value is higher than zero.

The reduction of national income (Y) by the volumes found gives a first
approximation of the activities level which, in line with the standards applied, is
sustainable. Needless to say the correction mentioned in Section 1 for double counting
must also be made. This amount is also deducted from the national income. We call
the sustainable level Y'. The difference between Y and Y' indicates, in money terms,
the sacrifice involved in attaining the desired sustainable use of the environment.

To be absolutely clear, it should be pointed out that this is a static approach. Effects
on other sectors of the economy as a result of taking measures and reducing activities
are not considered. Neither are developments that might be expected involved in the
approach. These are not taken into consideration, firstly because the exercise is aimed
at a correction of the figure of national income and not at the development of a
vision for the future, and secondly because, in the model to be used, a large number
of assumptions would have to be incorporated. The following may be regarded as an
exception to this.

Reduction of Y by double counting, by the costs of the measures needed to attain

a sustainable society and by the value added of the activities which have to be
reduced in volume, provides a first approach to Y'. The activities burdening the
environment which are to be discontinued will in most cases be replaced by
environmentally acceptable alternatives. The problem of the environment is in fact a
problem of allocation. It is a matter of shifting the patterns of activities into an
ecologically sound direction. This shift will be brought about both by increases in the
prices of products resulting from measures taken and by a direct reduction of the
activities that place a heavy burden on the environment.

The first estimate of Y' must therefore be raised by the value added of activities that
will come into being after reduction of the harmful activities and through the
disappearance of the compensation measures. An iterative calculation becomes
apparent here, for there are very few activities that do not adversely affect the
environment at all. The replacement activities may lead to the standards being
exceeded and thus to the need for further re-allocation. We propose postponing this
more dynamic approach until the second phase of the research and limiting the first
phase to an initial estimate of Y'.
 33
The familiar objection to entering restoration measures as final delivery instead of
as an intermediate (costs) does not apply to Y'. The objection is to the fact that the
environment remains outside the System of National Accounts when environmental
losses occur (loss is not written off), but is included, albeit partially, in the SNA
when the loss is restored (restoration is written up). This means in fact that a
comparison cannot be made between various years. Hence the continual pressure for
correction of double counting which has been going on for decades now. Because
environmental loss at Y' is written off, it is only logical that restoration should be
written up. Naturally, an increase in Y' will also occur as clean technology and flow
energy become cheaper, for then the deduction becomes less.

3. The necessary calculations are not new

The main reason to start with environment statistics at the Netherlands Central Bureau
of Statistics (CBS), 21 years ago, was to arrive at a correction of the national income
for environmental loss~s. The first publications resulted in estimates of the costs
incurred through measures for various degrees of restoration of function (the supply
curve). This kind of work was not continued at the CBS because it was impossible
to construct a demand curve and so obtain the shadow price (see above). Later the
CBS did undertake such calculations for the report of the Netherlands Scientific
Council for Government Policy (WRR) "The Next Twenty-Five Years" and for
scenario studies in the context of the Broad Social Discussion on the future energy
supply. This kind of estimate is now mainly made by other institutes such as the
National Institute of Public Health and Environmental Protection (RIVM). In the
chapter on costs and benefits of environmental measures in Concern for Tomorrow
(Langeweg et al., 1989) a number of 'supply curves' are included as illustrations. The
RIVM uses these data for scenario studies.
The static volume estimates proposed here, based on standards for sustainable use,
have the same character. The principal requirement is scientific and technical
knowledge. This applies both to setting the standards and to formulating the measures
needed to attain the standards. Assistence by the institutes working in the various
subsectors is indispensable for this.

4. Correction in stages

I. Inventory phase

1. Selection of activities causing most harm to the environment

In order to limit the number of calculations necessary, and because lesser
environmental problems are reduced automatically or disappear when the main
 34

problems are tackled, we propose a selection. The criteria for the selection are: the
physical scale on which the consequences occur and the size of the consequences. The
environmental problems to be dealt with will probably include: CO 2 emission, CFC
emission, emission of acidifying substances (S02' NO.. NH 3), release into the
environment of manure, heavy metals, pesticides and waste products, the consumption
of energy and natural resources, withdrawal of groundwater, construction of
infrastructure.

2. Compiling a framework (as yet uncomplete) for the calculation of the correction
Roughly speaking, the following stages will be necessary: establishing the level of the
present environmental burden, setting standards for the environmental burden in the
case of sustainable use of the environment, organizing measures for reducing the
environmental burden to the standard value based on cost-effectiveness and calculation
of the total correction. This means that for every environmental problem an inventory
will have to be drawn up that will contain the causes of the burden on the
environment, the effects of the burden, and possible measures for reducing the burden.
In order to gain a clear picture of depletion of energy and natural resources,
frameworks will be drawn up for "resource accounts", in which entries can be made
for stocks, possible growth and consumption of important energy sources and other
natural resources. This can largely be achieved by study of literature, but consultation
will also be necessary with the institutes working in the various subsectors.

3. Inventory of data requirements and availability

We can distinguish four categories of data:
- data on environmental burden and effects;
- data for determining levels of burden in the event of sustainable development
(setting standards);
- cost-effectiveness data on technical measures;
- data on the production size of activities.
Data sources are, apart from the CBS: government departments, universities, research
institutes and companies.

4. Moments of choice
Possibly a pragmatic selection of environmental problems on the basis of availability
of data; choice of orde,' of approach; choice of year to be surveyed.

II. Operational phase

Stages 5 to 11 have to be completed for each environmental problem. The order may,
however, be altered on the basis of experience in the inventory phase.
 35

5. Quantifying the environmental burden

Using CBS environmental statistics and other sources, the environmental burden will
be expressed in terms of emissions of substances, use of space, use of the soil and
consumption of energy and other resources.

6. Quantifying the scale of effects on the environment

Where it is considered necessary a quantitative description of the effects of the burden
on the environment will be given. It is not always necessary to have exact data on
effects, as in the case of the release of non-biodegradable plastics, for example. This
emission should be stopped completely from the point of view of sustainable use of
the environment, irrespective of the exact scale of the effects. Similar reasoning
applies to various natural resources for consumption over and above natural repletion.

7. Determining the level of environmental burden in the event of sustainable use of

the environment (setting standards)
It will often be necessary for this purpose to determine the extent of flows of
substances in a situation in which multifunctional use of the environment remains
possible. Data will have to be derived from studies of buffer capacities, balances of
substances, etc. For those international environmental problems to which the
Netherlands contributes, the standard will have to be related to the Dutch share in the
total environmental burden in the year surveyed.

8. Collecting cost-effectiveness data on technical measures

Building up a data bank filled with these data. Construction of cost-effectiveness
curves for the measures directed towards realization of the reduction in environmental
burden or the reduction in consumption of natural resources or energy deemed
necessary. Determination of total costs for reaching this level.

9. Determining the necessary reduction in activities

If it is proved that the lower level of environmental burden cannot be attained purely
by technical means, then a reduction in the related activities necessary to achieve the
standard (in addition to technical measures) will be determined.

10. Determining the value added of the activities that might no longer take place
This is relatively simple if we take the example of reduction of the livestock
population, but in the case of a reduction of car kilometres approximating methods
will have to be used. These data can also be incorporated in the cost-effectiveness
curves.

11. Totalization of the corrections, excluding duplications.

When the solution to one environmental problem requires a decrease in an activity,
this may mean a reduction in the measures required to solve another problem.
 36

12. Determining the extent of double counting in National Accounts concerning

environmental measures
TIle basis for this has already been laid in the report of in-service training with the
Environmental Costs Department of the Netherlands CBS. These calculations will have
to be supplemented and updated.

13. Relating the corrections to the System of National Accounts

A preliminary study is now being undertaken to examine this issue in detail.

5. Imperfections and advantages of the method

The drawbacks (or imperfections) are as follows.

1. The approach is strictly static. Only first-order effects are taken into accowlt
(see above). In reality, two additional movements will take place. The technical
measures and the reduction of activities will bring about an expansion or a
contraction of related industry, depending on the type of industry. Presumably
the net effect will be an increase in the correction amount. In the course of
time, however, the economy will adapt itself to the new situation. A second
movement takes place as a reaction to the reduction of polluting activities:
people are seeking environmentally acceptable alternatives. This will reduce
the correction amount. Both effects could in principle be estimated Witll the
aid of a dynamic model. However, the reliability of the results of such a
model would be questionable, because changing to sustainable use of the
environment means in actual fact a break in the trend, of which the
consequences exceed the predicting capacity of the current models.
2. The results of the approach do not indicate the state of the environment. If,
for instance, a cheaper anti-pollution technology is invented, the distance
between national income (Y) and the estimated sustainable activity level (Y')
becomes smaller. But if the technology is not or not generally applied, the
state of the environment changes only very slightly, or not at all. Furthermore
a decrease in costs does not necessarily run parallel with changes in physical
parameters. Therefore environmental statistics in physical wlits remain
indispensable.
3. The results of the approach do not represent individual valuations in the true
sense, as has been explained above. For, among other things, the intensity of
the preferences for a sustainable use of the environment cannot be measured.
However, this simultaneously implies that the intensity of the preferences for
the acceptation of the adverse effects and future risks involved in the present
growth pattern of production and consumption is equally unknown. Both of
these aspects should be clearly mentioned in the presentation of the results of
the method.
 37
4. The method ignores the loss of welfare suffered by those people who have
a strong preference for the survival of plant and animal species apart from
their role in the maintenance of the life support functions of our planet (which
is a prerequisite for sustainable development). These preferences could be
compared with the preferences for creating and maintaining art or churches,
which might not be considered indispensable for sustainable development and
yields, but the loss of which would constitute a decrease in welfare for those
who value them. A solution would be to class the diversity of species under
the function 'gene reserve'. Another conspicuous example of the same kind is
noise. Noise does not affect sustainability, but it can be very disturbing. A
solution would be to class this under the function 'quiet in the living area'.
Standards for noise levels are already operational in many countries. This, too,
should be mentioned in the presentation of the results.
5. For irreversible losses no measures can be formulated, of course. This holds
true for any method.
6. The method is laborious.

The advantages are as follows.

1. As far as we can see, the method is the only way to confront the national
income figures with the losses of environmental functions in monetary terms.
2. The method compels an exact definition of the term "sustainable economic
development". Without such a definition the term remains vague and not
operational in economic policy regarding the environment.
3. The physical data required for comparison with the standards come down to
basic environmental statistics which have to be made anyhow if a government
is to get a grip on the state of the environment. The formulation of the
measures required to meet the standards and the estimates of the expenditure
involved are indispensable for policy decisions. In other words: the work of
supplementing national income figures might be laborious, but it has to be
done if one wants to practise a deliberate policy with respect to the
environment.

References

Hueting, R. (1974). Nieuwe schaarste en economische groei. Agon Elsevier

Amsterdam/Brussel. English updated edition: New Scarcity and Economic Growth,
Amsterdam, New York, Oxford, 1980.
Hueting, R. (1989). Correcting National Income for Environmental Losses: Toward
a Practical Solution. In: Y.l. Ahmad, S.E. Serafy, E. Lutz, Environmental
Accounting for Sustainable Development. The World Bank, Washington, D.C.
This is a summary of the paper: Should National Income Be Corrected for
Environmental Losses? A Theoretical Dilemma. but a Practical Solution, 1988.
 38

Langeweg et al. (1989). Concern for Tomorrow. RIVM/Samson Tjeenk Willink,

Alphen aan den Rijn.
 40
GNP, as it is currently measured, basically captures the value added in.activities in
the monetized sectors of an economy. From the value of the domestic product
expressed in market prices, intermediate deliveries are deducted and corrections are
made for net income flows from abroad. Basically, what is left is composed of
various income components (wages, transfers, profits, etc.) and depreciation of
(produced) capital. Apart from exports, the national product is either consumed or
invested.

It has long been realized that this GNP does not fully capture economic welfare, and
that its development over time does not adequately reflect current or future welfare
changes (see, e.g., Hueting 1980, and, more recently, Leipert and Simonis 1988, Daly
1988, Pearce et al. 1988). Its defects in reflecting current welfare include:
1) the dimensional bias: not all activities and factors leading to welfare are
included as (with some exceptions) the assessment is restricted to the monetized
sectors of the economy;
2) the externalities bias: the aggregation of the contributions to welfare via net
added values (in market prices) may deviate from the activities' real net welfare
contributions (discrepancies between social and private costs, e.g., the costs of
environmental quality losses);
3) the imputation bias: determinants of changes in economic welfare other than
those related to net values added are ignored (e.g., changes in unpriced flows and
stocks such as most elements of environmental capital; and the value of owner used
assets);
4) the output anomaly: a variety of 'defensive activities' actually undertaken to
counter or neutralize environmental degradation (e.g., health expenditure in relation
to pollution damage) is counted as a contribution to GNP and hence interpreted as
adding to welfare) rather than as costs of attempts to cure or prevent welfare losses.
In addition, GNP does not capture impacts on future welfare, or losses of welfare
potentials;
5) the depreciation asymmetry: depreciation of environmental capital (e.g.,
replacement costs with respect to reductions in certain assets) is not incorporated.

Greening GNP

Various approaches have been suggested to correct at least some of the defects.
Hueting and Leipert (1987) mention:
a) leaving the national accounts as they are and providing counterbalancing
information through making explicit the changes in elements of environmental capital
through 'satellite accounts' and isolation of 'defensive expenditures';
b) separate presentation of an alternative national income figure corrected for
defensive expenditure;
 41

c) additional correction of other environmental damage through estimates of

people's willingness to pay for preventing environmental capital loss, etc.;
d) as b, with additional corrections of other environmental degradation by
estimating the cost of measures required to meet certain (e.g., 'sustainable') physical
standards with respect to environmental quality and resource stocks.

Leipert and Simonis (1988) concentrate on the 'output anomaly' or defensive

expenditure criticism of GDP as a measure of sustainable welfare. Economic growth
as registered by GDP in part is but a compensation for damage done to the
environment. They wish to examine the development of the share of these defensive
expenditures over time. A distinction is made between: a) costs of compensatory
economic activity, b) production, income and property losses of residual environmental
degradation, and c) damage to human health, flora, fauna, buildings, materials, etc.
Item a) is included in GNP in principle; the others are not. Actual environmental
protection investments and expenditure are analyzed for the FRG. Total cost of
protection is in the order of 1.5% of GNP, and rising (similar figures can be fowld
for other OECD countries). For an estimate of items b) and c) reference is made to
other studies reporting a figure of some 6% of GDP (including large elements of a
'willingness-to-pay' nature).

Estimates of defensive expenditures can indeed be made (and are increasingly being
made). As pointed out, these are but one of at least three elements of the full costs
of environmental degradation, the others being the value of environmental quality
loss, and the loss of environmental capital assets. Attempts to value the latter elements
are numerous, but have so far remained partial, unsatisfactory and hence unconvincing
(see, e.g., Kuik, Jansen and Opschoor 1989). Firstly, the relationships between
physical changes in stock levels or environmental quality and the resulting physical
damage is insufficiently known (and often challenged, as in the case of many health
effects). Secondly, even where estimates of total current damage (with reference to
some baseline) are available, their conversion to smaller annual changes in the
environment is highly dubious. Thirdly, especially in this category one needs to value
'subjective' elements such as the significance of species and ecosystems. In this area,
methods (e.g., 'Contingent Valuation Methods' for extracting estimates of people's
willingness to pay for maintaining certain standards) are not well tested and CarulOt
cover the ground required. Fourthly, if the amended GNP figure is to be a reflector
of welfare, corrections along the lines of applied welfare theory (including consumer
surpluses and willingness-to-pay estimates) would not only have to take place on the
environmental side, but also with respect to all other goods. Thus, attempts to arrive
at a corrected GNP figure along the lines of Hueting and Leipert's approach c) would
indeed be problematic - even though results of partial studies eliciting the willingness
of people to pay for certain environmental features may have policy relevance.
Perhaps points such as these explain why so far empirical attempts at establishing
'alternative GNPs' have not been very convincing. This approach would not result
 42
in an acceptable macro indicator for sustainability. It might be much more appropriate
to opt for presenting GNP separately, while providing in addition some measure of
defensive expenditures and other environmental costs.

An application of an approach at least partially along these lines has been reported
from Japan, where pollution and waste control costs consistent with preset quality
standards were deduced from GNP, among other corrections, to yield 'Net National
Welfare' (NNW). NNW is a modification of GNP based on recalculated values of
private and government consumption, services rendered by capital and durable
consumer goods, leisure time, extra-market activities, environmental pollution and
urbanization. Pollution control cost corrections were calculated to be in the order of
2-10% of GNP. Uno (1989) reports some more recent results with NNW related
pollution cost calculations: they rose sharply from 1955 to 1970, then dropped steadily
to the 1985 level of 2% of GNP. Cobb et ai. (unpublished) are attempting to correct
GNP for a wider range of environmental features (and other aspects) and to arrive
at an 'Index of Sustainable Economic Welfare'; as in the case of Japan, the index
shows lower growth rates, but the impact of environmental elements on the index is
small compared to, for instance, that of changes in income distribution - another
factor included in the calculation.

Hueting and Leipert favour method d) which aims at correcting GNP for the costs of
measures necessary to (hypothetically) ensure its sustainability, and which would
generate an estimate of 'sustainable income'. They do not demonstrate the method
empirically, but give some theoretical advantages and disadvantages. Advantages
include: (i) an explicit definition of 'sustainability' is required; (in reference must be
made to explicit physical standards and data; (iii) this method provides an adequate
(but sub-optimal) monetarization of sustainability. It requires the calculation of the
costs of (hypothetical) measures broadly defined to ensure the maintenance of resource
stocks and of environmental quality. Its disadvantages, according to the authors, are:
(i) by referring to SD standards, the link with individual valuation has been severed;
(ii) welfare losses beyond preservation of diversity for the sake of sustainability are
excluded; (iii) irreversible losses are not dealt with adequately, and (iv) the method
is laborious. Insofar as the method correctly captures expenses required for theore-
tically maintaining elements of environmental capital at prerequisite (and sustainable)
levels, it would in fact succeed in identifying the theoretical replacement costs of
using elements of environmental capital. Thus, it would generate a measure of
'potential sustainable income' of an economy.

Hueting and Leipert's avenue d) appears to be a promising one at first sight. If it

could be made operational, it would indicate changes over time of sustainable income
for any country, as long as the definition of what sustainability means in terms of
stock levels and environmental standards to be maintained remains the same, and as
long as costs of hypothetical preventive or compensatory activities are known.
 43

Attempts could be made to facilitate international comparisons. This would require (i)
an internationally agreed definition of items to include, and (ii) agreed methods for
calculating the costs of measures sufficient to maintain sustainability. Further
examination, however, gives rise to a number of questions. Firstly, the deductions
depend heavily on the threshold values chosen to ensure sustainability. Thoughts on
how to give operational content to sustainability have hardly beendeveloped, and
cannot be expected to stabilize in the near future. New standards are set regularly and
existing standards are frequently changed. As long as this remains the case, cost
estimates to meet SD thresholds will fluctuate accordingly and make time series and
their interpretation very problematic. Secondly, international agreement on the content
of sustainability will not rapidly emerge (one very sensitive area where views may
diverge is the area of the acceptability of nuclear energy production and waste
hazards). Comparisons of sustainable income across countries will be possible only
to the extent that these problems of definition and content have been solved. Thirdly,
thoughts on which methods to use to approximate these hypothetical costs will take
a very long time to converge, and as long as they have not, the outcomes of different
calculation exercises will vary highly and will not be strictly comparable. Such
calculations may have a high academic value (especially where uniform methodologies
are tried out ), but they are unlikely to influence policy very soon. FourtWy, problems
emerge with respect to cost curves. As technology is only beginning to develop
cleaner and less resource intensive processes, current cost estimates (based on 'add
on' techniques mainly) may overestimate medium-term cost curves and long-term cost
curves dramatically. On the other hand, it is difficult to see how the costs of
irreversible changes can be brought into the calculation. Fifthly, in some cases
teclmology will not be able to meet SD thresholds in the near future, as is
exemplified by the greenhouse gases (notably NzO and COJ. In many such cases, if
economic activities are to be made sustainable, their levels rather than the techniques
used will have to be controlled. In such cases, direct cost calculations fail: macro
models, general equilibrium models and input-output models would be needed to
generate alternative economic structures (with some activity levels bounded) and the
associated sustainable income levels.

Moreover, as the sustainable income measure proposed by Hueting and Leipert only
incorporates corrections based on hypothetical measures, this approach will at best
result in an estimate of potential sustainable income, in the sense that the income
level thus calculated is a fictitious one. It is not a measure of the actual economic
performance in terms of ecological sustainability.
 44

Conclusion

In evaluating the above in tenns of their significance as pointers for future research,
I suggest that attempts to correct or replace GNP by some 'gr~n' income-based
measure should continue, but I am afraid that
a) they will not be sufficient to influence policy, and
b) they will not be developed even to a second-best level of acceptability in the
near future.

Whilst this work goes on, I think it is extremely important that research is oriented
towards developing a small and consistent set of sustainability indicators (or of
environmental pressure and quality), to be produced in conjunction with GNP (or
other measures of economic activity and perfonnance), such as described in Chapter
2 (Opschoor and Reijnders). Together with "physical" environmental indicators, these
economic indicators could be regarded as 'arguments' or variables in a social welfare
function or policy objective function.

References

Daly, H. (1988). On Sustainable Development and National Accounts. In: D. Collard,

D. Pearce, D. Ulph (eds.), Economics. Growth and Sustainable Environments.
MacMillan Press, London.
Hueting, R. (1980). New Scarcity and Economic Growth. North Holland/Neth Central
Bureau of Statistics.
Hueting, R. and Chr. Leipert (1987). Economic Growth. National Income and the
Blocked Choices for the Environment. IIUG Discussion Papers 87-10, WZB,
Berlin.
Kuik, 0., H.M.A. Jansen and J.B. Opschoor (1989). Environmental Benefit Estimates
in Decision Making in the Netherlands. Paris: OECD (to be published by
Earthscan: 1991).
Leipert, Chr. and U.E. Simonis (1988). "Environmental Damage -Environmental
Expenditures: Statistical Evidence of the Federal Republic of Gennany". Int.
Journ. of Social Econ. 15,7: 37-52
Pearce, D.W., E.B. Barbier and A. Markandya (1988). Sustainable Development and
Cost Benefit Analysis. London: lIED/UCL.
Uno, K. (1989). " Economic Growth and Environmental Change in Japan: Net
National Welfare and Beyond". Archibugi F. and P. Nijkamp (eds.): Economy and
Ecology: Towards Sustainable Development. Dordrecht/London: Kluwer Ac. Publ.
307-333
 5

Natural Resource Accounting:

State of the art and perspectives for
the assessment of trends in sustainable development

Jaap Arntzen and Alison Gilbert

1. Introduction

Neglect of environmental considerations within economic planning is generally seen

as a major cause of the environmental problems which the world faces today (see
e.g., WeED, 1987; World Bank, 1987). Particularly in resource-dependent economies
such as most developing countries, trends in the quantity and quality of natural
resource stocks should be vital to economic planners and decision-makers.

As a result, a number of environmental modifications have been proposed for the

System of National Accounts (SNA) during its ongoing revision. This paper discusses
progress in this field of Natural Resources Accounting (NRA), assesses the method's
possibilities and restrictions and, more specifically, its potential as a monitoring tool
of sustainable development. The findings are based on a literature review and on
research into NRA conducted at the Institute for Environmental Studies (e.g., Gilbert
and Hafkamp, 1986; Perrings et al., 1988; Perrings et al., 1989; Gilbert et al., 1990).

2. National Accounts and Natural Resource Accounting

The System of National Accounts is an internationally used "comprehensive and

detailed framework for the systematic and integrated recording of the flows and
stocks of an economy" (Bartelmus, 1989, p.81). It provides an annually updated
infonnation framework, in the Keynesian tradition, of the perfonnance of national
economies.

SNAs have mostly been shaped by statisticians and are not easily accessible to users.
Repetto et al. (1989, p.10) compare national accounts with sausages: "there are many
consumers but few who want to know how they are put together". The best-known
output is Gross National Product (GNP), an aggregate indicator of national economic
 46
perfonnance or income during the year. GNP is frequently, but wrongly, used as the
only indicator of welfare (see Chapter 2 by Opschoor and Reijnders).

The SNA is an essential element of macroeconomic planning. The rigid structure

and monetary nature of the SNA contribute to unifonnity and facilitate comparisons
across countries and trend assessments. However, its general scope is limited due to
its emphasis on market-oriented flows of commodities and services expressed in
monetary values, and subsequent neglect of stocks and flows which cannot be so
measured. Nonmarketed flows are only included when they are directly comparable
to production on the market and their value can be reliably estimated (Repetto et al. ,
1989). This is an important limitation for developing countries with a large
subsistence sector which is usually poorly documented and understood.

Environmental shortcomings of the SNA are by now well documented (see e.g.,
Ahmad et al., 1988; El Serafy, 1990). We limit ourselve to a brief summary of the
major points. Firstly, natural resources are not, in practice, treated as economic
production factors. For example, whilst depreciation takes care of changes in the
quality of man-made capital and leads to changes in net product and income,
depletion of natural resource stocks, such as forests and range lands, does not. Their
extraction is considered as income only.

Secondly, pollution control and other types of environmental rehabilitation -"defensive"

expenditures- are often dealt with inconsistently. Defensive expenditures result in an
increase in the national product or income, which is then commonly interpreted as an
increase in welfare. It is not the defensive expenditures themselves which are related
to an increase in welfare. but the effectiveness of the measures financed. Consequent-
ly, the SNA does not generate appropriate insights into environmental damage and its
mitigation. In addition, pollution control by the source is considered an intennediate
input and as such does not lead to an increase in national product, while control of
the same pollution by another activity does (Repetto et al., 1989). This anomaly holds
for more areas of the accounts.

Finally, many environmental functions are not marketed and so are often neglected
by the accounts. The monetary bias and uncertainties regarding valuation methodology
makes it difficult, if not impossible, to include some of the most vital environmental
functions. Repetto et al. (1989) point out that the original concept of the SNA
explicitly included forests and subsoil assets, but that this has never been put in
practice. The SNA provides, at best, only partial insights into countries' assets and
subsequent economic stocks and flows.

The ongoing revision of the SNA has led to consensus on a number of issues. The
present hard core of the SNA will not be substantially altered - revisions will aim at
simplification and clarification (Luxembourg SNA-meeting 1989). However, the
 47
importance of natural resources for economic development and income generation
and the deficiencies of the present SNA in these respects are recognized, leading to
two follow-up activities proposed for inclusion in the new, revised SNA-handbook:
1. establishment of satellite accounts for the environment, making use of a
mixture of physical and monetary units; and,
2. a section with a clear review of shortcomings in dealing with the environment.
Among the potential disadvantages of such satellite accounts are problems of
aggregating, summarizing and interpreting the data and the risk of being ignored by
the target group, viz. macroeconomic planners and decision-makers (Repetto et al.,
1989). Nonetheless, paying systematic attention to the environment can be seen as an
important first step towards its integration in macroeconomic planning.

At present a consensus has emerged that NRA should take the form of satellite
accounts which cover the most important relationships between environment/natural
resources and economic activities (Ahmad et al., 1989). Such satellite accounts should
present information on the flows of environmental goods and services and their effects
on natural resource stocks. The accounts may be expressed in physical and/or
monetary units.

The potential exists for GOP to be adjusted for quantitative and/or qualitative stock
changes of natural resources. To what extent this can be achieved and how adjust-
ments for defensive expenditures should be made, are still contentious issues. A
disadvantage of adjustment would be the reduced intemational comparability of the
resulting accounting structure. The income of a country with an active environmental
policy and high defensive expenditures would appear lower than that of a similar
country without environmental control activities. As long as defensive expenditures
constitute only a small percentage of GOP, this disadvantage will be small.

The adjustment of economic accounts for defensive expenditures is further

handicapped. Data on the costs of environmental damage and the benefits from
pollution control are not yet sufficient for their inclusion in routine statistics such as
included economic or resource accounts. In addition, divergent opinions exist on the
most suitable depreciation method (e.g., users' costs or depreciation method).

3. Current and Planned NRA Activities

3.1. International Organizations

Most intemational activities are still in the early phases of implementation. Therefore,
only limited lessons can be drawn from their experiences. The World Bank and the
United Nations Statistical Office (UNSO) have developed a tentative framework for
satellite accounts, which will be included in a new NRA handbook (Bartelmus and
van Tongeren, 1988; Stahmer, forthcoming). The handbook will be modified on the
 48
basis of the results of a set of pilot studiesl . It should then become applicable for
general use and safeguard the required uniformity of, and consistency in, approach of
individual countries.

UNEP will provide inputs on specific aspects such as the classification, documentation
and valuation of natural resources, NRA experiences from industrialized countries and
treatment of environmental damage. A report on the first topic has been completed
(Gilbert et ai., 1990). The major conclusions of this report may be summarized as:
1. classification should be based on ecological/physical criteria; the detailed form
of documentation should also consider the type(s) of use made of natural
resources and issues which stem from that use;
2. priority in documentation should be given to so-called vital resources, defined
as those resources which are most important to a specific country's welfare;
3. no single, undisputed method for valuing stock changes is yet available, and
even for those resources with the best prospects for valuation (viz. non-
renewables), the picture is complicated by market distortions and external
costs; and,
4. the valuation method offering most promise can be identified by consideration
of whether changes in stocks are of a qualitative or quantitative nature, and
whether a market exists for the associated flows.

3.2. National Activities

Considerable experience with NRA has been gained at the national level, although
there is no uniformity in the approaches. The country applications differ in orientation,
and are often based on the specific requirements, and institutional and cultural factors
of the country in question. Theys (1989) identified the following major needs and
objectives behind these efforts:
1. analysis of the present and future state of the environment (Canada, France
and Norway);
2. analysis of economic and welfare impacts of environmental policies or their
absence (Indonesia, Japan and USA);
3. assessment of the value of exploitable natural resources and the conditions
for their regeneration (France and Norway); and,
4. identification of trade-offs between social, economic and ecological
development objectives.
The United States and Japan primarily aim at a correction of GDP by means of
monetary valuation of environmental functions and environmental damage. In other
countries (e.g., France) resource accounts make extensive use of physical units. Below
we discuss the experiences of Indonesia and Norway in more detail.

J Proposed counJries are, among others, Malaysia, Ivory Coas4 Ecuador, Mexico and Tunisia.
 49

Indonesia
This study (Repetto et al., 1989) aimed primarily at adjustment of the GDP and
Gross Domestic Investments figures for quantitative and qualitative changes of natural
resources stocks. Extraction of crude oil and timber, and the exploitation of soils for
crop production were selected because these activities and related resources are most
important for the country's economy. No attempts were made to correct GDP for
defensive expenditures; their size in proportion to revenues from the selected resources
and to total GDP can be assumed negligible.

The resource accounts include balance sheet accounts and flow accounts. Valuation
was based on the net price system, i.e. available market prices times quantitative
changes. The major appeal of the study lies in its clear findings. Over the period
1971 to 1984, the adjusted average annual growth rate of GDP was estimated at
4.0% instead of 7.1% (Figure 1(a». If investments are corrected for depletion and
depreciation effects, they prove to fluctuate strongly and even drop below replacement
investment level in 1980 (Figure 1(b».

The study also yields important sectoral findings (Table 1). This NRA method is
found to be extremely sensitive to new discoveries and revaluations of nonrenewable
resources such as oil, making it more difficult to identify clear trends in economic
performance. Oil represents the most extreme example, but the mechanism equally
applies to other nonrenewable resources (e.g., copper in Zambia). The new discoveries
and revaluations of oil have led to substantial increases of National Product and
Investments in 1973(74.

Stocks and flows of exploited renewable resources develop more gradually and here
this method may yield important planning insights. For example, depreciation costs
of soil fertility in terms of lost farm income are substantial; calculations show that
they equal the annual production increase. However, the raw data used for these
calculations are being debated. Recent data give reason to doubt the assumed
magnitude of soil erosion and the importance of human influences on the process.
 50

a) b)
14000 4000

12000 3000
s: s:
'g'". 10000 !2000
a: :::l
a:
c c 1000
8000
~ ~
iii iii
6000 0

4000 ·1000
1970 1975 1980 1985 1970 1975 1980 1985
Year Year

Figure 1. Comparison of a) GDP and 'NDP', and b) GDI and 'NDI', in constant
1973 Rupiah (adapted from Repetto et al., 1989).

Table 1: Comparison of GDP and 'NDP' in 1973 Rupiah (billions) (Repetto et aL, 1989).

Net Change in Natural Resource Sectors

YEAR GDP Petroleum Forestry Soil Net Change NDP

1971 5,545 1,527 -312 -89 1,126 6,671

1972 6,067 337 -354 -83 -100 5,967
1973 6,753 407 -591 -95 -279 6,474
1974 7,296 3,228 -533 -90 2,605 9,901
1975 7,631 -787 -249 -85 -1,121 6,510
1976 8,156 -187 -423 -74 -684 7,472
1977 8,882 -1,225 -405 -81 -1,711 7,171
1978 9,567 -1,117 -401 -89 -1,607 7,960
1979 10,165 -1,200 -946 -73 -2,219 7,946
1980 11,169 -1,633 -965 -65 -2,663 8,505
1981 12,055 -1,552 -595 -68 -2,215 9,840
1982 12,325 -1,158 -551 -55 -1,764 10,5661
1983 12,842 -1,825 -974 -71 -2,870 9,972
1984 13,520 -1,765 -493 -76 -2,334 11,186
Average
Annual 7.1% 4.0%
Growth
 51

Norway
The Norwegian system of resource accounts makes extensive use of physical units.
It distinguishes between accounts for material and environmental resources. Material
accounts are of two fonns: stock balances in physical units, and material flows of
environmental commodities through the economy in physical units, but overlapping
with flows measured in monetary units. The structure of the material flow accounts
is shown in Figure 2.

--"~~physical flows (material accounts only)

- ~ monetary and physical flows overlap

(both national and material accounts)

- ~ monetary flows (national accounts only)

,
I
I
I

, I
rL Consumption
(households)
'I II
I I II
I I II
I I II
I I II
The rest I I Other II
of the 1- sectors Investment
world
l-
I
I
I
I
I
I
I
L
I

Figure 2. The structure of the Norwegian material flow accounts (adapted from
Longva, 1981).
 52
Although they share a common outline, the material accounts for the various resource
commodities differ in specifics because:
1. the number of users for a specific resource differs (e.g., compare fish and
energy);
2. the time taken for regeneration determines the details of age structure of stock
variables and the collection frequency required;
3. resource mobility determines the spatial dimension; and,
4. if the exported or imported share is substantial, the international sector needs
more attention.

The environmental resource accounts are subdivided (see Figure 3) into land use
accounts and emission accounts. The former provide information about the amount
of land by identifying quality classes and their corresponding suitability for different
types of land use. Lack of data caused these accounts to be more limited than
anticipated. Emission accounts are based on a registration programme of air quality
and calculations of energy use, industrial statistics, etc. These accounts give
indications about the air quality and emission levels by type of pollutant.

a the emission accounts b the state accounts

Figure 3: Structure of the Norwegian environmental accounts (adapted from Longva,

1981).
 53

Alfsen et al. (1987) provide examples of how NRA could be used for policy support
purposes, usually via links to macroeconomic models. However, the long lead time
for account construction has delayed demonstration of the usefulness of the accounts
and the value added of an expanded accounting framework.

4. A preliminary evaluation

It is not possible to complete a thorough evaluation of most recent NRA initiatives

- NRA methodologies are still largely experimental. Considerable conceptual progress
has been made, data on natural resources and economic use have improved, but no
comprehensive application of an NRA methodology exists. The experiences of France
and Norway are valuable, but the theoretical and empirical details remain vague as
does the value added to economic planners and decision-makers. One may therefore
conclude that at its present state NRA is, at best, capable of monitoring certain
aspects of sustainable development, but still has to prove its particular usefulness for
policy support.

A number of choices have been made which appear vital for methodological and
empirical progress with NRA, in particular the use of satellite accounts based on
physical and - where possible - on monetary units. Schafer and Stahmer (1988) adopt
a pragmatic, staged approach in order to derive NRA from the SNA. The procedure
is summarized in Figure 4. In order to maintain comparison opportunities, they
propose to undertake sensitivity analyses when NRA and SNA use different valuation
methods.

SATELLITE SYSTEM

monetary data non-monetary data

from
from
national accounts
national accounts
---------- ~P':~~rl! ~':9~g.:~
additional data
.
additional data

Alternative
computations

Figure 4. Data blocks in the System of Satellite Accounts (adapted from Schafer and
Stahmer, 1988).
 54

In spite of the consensus on general principles and objectives of NRA, opinions still
differ on definitions, the scope of NRA and on selected issues such as correction of
GDP for defensive expenditures. A central point in NRA discussion is the
(in)compatibility between strict SNA objectives and the demand for internal
consistency, and optimal treatment of natural resources from the perspective of
sustainable development. Possibilities to develop a single welfare indicator based on
sustainable development (e.g., "green" GDP) are presently the subject of debate; the
feasibility and meaning of such an indicator is widely questioned. In our view,
satellite accounts provide a useful framework for the development of a set of
indicators of sustainable development (e.g., a depletion index based on stock
mutations) which could complement existing SNA-based indicators. The choice of
indicators and their development should interact with the development of NRA.

Documentation of the pollution issue -most significant for the Netherlands and other
industrialized countries- appears to be more difficult than that of natural resource
depletion. The NRA output is expected to be higWy aggregated. Local or regional
factors are important determinants of pollution impacts (Blades, 1988), and should be
taken into consideration when interpreting NRA output. Issues regarding incorporation
of defensive expenditures have been mentioned previOUSly. Finally, the incorporation
and valuation of changes in environmental quality are not yet possible.

Valuation is a crucial element and problem in the entire NRA discussion. The choice
of valuation method has to be based on specific NRA requirements (e.g., internal
consistency) and on the theoretical and empirical strengths and weaknesses of the
various methods - e.g., willingness-to-pay, replacement value, net present value or
market net price value based on market prices.

NRA makes heavy data demands, and data availability is limited. In practice, various
assumptions must be made -as in the Indonesian study-, the data have to be left out,
and/or data collection programs must be stimulated. The construction of a full set of
natural resource accounts is a lengthy and costly exercise. The Indonesian study is the
only one carried out with limited inputs l .

The simultaneous use of physical and monetary units impairs direct comparability.
It weakens the link with traditional macroeconomics and consequently may be by-
pass planners and decision-makers. These disadvantages must, in order for NRA to
be implemented, be outweighed by advantages of a superior treatment of the environ-
ment whilst at least making a start towards integration of environment and macro-
economics.

/ By posigraduale students under supervision of stafffrom the World Resources Institute.

 55

NRA may be applied at different spatial levels. Regional, subnational applications

representing a smaller set of resources and resource using activities may provide a
better opportunity for testing NRA methods. Sectorial or resource oriented applications
offer even better prospects, particularly in the short term. Finally, it is doubtful
whether NRA is the most efficient and effective method for monitoring sustainable
development in the short term, and perhaps even in the long run. The situation would
be made more clear if an inventory was made of existing environment-economic
models (e.g., extended input-output analyses), data sources and bases, and environ-
mental statistics in general.

5. Conclusions

The value of NRA, particularly in comparison with the indicators discussed in this
document, lies in its strong macroeconomic orientation and in its more comprehensive
and systematic approach. Simultaneous use of physical and monetary units introduces
aggregation problems, but these problems must also be faced with regard to indicator
construction.

Satellite accounts present possibilities to develop NRA and indicators jointly and as
complements to the SNA. NRA could provide a necessary systematic basis for
indicator construction and calculation, and could also guide the choice of indicators
for sustainable development. It also offers a comprehensive framework within which
individual indicators can be fitted and presented. During the gradual construction of
NRA, composite indicators may be derived.

However, we are of the opinion that, in the short term, the monitoring of sustainable
development can be more successfully pursued by indicators. The prospects for NRA
are best in the longer term, and require an inventory and analysis of existing material.
We recommend an evolutionary approach for construction of natural resource
accounts, beginning with a specific sector or resource issue, such as use of
groundwater, and based on existing models and data.

References

Ahmad, YS., S. EI Serafy and E. Lutz (eds.) (1989). Environmental Accounting for
Sustainable Development. Proceedings of a UNEP/World Bank Symposium.
Alfsen, K., T. Bie and L. Lorentsen (1987). Natural Resource Accounting and
Analysis: The Norwegian Experience 1978-1986. Central Bureau of Statistics of
Norway.
Bartelmus, P. and J.W. van Tongeren (1988). SNA Framework for Environmental
Satellite Accounting. World BanklUNSO.
 56

Bartelmus, P. (1989). Environmental Accounting and the System of National Accounts.

In: Ahmad et al. (1989). Environmental Accounting for Sustainable Development.
Proceedings of a UNEP/World Bank Symposium. pp. 79-87.
Blades, D.W. (1989). Measuring Pollution within the Framework of the National
Accounts. In: Ahmad et al. (1989). Environmental Accounting for Sustainable
Development. Proceedings of a UNEP/ World Bank Symposium. pp. 26-31.
El Serafy, S. (1989). Natural Resource ACCOunting: An Overview. Paper to an ODl-
Conference on Environment and Development. March 1990, London.
Gilbert, AJ. and W.A. Hafkamp (1986). Natural resource accounting in a multi-
objective context. The Annals Of Regional Science. Special Edition -
Environmental Conflict Analysis, vol XX(3):10-37.
Gilbert, A., O. Kuik and J.W. Arntzen (1990). Natural Resource ACCOunting: Issues
related to Classification and Valuation of Environmental Assets. Inst,itute for
Environmental Studies E-90/1; Report prepared for UNEP.
Longva, P. (1981). A system of natural resource accounts. Rapporter 81/9, Statistisk
Sentralbyra, Norway.
Perrings, c., J.B. Opschoor, J.W. Arntzen, A. Gilbert and D.W. Pearce (1988).
Economics for Sustainable Development in Botswana. Report for the National
Conservation Strategy.
Perrings, c., A. Gilbert, D.W. Pearce and A. Harrison (1989). Natural Resource
Accounts for Botswana: Environmental Accounting for a natural resource-Based
Economy. LEEC-Paper 1989-11, London.
Repetto, R., W. Magrath, M. Wells, C. Beer, F. Rossini (1989). Wasting Assets:
Natural Resources in the national Income Accounts. World Resources Institute.
Schafer, D. and C. Stahmer (1988). Conceptual Considerations on Satellite Systems.
Working Party on National Accounts and Balance, Geneva.
Theys, J. (1989). Environmental Accounting in Development Policy: The French
Experience. In: Ahmad et al. (1989). Environmental Accounting for Sustainable
Development. Proceedings of a UNEP/ World Bank Symposium. pp. 40-53.
Warford, J. and Z. Partow (1989). World Bank Support for the Environment: Progress
Report. Seminar on the Economics of Environmental Issues. OECD Paper No. 11,
Paris.
World Commission on Environment and Development (WCED) (1987). Our Common
Future. Oxford University Press, Oxford.
 6

The predictive meaning of sustainahility indicators

Leon Braal

1. Introduction

All people use indicators in making day-to-day decisions. For example, in deciding
on what type of clothes to wear, the cloud cover, sunlight, and outside temperature
are rapidly evaluated. Even faster evaluations of numerous indicators take place when
driving a car in traffic. The indicators used in daily life are selected, often not even
consciously, for their known or assumed information content and its easy digestibility.
The information is in many situations also implicitly assumed to have predictive
power, an assumption which is rarely tested for its accuracy. Over time, with proper
training and growing experience, most people manage, however, to sustain their lives
based on such uncertain information flows.

Compared with the complexity of our daily lives, the complexity in man-environment
systems is commonly referred to as enormous. This is partly because our
understanding of the dynamics of these systems is still quite limited and partly
because they generally do not follow the rules planners try to impose on them. The
idea in economic and environmental planning is, however, more or less the same as
in everyday life, i.e., to develop proper training, acquire experience and develop
methods for quick but reliable assessments of looming problems and opportunities.
The challenge for the planners is to find indicators which can be assessed as easily
and reliably as temperature on a thermometer.

In this paper, sustainability indicators are defined as indicators which provide

information, directly or indirectly, about the future sustainability of specified levels
of social objectives such as material welfare, environmental quality and natural
systems amenity. I distinguish two types of sustainability indicators. Firstly, the
predictive indicator. This type of indicator provides direct information about the future
state and development of relevant socioeconomic and environmental variables. This
information constitutes the basis for anticipatory planning and management. The
predictive power is based on mathematical models of the man-environment system.
The second type is the retrospective indicator. This type includes the traditional policy
evaluation and historical trend indicators. They provide infonnation about the
 58

effectiveness of existing policies or about autonomous developments, respectively.

From these indicators decision-makers may learn and improve policy effectiveness.
In this way, retrospective indicators may provide indirect information about future
sustainability. They are usually quantified by a combination of measured data and
reference values (e.g., historical situations, economic targets, health standards).

If retrospective indicators were the only guidance available, policy-making would be

bound to remain an 'after-the-fact' activity locked in a trial-error-evaluation-new-trial
cycle. To avoid this trap, predictive indicators with direct information about future
sustainability are necessary. However, useful as they are in principle, the predictive
indicators are in fact scientifically disputable and therefore risky in practical
management. Scientifically reliable information is only obtainable through retrospective
methods. Concerning the future, the best we can do is generate plausible information.

So, there are basically three options in developing effective sustainability indicators.
First, we can focus our efforts on improving our knowledge of how policies work and
systems develop, for example, by processing more monitoring network data and by
faster policy response mechanisms. In this way we may hope to improve the
suitability of the trials, minimize the errors and accelerate the evaluation phase. The
second option is to direct our energies towards developing the best possible models
as a basis for predictive indicators. And thirdly, we can distribute our resources over
both strategies, hoping that the intensified modelling will benefit from the increased
data collection and policy effectiveness evaluations.

In this paper I discuss the problems and perspectives of these options for the
development of adequate and effective sustainability indicators. As a background for
this discussion, a brief review of indicators in general is given in Section 2. In the
last few years an avalanche of publications on sustainability has hit the scientific and
policy community, a summary of which is given in Section 3. Against this
background the predictive properties of various types of indicators are discussed in
Section 4.

2. Indicators

Indicators represent components or processes of real world systems. As a consequence

of this definition, indicators are defined as models and therefore have all of their
possibilities and limitations. The numerical values of indicators tend to have special
meaning to particular observers, a meaning that goes beyond the numerical value
itself. For example, the number of predator birds may be used to represent the vitality
of a whole forest ecosystem.
 59

In principle, infonnation embodied in indicators is used in decision-making. Good

indicators are therefore defined as indicators which are consistent in the representation
of complex processes while using a fonnat which is psychologically attractive, so that
they may aid decision-makers rather than confuse them. Let us see what this means
in tenns of requirements as to fonn and content.

Indicators may be distinguished as a special class of models by the following

requirement (see Vos et al., 1985):

1. The infonnation must be presented in an attractive fonnat

Effective indicators have a fonnat which is designed with an explicit target group
in mind. Three types of target groups for sustainability indicators are distinguished,
based on quantity of infonnation bits incorporated in the indicator (see Figure 1):

INDICATORS
INCREASING FOR THE
CONDENSATION PUBLIC
OF DATA

INDICATORS
FOR
POLICY MAKERS

INDICATORS
FOR
SCIENTISTS

..
TOTAL QUANTITY OF INFORMATION

Figure 1. Relationships between indicators, data and infonnation.

 60
(A) Professional analysts and scientists.
This group is most interested in raw data which can be analyzed
statistically. They prefer many information bits per message conveyed.
(B) Policy-makers.
Policy-makers prefer data which are related to policy objectives, evaluation
criteria, target and threshold values. The information should be condensed
to a few bits per message.
(C) The public.
This group, miscellaneous as it is, is assumed to prefer unambiguous
messages, free of redundancy, in a single bit of information.
So for each sustainability indicator a format must be chosen which is attractive to,
and has meaning for, the particular target group.

The following three requirements pertain to the scientific adequacy of indicators, i.e.
the degree to which an indicator represents the structure of a system accurately, and
the dynamic behaviour of that system consistently (see, for example, Ackoff, 1962;
Bennett & Chorley, 1978; Spain, 1982).

2. The indicator must be representative for the chosen system

Representativeness must be defined in terms of accuracy in either the behavioural or
tlle structural description, or both, including explicit margins of uncertainty.

3. The indicator must have a scientific basis

The indicator must preferably be based on an empirically quantified, statistically
tested, causal model of the system it represents. Discrepancies between this ideal and
the actual empirical and theoretical basis of the indicator must be explicit, i.e., it must
not be hidden by the format of the indicator.

4. The indicator must be quantifiable

The data on the structure or behaviour of the system represented by the indicator
must be available or obtainable with the present technology.

In addition to these general form and content requirements, Liverman et al. (1988)
propose a number of special requirements for indicators of sustainability. They
include:

5. The indicator should include reference or threshold values

This is, of course, relevant and necessary for the identification of the degree of
sustainability of a particular development, but it implies that the relevance and
appropriateness of the reference values should be beyond doubt and dispute.
 61

6. The indicator should provide information without social bias

Sustainability indicators should not be formulated from a narrow ethical-theoretical
framework. As the target group may include decision-makers from various political
and ethical convictions, noncompliance with this requirement makes an indicator
immediately unappropriate.

7. The indicator must represent reversible and manageable processes

This requirement is sensible from the planner's point of view but disputable from that
of the scientist. Some irreversible and unmanageable processes may provide useful
indicators, although their predictive meaning may only be indirect. For example, the
number of extinct species is not reversible and the frequency of typhoons is not
manageable, but the information, especially when available as time series data, may
still be useful in avoiding disastrous courses of action.

8. The indicator should have predictive meaning

As stated before, without direct predictive meaning the usefulness of an indicator is
limited to retrospective values and can only have indirect meaning for sustainable
development planning. We shall return to this in Section 4.

3. Sustainability

Sustainable development has been the key concept of the 1980's in economic planning
and environmental management. It has led to a proliferation of ideas and publications
(e.g., Clark & Munn, 1986; WCED, 1987; Opschoor, 1987; Turner, 1988; Pezzey,
1989; Archibugi & Nijkamp, 1989; Costanza, 1991). In the view of the WCED, the
concept combines two basic notions: economic development and ecological sustain-
ability. Ecologically sustainable economic development can be thought of as the
process of related changes of structure, organization and activity of an economic-
ecological system, directed towards maximum welfare. which can be sustained by the
resources to which that system has access. I shall further refer to this with the
abbreviated term sustainable development.

Many of the definitions of sustainability available in the literature are variations on

an essentially anthropocentric welfare optimization theme. Most authors take survival
for granted and identify the objective to be maximized as welfare for the present
generation. The ethics-driven, self-imposed constraints include welfare distribution
within the present generation as well as between the present and future generations.
Externally determined constraints are the stocks of exhaustible resources and the
sunlight energy supply rate. Constraints which are manageable, but subject to
thermodynamic laws, are the economic production factors, including technology, and
environmental quality, natural system amenity and regenerability of ecological resource
systems (see Braat, 1991 for more details). The objective function and all constraints
 62

are highly dynamic, and can only be influenced partly by well-intended policies and
management strategies.

The next aspect to be addressed is the shape of sustainable development. To obtain

an idea of the variety of possible shapes, we shall first turn to documented history
and then to simulated futures.

Sustainability patterns in history

Most people who try to imagine what (ecologically) sustainable (economic)

development looks like are confronted with the problem that they have never seen
anything by that name. Ecologists and historians, in general, do not have this problem.
In Figure 2 examples of sustained development in natural ecosystems are shown
which are drawn from ecology textbooks. When transferred to ecosystems, the concept
of sustainability does, of course, not refer to welfare as objective function, but, for
example, to total biomass as an indicator for population or ecosystem survival. The
constraints have all been condensed in the concept of carrying capacity.

The examples illustrate that sustainable development can have many different shapes.
The presence of biomass may be sustained, although the quantities may change
through time and in some patterns even dramatically collapse before recovery. The
graphs show that in evaluating the sustainability of a system by means of the
development pattern of an indicator, the time horizon considered is essential.

Natural communities and human economies are both highly dynamic types of systems.
In ecological succession species colonize an area, may dominate for some time,
several are outcompeted and in the long run many become locally extinct (see e.g.,
Odum, 1971; Miles, 1987). Similarly, in the histories of human economies, social
groups successively conquer a region, establish a political empire and an economic
organization, dominate for decades or centuries and are eventually expelled, absorbed
by successors or exterminated (e.g., Sumerians, Romans, Aztecs, the medieval feudal
system; see e.g., SHcher van Bath, 1960). All of these examples show that sustained
development of an ecosystem or a socioeconomic system, respectively, involves
sequences of growth, decline and replacement of system components.

A case study of the 19th and 20th century history of a region in the Netherlands
(Braat and Steetskamp, 1991) illustrates both the variety of possible development
shapes and the sequences of subsystems. The study documents a sustained
development of the regional economy with clear successive introduction and
disappearance sequences of various economic sectors and resource systems. By
moving from a predominantly autarkic economy in the early 1800's to an open
economy in the 1900's, the sustained development process has led to increased
 63

welfare for an increased population at the cost of economic dependency on the

surrounding regions and disappearance or pollution of most of the natural ecosystems.

SUSTAINED DEVELOPMENT SUSTAINED DEVELOPMENT

IN ECOSYSTEMS 1 IN ECOSYSTEMS 2

CARRYING CAPACITY CARRYING CAPACITY

ECOSYSTEM
BIOMASS

SUSTAINED DEVELOPMENT SUSTAINED DEVELOPMENT

IN ECOSYSTEMS 3 IN ECOSYSTEMS 4

CARRYING CAPACITY CARRYING CAPCITY

ECOSYSTEM
BIOMASS

Figure 2. Examples of sustainable development in ecosystems.

Sustainability patterns in simulated futures

The world model developed by Meadows and co-workers in the early 1970's is quite
likely the most widely known of all simulation models. The futures generated by this
model also show a number of shapes of sustainable development. The standard run
 64

of World #3 (Meadows et ai., 1972) shows a rise-and-fall pattern for the major state
variables, but population, food and industrial production have levelled off by the year
2100, albeit at rather low levels. The simulations have indicated that the human
species can at least survive and be p.-esent in a distant future. This happy prospect
was, however, completely ignored by the public because of the collapse of the world
economy with extensive starvation, widespread pollution and exhausted fuel resources
at a much nearer moment in time.

These and many other future studies (see for a review Meadows et ai., 1982; Odum,
1987) suggest that the inherent dynamics of the global ecological-economic system
lead to a growth-peak-decline pattern, which ends in a relatively small stock of natural
and economic capital, which can subsequently be sustained indefinitely on the steady
sunlight energy supply and recycled materials. Odum has pointed out, however, that
if viewed on a longer time scale, this pattern may be repetitive, i.e., create a pulsing
system (Odum, 1983).

Conclusion

These examples of possible shapes of sustainable development from history and

simulated futures lead to the conclusion that there are many different patterns of
development of the major variables of man-environment systems which could be
called sustainable. For sustainability indicators this means that in order to be
operational a reference set of acceptable sustainable development patterns must be
defined. By a pattern I mean a cluster of specified development curves for indicator
variables such as expendable income, environmental quality, etc. A reference set
would include several acceptable patterns. In view of the phenomena of replacement
of system components and fluctuations in quantities of system components the
specification must also identify the system aggregation level, for example, global,
continental, national, or regional.

Section 3 has thus provided a view of the complications of representing sustainability

by means of indicators. Rise-and-fall patterns, succession series of system components
and relationships with other man-environment systems are phenomena which should
be captured by sustainability indicators if they are to produce an adequate view of
possible and likely futures. To assess the sustainability of such futures, they have to
be compared with the reference set of sustainable development patterns. I will not
deal with the problem of specifying such a reference set but leave that to the political
system, and instead tum to the predictive meaning of various types of indicators.
 65

4. Predictive meaning

Predictive indicators

A predictive indicator is defined as an indicator of which the numerical value has an

immediately clear, predictive meaning. It provides direct information about a possible
or likely future of a (part of the) system represented by the indicator. As will be
illustrated in this section, the predictive meaning of an indicator is directly dependent
on its format. However, as was pointed out in the previous section, no indicated
future can by itself generate the assessment that it constitutes a sustainable future.
This is only possible if the indicated future is referenced to the reference set of
sustainable development patterns.

When designing a predictive indicator, a series of choices must be made with regard
to the format of the indicator and the technique of generating future numerical values.
The first choice concerns the types of system variables used as format of the
indicator. An indicator may represent the complex system by means of a
(a) Stock,
(b) Flow (measured as change in stock value), or
(c) Ratio between stocks, flows or stock and flow.

By themselves, single stock, flow or ratio values have no relevant predictive meaning.
However, a decision-maker may attach. predictive meaning to such singular values
through implicit interpretation based on assumptions, or hidden knowledge of a time
series of indicator data, of reference situations, of the dynamics of the system, and
of values of other indicators of the system's performance. This is in fact what most
people do when they make decisions. Although common and to some extent possibly
effective, the hidden interpretation approach is at least unscientific, i.e., it cannot be
tested and falsified, and in complex man-environment systems possibly quite risky.
An obvious improvement is to make the assumptions explicit, i.e., the time series,
reference situations, and the other aspects included in the interpretion process. This
leads us to the next choice.

In contrast to a single value, a trajectory of future values does have predictive

meaning without reference values. In fact, three options seem interesting as to the
dimension of the indicator:
(a) single future value (as in choice 1).
(b) a series of future values (trajectory through time).
(c) three-dimensional projections through time.
A graphic picture containing the trajectories of key variables of the man-environment
system (e.g., population size, per capita expendable income and environmental quality)
may yield a comprehensive and intelligible predictive indicator. If the second or third
option is chosen, a reference value, other than the situation at the start of the
 66

trajectories, is not necessary for a predictive meaning, but in some cases a historical
or target value may support the message of the indicator. Combined with the curves
from a sustainable development reference set, the picture may provide a useful
sustainability indicator. With a three-dimensional picture, the simultaneous develop-
ment of two variables and time can be shown, and often this format produces
attractive indicators. More than three dimensions are unlikely to yield intelligible
indicators, so they are dismissed here.

If the planner has chosen to use a single future value indicator, the predictive
meaning of this value, as stated above, is dependent on the reference value included
in the indicator. Without reference value there is no predicative meaning. The use of
the present situation can be thought of as default choice, but should be explicitly
included in the indicator anyway. The options can be summarized as:
(a) The current value (default choice).
(b) A historic value, representing a system condition which was followed by
a sustained development of key variables.
(c) A subjective value, representing a condition assumed necessary to achieve
a sustained development.
The alternatives to the current value are basically in two classes: historical values
and chosen values, either targets or thresholds. Targets are defined as subjective
reference values, e.g., a desired per capita expendable income, while thresholds are
defined as analytically based reference values, for example, a maximum allowable
ambient concentration of sulphur dioxide. In both options (b and c) the assumption
is that the reference value represents a system condition which is necessary to achieve
sustainability.

Future values of the indicator must be generated. This requires a forecasting

technique, of which there are many. Each technique has special advantages and
disadvantages, for example concerning data requirements, reliability and transparency
(for a review and evaluation see e.g., Jantsch, 1967; 1972; van Doom and van Vught,
1978; ERL, 1984). The fourth choice therefore regards the type of technique with
which the future values of the model are generated. From the many available
techniques, three quite popular options are mentioned here:
(a) Trend extrapolation: future values based on past values of the indicator
only.
(b) Regression model in which the indicator is a dependent variable.
(c) Theory based simulation model, with endogenous, mutually interdependent
variables, feedback loops and independent exogenous variables.
The predictive power of the simulation model (option c) may be relatively large if
the theory is firm and historical data are available against which the model predictions
can be tested. Regression techniques have little predictive power outside the range
of the original data. Extrapolating a historical trend is a common technique, but has
 67
only predictive power if the dynamics of the system are known, and the mathematical
fWIction obtained through fitting a curve to the historical data, has a theoretical basis.

The conclusion of this discussion of predictive indicators may be that the scientifically
most appropriate way is to develop simulation models to generate trajectories of future
values for selected socioeconomic and environmental variables, with explicitly defined
reference values. And whether to choose stock, flow or ratio variables as indicator
variables is more a matter of psychological attractiveness and easy intelligibility than
of scientific adequacy.

Retrospective indicators
In Section 1 a second type of indicator was identified, the retrospective indicator,
which can either be a policy evaluation or a trend indicator.

(1) Policy evaluation indicators

Information obtained from monitoring networks and special studies may serve as a
basis for retrospective evaluation of currently operative policies in achieving the
policy objectives. This type of indicator must closely resemble the character of the
policy objectives to be evaluated.

(2) Trend indicators

The same type of data may be used to detect historical development trends or sudden
shifts in the past, i.e. retrospective evaluation in a more general sense. In situations
where policies are absent or a number of policies are interacting, data series for
sensitive indicators may be useful. The selection of these indicators should primarily
be based on the identification of sensitive social groups and natural ecosystems.

Retrospective indicators can be useful as sustainability indicators in two ways. Firstly,

in a direct manner, their numerical values may point out whether welfare,
environmental quality and amenity levels have in fact been developed towards, or
maintained at, desired levels in the historic period considered. To assess the
effectiveness of a particular policy in that period, the measured data of the
retrospective indicator should then be contrasted with reference values (see the
argwnent given above). Such an exercise leads to a purely retrospective assessment.

This then can be turned into a predictive assessment if the discrepancy between
measured data and reference value leads to an adaptation of the relevant policies. To
predict the future condition resulting from this effort a model is required which
calculates the impact of adapted management on the discrepancy between the policy
effectiveness indicator value and the reference value(s). For illustrative purposes,
assume that there is a yearly "evaluation and adaptation" cycle and that the shape of
 68
the policy effectiveness indicator is a ratio. A generic fonnulation of a retrospective
indicator with predictive meaning then becomes:

VALUE OF STOCK X (T+I) =

[11 POLICY EFFECTIVENESS (T)) • VALUE OF STOCK X (T)

based on

POLICY EFFECTIVENESS (T) =

MEASURED VALUE of STOCK X (T)/REFERENCE VALUE of STOCK X

where (l/POLICY EFFECTIVENESS(T)) represents the adapted management effort.

The predictive power at any moment in time is detennined by the probability that
the adapted policies do indeed decrease the discrepancies. A number of factors affect
this probability, for example, bureaucratic inertia, lack of effective instruments, and
external factors affecting the processes. The predictive power may in fact be greater
if time series data provide a statistical basis for the model used to generate the degree
of improvement that can be expected per unit of adapted management effort. If,
instead of a single comparison between the current stock value and a reference value,
a historical time series of stock characteristics is used, correlated with characteristics
of the actual policy implementation, an empirical model can be developed as a basis
for predictions.

Conclusion

Individual sustainability indicators may thus be quantified and, with limitations, have
a degree of direct or indirect predictive power. The problem in sustainable develop-
ment modelling is, however, that indicators such as material welfare, environmental
quality, renewable resources, and amenity are all very much interdependent, so that
the only good model to base predictive indicators on, is a comprehensive man-
environment systems simulation model. This means coupling indicator variables with
well-measurable system and policy variables.

The robustness and reliability of such a model can be determined by historical tests,
and economic and environmental policy plans can be evaluated for their contributions
to the future sustainability of the whole man-environment system. Such a systems
model makes it also possible to evaluate natural developments and disastrous events
for their impact on the planned development track. Of course, the systems simulation
models have limitations, but at least they allow us to use the monitoring network data
 69

for more than only retrospective evaluations of trends and policy effectiveness. On the
other hand, retrospective indicators as such seem very useful in providing a basis for
creating a better understanding of which development patterns of man-environment
systems at various spatial levels can be sustained.

References

Ackoff, RL. (1962). Scientific method: optimizing aPJllied research decisions. Wiley
& Sons, New York.
Archibugi, and P. Nijkamp (1989). Ecology and economics: towards sustainable
development. Kluwer Academic Presss, Dordrecht.
Bennett, RJ. & RJ. Chorley (1978). Environmental systems: philosophy. analysis and
control. Methuen & Co., London.
Braat, L.C. (1991). Systems ecology and sustainable development: links on two levels.
In: CAS. Hall (ed.) (1991). Maximum power (forthcoming).
Braat, L.C. & I. Steetskamp (1991). Ecological-Economic analysis for regional
sustainable development. In: Costanza, R. (ed.) (1991, forthcoming).
Clark, W.e. & R.E. Munn (1986). Sustainable development of the biosphere.
Cambridge University Press, Cambridge.
Costanza, R (ed.) (1991). Ecological Economics: an introduction (forthcoming).
Doom J. van, and F. van Vught (1978). Forecasting: methoden en technieken voor
toekomstonderzoek. Van Gorcum, Assen.
Environment Resources Limited (1984). Prediction in Environmental Impact
Assessment. SOU, Den Haag.
Jantsch, E. (1967). Technological forecasting in perspective. OECD, Paris.
Jantsch, E. (1972). Technological planning and social futures. Cassell/Ass. Business
Programmes, London.
Liverman, D.M., M.E. Hanson, BJ. Brown and R.W. Meredith, Jr. (1988). Global
sustainability: toward measurement. Environmental Management Vo1.12, n02,
pp.133-143.
Meadows, D.H., D.L. Meadows, J. Randers and W.W. Behrens (1972). Limits to
Growth. Universe Books, New York.
Meadows, D.H., J. Richardson and G. Bruckmann (1982). Groping in the dark: the
first decade of global modeling. Wiley, New York.
Miles, J. (1987). Vegetation succession: past and present perceptions. In: Gray, A.J.,
MJ. Crawley and PJ. Edwards (eds.). Colonization, succession and stability.
Blackwell, Oxford.
Odurn, E.P. (1971). Fundamentals of ecology. Saunders, Philadelphia.
Odum H.T. (1983). Systems ecology: an introduction. Wiley., New York.
Odum H.T. (1987). Models for national. international and global systems policy. In:
Braat, L.e. and W.FJ. van Lierop (1987). Economic-ecological modeling. North-
Holland, Amsterdam.
 70

Opschoor, J.B. (1987). Duurzaamheid en verandering. Free University Press,

Amsterdam.
Pezzey, J. (1989). Economic analysis of sustainable growth and sustainable
development. World Bank Environment Department Working Paper 15.
Washington, D.C.
Slicher van Bath, B.H. (1960). De agrarische geschiedenis van West Europa 500-
1850. Spectrum, Utrecht.
Spain, J.D. (1982). BASIC Microcomputer models in Biology. Addison-Wesley,
London.
Turner, R.K. (1988). Sustainable environmental management. Belhaven Press, London.
Vos, J.B., J. Feenstra, J. de Boer, L.C. Braat and J. van Baalen (1985). Indicators
for the State of the Environment. Instit. for Env. Stud. Report 85/1, Amsterdam.
 7

The AMOEBA approach as a useful tool for establishing

sustainable development?]

Ben ten Brink

1. Summary

The concept of sustainable development is a political rather than a scientific concept.

It is a balance between the environment and human uses. If policy-makers want to
make rational choices concerning sustainable development, they have to define this
concept and formulate verifiable ecological objectives, and moreover they will need
to possess adequate economic and ecological information. The need for verifiable
ecological objectives and adequate ecological information gave rise to the AMOEBA
approach. This chapter reflects on the concept of sustainable development and gives
a short description of the AMOEBA approach. Finally, the possibilities of the
AMOEBA approach for finding quantitative indicators for sustainable development are
evaluated.

2. Sustainable development

Sustainable development is a political concept

Two years after the World Commission on Environment and Development (Brundtland
Commission) had published "Our Common Future" (1988), the concept of "sustainable
development" has been generally accepted as a leading policy concept in most
countries. The question arises which indicators can be used to bring this concept into
practice. To give an answer to this question we first have to consider the concept of
sustainable development in more detail. The Brundtland Commission stated:
"Humanity has the ability to make development sustainable - to ensure that it
meets the needs of the present without compromising the ability of future
generations to meet their own needs. (..) Thus sustainable development can only
be pursued if population size and growth are in harmony with the changing

/ A number offigures in this contribution will be published in Marine Pollution Bulletin. Pergamon Press
has kindly granted pennission to reprint them here.
 72

productive potential of the ecosystem. Yet in the end, sustainable development

is not a fixed state of harmony, but rather a process of change in which the
exploitation of resources, the direction of investments, the orientation of
technological development, and institutional change are made consistent with
future as well as present needs. We do not pretend that the process is easy or
straightforward. Painful choices have to be made".

To bring this abstract concept into practice, many political choices have to be made.
What is meant by needs of the present and needs of future generations? Needs to
what extent, and are these needs the same for the first, second and third world? For
how many generations, etcetera? As the Brundtland Commission said: "It is not a
fixed state ... painful choices have to be made". Consequently the nature of
sustainable development cannot be determined by scientists doing scientific research.
It requires a political choice, which must be continuously adjusted as a result of new
knowledge, changing social requirements or unforeseen developments in the economic
and ecological system itself. It is a kind of balance between the environment and
human uses, a balance between the ecological and the economical system.

Adequate ecolo&ical information is missin&

If we accept that sustainable development is a political rather than a scientific

concept, it follows that policy-makers first have to define this concept and formulate
verifiable ecological objectives next to the existing economic ones. A policy which
achieves both gives rise to sustainable development, and is a coherent economical and
ecological policy. If policy-makers want to make rational choices to achieve
sustainable development, they will also need adequate economic and ecological
information. Information is "adequate" when it:
- gives clear indications whether the objectives will be met,
- is information on the system as a whole,
- is of a quantitative character,
- is understandable for non-scientists,
- contains parameters which can be used for 1O-20-year periods.

The regular flow of economic information on, for example, economic growth, the
debt burden, the balance of payments, inflation, employment, etc., which is published
annually by the OECD, is familiar to everyone. But until now, there has been a
serious shortage of ecological information of the same kind (Fig. 1).
 73

Economical Ecological

QJ
Dow Jones
Information: GNP
Growth
etc.

0°
[J
Policy: Decisions
a
a
a
a
<)
• Interest Raise
Measures: • employment plan ?
• etc.

Result: " Non-substalnable"

Development

Figure 1. Shortage of ecological infonnation.

In fact, ecological infonnation is available in abundance, ranging from indications of

fish diseases and the disappearance of the otter and seal, to the pollution of water and
the sediment layer. This infonnation is, however, always fragmentary, often qualitative
and very detailed. The most important problem is not so much the shortage of
scientific infonnation, but the lack of coherence and the difficulties of placing it in
a practical context for use by policy-makers and the public. Policy-makers cannot
get an overall view of the ecological situation. Moreover, the process of
communicating it to society takes a long time.

Management by accident

In practice it can be seen that in the development of environmental policy calamities

are taking over the role of science: "management by accident" instead of
"management by knowledge and prediction" (Fig. 2). Numerous examples can be
given:

- strict emission regulations for drins, cadmium, mercury and PCBs after cases
of lethal poisoning (Japan and the Netherlands).

- strict shipping regulations (MARPOL) after disasters with the Torrey Canyon
and the Amoco Cadiz.

- 50% reduction in the emissions of certain pollutants as a result of the Sandoz

affair (the Rhine Action Programme).
 74
- 50% load reduction for certain fX>llutants and the acceptance of the
"precautionary principle" at the Second North Sea Ministers' Conference in
resfX>nse to such alanning phenomena as fX>isonous algae, deoxygenation,
disease and large-scale mortality among fish (the North Sea Action
Programme).

Scientists are waiting for imfX>rtant fX>litical issues raised by fX>licy-makers, while
fX>licy-makers are waiting for imfX>rtant ecological issues and ecological indicators
raised by scientists. This is a fatal deadlock (ten Brink, 1989).

Calamities

Scientists
..
,, ,
,
,I "
-,"\ "
-ns ""
~ \,
e:o "' . -0
~

:.
CI')

:e •
u .'
"0.'
I
Q)
~
en
~ ~
o " 0'
Q.
,,' " .-
"
enI ' " , / !!.
" '.:
Policy makers Public
Opinion
"
II

.
,-,,
II

Measures

Figure 2. "Management by accident"

A continuous flow of ecological information, similar to that of the annual economic

figures, must be created. To this end, a convenient number of variables or indicators
will have to be specified, which can be used to give a regular indication of the state
of the environment. These ecological variables will resemble those used for the
economic information.

A similar problem, the need for verifiable ecological objectives and adequate
ecological information for the Dutch management of the North Sea gave rise to the
development of the AMOEBA approach.

3. The AMOEBA approach

The Dutch Water Management Plan

The AMOEBA approach has been developed for and applied to the Dutch Water
Management Plan (Ministry of TransfX>rt and Public Works, 1989). "AMOEBA" is
 75

the Dutch acronym for "a general method of ecosystem description and assessment".
It is a conceptual model for the development of quantitative and verifiable ecological
objectives, and gives the possibility for a quantitative description and assessment of
ecosystems as well. It is based on the concept of sustainable development.
This section gives a short description of this model. In the next section the suitability
of the AMOEBA approach for finding quantitative indicators for sustainable
development is evaluated. A complete, more detailed, description of the model can
be read in ten Brink et at. (1991).

The absence of verifiable ecological objectives

The main objective underlying the Dutch management of the North Sea and inland
waters could be loosely defined as:

The attainment and the maintenance of a water quality level in order to

preserve the ecological values in relation to desired uses of the water system.
This policy is in accordance with the view of the Brundtland Commission on
sustainable development, which at present forms the starting point of
government policy in the Netherlands.

Just like the concept of sustainable development, this objective is difficult to quantify
and to verify. What is acceptable and what is not? Which are the "ecological values"
to be preserved? Are they seals, algae, bacteria or herring gulls? Neither the species
nor the desired numbers are stated explicitly. Yet choosing species A can lead to a
totally different policy than a choice for species B or C (Fig. 3).
 76

,
Ecological objective: "conservation of ecological values"

1\
,
Concrete objective:

Measures:
/\/\
A
Seals or

B or C
Algae

D ?
?

" Amoeba" stands in Dutch for" general method for

ecosystem description and assessment".

Figure 3. What must be preserved: algae or seals? This makes a big difference.

Fundamental ecological values

In the AMOEBA approach the term "ecological values" is defined as: "The desired
state of the biotic component of the water system. This desired state is to be
determined by the government". The subject of study is therefore the biological
component of the water system. The physical and chemical components are considered
means by which to reach the ecological objectives. The landscape aspect is not taken
into consideration (Fig. 4).
 77

Shore

Water

Biological
Bottom Chemical
Physical

I"· \ I Subject of study

Figure 4. The subject of the AMOEBA approach is the biological component of the
water system.

In order to establish precisely what is meant by these ecological values, the most
fundamental values man attributes to plant and animal life are examined. There are
three categories of valuable characteristics whose sustainability is desirable:

1. Production and yield

These are valuable for functional reasons. This category is a prerequisite for man's
existence, e.g., fisheries. These values are closely associated with the abundance of
species, the production of oxygen and the self-purifying capacity.
2. Species diversity
This is valuable for ethical and aesthetic considerations. It involves concepts such as
the preservation of species, rarity and completeness.
3. Self-regulation
Self-regulation has ethical, aesthetic/recreational and economic considerations, which
are closely related to concepts such as naturalness, stability, intactness, authenticity
and visual integrity. Moreover, self-regulating ecosystems have low management costs.

A guaranteed sustainable ecosystem: the reference system

Which ecosystems and kinds of uses and measures provide guarantees for sustainable
production, diversity and self-regulation? It has become clear that unrestricted
discharges of contaminants or fishery threaten sustainable production and diversity.
 78
A system which has not at all, or only slightly, been influenced by human activities
may provide clues to define parameters and processes essential for sustainability. Such
a system contains the conditions for the evolution and survival of organisms, including
man, living in and around it for millennia.

The assumption is made that an ecosystem which is hardly or not at all manipulated,
offers the best guarantee for preservation of these fundamental values: the
REFERENCE system. The closer one comes to the point of reference, the larger the
guarantee for ecological sustainability, and vice versa. Society chooses her objectives
somewhere between zero and the point of reference. This is, in essence, making a
choice between the direct costs of measures and the loss of guarantees for
sustainability in the long term (Fig. 5). The search for a concrete ecological objective
can therefore be reduced to the question: "what is the maximum acceptable distance
to the point of reference"?

Present Objective

Bad Good
, I

-------1 ~.....M·.·a·s·u·r.·s......

. .-1-.·
o Reference

Costs Costs

Long term? Short term

Figure 5. Society chooses the ecological objective somewhere along the axis between
zero and the reference point.

The introduction of a reference system provides a standard by means of which an

assessment can be made of the ecological condition of a system. The ecological
objective, however, does not necessarily coincide with the reference system. Once a
government decides on the maximum acceptable distance from the reference point, a
verifiable ecological objective is established.

Species as target variables

A quantitative comparison between the reference system and the present-day

ecosystem has been made for 60 plant and animal species in terms of their:
 79

- numbers
- distribution
- health
They are called the target variables. Together they are the parameters by which the
ecological objective is expressed.
Six criteria have been used for choosing the target variables, including those
mentioned under the heading "Adequate ecological information is missing" (above).

Three sources are used to determine the reference system:

- old inventories
- comparative research involving other systems
- ecological theory.

The AMOEBA anvroach in practice

In the Water Management Plan the AMOEBA approach is applied to the North Sea
and major rivers (Rhine and Meuse). Reference numbers and actual numbers are given
for 60 selected target variables, most of them species. Since water authorities and
policy-makers require a clear and simple presentation, a "radar diagram" has been
used. The target variables are arranged in systematic order in the form of a circle.
The distance from the edge of the circle to the centre represents the numbers in the
reference situation for each species, for example, 2000-3000 breeding pairs of
cormorants, 16,000 hectares of sea grass, or 4000-6000 seals (=100%). The actual
numbers are superimposed on this circle, for example, 500 breeding pairs of
cormorants and 3000 hectares of sea grass. For visualization purposes all points are
connected by a line, which produces the two amoeba-like figures. These figures, the
"AMOEBAs" for the sea and the major rivers, present a relatively simple picture of
the actual situation of these ecosystems (Fig. 6).

Impact AMOEBAs

To approach the reference system a strategy has been drawn up. Conceptual and
mathematical models have been developed to assess changes in all species in the
"AMOEBA" as a result of changing variables such as water quality, fish catches,
restoration of biotopes, etc. The impact of six policy alternatives (packages of
measures) was calculated (Ministry of Transport and Public Works, 1990) and
illustrated by "impact AMOEBAs". Two "impact AMOEBAs" for the sea are shown
in Figure 7.
 80

_.oIIol
Situation 1988 Situation 1988
cr........
CJ ....... CJ ...
~ -"(till) ~ 0'(1_

-
A1lntl'

... -...._.....-... - linda,

"""lIlp
CommDft Dot wtl.11l.

Mlntn, LobItt, "um., Anemone

Figure 6. The selected target variables showing the present ecological situation in the
sea (A) and the major rivers (B).

Policy-option 4 Policy-option 2A+

(......motbt.)
<_.ok)
90% reduction of 50% reduction + 90%' 99%
reduction of PCB, TST, PAC +

--- .
pollutanla I nutrlenla
auppllll11_ry muaurH
e:tfKt2010
!l'ttet20.t0

Improvtmllnt ""'01 2010

OtlwlorlUon tf'rtCt 2010

""""'U(1NO)

Figure 7. Expected ecological impacts of two policy alternatives: the "impact

AMOEBAs".
 81

Ecological Dow Jones Index

To simplify and quantify the comparison between the effect AMOEBAs, the
AMOEBAs are expressed in terms of an index number. This was achieved by adding
up all of the distances (32) to the reference circle. The resulting figure has been
named the "ecological Dow Jones index". The smaller the total distance from the
reference situation, the better the policy alternative. No distinction has been made
between the values for the various species. If these figures are arranged on a scale
from 3200 (32 x 100% distance) to 0 (32 x 0% distance), the result is as shown in
Figure 8.
Ecological" Dow Jones index"
in ~ reference %
Predicted situation in 2010 (2030)
Present

I~
Extinct
or
Doubled 2 ....
hl\
...,....
1 2 • 0 ... 75...

0 00 JJ800
III II II!
I
3200 2400

~
Figure 8. The "ecological Dow Jones index" of six "impact AMOEBAs".

4. AMOEBA as indicator for sustainable development?

To what extent can the AMOEBA approach play a role in this flow of ecological
information? As is shown above, first there has to be a verifiable ecological objective.
The Dutch Water Management Plan to which the AMOEBA approach is applied
states:

"In accordance with the recommendations of the Brundtland Commission,

policy in the long term is aimed at achieving sustainable development. A
sustainable development situation will have been achieved when the numbers
of the species included in the sea and river AMOEBAs approach those of the
reference situation. It is not necessary to return to the reference situation
completely. For saltwater areas, the aim is to reach between 75% and 200%
of the numbers in the reference situation (the target AMOEBA, Figure 6). For
freshwater areas, the corresponding percentages are 50% and 200%. The
minimum level is lower for freshwater areas than for saltwater areas, because
 82
a higher human interference is accepted in the freshwater systems by
industrialization and the population density in the Rhine, Meuse and ScheIdt
river basins. At these levels, there are acceptable guarantees on sustainable
ecosystems and sustainable use".

Figure 9 shows this ecological objective in greater detail for the sea. This is called
the target AMOEBA.

ObJec:ttve 2AP40BJZ.DRW
..-l

-
-- ... - ...-
----
~-_ .......,..

Figure 9. The ecological objective for the North Sea, the target AMOEBA.

With this objective a political choice has been made. Although this objective will
undoubtedly be adjusted in the future, policy-makers have defined sustainable
development in a verifiable way: the target AMOEBA. Every four or five years new
sea and river AMOEBAs will give a description and assessment of the situation. As
long as the situation as shown in the target AMOEBA has not been reached it has
to be considered whether supplementary measures have to be taken to achieve
sustainable development. In this case the AMOEBA can serve as an adequate
indicator for sustainable development by definition because of the reason stated above
and because it meets the demands of adequate information.

With the aid of the AMOEBA a balance can be found between sustallllllg the
ecosystem and its use in a rather systematic way. AMOEBA and the economic figures
are the two compasses by 'vhich policy-makers can plot a course towards sustainable
 83

development (Fig. 10). The success and failure of four years of government policy
can simply be read from these compasses.
INTEGRAL WATER MANAGEMENT Sustainabitity

A SYSTEMATIC SEARCH
INTO
" The 7 STEP PLAN "
Drinking
Water

Sand or grovel

Shlppmg

USE BALANCE SYSTEM

Figure 10. Sustainable development is a balance between safeguarding the ecosystem

and its economic use.

5. Evaluation

From the above it can be concluded that the AMOEBA may be a useful indicator for
sustainable development. But there are still some unanswered questions:

Ouestions on sustainable development

- Can the reference system give guarantees for sustainable development?

- Is it possible to reconstruct a reference system that faithfully reflects the
uninfluenced situation?
- Is the target AMOEBA really sustainable?
- Can every deviation from the reference system be dismissed as nonsustainable?

Answers

- By definition, the reference system guarantees sustainable production, diversity

and self-regulation. The very existence of man and the species of plants and animals
which we know today is proof of this.

- Selecting target variables and reconstructing the reference numbers are not
one-off activities. It may take some time before general scientific agreement is
 84

reached. When choosing a reference year, for example, 1930 for the North Sea, a
correction will be necessary for human influences for several species. The reference
system is reconstructed step by step using individual components in much the same
way as archaeologists reconstruct the past.

In the first few years the determination of sustainable development in terms of

verifiable objectives will proceed on a trial-and-error basis. This is not to say that
it will be done at random, but rather that each possibility will be tested systematical-
lyon the basis of the following considerations:
* Which objectives are both practically and economically feasible?
* Which objectives offer reasonable guarantees for sustainable use and
sustainable ecosystems?
* Which measures give high returns at low costs?
* What kind of time scales are involved?
This is a learning process, in which provisional ecological objectives become more
definite as time passes. Adjustments to ecological policy are introduced as a result
of new knowledge, changing social requirements or unforeseen developments in the
system itself. These are the factors behind the continuous "dialogue with the water
system". A similar trial and error approach is standard practice in economic policy-
making. The OECD publishes its economic predictions on the basis of the latest
figures and perceptions. These predictions are adjusted annually, and form the basis
of financial and social-economic policy. I recommend an ecological policy that is
based on a similar systematic approach (ten Brink, 1989).

- The more extensively a water system is used, the more the ecosystem is
affected, and vice versa. These are the two extremes of a balance which is tilting in
the wrong direction (Fig. 10). The problem is to find a way to optimize both sides,
so that acceptable guarantees can be provided for both the sustainability of the water
system and its use. This balance cannot be calculated, but will have to be determined
by experiment; firstly, because sustainable development is a political concept and,
secondly, because the relation between the system itself and its use is too complex
to be ever completely disentangled.

As with economic figures, uncertainties are unavoidable with ecological figures and
there are no definite answers. It is up to policy-makers to identify acceptable
uncertainties. The only real solution to great uncertainties is to report on the situation
regularly and modify policy accordingly (Fig. 11) (ten Brink, 1989).

- Not every deviation from the reference system can be dismissed as

nonsustainable. Even in systems which deviate from the reference it is conceivable
 85

that all the species will survive in the long tenn. Our perception of the maximum
acceptable distance from the reference will change, as we acquire more knowledge.

Advise Ate

t
time
Advise A'

Advise A

Knowledge

Figure 11. Ecological uncertainties and increasing know-how induce a continuous

adjustment of environmental policy and advice.

Ouestions about the indicative character

- Can the AMOEBA act as an overall thennometer for the entire system?
- Did the use of the AMOEBA approach lead to an oversimplification of the
real situation?
- Will there not be another AMOEBA when choosing another set of target
variables?

Answers

- The danger with the AMOEBA approach is that the target variables may
acquire a life of their own, and that the ecosystem as a whole will suffer neglect.
This can easily lead to an oversimplification of the real situation. A solution to this
is to increase the number of target variables, so that a broader range of system
 86

components is included in the decision-making process. For example, in addition to

the target variables of the sea AMOEBA, 40 biological target variables have been
used for the assessment of the ecological situation. Furthermore, it is advisable to
supplement chemical and physical target variables when carrying out this analysis
and assessment.

- Different choices of target variables are possible within the limits imposed
by the above mentioned six criteria. Undoubtedly this leads to a different AMOEBA.
The ultimate aim, however, is not a particular shape of the AMOEBA, but merely to
bring out the changes which occur over a certain period and the measures to be taken.
We do not expect that these differences will lead to measures which differ in
principle. Moreover, choices have to be made for formulating objectives, as was done
in, among others, the world of economy and the medical world.

Ouestion about the costs

- Many countries will not posses the financial and technical means to research
and monitor so many species. Does this mean that the AMOEBA approach
will only be an instrument for rich countries?

- No! New species may be added to the AMOEBA anytime. One might start
with species which are most critical and easy to monitor. Some information is better
than none at all. In this manner a framework of managing aquatic ecosystems can be
constructed gradually.

Conclusion

It is of the highest importance that the natural resources of the planet are preserved
in order to reach a state of sustainable development. The AMOEBA approach can
serve as a valuable instrument for policy-makers for the management of their
countries' natural resources, our common heritage.

References

Brink, BJ.E. ten (1989). Systematically in search of sustainable development.

Memorandum GWWS 89.006, Tidal Waters Division, The Hague.
Brink, B.J.E. ten, Hosper, S.H., Colijn, F. (1991). A quantitative method for
description and assessment of ecosystems: The AMOEBA approach. Proceedings
 87
of the International Conference on the Environmental Management of Enclosed
Coastal Seas, 1990, Kobe, Japan. In press. Marine Pollution Bulletin.
Ministry of Transport and Public Works (1990). Ecological impact assessment on
coastal waters. Tidal Waters Division. Note GWWS 90.009. The Hague (in
Dutch).
Ministry of Transport and Public Works (1989). Third National policy document on
water management. A time for action. The Ha&ue.
World Commission on Environment and Development, Commission Brundtland (1988).
Our Common Future. Univ. Press. Oxford.
 8

TOWARDS SUSTAINABILITY: INDICATORS OF

ENVIRONMENTAL QUALITY}

Helias Udo de Haes, Maarten Nip, Frans Klijn

1. Introduction

In environmental policy the concept of sustainability refers to a sustainable

relationship between society and environment (see Figure 1).

1
<

society environment

2
>

Figure 1. Sustainability concerns sustainable relationships between society and

environment.
1 = resource functions of the environment for society;
2 = impacts of society on the environment.

In fact, such sustainability is achieved when there is a certain stability, or balance,

in the relationships. This stability, however, may be achieved at a level of very poor
quality of either society (development level), e.g., in developing countries, or the
environment (environmental quality), e.g., in highly polluted industrial areas, or both,
such as in eroding mountainous areas in developing countries or slums around large

/ I1lis contribution is largely based on research in the context of 'area-oriented integration', a research
project of the RlVM, which was worked on by: J.J. Hofttra. J.B. Latour and P.K. Koster (RlVM), M.M.H.E. vall
dell Berg, C.L.G. Groe,~ F. Kliin, M.L Nip and H.A UM de Haes (eMU
 90

cities in the industrialized world. This means that sustainability cannot be defined,
nor measured, without relating it to quality standards for society as well as for the
environment. In the second instance, the desired 'equilibrium' between the two must
also be defined.

The quality of societies is often assessed by means of economic standards, although

these cover only one aspect of well-being. For determining environmental quality no
generally accepted methods are available yet, although there are standards for some
aspects of the environment, such as for concentrations of toxic substances. In this
contribution, emphasis is on environmental quality, which is a normative concept, of
course. A method of assessing environmental quality will be presented, based on
research being carried out by the Centre of Environmental Science (CML) and the
National Institute of Public Health and Environmental Protection (RIVM) in the
context of 'region-oriented integration' (Latour et al., 1990; Klijn et al., 1990a;' Klijn
et al., 1990b). Environmental quality, in this context, is defined against the desire
from society that the environment will sustainably function as a resource for various
kinds of land utilization, as well as support intrinsic nature values (species diversity
a.o.).

2. Ecosystem approach and environmental quality assessment

In this contribution the physical environment is regarded as an ecosystem, i.e., an

interdependent system of abiotic and biotic components. Such an ecological approach
to the environment has received much attention in recent years. In water policy in
particular the ecological approach has been used extensively (see, for example, the
National Policy Document on Water Management, Ministry of Transport and Public
Works, 1989). This resulted in a comprehensive set of ecosystem parameters, both
biotic and abiotic, to be used in assessing surface water quality (CUWVO, 1988).
Another recent development in this field is the so-called AMOEBA approach (ten
Brink and Hosper, 1989) in which the present abundance of plant and animal species
as well as that of abiotic features in the environment are compared with a historical
reference.

For terrestrial areas, too, a coherent ecological framework is being established for
policy-supporting analyses. So far this has resulted in an ecodistrict classification and
mapping (Klijn, 1988; Klijn and Udo de Haes, 1990), which has served as a
geographic basis for the Dutch National Environmental Survey Concern for Tomorrow
(RIVM, 1989). A next step in this research programme concerns the development of
a comprehensive method of assessing environmental quality of both aquatic and
terrestrial ecosystems, specifically for regional ecological characteristics (i.e.,
ecodistricts).
 91

In ~s contribution the first outlines of this method are presented. Firstly, the spatial
scale for which the environmental quality is to be assessed is discussed. Secondly,
the fact that different land utilization types require a different environmental quality
is dealt with; a possible solution to this dilemma is suggested. Then, the significance
of defining a consistent environmental quality objective is discussed. This objective
must be defined in general qualitative terms, connected with the desired land
utilization types including nature conservation. The next section concerns the
specification of the objective in terms of measurable environmental quality parameters.
The criteria which play a role in selecting such parameters are discussed. Then, the
problem of setting quantitative standards for the selected parameters in terms of target
values, and the problem of quantifying the present value of each parameter, is looked
at briefly. Next, a comprehensive method of presenting the results for policy-makers
will be dealt with. Finally, an example is given as to how the presented method may
work out for the Lowland Peat area in the Netherlands.

3. The choice of areas

Firstly, it has to be decided on which spatial scale environmental quality has to be

assessed. In this context we plead for quality assessment at a scale level which suits
the policy questions, but which can also be differentiated easily if required. For
national environmental policy quality assessment at a regional level may yield the
most applicable results, in our opinion. It allows for a region oriented environmental
policy with differentiated policy objectives and measures (Klijn and Laansma, 1990).

In this context it is considered worthwhile to link up with the hierarchical

classification of ecosystems of Klijn (1988; Klijn and Udo de Haes, 1990) (see Figure
2), which allows elaboration at different spatial scales. The classification concerns six
spatial scale levels. At each scale level the mapping characteristics are chosen in
connection with ecological processes operating at that very level. It is proposed that
the ecodistrict classification should be used for the Dutch national environmental
policy. The ecodistrict classification has an indicative scale range of 1:500,000 to
1:2,000,000. In the Netherlands, 26 terrestrial ecodistrict types with a total of 83 areas
(ecodistricts), and 11 aquatic ecodistrict types with a total of 16 areas, are
distinguished.
 92

MAPPING SCALES BASIC MAPPING UNIT

ECOZONE (1: > 50,000,000) >62,500 sq.km

ECOPROVINCE (1: 10,000,000 - 50,000,000) 2,500 - 62,500 sq.km
ECOREGION (1: 2,000,000 - 10,000,000) 100 2500 sq.km
ECODISTRICT (1: 500,000 - 2,000,000) 625 - 10,000 ha
ECOSECTION (1: 100,000 - 500,000) 25 625 ha
ECOSERIES (1: 25,000 - 100,000) 1.5 25 ha
ECOTOPE (1: 5000 - 25,000) 0.25 1.5 ha
ECO-ELEMENT(l: < 5000) < 0.25ha

Figure 2. Proposed tenninology for a hierarchical system of ecosystem classification

on different spatial scales.

4. Multifunctional areas and environmental quality requirements

Most areas fulfil different functions, or, in other words, carry different land utilization
types. These may require different environmental characteristics. For example, agri-
culture in the Lowland Peat area in the Netherlands requires a relatively low
groundwater level for reasons of carrying capacity for both agricultural machines and
cattle. Nature conservation in this area, on the contrary, requires high groundwater
levels for the protection of characteristic plant and bird life. How to deal with such
contradictory requirements when detennining environmental quality?
Various sectors in environmental policy have dealt with this problem differently. In
water policy a basic quality was defined for all functions. For specific areas more
strict 'ecological' standards were set. In fact, these are standards in behalf of nature
conservation. In soil policy the concept of multifunctionality is applied. This implies
the protection of soil characteristics for all present and future land use functions,
taking into account a certain period of recovery. In environmental hygiene policy the
concepts of general and specific environmental quality have been introduced (GEQ
and SEQ respectively). Finally, in the National Environmental Policy Plan (Ministry
of Public Housing, Physical Planning, and Environment, 1989) it is stated that general
environmental quality is the same as basic quality or multifunctionality. The concepts
of general and specific environmental quality have become the standard concepts for
(integrated) environmental policy as a whole.

Thus, two essentially different quality levels are recognizsed in environmental policy,
viz., general environmental quality (GEQ) and specific environmental quality (SEQ).
These quality levels are defined as follows (Ministry of Public Housing, Physical
Planning, and Environment, 1989):
 93

General environmental quality (GEQ) is: 'such a quality of the environment ....
that the health and well-being of man, as well as the preservation of wildlife
(animals and plants), goods and land use types are safeguarded in general'. It
relates to a basic quality level which concerns the entire country. If desired it can
be specified on a regional basis.
Specific environmental quality (SEQ) provides additional or more stringent quality
demands for the preservation of susceptible land utilization types or vulnerable
culture or nature values.

These concepts can be used as guidelines. From the definition it follows that GEQ
standards primarily concern economic land utilization types (multifunctional), but also
safeguard a basic nature value. Because of the different requirements of different land
utilization types, compromises may be necessary.

SEQ standards, in our opinion, should primarily concern nature conservation. In

addition, however, other land utilization types may be fulfilled in a SEQ area, as long
as they do not conflict with nature conservation (for example, outdoor recreation,
education). This interpretation of SEQ slightly deviates from the definition cited
above, which also recognizes SEQ standards for susceptible land utilization types. This
deviation is based on our opinion that, in the end, GEQ standards should provide
enough protection for all economic land utilization types, even if present GEQ
standards may not suffice yet. Hence, GEQ standards should also protect forestry or
groundwater exploitation for public water supply, although it is recognized that for
the latter at present only the so-called strategic water resources are protected in SEQ
areas.

To summarize, GEQ and SEQ are interpreted as follows:

GEQ implies:
- conditions enabling sustainable economic land utilization to an extent that the
various land utilization types do not conflict; the land utilization type which
has the strictest requirements determines the quality standards;
the presence of basic nature values and related conditions; the GEQ standards
hence also comprise standards concerning the actual presence and abundance
of organisms;
the possibility of achieving specific environmental quality, i.e., meeting higher
standards for nature conservation, within a limited period (protecting the
'potential');
such a quality that nature values in adjacent SEQ areas are not threatened.

SEQ implies:
- the presence of high nature values and related conditions;
- conditions enabling sustainable land utilization to an extent that these do not
negatively affect nature values.
 94
With this distinction, however, the protection of nature values is not sufficiently
assured yet. Firstly, it may be desirable to preserve the present quality when this is
better than the standards prescribe. In such a case, the principle of a 'steady state',
Le., no further deterioration of the present quality, may apply. Do note that this is not
exactly the same as the well-known 'stand-still' principle, which primarily concerns
the societal activity which causes the environmental effects. Secondly, no minimum
size of SEQ areas, nor their exact location, are determined yet. These would, however,
appear to be primarily a concern for nature policy and physical planning.

5. Determining an environmental quality objective

A focal point in assessing environmental quality concerns the definition of

environmental quality objectives. For most economic land utilization types the
objectives would be the same for all ecodistricts; for others a regional specification
may be necessary, i.e. per ecodistrict (type). This goes especially for nature
conservation objectives, which have to be in tune with the potential.

Determining an environmental quality objective is possible only in relation to more

general policy objectives for the area. Such a general policy objective implies a
decision on the parts of the area which are to be multifunctional (GEQ) on the one
hand, and the parts which primarily have a nature conservation function (SEQ) on the
other hand. This policy objective will then have to be specified in terms of desired
environmental characteristics, both abiotic and biotic, including the kind of nature in
terms of desired plant and animal communities. This specification in terms of
environmental characteristics forms the true 'environmental quality objective'.
However, the specification is still in general qualitative terms.

As a next step the objective must be translated into measurable and quantifiable
parameters, for which numerical standards (target values) can be set. The choice of
parameters and the defining of target values will be treated in the following two
sections. Since policy objectives for specific ecodistricts or ecodistrict types are often
either entirely lacking or insufficiently defined, it may be necessary to deduce them
from a variety of policy documents. For the Netherlands, for example, policy
documents on nature conservation and development (Ministry of Agriculture and
Fisheries, 1989), on large surface waters (Ministry of Transport and Public Works,
1989), and on physical planning (Ministry of Housing, Physical planning and the
Environment, 1990) have been used. This will yield a guideline for deciding for GEQ
or SEQ, as well as for the relevant and foreseeable land utilization types.

When the decision on GEQ or SEQ has been made, the translation into environmental
quality Objectives must be achieved. At this stage we are confronted with the problem
of how to define the quality objective: are we to strive for historic situation
 95
(reference), or should the objective be something totally different from the former
situation because of either irreversible changes in the environment or changed
requirements as to the character of nature reserves? In background documents to the
Nature Policy Plan on nature development it is outlined what undisturbed nature
would look like in various parts of the country. In a comparable manner a
reconstruction of the environmental characteristics of the North Sea in 1930 has
served as a reference in the AMOEBA method (ten Brink and Hosper, 1989). The
quality objective for the North Sea has been defined as a compromise between this
reference and the present state.

For terrestrial nature reserves the quality objective may be defined as resembling a
past reference or a reference elsewhere. For GEQ areas with other land utilization
types dominating over nature, however, it is less simple to define the objective. In
any case, a 'monofunctional nature' reference has little significance, since deviations
from this reference result automatically from the desire of multifunctionality, with
nature subordinated. So, an environmental quality objective for GEQ areas will have
to be defined independently. Requirements from the point of view of sustainable
economic land utilization must have primacy in this. As far as basic nature values are
concerned, past references may give a clue as to the objective, but nature development
under the present abiotic circumstances may result in an entirely different outcome.
The quality objectives for both GEQ and SEQ will have to be defined for each
ecodistrict type. It may be necessary to differentiate between individual ecodistricts
or even separate nature reserves.

6. The selection of quality parameters

An environmental quality objective will predominantly be defined in qualitative terms.

This does neither allow determining the present quality of the environment in
comparison with the objective nor the monitoring of changes in this quality.
Therefore, the objective must be translated into measurable quality parameters.

In selecting quality parameters the following four criteria are important:

- relevance to environmental policy
- sensitivity
- detectability
- appeal

Relevance to environmental policy is the central criterion: the quality parameters have
to be good measures, or indicators, for the sustainable resource functions of the
environment. Hence, they have to relate to the requirements of land utilization types,
such as agriculture, forestry, public water supply, recreation and nature conservation.
This implies a direct relationship with the quality objective mentioned above.
 96
Parameters must be sensitive in the sense that they reflect environmental changes due
to human activities and preferably react immediately. This implies that they must react
on environmental processes such as acidification, eutrophication, groundwater lowering,
disturbance by noise, intoxication, etc. Changes in the value of such parameters can
then be contributed to these processes and their causes, giving direction to policy
actions. Furthermore, knowledge on the process-response relationship with respect to
such parameters allows for predicting changes with different policy scenarios.

Detectability refers to the ease with which the parameters can be determined and
quantified. A parameter which can be determined with aerial photos has an advantage
over a parameter which requires field measurements or even laboratory work.

The appeal of the parameters can be a real support for environmental policy since it
may enlarge the support by society of a policy focused on quality improvement.
Secondly, it enables laymen or policy-makers to detect changes in environmental
quality themselves. In this context, especially the so-called 'ambassador species',
predominantly mammals or birds which have a high 'caressing factor', may help. Of
course, appeal is not required for all parameters.

It is proposed to select abiotic (physical, chemical) and biotic quality parameters.

Because the National Environmental Policy Plan clearly stated that GEQ must also
safeguard basic nature values (animals, plants and ecosystems), it is considered
necessary to measure these directly, i.e., with biotic parameters. For SEQ areas the
biotic parameters are closely related to the quality objective itself, which is defined
in terms of desired plant and animal communities. In these areas, abiotic parameters
provide a kind of safety net: only the combination of desired abiotic conditions and
the actual occurrence of relevant species suffices to indicate a sustainable quality.

In Figure 3 the relationship is shown between ecosystem components from which

quality parameters may be derived and those land utilization types to which they may
be relevant.
 97

LAND UTllJZATION TYPES:

Agri- Forestry Fisheries Water Recreation Nature

culture Supply

PARAMETERS:

ABIOTIC
Geomorphology * * * * * *
Soil * * * *
Groundwater * * * *
Air * * * * *
Water * * * * * *
BIOTIC
Vegetation str. * * * * *
Flora * * * *
Fauna * * * *
Figure 3. The significance of some ecosystem components for land utilization.

As far as plant communities (flora) are concerned, it is practicable to link up with the
ecotope classification (Stevers et ai., 1987). Ecotopes are defined as land units which
are homogeneous as to vegetation structure, succession stage, as well as the main
abiotic site factors which are relevant to plant growth. In this classification system
species are classified in ecological groups belonging to a certain ecotope type. Firstly,
those ecotope types are selected which are characteristic for an ecodistrict type, with
a further subdivision for the GEQ and SEQ areas. Then, a number of plant species
are selected from each ecological species group belonging to an ecotope type.
A similar procedure can be followed for the fauna. Instead of ecotope types, habitat
types are used.

7. Quantification

For each parameter a method for quantification has to be established, both for the
current value and the target value (standard).

The current situation is quantified by using existing data. Field inventories and
monitoring networks have been used in this process. Also, new monitoring networks
have been proposed, focusing on the selected parameters.

The target value of most abiotic parameters is based on existing standards. For biotic
parameters historical data (as a reference), data from comparable areas and expert
judgment are used in setting a standard.
 98

For economic land utilization types the point of departure is sustainable economic
functioning, without affecting the carrying capacity of the environment. For nature
conservation, the preservation of naturalness and of species diversity are the main
objectives. For plant species parameters, for example, target values are defined
with regard to the presence, abundance and minimum share of species per group per
area unit (for example a grid cell of 1 km2).

The ultimate goal concerning target values is that numerical standards are set for all
parameters. These need not be static, since neither our requirements on the
environment nor the environment itself are static. Therefore, it is desirable to be able
to change the parameter, or the numerical standard once set.

8, Presentation

The result of quality assessment can be presented on three levels. Firstly, we may
distinguish changes on the level of separate species and separate abiotic parameters.
These can be documented in a list. Secondly, for policy-makers a graphic presentation
is useful, preferably in a form which resembles the AMOEBA presentation (ten Brink
and Hosper, 1989). In the original AMOEBA presentation three values are plotted in
a circular figure for each parameter: reference values, target values and current
(measured) values. The circle represents the reference values, i.e., the undisturbed
situation. Target values may both exceed or fall short of the reference values,
depending on the parameter. The compromise between quality objective and historic
reference situation is visualized by the discrepancy between the two values. The
difference between the target values and the current values, in their turn, indicates the
quality deficit. The strength of the AMOEBA approach is mainly its comprehensible
presentation of data and objectives.

For the quality assessment of ecodistricts it is proposed to present two'AMOEBAs'

per ecodistrict: one for the GEQ parts, and one for the SEQ parts. Both compose a
reproduction of aggregate parameters. The aggregate results from combining species
into species groups and abiotic factors into composite parameters (e.g., all heavy
metals together). Additional information is required which compares the surface of the
SEQ area (nature conservation) with the target value for this surface.

The proposed presentation deviates on some points from that in the original
AMOEBA approach. Because of the irrelevance of reference values for GEQ areas,
as argued above, target values are plotted on the circle instead of reference values.
Another point of difference is the significance of exceeding or falling short of the
target line. In the original AMOEBA all deviations from the reference are considered
a sign of imbalance and hence undesirable. This may hold for a system with strong
feedback mechanisms between the parameters. In the case of terrestrial areas
 99

influenced by man such balance situations are not to be expected. Exceeding or

falling short of the target values cannot simply be characterized as negative.
Therefore, the difference between negative and positive (or neutral) deviations is
marked (hatched and white respectively; see Figure 5).

Finally, a further aggregation to one environmental quality index is possible, though

only at the cost of information. In the present research project no such aggregation
has been attempted yet. Despite the many problems to be encountered in such an
aggregation, aggregation as such appears useful.

9. An example: the Dutch Lowland Peat area

In this section, an example will be given for the Lowland Peat area (Ecodistrict type
H5, see Figure 4) to illustrate the procedure of assessing environmental quality.

At the time of preparing this contribution no target values had been defined yet, nor
had the current values of the parameters been determined. Both will be elaborated
in behalf of the State of the Environment Report which is due for 1991 (RIVM, in
prep.). In this example a tentative quality judgment is given based on educated
guessing. It primarily concerns the definition of a quality objective with regard to the
nature value of the ecodistrict type, since the requirements of the economic land
utilization types are considered to be too well-known already. The selection of
environmental quality parameters is exemplified only for the GEQ parts of the
ecodistrict type.
 100

Figure 4. Ecodistrict type H5, distribution of the Lowland Peat Area in the
Netherlands

GEO and SEO areas

The Lowland Peat area is predominantly used for agriculture, especially dairy farming.
This GEQ part has secondary functions for recreation and nature conservation. It
consists of headlands with ditches and canals. The rich vegetation and bird life
(almost 3/4 of the European population of the Black-tailed Godwit breeds in the
Lowland Peat area) is related to the low-intensity dairy farming.

Some parts of the Lowland Peat area have high nature values. They are partly
protected as reserves. For these areas, the swamps (including marshwoods), lakes and
reed marshes, SEQ is necessary.

GEO objective

The spatial pattern of the landscape will be grafted onto the historic allotment of the
reclaimed land. The groundwater level will remain sufficiently high to counteract
oxidation of the peat, especially in the bog centres. On the other hand, the height of
the groundwater level must not restrict agriculture too much. This implies an optimal
use of the current differentiation between relatively dry and relatively wet lands. The
concentrations of heavy metals and other toxic matter in the soil will not increase and
will stay below existing soil standards, while the quality of groundwater and surface
water will be such as to allow drinking water production through bank infiltration.
 101

Nature will continue to be an important function of the GEQ parts of the Lowland
Peat area. It concerns the preservation and development of nature values connected
with dairy farming: ditches with a rich flora (in places dependent on upward seepage
of groundwater) and a sustainable population of meadow birds and mammals. The
ditch banks will carry species-rich herbaceous vegetations of sites with a moderate
nutrient availability. The plant growth will sustain a rich fauna, including various
species of butterflies. The ditches themselves will be partly covered with macrophytes
(broad-leaved waterplants) or bordered by reed fringes in slightly brackish areas. All
surface waters are clear to considerable depth, and are not dominated by duckweed,
algae or Azolla. The water vegetation is such as of water of moderate nutrient
availability. In the reed fringes, birds find nesting opportunities and food. Amphibians
and macrofauna are abundant.

SEQ objective

The parts of the Lowland Peat area for which the nature function is of primacy are
the lakes, reed-marshes, swamplands (including marshwoods) and also some meadows.
These SEQ areas comprise the so-called 'ecological main structure' of the Dutch
Nature Policy Plan (Ministry of Agriculture and Fisheries, 1989), which is the spine
of the SEQ areas in the Netherlands. Land utilization types which either depend on
high nature values or do not negatively affect nature, may be allowed as co-functions
in SEQ areas. These include small-scale outdoor recreation, public water supply,
fisheries, and cultivation of thatch.

In the shallow lakes, typical of the Netherlands and originally with a moderate
nutrient availability, plant and animal communities rich in species and with all trophic
levels represented are found. This implies that predators, such as Qtters and Black
Terns, will be abundant. No symptoms of eutrophication are manifest. In fens and
bogs plant communities are found which belong to various stages of peat formation.
The lakes provide attractive places for water recreation and fishing (quality objective
as in the National Policy Document on Water Management (Ministry of Transport and
Public Works, 1989».

The reed marshes and swamps house a rich avifauna and insect life. The groundwater
level is very high everywhere. Individual SEQ areas are connected by means of a
well-developed ecological infrastructure.

Choice of parameters

The quality objectives have been translated into measurable abiotic and biotic quality
parameters. In this example, which is merely an illustration of the procedure, a
 102

number of quality parameters will be mentioned for the GEQ area only (see below).
For SEQ areas the procedure is the same, but either more sensitive parameters are
selected or the standards are set more strictly. For GEQ areas the following
parameters can be thought of as potentially relevant:

Biotic parameters (Flora):

- 'Marsh marigold (Caltha palustris) group': species of grasslands on wet sites
with moderate nutrient availability;
'Meadowsweet (Filipendula ulmaria) group': species of tall herbaceous
vegetation on wet sites with moderate nutrient availability;
'Burreed (Sparganium emersum) group': species in waters with moderate
nutrient availability;
'Fringed Water Lily (Nymphoides peltata) group', species in waters with
moderate nutrient availability.

Biotic parameters (Fauna):

- Birds: 'Black-tailed Godwit group';
'Reed Bunting group';
- Mammals: 'Ermine group';
'Hares';
- Butterflies: 'Small Pearl-bordered Fritillary group';
- Amphibians: 'Green Frog group';
- Macrofauna: 'Great Diving Beetle group';
- Fishes: 'Pike group'.

Biotic parameters (vegetation structure):

- Algae growth: percentage of ditch with algal growth, closed surface of
duckweed or Azolla;
- Intact belts and wood lots: length of intact hedgerows per km2•

Abiotic parameters (physical):

- Seepage zone: surface with lithotrophic seepage along rivers and adjourning
elevated ecodistricts;
- Groundwater level: surface with groundwater level fib (highest winter level:
25-40 cm, lowest summer level: 50-80 cm below the surface);
- Depth view: depth view in surface waters;
- Depth of ditches: total length of ditches containing water.

Abiotic parameters (chemical):

- Electric conductivity: electric conductivity of surface waters;
- Organic pollutants: concentration of biocides and other organic pollutants in
soils and water;
- Phosphate: concentration of total phosphate in surface water;
 103

- Heavy metals: concentration of heavy metals in soils.

A specification of the quality objective in tenns of quality parameters such as the

above makes quantification possible and hence a quality judgment of the state of the
environment. This quality judgment is expressed as a discrepancy between target
values and the current values for each parameter. All these discrepancies can be
graphically expressed in an 'AMOEBA'. In Figure 5 a tentative quality judgment is
given for the above example.

10.

Marsh marigold group

Black·tailed godwit
group
Reed bunting group

Ermine group

Small pearl· bordered

fritillary group
~-.L_- Green frog group
Gre.1 diving beetle group

5.

Figure 5. AMOEBA for the general environmental quality (GEQ) of the Lowland Peat
area (tentative judgment of the current state, no quantified target values).

10. Summary and conclusions

After having argued that sustainability is only a useful concept if related to quality
criteria, we have proposed a comprehensive method to assess environmental quality.
This method, which in essence is applicable independent of spatial scale, was
elaborated for ecodistricts. Ecodistricts are supposed to be an appropriate scale level
in defining environmental quality objectives at the national policy level. Recent
developments in quality assessment of aquatic ecosystems in the Netherlands,
especially the AMOEBA approach (ten Brink and Hosper, 1989), are considered very
inspiring; there is a close resemblance with the method we have proposed.
 104

Specific characteristics of the method presented in this contribution are:

1. The environment is considered an ecosystem, i.e., a network consisting of
interrelated abiotic and biotic components.
2. Environmental quality is assessed in behalf of both economic land utilization
types and nature.
3. It is proposed to assess the environmental quality at the spatial scale level of
ecodistricts; to this end quality objectives must be specified for each ecodistrict
type. The procedure as such is valid for whatever spatial scale.
4. In defining the quality objective a distinction is made between areas where
economic land utilization types are dominant, for which general environmental
quality requirements (GEQ) suffice, and nature conservation areas, for which
specific environmental quality standards (SEQ) must be set.
5. In selecting parameters to quantify the environmental quality, four criteria have
to be met: relevance to environmental policy, sensitivity, detectability and
appeal.
6. Numerical standards for each quality parameter (target values) are needed to
sustain environmental quality. In GEQ areas this implies standards which
ensure a sustained carrying capacity for land utilization, a basic nature value,
as well as a potential to achieve specific nature values within a well-defined
period. In SEQ areas requirements of nature conservation should determine the
standards to be set.
7. A graphic presentation of the results of environmental quality assessments
analogous to the one designed in the AMOEBA method (ten Brink and
Hosper, 1989), i.e., in the form of 'AMOEBAs', is highly esteemed by policy-
makers. In our opinion it is a good example of getting the message across
from science to policy-makers. A slightly adapted version, without the few
remaining inconveniences as to legibility, has been shown in figure 5.
8. As an example the procedure of assessing environmental quality has been
illustrated for ecodistrict type 'Lowland Peat area', with a tentative quality
judgment for the GEQ part of it.
9. Within the outlined framework a quantitative elaboration is now in progress
for two case studies in behalf of the State of the Environment Document
(collaboration of RIVM and CML).

References

Brink, BJ.E. ten and S.H. Hosper (1989). Naar toetsbare ecologische doelstellingen
voor het waterbeheer: de AMOEBE-benadering. H20 1989(22)/20: 612-617.
CUWVO (Commissie Uitvoering Wet Verontreiniging Oppervlaktewateren) (1988).
Ecolo&ische normdoelstellin&en voor Nederlandse oppervlaktewateren. Den Haag.
Klijn, F. (1988). Milieubeheer&ebieden. CML-rapport 37, Leidenj RIVM-rapport
758702001, Bilthoven.
 105

Klijn, F. and H.A. Udo de Haes (1990). Hitkarchische ecosysteemclassificatie: voor-

stel voor een eenduidig begrippenkader. Landschap 1990/4.
Klijn, F. et al. (1990a). De milieukwaliteit van ecodistricten. Deel 1. Ecologische
normstelling en milieukwaliteitsbepaling. CML-rapport 62, Leidenl RIVM-rapport
751901002, Bilthoven (Eng. summary).
Klijn, F. et al. (1990b). De milieukwaliteit van ecodistricten. Deel 2. Methode en
aanzet tot uitwerking. CML-rapport 63, Leidenl RIVM-rapport 751901003,
Bilthoven (Eng. summary).
Klijn, F. and A. Laansma (1990). Gebiedsgericht milieubeleid: theorie en praktijk en
aanzet tot onderzoeksprogrammering. CML rapport 61, Leiden (Eng. summary).
Latour, J.B., J.J. Hofstra and M.I. Nip (1990). De toepasbaarheid van de AMOEBE-
benadering op terrestrische ecosystemen. RIVM-rapport 711901001, Bilthoven.
Lexmond, Th. M. and Th. Edelman (1987). Huidige achtergrondwaarden van het
gehalte aan een aantal zware metalen en arseen in grond. 3 HBM bodembescher-
mingo The Hague.
Ministry of Agriculture and Fisheries (1989). Nature Policy Plan. The Hague.
Ministry of Transport and Public Works, (1989). National Policy Document on Water
Mana~ement. The Hague.
Ministry of Public Housing, Physical Planning, and Environment (1989). National
Environmental Policy Plan. The Hague.
Ministry of Public Housing, Physical Planning, and Environment (1989). Vierde nota
over de ruimtelijke ordening EXTRA. Deel 1: ontwerp-PKB. The Hague.
Moller Pillot H.K.M. (1971). Faunistische beoordeling van de verontreiniging in
laaglandbeken. Proefschrift Nijmegen. Pillot-Standaardboekhandel, Tilburg.
RIVM (ed. F. Langeweg) (1989). Concern for Tomorrow. A national environmental
survey 1985-2010. National Institute of Public Health and Environmental
Protection, Bilthoven.
Stevers, R.A.M. et al. (1987). Het CML-ecotopensysteem, een landelijke ecosysteem-
typologie toegespitst op de vegetatie. Landschap 1987(4)f2: 135-150.
 9

Contours of an integrated environmental index for

application in land use zoning

Joop de Boer, Harry Aildng, Ella Lammers, Vera Sol and Jan Feenstra

This paper focuses on the development of an integrated environmental index for use
in land use zoning. Such a composite indicator, which integrates various aspects of
environmental conditions into one index, may be a useful tool for measuring
sustainability (Liverman, Hanson, Brown and Merideth, 1988). It is clear, however,
that great care needs to be taken in the valuation and scaling of the components in
such indicators in order to avoid the familiar problem of combining apples and
oranges. The use of composites may also be difficult to communicate and explain to
the public and to policy-makers, especially at the local level.

The main features of our approach to these problems will be presented in this paper.
It is based on a feasibility study on the development of an integrated environmental
index for use in land use zoning (e.g., Aiking et al., 1990). In view of the many
problems which have to be dealt with in order to measure sustainability, the scope
of our contribution is rather limited. Apart from this limited scope, however, at least
two points should be emphasized. Firstly, the strategy that we propose, viz., to
combine empirical data with value judgments may find more general application.
Secondly, our work shows how the development of an index is shaped by the
particular (policy) context for which it is intended.

In the following sections we shall start with an introduction on the purpose of the
index and the context of its intended use. Against this background we shall next
discuss both the scientific and the policy oriented criteria that the index should meet.
Then we shall take a closer look at the protection of human health by means of
environmental quality standards. As this protection is far from complete, two different
approaches will be considered that may lead to an environmental index by combining
information on the quality of living conditions. The first method is keyed to the
specific target organ of each individual pollutant. The second method is based on
environmental standards. Both methods will be discussed in detail.
 108

Purpose and context

In developing an environmental index one must always choose the proper balance
between simplicity and completeness. As Ott (1978) has emphasized, an index is
primarily designed to simplify. In the process of simplification the loss of information
should be kept to a minimum, however. Moreover, the simplification can only derive
its legitimacy from the context for which the index is intended. It follows that index
development must start with a carefully defined concept of the purpose of the index.
It also implies that the original purpose must be respected when the index is being
used.

In our study the index is meant to assist local and regional administrators in
comparing environmental conditions at different locations in the course of land-use
zoning. Land use zoning is an instrument aimed at striking a balance between the
development of activities with adverse effects on the local environment and the
protection of land uses which are sensitive to these effects. With the help of
environmental legislation and town and country planning, attempts are being made
to keep sensitive land uses, such as housing, out of the range of influence of polluting
activities, such as road traffic and industry.

To improve the environmental quality around large industrial complexes, an integral

approach is being developed for situations in which noise, odour, and the risk of
calamities may occur together. Inclusion of exposure to local air pollution is also
being considered. Most of these pollutants are regulated by environmental standards
or will be regulated in the near future. With a few exceptions, however, these
standards are not tuned to situations with combinations of sources and/or types of
pollution. For purposes of planning, therefore, the question has arisen whether it is
feasible to combine and to integrate all of the available information about these local
environmental influences on human health and well-being. This was the question
central to our study.

Criteria

In studying the feasibility of an environmental index one has to consider both

scientific and policy oriented criteria. From a scientific point of view the index should
give a valid representation of the combined influence of local environmental stimuli
on the quality of living conditions. In addition, the numerical value of the index
should be reliable and the procedure of index development should be as transparent
as possible. Validity is a major problem as the scientific knowledge of environmental
effects, and combined effects in particular, is very limited. This is the reason for
stressing the importance of transparancy of the procedure.
 109
From a policy oriented point of view, the main criteria all refer to the compatibility
with existing purposes and common practice. First of all, the index should be
compatible with the policy of zoning. It should add to an integral approach by
pointing to comprehensive information on environmental quality. Furthermore, the
index has to be sensitive to the distribution of conditions over a certain geographic
region. It is also critical that the index can be easily analyzed to reveal details such
as the sources of pollution contributing the most.

A major issue is the consistency of the index with the existing standards of
environmental regulation. It would not be acceptable that the index might compensate
for any violation of the limit values that are set for the individual pollutants. In
addition, the index should be balanced in a way that reflects the different values
which society puts on certain environmental effects, such as annoyance and cancer
risk. These values are incorporated in the environmental standards, but not without
regard to the consequences of legislation, as different standards may apply to existing
and to new situations.

To improve implementation a suitable index should have an operational character.

This criterion puts restrictions on the units of measurement that can be used to
characterize exposure levels. For example, it is common practice in the Netherlands
to express noise levels in a single figure like the equivalent A-weighted sound level
(dB (A» corrected for the time of day. It is not recommendable to reopen the
discussion on this measure.

The above criteria have several implications for the way in which an index may be
developed. The main restraint results from the demanded consistency with existing
standards. As no exceeding of the standards will be allowed, there is a restriction
on the range of environmental effects which have to be considered in the development
of a suitable index. However, this restriction of range cannot be exactly estimated, as
the individual pollutants may have combined effects. So the range of possible values
is not known in advance. Moreover, expected progress in environmental policy makes
it desirable to develop an index that can be extended by adding newly regulated
substances.

Environmental health perspective

In view of the purposes of policy-makers and the context for which the index is
intended, it may seem recommendable to develop an index based on environmental
standards. Before it may be concluded that this is the obvious way of index
development, however, one should consider this matter from an environmental health
perspective. This is especially important, as the index has to integrate information on
 110

various kinds of effects on health and well-being. Therefore, we shall take a closer
look at the protection of human health by means of environmental quality standards.
Health may be conceptualized as a dynamic system of interacting subsystems which
are more or less sensitive to specific environmental conditions. Without going into all
the complex interactions and ,feedback loops, one may say that these subsystems can
adapt to an adverse environmental condition or stressor as long as their time-
dependent threshold levels have not been exceeded. When different polluting agents
are interacting with the same subsystem or target organ, their contributions may be
additive, synergistic or antagonistic. Moreover, as adaptation and resilience of the
subsystem depend on the functioning of the other subsystems, it will make a
difference whether or not these are being stressed as well. Consequently, one has to
be very careful when assessing any combined effects.

In order to protect residents against polluting agents that will be considered in land
use zoning, it is first of all necessary to identify the subsystems or target organs
which may be sensitive to these agents. After analyzing the relationship between dose
and response for each subsystem, one should take a closer look at the functioning of
the whole system. This approach may seem rather futile, as it is clear from the
beginning that there will be very little information on dose-response relationships.
Moreover, there is no analytical model to integrate this information on a level that
covers the whole system of human health and well-being. In spite of these problems,
however, it should be noted that the feasibility of this approach is partially dependent
on the range of health effects which may be expected in view of the existing
standards.

The range of possible health effects may be as diverse as the set of polluting agents
that will be considered in land use zoning. These agents include noise and odour from
different sources, local air pollution by fifteen so-called high-priority substances, and
a variety of hazardous substances that can cause major accidents. Some rearrangement
may simplify these matters, however. Noise and odour are primarily important because
of their psychological effects in terms of annoyance. The local air pollutants can be
distinguished into eight toxic and seven carcinogenic substances. Finally, the effects
of hazardous substances can be categorized as the risk of dying by accidental
exposure to explosions, heavy fires, or toxic fumes.

The existing standards protect the general population against very high levels of
annoyance by noise from road traffic, rail traffic, aircraft and industry. However, there
is a considerable range between the target values and the limit values of these
standards. As a representative of the Ministry admitted, little is known about the
arguments that led to the actual limit values (Eggink, 1987). With the present
knowledge some estimates can be given of the percentage of the population that may
be highly annoyed even if the limit values have not been exceeded. The Noise
Nuisance Act appears to have a target level of between 0 and 12% of the population
 111
being (highly) annoyed. It tolerates maximum levels of 15% to 25% for new situa-
tions and an absolute maximum, for all noise sources together, of about 40% (Eggink,
1987). Other kinds of health effects, for example on blood pressure, have not been
very well documented, but should not be dismissed completely.

Annoyance by odour will be regulated in the near future by air quality standards.
The main policy objective is to reduce odour annoyance to a level which is
considered to be negligible. For residential areas in the vicinity of large industrial
complexes this level will be extremely difficult to achieve, even in the long term.
According to Eggink (1987) a limit value will be set that probably tolerates
annoyance among 5% of the population. The problems in standard setting include
the assessment of odour concentration in ambient air. It is common practice to
calculate odour concentration on the basis of odour emission. This method is relatively
simple if the number of sources is limited. In areas with industrial complexes,
however, the assessment of odour concentration still has to be worked out.

With regard to the other air pollutants a distinction should be made between classes
of toxicants which are known or assumed to have a threshold of response, and classes
of toxicants which are known or assumed to have no threshold of response. In the
context of land use zoning eight substances are considered which belong to the first
class. The existing standards protect the general population against adverse effects
from each individual substance. This means that the limit values,of these standards
are less than or equal to the adjusted no-observed-adverse-effect levels (NOAEL.) of
these substances. Some target values are a factor 100 lower than the limit values, but
this is not uniformly so. As residents are not exposed just to the local air pollutants
that are considered in land use zoning, there is no guarantee that they are protected
against any combined effects of substances interacting with the same target organ.

The nonthreshold toxicants are known or assumed to incur some risk of adverse
response at all doses above zero. In the Netherlands environmental quality standards
will limit the excess risk of cancer at a level of one in a million per year for each
individual substance. Some of these standards have target values which are a factor
10 lower than the limit values, but this is not a uniform pattern. The total risk due
to all carcinogenic substances has been set to a limit value of 10-5 per person per
year.

Risk standards will also be used to protect the population against large industrial
hazards. Managers of installations that can cause major accidents, such as those
covered by the Seveso Directive, are required by law to produce a quantitative risk
analysis. The results of this analysis should include risk contours, mapping areas with
stated chances of death to an individual permanently located in a specific place. The
results should also include a F-N (frequency of N or more fatalities) curve indicating
the chances of killing more than a stated number of people in one accident. On the
 112

basis of these risk calculations standards will be developed which will be applied to
each individual installation. It is proposed to set the limit value of individual risk to
residential populations at 10-6 and the target value at 10-8 per year. The proposed
standards will also refer to the societal risk.

Although we have to restrict ourselves to this brief overview of environmental

standards, it may be concluded that there is reason to be concerned about possible
health effects. Moreover, it is clear that the target values of the quality standards
will usually not be achieved in the vicinity of large industrial complexes. In view
of these conclusions, at least two ways can be distinguished for the development of
an index for use in zoning.

A basically scientific approach focuses on the target organs of the environmental

pollutants. By combining information on possible adverse effects, ranging from
annoyance to death, an estimation can be made of the total environmental influence
on human health.

Alternatively, the second approach focuses on the target values of environmental

standards. In this case, the index is composed by combining information on relative
distances between the conditions in the designated area and the target values of the
quality standards. In the next sections both approaches will be considered in more
detail.

First approach: estimating possible health effects

The first approach to index development focuses on possible health effects. As the
scientific knowledge on this subject is very limited, it will be necessary to combine
empirical data with value judgments. As this may lead to an obscure combination,
great care needs to be taken in carrying out a step-by-step method. In general, the
following steps can be distinguished.

Identification: The polluting agents have to be arranged according to their presumably

most sensitive target organs, for example, by using a matrix such as Figure 1.

Assessment: The interaction of each agent with the target organ has to be described
and estimated on the basis of dose-response relationships.

Summation: The comparable effects on health have to be combined for agents acting
in an identical way on the same target organ, such as two air pollutants that damage
the lung by affecting the lung membranes.
 113

Indexation: The combined effects have to be transfonned into dimensionless units

indicating the magnitude or the chance of a certain kind of effect.

Valuation: The dimensionless units have to be valuated on a numerical scale ranging

from acceptable to not acceptable for human health and well-being.

Aggregation: The resulting scales or subindices have to be merged into the integrated
environmental index using a combination rule which preserves their meaning in tenns
of impacts on human health.

The core of this method is the designation of comparable influences on health within
the range of effects which may be expected in view of the existing standards. As
mentioned before, at least three categories of effects may be distinguished, namely
annoyance, toxic effects and mortality, including cancer. Comparable health effects
have to be combined by summation and indexation, leading to a subindex for each
category of effect (Figure 1). First of all, however, one has to identify the effects
which may be seen as comparable.

The category of annoyance may seem relatively simple. Noise from road traffic,
railroads, airplanes and industrial sources, as well as odour from industrial sources,
may be conceived as unwanted stimuli, causing comparable psychological reactions
in tenns of aversion and annoyance. Some studies have shown, however, that people
have difficulties in answering questions about the total annoyance by environmental
stimuli. They find it much easier to express their annoyance by specific sources. Their
"total" annoyance, as assessed by direct questions, appears to be less than or equal
to the annoyance caused by the dominating source. The difference depends on the
annoyance by the source that is not dominating. On the basis of empirical data and
some reasonable assumptions, Miedema (1990) has proposed a fonnula for combining
dose-response relationships for individual sources. Taking into account that these
relationships may show different slopes and intercepts, his combination rule results
in a subindex which is essentially a dose equivalent measure of annoyance. For this
particular type of dose-response relationships Miedema (1990) has proposed the
following equation:

p a
L = 10 log [}: exp [(a;l, - d)/a]]
i=l

where exp X = 10x/1O

L = dose-equivalent measure of annoyance
L; = dose of source i (in dB(A))
a; = slope parameter of source i
d; = intercept parameter of source i
 114

a = weighted sum of slope parameters

Without investigating this matter in depth, it may be concluded that annoyance from
different sources can be combined. It should be noted, however, that not all the
parameters are known yet. More research is necessary, especially for odour.

a. b. c.
EFFECT annoy- toxic morta-
ance effects lity

AGENT

I. NOISE
road x
rail x
air x
industrial x

II. ODOUR x

III. AIR POLLUTION

trichloroethene x
toluene x
T tetrachloroethene x
0 dichloromethane x
X styrene x
tetrachloromethane x
chloroform x
phenol x

epichlorohydrine x
benzene x
C vinylchloride x
A 1,2-dichloroethane x
R propylene oxide x
C acrylonitril x
ethylene oxide x

IV. RISK x

SUBINDICES ANNOY- TOXIC MORTA- INTEGRATED

ANCE EFFECTS LITY ENVIRONMENTAL
INDEX

Figure 1. Matrix representation of polluting agents and possible effects (x). The
bottom row shows the subindices for each individual category of effects.
 115

The category of toxic effects is rather diverse in theory, but a closer look at the eight
substances in Figure 1 shows that the situation is less complicated than it might have
been. All 15 substances that will be considered in zoning happen to be organic
compounds. The eight toxicants display a non-zero threshold dose-response
relationship. In all eight cases exposure levels above the NOAEL will first of all
cause irritation of the mucous membranes and pre-narcotic effects on the central
nervous system. These organs may be conceived as the toxicants' most sensitive
targets. Therefore, it may be concluded that the eight toxicants act in a more or less
identical way on the same target organs, as long as exposure levels will not increase
beyond those at which mucous membrane irritiation and pre-narcotic effects occur.

In other words the eight toxicants will have a combined effect that is greater than
the effect of each individual substance. It should be noted, however, that our
knowledge of these NOAELs is not complete. In practice it is often necessary to use
values derived by extrapolation, using more or less arbitrary uncertainty factors, often
misleadingly called safety factors. NOAELs may be based on different types of data,
such as epidemiological studies or animal experiments, and the safety factors may be
incomparable. The fact that different agencies do not always propose the same value
for such an adjusted NOAEL is an example illustrative of this problem. Nevertheless,
as this is the best available knowledge, it is justifiable to use the adjusted NOAEL
(denoted by NOAEL.) in combining the possible effects of the toxicants. This can 1:>e
done by summing the ratios of the air concentrations of the individual substances and
their corresponding NOAEL•.

From the list of substances in Figure I, it can be seen that seven of the air pollutants
are carcinogens. These nonthreshold toxicants are known to incur some risk of adverse
response at all doses above zero. This risk, however, is not easy to assess because the
exposure levels in ambient air will fall well below the lowest experimental levels at
which a response has been observed. As the behaviour of the dose-response curve
cannot be predicted in the low-dose range, a model must be assumed for the
calculations in this range. In general it is recommended to use a linear interpolation
between the response observed at the lowest experimental dose and the origin at zero
dose and zero response (e.g., Hallenbeck and Cunningham, 1986).

Although the seven carcinogens will cause different kinds of cancer, they are assumed
to have the same consequence, namely death. Therefore, it is justifiable to combine
their levels of excess risk by adding the chances. In the context of zoning, these risks
may also be combined with the individual risk due to major accidents in industrial
sites. Because the assessment of fatality risks is a very technical matter, it will not
be discussed here. The main point is that all these small risks can be added. TIle
result gives a total excess chance of mortality per person per year.
 116

Valuation and aggregation

The first steps of index development may result in three subindices, namely the dose-
equivalent measure of annoyance, the summed ratios of concentration and NOAEL..
and the chance of dying per person per year. These subindices are incommensurable
as long as they are expressed in their own units. Moreover, there is no analytical
model to assess the meaning of these subindices for the lifetime health of the
population. Consequently one needs some kind of expert judgment to transform the
information contained in the subindices into a meaningful integrated index. Several
methods are described in the literature on decision analysis and value measurement
(e.g., Von Winterfeldt and Edwards. 1986). We shall focus here on value judgment,
without entering into the discussion whether it makes sense to distinguish value from
utility.

To find a common measure for the subindices, one may use a value function that
transforms the so-called natural scale of each subindex into a numerical value scale.
This scale has two endpoints indicating the acceptable and the wlacceptable level, for
example, with a range from 0 to 100. The two corresponding anchors on the natural
scale have to be chosen by the experts. In doing this they might consider the limit
values of the environmental standards, which must never be exceeded.

When working with sophisticated experts, one may ask them to draw a value curve
between the two anchors on the natural scale. Alternatively one may ask them to
select functional forms of value curves from a given set. In creating a value function
it is important to consider the arguments that may restrict the shape of the curve. For
example. there may be a priori reasons to assume that the value function is concave,
convex, or linear. Von Winterfeldt and Edwards (1986. p. 237) have taken the
position that a value function will be linear in that natural scale that most closely
reflects the value concerns to which it is related. Additional concerns may produce
non-linear value functions, for example, when health and policy concerns are mixed
together. In our case. the subindices seem close to most value-relevant considerations
about health impacts. Therefore, the two examples presented in Figure 2 might be
realistic from an environmental health perpective. In Figure 2A the dose-equivalent
measure of annoyance on a logarithmic loudness scale is linearly related to a value
scale, which is also logarithmic. A similar function is presented in Figure 2B with
respect to the excess chance of dying per person per year. It should be emphasized
that these examples are not based on a definitive judgment of experts.
 117

ANNOYANCE RISK
100 100
V V
A A
L L
U U
E E
10 10
5 5
C C
A A
L L
E E
1
50 55 60 65 0001 .001 01
-S
dBIA) CHANCE 10

Figure 2A and 2B. Examples of value functions for dose-equivalent annoyance (left)
and excess chance of dying per person per year (right). All scales
are logarithmic.

After the natural scales have been transformed into value scales, the question arises
how these valued subindices can be combined into an integrated index. In fact two
questions have to be considered. Firstly, one may ask whether it is desirable to weight
the subindices before they are combined. The experts may feel that the range from
o to 100 means something different in the case of annoyance than in the case of
lethal risk. It should be emphasized that this difference is not just a matter of death
being more important than annoyance. Von Winterfeldt and Edwards (1986, p. 285)
have argued that weights should be controlled by the range of the scale over which
the value function is defined. In particular, weights should increase as the range of
the natural scale increases (e.g., Keeney and Raiffa, 1976).

The second question refers to the properties desirable for combination rules. The
subindices may be thought of as substitutes for each other, provided that the legal
standards have not been exceeded. The experts may feel that a high (i.e.,
unacceptable) score on one of the subindices might be counterbalanced by low scores
on the others. Alternatively they might want to emphasize the highest score, arguing
that a chain is only as strong as its weakest link. A comprehensive discussion of
combination rules and aggregation functions was given by Ott (1978). His analysis
shows that the root-sum-power has some desirable properties. The root-sum-power is
a nonlinear aggregation function of the following form:
 118

I = (1/ + II + 1l)1/p
where I = integrated index
I» 12 , 13 = subindices
p = exponent

Several aggregation functions are special cases of the root-sum-power, depending on

the value of p. If p=l. a straight line results. which is the linear sum aggregation
fWlction. If p=2, the root-sum-square results. In our case, this function plots in a
three-dimensional value space as a series of concentric spheres of radius I from the
origin. For larger values of p the index approaches the value of the largest of the
three subindices, as can be seen in Table 1. This limiting case is less suitable,
however, when fine shades of environmental quality are to be distinguished. Another
possibility would be to use the aggregation rules that are proposed in the literature
on decision analysis and multiattribute utility theory (e.g., Keeny and Raiffa. 1976;
Von Winterfeldt and Edwards, 1986). In concluding this section we may say that it
is important to be very explicit about the range of the subindices and the properties
of the aggregation rule. In order to get a grasp of the possible outcomes it will be
prerequisite to do a sensitivity analysis.

Table 1. Some examples of index I by aggregating the subindices I» 12 and 13, for
selected values of p in a root-sum-power function.

I) 12 13 P I II 12 13 P I

80 30 10 1 120 40 40 40 1 120
80 30 10 1.5 94 40 40 40 1.5 83
80 30 10 2 86 40 40 40 2 69
80 30 10 3 81 40 40 40 3 58

Second approach: combining distances from target values

The preceding discussion applies to index development from an environmental health

perspective. An obvious alternative is to develop an index on the basis of target and
limit values of legal standards. Such an index would reflect the aggregated distance
between the actual exposure and the target level for all pollutants considered in
zoning that are regulated by legal standards. For each individual pollutant this distance
can be expressed in units that indicate the relative position of the actual exposure on
a scale bounded by the target and limit values of its standard. A simple formula is:
 119

where D; = distance for pollutant i

100 = arbitrarily assigned number
X; = actual exposure for pollutant i
T; = target value for pollutant i
L; = limit value for pollutant i

The individual distances may be aggregated into a subindex for each type of polluting
agent, such as noise, odour, air pollution, and major accident hazards, and
subsequently aggregated into the integrated environmental index. The root-sum-power
mentioned above may also be useful in this case. By using this method it will be
easy to analyze the index in order to reveal the sources of pollution contributing the
most. However, it will not be easy to get a grasp of the range of the subindices,
because quite a few standards have to be considered, expecially for air pollution. It
should be noted that aggregation and valuation cannot be distinguished in this case,
because valuation has already been done implicitly in setting the standards. Therefore,
it may be concluded that this method is less transparent than the method that focuses
directly on health.

In addition, it should be emphasized that this method will lead to meaningful results
only if the standards are based on homogeneous considerations. This appears to be
the aim of environmental policy-makers. However, as was mentioned in the section
on health protection, this aim has not yet been achieved, especially not for annoyance
and toxic effects. Moreover, the existing standards are not tuned to situations with
combinations of sources and/or types of pollution. On the basis of this it can be
concluded that integration will be a weak point of this method as long as integrated
standards are lacking.

Concluding remarks

In our study we concluded that it is feasible to create an integrated environmental

index, but that it would be premature to choose a particular method without additional
research (e.g., Aiking et ai., 1990). From a scientific point of view it is preferable
to focus directly on possible health effects. It will be clear, however, that we should
not be deceived by all kinds of simplifications. For example, it was noted that the
eight toxicants considered in zoning are all found to share the same most sensitive
target organ (mucous membrane). In view of this one may wonder whether it is
justifiable to omit any exposure to substances that will influence the same target organ
(e.g., 802 , NO.), but are not considered in zoning. This will be an important point
when valuation and aggregation are discussed with experts.
 120

From a policy oriented point of view it is preferable to focus on the target levels of
legal standards. This method will be the most compatible with the purpose and the
context of zoning. For example, such an index can be easily analyzed to reveal the
sources of pollution contributing the most. However, one may doubt whether such an
index would give a meaningful representation of the combined influences on the
quality of living conditions. Therefore, it will be necessary to evaluate these standards
in view of situations with combinations of sources and/or types of pollution.

Literature

Ailing, H., J. de Boer, V.M. Sol, P.E.M. Lammers and J.F. Feenstra (1990).
Haalbaarheidsstudie Milieubelastingsindex. Leidschendam: Ministerie van VROM,
Reeks Integrale Milieuzonering 8.
Eggink, E. (1987). Limits for environmental annoyance. In: H.S. Koelega (ed.),
Environmental annoyance and the setting of standards: Proceedings of a discussion.
Utrecht: University of Utrecht, Psychological Laboratory.
Hallenbeck, W.H. and K.M. Cunningham (1986). Ouantitative risk assessment for
environmental and occupational health. Chelsea, MI: Lewis Publishers, Inc.
Keeney, RL. and H. Raiffa (1976). Decisions with multiple objectives: Preferences
and value trade-offs. New York: Wiley.
Liverman, D.M., M.E. Hansen, BJ. Brown and RW. Merideth, Jr. (1988). Global
sustainability: Toward measurement. Environmental Management, 12, 133-143.
Miedema, H.M.E. (1990). Een index voor milieukwaliteit. Deel 1: Afleiding en
toepassing voor geluid en geur. Leiden: Nederlands Instituut voor Praeventieve
GezondheidszorgffNO.
Ott, W.R (1978). Environmental indices: theory and practice. Ann Arbor, MI: Ann
Arbor Science Publishers.
von Winterfeldt, D. and W. Edwards (1986). Decision analysis and behavioral
research. Cambridge: Cambridge University Press.
 121

Notes on the contributors

Dr. Harry Aiking studied biochemistry at the University of Amsterdam (1967-1973),

where he completed a Ph.D. in microbiology in 1977. He worked as a research
associate at Indiana University in Bloomington, USA (1978). the Central Bloodbank
Laboratory in Amsterdam (1979) and joined the Institute for Environmental Studies
of the Free University in Amsterdam in 1980. Since 1987 he has been responsible for
IES research on Environmental Toxicology. He has been Interim Head of the
Chemistry. Toxicology and Ecology Division since 1989.

Dr. Jaap Arntzen is an environmental economist. presently employed as coordinator

of the research programme Environment and Third World at the Institute for
Environmental Studies, Free University Amsterdam. His research experience in the
Netherlands and developing countries includes rural natural resource management,
debt-for-nature swaps, environmental profiles. conservation strategies and
environmental aspects of development assistance.

Joop de Boer is research coordinator at the Institute for Environmental Studies in

Amsterdam. He has a M.A. in psychology from the University of Amsterdam. His
current research includes community response to environmental stressors and threats,
especially in the field of risk perception and risk communication.

Peter Bosch, M.Sc., (1957) received his professional training as a physical

geographer. Since 1988 he has been head of the Study Department for Environmental
Statistics at the Netherlands Central Bureau of Statistics. He is currently working on
the development of a satellite account for the environment to the Netherlands System
of National Accounts.

Leon Braat, M.Sc., studied systems ecology at the University of Florida and
environmental science and environmental economics at the Free University of
Amsterdam. He worked at the Institute for Environmental Studies of the Free
University Amsterdam from 1979 through 1990, concentrating on economic-ecological
simulation models. In December 1990 he joined the National Institute of Public Health
and Environmental Protection to head the department of Integrated Environmental
Modelling.

Ben ten Brink, M.Sc., was born on 30 April 1954 in The Hague, Holland, and has
a Master's degree in marine biology and environmental ecology with physical
 122

planning, environmental law and scientific journalism, University of Leyden. Author

of the "Policy and Management Plan for the Biesbosch National Park", environmental
policy in policy development for road infrastructure, and project leader of "nature
conservation in water management", of the National Policy Document on Water
Management", and of "Macro Aquatic Assessment" for the national government of
the Netherlands.

Jan Feenstra, M.Sc., studied organic and clinical chemistry at the University of
Leyden (1962-1970). In 1978 he joined the Institute for Environmental Studies. In
the Chemistry, Toxicology and Ecology Division, he coordinates desk studies that
support international, national and regional environmental policy. Since 1985 he
has been responsible for this line of research at the Institute.

Alison Gilbert, M.Env.Stud., is a graduate of Macquarie University in Australia.

Although originally trained in biology, she has worked in close cooperation with
economists since 1982. The focus of her research has been the analysis of economic-
environmental interactions using a variety of techniques including simulation
modelling. Her work on natural resource accounting began in 1985; this initial work
was supported by both the Australian and the Dutch governments. Since 1987 she has
been a Research Associate at the Institute for Environmental Studies, Free University
of Amsterdam. Her main activities have been part of a universitysponsored project
entitled "Ecologically Sustainable Economic Development"; she is currently completing
a Ph.D. dissertation entitled "NRA and Sustainable Development".

Dr. Leen Hordijk studied econometrics at the Erasmus University in Rotterdam where
he finished his studies in 1973. Since 1987 he has worked at the National Institute
of Public Health and Environmental Protection (RIVM, Bilthoven), where he has held
several positions; he currently heads the Environmental Projections Office, which is
in charge of inter alia the next issue of the book Concern for Tomorrow. In the
spring of 1991 he will join the Agricultural University Wageningen as a professor in
environmental sciences and director of the Centre for Environmental Studies. Leen
Hordijk has published many papers in the fields of econometrics, regional economics,
environmental economics and acid rain. He recently he co-edited the book The RAINS
Model ofAcidification. Science and Strategies in Europe (Kluwer Academic Publishers,
1990).

Dr. Roefie Hueting (1929) is head of the Department of Environment Statistics at the
Netherlands Central Bureau of Statistics. He founded this department in 1969. In
1974 Mr. Hueting obtained his Ph.D. in economics on the thesis "New Scarcity and
Economic Growth". He has published numerous papers on the relation between
economy and the environment.
 123

Frans Klijn, M.Sc., (1958) acquired a degree in physical geography in Amsterdam

(1984). He specialized in landscape ecology. After short-term research and teaching
jobs in both environmental science and landscape ecology at the University of
Amsterdam. he has been working at the Centre of Environmental Science of
the University of Leyden since 1987.

Onno Kuik, M.Sc., studied agricultural economics at the University of Wageningen

(1974-1982). He joined the Institute for Environmental Studies of the Free University
in Amsterdam in 1985. His main research interests are the valuation of environmental
change. economic instruments of environmental policy and the interactions between
agriculture and the environment.

Ella Lammers. M.Sc., studied chemistry at the University of Utrecht (1981-1988). For
a short time she worked for the local government as an official for environmental
affairs. She has worked at the Institute for Environmental Studies of the Free
University of Amsterdam since 1989. She is mainly working on several studies to
support the national and provincial environmental policy.

Maarten Ivo Nip. M.Sc., (1963) acquired a degree in biology in Amsterdam (1989).
He specialized in landscape ecology and environmental biology. Since 1989 he has
been a Ph.D. student at the Centre of Environmental Science of the University of
Leyden. doing research in the field of environmental quality assessment.

Professor dr. Hans (Johannes Baptist) Opschoor was born on 22 May 1944 in
Zwijndrecht. He graduated from the Erasmus University. Rotterdam (cum laude) and
obtained a doctorate degree (1974) from the Free University in Amsterdam. He was
a research fellow at the Rotterdam School of Economics. the Free University in
Amsterdam. the Netherlands Institute for Advanced Studies in Wassenaar and the
University of Botswana. From 1982 to 1990 he was director of the Institute for
Environmental Studies at the Free University. In 1987 he was appointed professor
at the Free University and in 1990 he was elected Chairman of the Advisory Cowlcil
for Research on Nature and Environment. His research interests include: fundamental
aspects of environmental economics. institutional and instrumental change and
environmental policies. environment and (sustainable) development. environment and
economic systems.

Professor dr. Lucas Reijnders (1946) is part-time professor of environmental science

at the University of Amsterdam. His main research interests are in the fields of waste
minimization and energy efficiency. He is also working for a private environmental
organization Stichting Natuur en Milieu and is a member of a number of advisory
councils.
 124
From 1971 till 1980, dr. Vera M. Sol studied chemistry at the University of
Amsterdam. In 1989 she completed her Ph.D. in chemistry at the University of
Leyden. In 1989 she joined the Institute for Environmental Studies of the Free
University in Amsterdam. She is working on studies supporting environmental policy,
with special attention to integration of environmental impacts.

Professor dr. Helias A. Udo de Haes studied biology at the University of Leyden.
He did his Ph.D. research on an ethological subject at the "Max Planck Institut fiir
Verhaltensphysiologie" in Seewiesen (near Munich). He has been director of the
Centre of Environmental Science of the University of Leyden since 1977. Since 1987
he has had a chair in environmental science.

Dr. Harmen Verbruggen is head of the department of Economics, Technology and

Social Sciences of the Institute for Environmental Studies, Free University Amsterdam.
Since 1990 he has also been dept. director of this Institute. His major field of
specialization is the interplay between development economics, international economic
relations and environmental studies.
 125

List of workshop participants

The workshops on "Development of Indicators of Sustainable Development" were both

held in Utrecht, the Netherlands; the first one on October 30, 1989, the second one
on February 22, 1990. The following persons attended one or both workshops.

J.W. Arntzen, Institute for Environmental Studies, Free University, Amsterdam

J. Bakkes, National Institute of Public Health and Environmental Protection, Bilthoven
H. de Bliek, Ministry of Economic Affairs, The Hague
J. de Boer, Institute for Environmental Studies, Free University, Amsterdam
P.R. Bosch, Central Bureau of Statistics, Voorburg
L.C. Braat, Institute for Environmental Studies, Free University, Amsterdam
BJ.E. ten Brink, Ministry of Transport and Public Works, Tidal Waters Division,
The Hague
R. During, Netherlands' Organization for Applied Scientific Research, Delft
J.F. Feenstra, Institute for Environmental Studies, Free University, Amsterdam
R. Fredriksz, Ministry of Housing, Physical Planning and Environment, Leidschendam
OJ. van Gerwen, Advisory Council for research on Nature and Environment, Rijswijk
AJ. Gilbert, Institute for Environmental Studies, Free University, Amsterdam
H. de Graaf, University of Leyden
L. Hamers, Federation of Netherlands' Industry, The Hague
L. Hordijk, National Institute of Public Health and Environmental Protection,
Bilthoven
H.M.A. Jansen, Institute for Environmental Studies, Free University, Amsterdam
F. Klijn, Centre of Environmental Science, University of Leyden, Leyden
P. Koster, National Institute of Public Health and Environmental Protection, Bilthoven
H. de Kruijf, National Institute of Public Health and Environmental Protection,
Bilthoven
OJ. Kuik, Institute for Environmental Studies, Free University, Amsterdam
P.J.A. van de Laak, Netherlands' Organization for Applied Scientific Research, Delft
F. Langeweg, National Institute of Public Health and Environmental Protection,
Bilthoven
R. van de Lee, Ministry of Housing, Physical Planning and Environment,
Leidschendam
RJ.M. Maas, National Institute of Public Health and Environmental Protection,
Bilthoven
GJ. van der Meer, University of Amsterdam, Amsterdam
E. Meijers, Ministry of Housing, Physical Planning and Environment, Leidschendam
 126

M. Nip, Centre of Environmental Science, University of Leyden, Leyden

J.B. Opschoor, Institute for Environmental Studies, Free University, Amsterdam
R. Reiling, National Institute of Public Health and Environmental Protection, Bilthoven
C.K. Spiegel, Institute for Environmental Studies, Free University, Amsterdam
H. Udo de Haes, Centre of Environmental Science, University of Leyden, Leyden
H. Verbruggen, Institute for Environmental Studies, Free University, Amsterdam