Академический Документы
Профессиональный Документы
Культура Документы
Valuation
Professor Clive L. Spash holds the Research Chair in Environmental and Rural
Economics at the University of Aberdeen, UK. He is head of the Socio-Economic
Research Programme at the Macaulay Institute and President of the European
Society for Ecological Economics.
1 Greenhouse Economics
Values and ethics
Clive L. Spash
Edited by
Michael Getzner, Clive L. Spash and
Sigrid Stagl
First published 2005
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Simultaneously published in the USA and Canada
by Routledge
270 Madison Ave, New York, NY 10016
Routledge is an imprint of the Taylor & Francis Group
This edition published in the Taylor & Francis e-Library, 2005.
“To purchase your own copy of this or any of Taylor & Francis or Routledge’s
collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”
© 2005 Michael Getzner, Clive L. Spash and Sigrid Stagl for selection
and editorial matter; the contributors for individual chapters
All rights reserved. No part of this book may be reprinted or reproduced
or utilised in any form or by any electronic, mechanical, or other means,
now known or hereafter invented, including photocopying and recording,
or in any information storage or retrieval system, without permission in
writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
A catalog record for this book has been requested
Contents
PART I
Extending the environmental valuation approach 21
PART II
Taking multiple criteria into account 97
PART III
Deliberation, participation and value expression 185
Index 290
List of figures vii
Figures
Tables
Contributors
Preface
This book arose from on-going debates occurring amongst European ecological
economists who have been reflecting upon extensions of the ‘standard’ economic
model of environmental valuation. Such extensions try to account for citizens’
attitudes, institutional framing and social context as they affect any given
environmental change being valued. Three broad approaches to valuation can
then be identified and are pursued here. First, economists have been looking to
learn from social psychology. Thus, for example, the concept of reasoned action
has been one way in which qualitative and quantitative factors have begun to be
linked in some contingent valuation studies. At the same time the process of learning
from social psychology challenges standard approaches and requires that new
alternatives be developed. A major set of alternatives to environmental cost–benefit
analysis, which form the second broad approach, focus upon multiple criteria for
understanding and improving decisions. The work reported here explores both
the potential, as well as the pitfalls of multidimensional and non-monetary
techniques. The final approach deals with innovation in the area of vested interest
group and citizen participation in processes of environmental valuation. One focus
of research in this area has been in terms of combining economic valuation and
participatory deliberative approaches arising from political science. Work is
presented comparing citizens’ juries and contingent valuation and reflecting upon
potential combinations termed deliberative monetary valuation. Stakeholder or
vested interest group participation is more common than trying to represent the
general public but, as discussed, can also prove difficult. Overall, this book shows
a range of approaches being explored by ecological economists and how they are
open to interdisciplinary critiques, often dismissed by narrowly focused disciplines.
In particular, European ecological economists have been engaging with traditional
economics approaches while also learning from applied philosophy, geography,
political theory, social psychology and sociology.
xiv Acknowledgements
Acknowledgements
Michael Getzner would like to pay tribute to Professor Egon Matzner (Vienna),
who sadly passed away prematurely in October 2003, and who he will miss as
both an outstanding teacher and a colleague.
We all acknowledge our debt to the European Society for Ecological Economics
which has provided an active forum for discussion on such matters as are covered
in this volume. There is now a great deal of networking that takes place across
Europe amongst those addressing the socio-economic aspects of ecological
problems. This includes both those involved in policy processes and academics.
The area of science and society interactions is a general and key concern with
specific realisations in new technologies. Friends and colleagues who have stimulated
our thoughts in these areas with respect to the topics in this volume include: Anders
Biel, Jacquie Burgess, Claudia Carter, Mick Common, Silvio Funtowicz, Colin
Green, Alan Holland, Sybille van den Hove, Joan Martinez-Alier, Martin
O’Connor, John O’Neill, Ortwin Renn, Peter Söderbaum and Arild Vatn. There
are also several people outside Europe with whom we interacted and acknowledge
as active in these debates including: Edmundo Claro, John Gowdy, Jack Knetsch,
Michael Lockwood, Dick Norgaard, Bryan Norton and Sabine O’Hara. Of course
as editors we also thank all the contributors to this volume.
Abbreviations
Introduction
The method of inclusion of environmental resources and ecosystem services in
decision processes determines how far they are taken into account with results
affecting the quality of our lives and those of future generations. A persistent
argument has been that monetary valuation is essential if the ‘environment’ is to
have any chance of being included in government and business decisions.
Environmental cost–benefit analysis (CBA) was developed by environmental
economists in order to achieve this monetisation of environmental entities so that
the prices in market economies might be adjusted. A range of methods were
developed including travel cost, hedonic pricing, production function analysis,
contingent valuation and choice modelling (see Hanley and Spash 1993; Spash
and Carter 2002). The overall aim has been to select project options on the basis
of their welfare impacts and to support government taxes which reflect the social
costs of environmental degradation. This approach has met with some success in
that various national and international agencies have been interested in performing
monetary valuation exercises as part of their overall assessment of projects. The
idea of environmental taxation has also risen on the political agenda if remaining
limited in practice.1
However, there has also been criticism of environmental CBA, or more
specifically of some of the studies conducted under that guise. Critiques can be
broadly grouped into those concerned with the theoretical foundations of economic
values, and those looking at the validity of specific numbers being produced and
the tools employed. In the former case, Kapp (1950) provides an isolated early
example which shows the limits to monetary as opposed to other values in society.
More recently, environmental philosophers have produced a whole body of
literature showing the narrowness of value theory typically expressed by economists
and the failure of economic training to place this theory in the context of wider
social values (see O’Neill 1993). Reflecting upon this literature requires realising
that most economics concerns consequentialism using a form of preference-based
utilitarianism, which is a very specific philosophy of value rather than a generally
accepted meta ethic which can be universally applied. In trying to address the
environmental problems of the 20th century, economic value theory has shown
2 Clive L. Spash, Sigrid Stagl and Michael Getzner
itself lacking in several areas, such as the treatment of time, complexity, strong
uncertainty, political systems, rights and social norms. The enhanced Greenhouse
Effect raises all these issues and exposes the failures of discounting, partial
equilibrium analysis, risk assessment and assuming that efficiency is the key social
goal (Spash 2002b).
The other area of critical analysis abstracts from the first set of problems and
tends to be a debate on validity which has been largely internal to the economics
profession (although implications of the value debate are also related via theoretical
foundations). Validity concerns result in specific applications being criticised for
their failure to respect microeconomic and welfare theoretic constraints. Those
working within decision making government institutions, such as environmental
protection agencies or the treasury, may see such theoretical restrictions as academic
and going beyond what is required for practical policy making (for example, see
comments by Burney 2000). However, once valuation becomes divorced from its
theoretical roots, numbers can be produced which have little content or meaning,
and are defensible only in terms of their political role rather than their theoretical
basis. In this respect the popularisation of environmental CBA has been a problem
as studies proliferate, numbers are transferred out of context and applications
seek to meet policy desires. Excessive aggregation is exemplified by attempts to
value regional, national and even the world’s ecosystems, although what is meant
by a trade price in these contexts is mystifying, and asking who might be the ‘seller’
and ‘purchaser’ exposes the fallacious use of economic value theory. Furthermore,
the range of reasonable numbers stemming from monetisation exercises is often
very broad, which facilitates the arbitrary use of valuation results.
The contingent valuation method (CVM) has become the most widely conducted
CBA tool. The main advantage attracting this attention is the ability of CVM to
estimate what are termed option, ‘existence’ and bequest values in addition to
direct use values. These are amounts reflecting: the preservation of an option to
use a resource in the future, the maintained existence of a resource or entity
regardless of personal use, and the desire to pass the resource to one’s descendents.
The sum of these indirect or passive use values can be large compared with the
direct use values associated with non-market goods to which the other CBA methods
are solely restricted. The popularity of the CVM is also based upon the apparent
simplicity of asking members of the general public the maximum they are willing
to pay (WTP) for an environmental improvement or, less commonly, the least they
would be willing to accept (WTA) in compensation for environmental degradation.
Considerable research energy has gone into refining CVM and related stated
preference approaches. The result has been to increase the complexity and cost of
conducting a CVM if it is to attain the highest standards of validity. In the United
States of America this is the legally defensible standard which can cost millions of
dollars for a nationally representative random sample including a range of pre-
tests, focus groups and internal consistency and validity checks in the survey, as
well as several variations on survey design.
The litigation surrounding the oil spill from the tanker Exxon Valdez led to a
set of such standards being produced by a panel of experts commissioned by the
Exploring alternatives for environmental valuation 3
National Oceanic and Atmospheric Administration (NOAA). This panel consisted
of a number of famous economists, such as Nobel Prize winners Kenneth Arrow
and Robert Solow. The panel was also asked to investigate the concept of ‘existence’
value. The panel’s report gave a qualified support for the CVM (NOAA 1993).
The panel stated that ‘existence’ values are a theoretically meaningful aspect of
value. As regards the ability of the CVM to estimate such values, the panel noted
five main problem areas: inconsistency of responses with economic models of
rational choice; mental account bias, especially when respondents are inadequately
informed as to substitutes; aggregating benefits; providing information; and warm
glow effects. However, the panel felt that so long as certain guidelines were followed,
CVM results could be judged as both meaningful and useful indicators of natural
resource damages. These guidelines are that (i) WTP should be used in preference
to WTA; (ii) mail surveys should be avoided; (iii) respondents should be given full
information on the resource change (including information on substitutes), and be
asked how well they understand this information; (iv) open-ended responses should
be rejected in favour of closed-ended referendum formats; (v) a random population
sample is required; (vi) respondents should be reminded about the need to reduce
expenditure on some item of their budget in order to be able to pay their stated
bid; and (vii) careful pre-testing should be carried out.
Such rules tend to be excessively prescriptive and ignore the specific context in
which a valuation exercise must take place. There are certainly general design
features in any contingent valuation study which are desirable. These include clear
description of the institutional context, explaining the consequences and expected
benefits of payment, being aware of various biases which can be introduced in
design, such as use of a contentious bid vehicle, and generally producing a realistic
scenario. Without addressing such issues any research which claims to be relevant
to the economic debate will be seen as inadequate and may also be misleading.
However, employing rules which prevent hypothesis testing and experimental design
is more worrisome. The NOAA Panel holds an implicit belief in a single universal
standard of best practice which ignores cultural, legal and institutional variation.
People trade differently in different countries (e.g. haggling vs. accepting the stated
price), property rights vary (WTP vs. WTA), areas of trade which are taboo differ,
and truly random sampling is difficult at best and impossible in industrially
developing countries. A serious concern here is the extent to which research designs
which fail to conform to the rules are branded as poor practice when they may be
valid research aiming to question the underlying tenets of the model. Blind use of
highly specific guidelines can remove the room for original research, especially in
the areas of economic psychology and human behaviour.
CBA, along with mainstream microeconomics, has been increasingly criticised
for building upon an unsound theoretical base in terms of human behaviour and
the CVM has made this even more evident. That is, the microeconomic axioms of
choice make assumptions which are at best inconsistent with theories of modern
psychology and empirical evidence. In the valuation area, Kahneman and Tversky’s
(1979) prospect theory revealed that people value gains and losses asymmetrically
and this can explain the observed gap between WTP and WTA measures. Knetsch
4 Clive L. Spash, Sigrid Stagl and Michael Getzner
(1994, 1995, 1999; also Knetsch and Sinden 1984) has shown the frequent
occurrence of behaviours which are inconsistent with accepted economic norms
but commonly dismissed by economic models. The refusal to make trade-offs has
been shown to arise in CVM studies both amongst those who protest against the
use of monetary valuation of the environment and also those prepared to make a
contribution (Spash 2000b, 2002a, 2002c). The results can be linked to rights-
based beliefs which contrast with those of the economic utilitarian (Spash and
Simpson 1994; Spash 2000c). In terms of the economic model, the only way in
which such behaviour can be characterised is as lexicographic preferences which
are deemed, at best, to be held only by a strange minority (Spash 1998, 2000a;
Spash et al. 2000). A range of techniques is therefore required to understand human
psychology in terms of the attitudes, ethical beliefs and social norms which motivate
behavioural responses.
Concerns raised about the practice of attaching monetary values to all
dimensions of socio-economic and biophysical systems have led to the call for
alternative valuation methods (Martinez-Alier et al. 1998). There has been a
powerful argument for shifting away from the sole focus on outcome towards the
quality of decision processes (Funtowicz and Ravetz 1990; O’Connor et al. 1998).
In particular, multiple criteria assessment has become a more widely used method,
offering the potential for the integrated assessment of local, national and inter-
national policies and as a means for combining different perspectives associated
with sustainability goals. Sustainability raises a set of issues based on civil rights of
current and future generations as well as respect for ecological systems. Under-
standing that biophysical as well as human systems are complex, and will never be
fully understood, has led to the development of approaches which favour adaptive
behaviour and learning processes over optimal solutions. Closely linked to this
view is the acknowledgement of uncertainty of future events and their impacts on
human and non-human systems. All these factors redefine the role of experts, the
meaning of knowledge and how decision processes need to be designed to make
more effective policy.
In the environmental policy arena, and elsewhere, there has been a push for
greater public participation, e.g. in Europe the Aarhus Convention (European
Commission 1998), and the inclusion of non-governmental stakeholders in project
appraisal. Researchers are meeting this challenge both through interdisciplinary
discourse and also by contributing to the formation of novel democratic insti-
tutions. There is both innovation in traditional environmental CBA and in terms
of new political institutions for addressing valuation issues. In the former case are
the range of techniques termed deliberative monetary valuation (Spash 2001),
and in the latter such tools as the citizens’ jury. Research into the operational
criteria for effective public participation is at an early stage of development.
The search for validity in applying CBA methods has lead to a variety of appeals
to interest groups and/or members of the public in an attempt to supplement the
normal information content of prices placed on the environment. Thus, the travel
cost method may be combined with an interview approach in order to sustain
assumptions of how individuals behave, value time and relate to the environment.
Exploring alternatives for environmental valuation 5
Hedonic pricing has sought corroboration from estate agent valuations as
representing ‘informed’ preferences. Focus groups have been used in conjunction
with the CVM to test survey design on the basis that group deliberation could
validate the information content and help identify design biases. This last approach
is most clearly where deliberative practices have begun to enter. For example, the
largest CVM study in the UK was conducted on environmental impacts associated
with aggregates (Department of the Environment, Transport and the Regions
1999). An interesting feature of this work in the current context was the informal
use of vested industrial interests (stakeholders) in the first part of the study and the
use of public focus groups in the design stage of the second, although the feedback
from the public proved problematic by diverging from economic assumptions (e.g.
the expressed desire for community compensation unrelated to the individual),
and neither process was formally reported.
Thus, two broad approaches to combining deliberation and monetary valuation
can be identified. The first regards monetary valuation as basically sound but
being able to benefit from supplementary, and often informal, processes borrowing
elements of deliberation. The second sees the use of deliberative approaches as a
new method allowing the (collective or individual) production of a monetary
valuation for environmental goods and services. Under the first approach a variety
of alternatives exists, and monetary valuation may be either followed or preceded
by some element of deliberation. Stakeholder participation, as mentioned above,
may be employed to validate outcomes. The implication being that ex post
deliberations can be used in some way to adjust valuation results or their
presentation. Deliberative processes and environmental valuation may also be
sequential, e.g. selecting a sub-sample of participants from a CVM survey for a
citizens’ jury on the same environmental issue. Ex ante deliberation has been
employed in designing CVM surveys with the use of focus groups to test the wording
and respondents’ understanding of survey questions. Deliberation is then regarded
as useful in providing insight into the processes by which respondents produce
their WTA or WTP bid. This may be extended to allowing a deliberative process
to determine the options or institutional context to be valued in the survey.
The second approach is what Spash (2001) has termed deliberative monetary
valuation (DMV) as advocated by, for example, Brown et al. (1995), Jacobs (1997),
Ward (1999) and Kenyon et al. (2001). DMV is the use of formal deliberation
concerning an environmental impact in order to express value in monetary terms
for policy purposes, and more specifically as an input to CBA. For example, consider
a proposal to build a new road through a wilderness area and so destroy the habitat
of a number of rare or threatened species. A group of citizens would be selected
and meet to discuss information about these environmental damages associated
with the development. Known costs and benefits (discounted) would be presented,
while those pertaining to environmental damages would be deliberated. The citizens
would form a jury aiming to provide a monetary value for environmental damages
which might be in terms of an individual WTA to allow the project to proceed.
The result would then be incorporated into a net present value calculation to
determine the viability of the project.
6 Clive L. Spash, Sigrid Stagl and Michael Getzner
The meaning of such values remains contentious, as they are mediated individual
values. In order to address the increasingly evident fact that preferences are formed
and that there is a lack of arbitrage in valuing environmental goods and services,
economists are appealing to methods from political science. However, approaches
such as citizens juries and contingent valuation, or more generally deliberative
forums and monetary valuation, differ in fundamental ways. This has been pointed
out by Niemeyer and Spash (2001) as involving the approach taken to theoretical
factors (individual and social ontology, preference basis, rationality theory), practical
factors (justification, framing, value representation, institutional setting), and
political factors (manipulation, representation, social impact). The variation
between approaches to each of these factors means the very concept of DMV is
brought into question. In simplified terms, can DMV take the best of both monetary
and deliberative methods as advocates hope or does it merely create a messy
confusion as to the values being expressed?
Clearly, research into valuation needs to span different disciplines such as social
psychology, applied philosophy, political theory and economics. While interest in
doing so seems to have increased, the quality of the applications can be questionable
(Spash 2000b). Part of the problem is for economists to understand the requirements
of other disciplines, while practitioners in those other disciplines must similarly
have a good practical knowledge of what economists are trying to do. Thus,
applying attitudinal measures from social psychology may prove insightful with
respect to WTP, but only if researchers also understand welfare economics and
CBA tools so that their work can address the economics literature. This is a
challenging research agenda but one which, as several contributions in this volume
show, is now being taken on at several different levels.
Note
1 Revenue raising is a key objective of taxation rather than designing environmentally
efficient and effective taxes. This means taxes which may be described as environmental
in fact fail to address the pollution problem to which they are supposedly related. For
example, a tax justified as reducing the Greenhouse Effect may select carbon dioxide
emissions regardless of the other gases needing control and the tax may then relate
only to industrial emissions from point sources, ignoring a whole range of non-point,
household and public sources, or select only one specific fuel. Such decisions are
taken on the basis of ease of administration and revenue-raising potential rather
than their economic efficiency at controlling pollution and matching marginal social
costs of environmental use.
References
Ajzen, I., T.C. Brown and L.H. Rosenthal (1996) ‘Information bias in contingent valuation:
effects of personal relevance, quality of information and motivational orientation’,
Journal of Environmental Economics and Management 30(1): 43–57.
Brown, T.C., G.L. Peterson et al. (1995) ‘The values jury to aid natural resource decisions’,
Land Economics 71(2): 250–60.
Burney, J. (2000) ‘Is valuing Nature contributing to policy development’, Environmental Values
9(4): 511–20.
Department of the Environment, Transport and the Regions (1999) The Environmental Costs
and Benefits of the Supply of Aggregates: Phase 2, London, Department of the Environment
Transport and the Regions: 208.
European Commission (1998) Aarhus Convention on Access to Information, Public Participation in
Decision Making and Access to Justice in Environmental Matters, Brussels.
Funtowicz, S.O. and J.R. Ravetz (1990) Uncertainty and Quality in Science for Policy, Dordrecht,
Kluwer Academic Publishers.
Hanley, N. and C.L. Spash (1993) Cost–Benefit Analysis and the Environment, Aldershot, Edward
Elgar.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature? Economics, Ethics and Environment,
London, Routledge: 211–31.
Kahneman, D. and A. Tversky (1979) ‘Prospect theory: an analysis of decision under
risk’, Econometrica 47(2): 263–91.
Kapp, K.W. (1950) The Social Costs of Private Enterprise, New York, Shocken.
Kenyon, W., C. Nevin and N. Hanley (2001) ‘Citizens’ juries: an aid to environmental
valuation?’ Environment and Planning C 19: 557–66.
Knetsch, J.L. (1994) ‘Environmental valuation: some problems of wrong questions and
misleading answers’, Environmental Values 3(4): 351–68.
Knetsch, J.L. (1995) ‘Assumptions, behavioral findings, and policy analysis’, Journal of Policy
Analysis and Management 14(1): 68–78.
Knetsch, J.L. (1999) ‘Environmental valuations and standard theory: behavioural findings,
context dependence, and implications’, in T. Tietenberg and H. Folmer (eds) The
Exploring alternatives for environmental valuation 19
International Yearbook of Environmental and Resource Economics 2000/2001, Cheltenham,
Edward Elgar: 41.
Knetsch, J.L. and J.A. Sinden (1984) ‘Willingness to pay and compensation demanded:
experimental evidence of an unexpected disparity in measures of value’, Quarterly Journal
of Economics 99(3): 507–21.
Martinez-Alier, J., G. Munda and J. O’Neill (1998) ‘Weak comparability of values as a
foundation for ecological economics’, Ecological Economics 26(3): 277–86.
Niemeyer, S. and C.L. Spash (2001) ‘Environmental valuation analysis, public deliberation
and their pragmatic syntheses: a critical appraisal’, Environment & Planning C: Government
& Policy 19(4): 567–86.
NOAA (1993) ‘Natural resource damage assessment under the Oil Pollution Act of 1990’,
Federal Register 58(10): 4601–14.
O’Connor, M., S. Funtowicz, F. Agliera-Klink, C.L. Spash and A. Holland (1998) Valuation
for Sustainable Environments: The VALSE Project Full Final Report. Ispra, European
Commission, Joint Research Centre: 395.
O’Neill, J. (1993) Ecology, Policy and Politics: Human Well-Being and the Natural World, London,
Routledge.
O’Neill, J. (2001) ‘Representation’, Government and Policy 9(4).
Ostrom, E. (1998) ‘A behavioral approach to the rational choice theory of collective action’,
presidential address, American Political Science Association. American Political Science
Review 92(1): 1–22.
Royal Commission on Environmental Pollution (1998) Setting Environmental Standards,
London, Her Majesty’s Stationery Office: 232.
Spash, C.L. (1998) ‘Investigating individual motives for environmental action: lexicographic
preferences, beliefs and attitudes’, in J. Lemons, L. Westra and R. Goodland (eds)
Ecological Sustainability and Integrity: Concepts and Approaches, Dordrecht, Kluwer Academic.
Spash, C.L. (2000a) ‘Ecosystems, contingent valuation and ethics: the case of wetlands re-
creation’, Ecological Economics 34(2): 195–215.
Spash, C.L. (2000b) ‘Ethical motives and charitable contributions in contingent valuation:
empirical evidence from social psychology and economics’, Environmental Values 9(4):
453–79.
Spash, C.L. (2000c) ‘Multiple value expression in contingent valuation: economics and
ethics’, Environmental Science & Technology 34(8): 1433–8.
Spash, C.L. (2001) ‘Deliberative Monetary Valuation’. 5th Nordic Environmental Research
Conference, University of Aarhus, Denmark.
Spash, C.L. (2002a) ‘Empirical signs of ethical concern in economic valuation of the
environment’, in D.W. Bromley and J. Paavola (eds) Economics, Ethics, and Environmental
Policy: Contested Choices, Oxford, Blackwell: 205–21.
Spash, C.L. (2002b) Greenhouse Economics: Value and Ethics, London, Routledge.
Spash, C.L. (2002c) ‘Informing and forming preferences in environmental valuation: coral
reef biodiversity’, Journal of Economic Psychology 23(5): 665–87.
Spash, C.L. and C. Carter (2002) ‘Environmental valuation methods in rural resource
management’, in F. Brouwer and J. van der Straaten (eds) Nature and Agricultural Policy in
the European Union: New Perspectives on Policies that Shape the European Countryside, Cheltenham,
Edward Elgar: 88–114.
Spash, C.L. and N. Hanley (1995) ‘Preferences, information and biodiversity preservation’,
Ecological Economics 12(3): 191–208.
Spash, C.L. and I.A. Simpson (1994) ‘Utilitarian and rights-based alternatives for protecting
Sites of Special Scientific Interest’, Journal of Agricultural Economics 45(1): 15–26.
20 Clive L. Spash, Sigrid Stagl and Michael Getzner
Spash, C.L., J. van der Werff ten Bosch, S. Westmacott and J. Ruitenbeek (2000)
‘Lexicographic preferences and the contingent valuation of coral reef biodiversity in
Curaçao and Jamaica’, in K. Gustavson, R.M. Huber and J. Ruitenbeek (eds) Integrated
Coastal Zone Management of Coral Reefs: Decision Support Modeling, Washington, DC, World
Bank: 97–118.
Ward, H. (1999) ‘Citizens’ juries and valuing the environment: a proposal’, Environmental
Politics 8(2): 75–96.
A framework for valuing nature 21
Part I
Extending the
environmental valuation
approach
22 Michael Getzner
A framework for valuing nature 23
Introduction
The contingent valuation method (CVM) seems to divide the economics profession.
While there might be higher agreement among a core of (neoclassical)
environmental economists on the theoretical foundations of valuing public goods
with that specific method, applying the CVM is an important point of scientific
discourse. Ecological economists not only doubt empirical applications of the CVM
in many respects, but also question its welfare theoretic foundations and more
generally those of cost–benefit analysis (CBA). A crucial point in that debate is the
understanding of markets as social institutions, and the behaviour of individuals
as embedded in a specific social and institutional context.
An important element in that debate is also the critique of the concept of Homo
oeconomicus who is assumed to have well-behaved, stable preferences, which are
‘home-grown’ and can be elicited by means of CVM surveys – provided several
guidelines are followed (e.g. Arrow et al. 1993; Department of the Interior 1994).
An aggregation of these individual values is said to point to the (societal) ‘value’ of
a natural good, and compared with the costs of protecting the respective natural
good, it is assumed that society’s welfare is in principle not much more than the
sum of individual welfare aggregated in a social welfare function.
Underlying preconditions for this approach are the notion that individuals are
in principle willing to trade money for natural goods. That means that even in the
case when individuals hold ethical beliefs or commitments often associated with
natural goods, they can be ‘bribed’ to give up their beliefs for money. It has to be
assumed that individuals who state a willingness to pay (WTP) base their decisions
on similar foundations and considerations regardless of the nature and the
characteristics of the good in question. Thus, using the CVM in a CBA for societal
decisions assumes that individuals do not take different viewpoints when valuing
private or public (natural) goods because contingent valuation puts individuals in
a situation where they decide on public goods in a market-like setting. Another
element in this debate is whether the hypothetical bias can have such an influence
that results of CVM studies are fundamentally flawed. The main argument is that
people behave differently when actually forced to ‘realize’ their stated WTP.
24 Michael Getzner
The ‘standard’ and ‘extended’ model of valuation
Figure 2.1 shows the ‘standard’ model of valuation and an ‘extended’ model that
includes a number of additionally influential factors regarding an individual’s stated
behaviour (stated WTP). Normally, studies explaining WTP in a bid function try
to relate a number of individual factors (socio-economic characteristics like age
and income, knowledge, perception of environmental quality) to respondents’ WTP
bids.1 The explanatory power of such models is often very low (Mitchell and Carson
1989, argue that models should at least explain 15 per cent of the dependent
variable’s variation). A low explanatory power might indicate omitted variables
bias indicating that those individual variables chosen to explain stated behaviour
are not the variables which fully influence a respondent’s stated WTP. Following
that argument, other variables might be able to reasonably explain stated behaviour
with a higher degree of accuracy. Recently, a number of authors have stressed the
importance of the social and institutional context of decisions, especially when
these decisions refer to public goods (e.g. Hodgson 1997; Hayden 1993; see with
reference to bathing water quality: Georgiou et al. 1998).
An extension of the ‘standard’ model of valuation2 to account for additional
factors is presented in the lower part of Figure 2.1. First, extending individual
factors may play a vital part in a valuation survey when respondents are not willing
to trade natural goods for money (lexicographic preferences, see this section). The
institutional context of the survey is on the one hand, of course, set up by the
payment vehicle and the valuation questions themselves.3 On the other hand,
respondents’ beliefs about and perceptions of the CVM instrument may alter their
behaviour (next section). The social context comprises both beliefs about the referen-
dum and about assumptions on others’ contributions, and the consumer–citizen
dichotomy (third section). Figure 2.1 also shows the labels of variables that
operationalize the different additional factors (to be used in the econometric model
in the empirical part of this chapter).
The central questions for which answers are sought by means of the empirical
survey in the third section are: how important is the notion of lexicographic prefer-
ences and what is the effect of the respondents’ beliefs about institutional and
social context variables in explaining WTP bids for a regional biodiversity
programme in Carinthia (Austria)?
The structure of this chapter is as follows. First, the elements of the ‘extended’
model of valuation are briefly discussed, presenting a focused literature review of
the specific issues which build the basis of the empirical study presented in the
second part of the chapter. Second, results of a valuation study designed to include
‘extended’ factors into the bid function are presented.
history tells us an important fact, namely, that goods and services can be
substituted for one another. … There is no specific object that the goal of
sustainability, the obligation of sustainability, requires us to leave untouched.
… Sustainability doesn’t require … that any particular species of fish or any
particular tract of forest be preserved’ (emphasis in the original).
Empirical results17
where Pr(WTPi > 0) is the probability that respondent i exhibits some positive
WTP. WTPi refers to respondent i’s WTP bid (in Austrian schillings).
Four groups of variables are included in the analysis. The first group S describes
the socio-economic characteristics of respondent i (e.g. income, gender, age, job);
P denotes respondent i ’s preferences and behaviour regarding natural goods in
general (e.g. knowledge concerning biodiversity, donations to environmental
organizations, memberships). I is a set of institutional context variables, which
describe not only ‘objective’ arguments (e.g. whether hypothetical bias has been
explicitly mentioned in the survey by ‘cheap talk’) but also respondent’s beliefs
about the CVM setting. Finally, C comprises arguments regarding ‘subjective’
perceptions and citizen’s values about the social context of the survey. These
variables include the respondent’s opinions about the outcome of the referendum
and the consumer–citizen dichotomy. The WTP function increases or decreases
in arguments18 (for a complete list of variables, see Table 2.3).
A framework for valuing nature 33
Design of the survey
The survey for testing the relevance of extending the ‘standard’ valuation model
took place in December 1997 and January 1998 at the University of Klagenfurt
(Austria). Subjects were mainly undergraduate students of business administration
as well as sociology, English and French language. The surveys were held in the
classes; about 20 to 25 minutes were necessary to complete the questionnaire
(sample size 189).
The ‘design guidelines’ of Arrow et al. (1993) regarding the appropriate design
and setting of the CVM survey were followed as closely as possible. After a number
of pre-tests (in classes and with focus groups), the survey was finally structured as
follows. After filling out a number of questions regarding socio-economic variables
(age, gender, studies/courses enrolled, income, children), a set of indicators dealt
with the respondent’s attitudes towards the environment (membership of
environmental organizations, preferred activities in recreational areas, e.g. sports)
and with their knowledge of and familiarity with the concept of biodiversity (using
open-ended as well as closed questions). After these introductory ‘warming-up’
questions, the sample was divided into two groups: one group went straight on
answering the following valuation questions without any further explanation from
the instructor. The other group was asked to take a short break and listen to the
instructor’s short presentation of the problems associated with the hypothetical
bias of such CVM surveys.
The next set of questions asked for the respondent’s WTP for a nature
protection programme in Carinthia.19 These open-ended questions were worded
as conservatively as possible,20 including references to the respondent’s budget
constraint. The payment vehicle was an earmarked contribution to a nature
protection fund. The payment motives as well as reasons for protest or zero bids
were asked for. Additionally, a question on the respondent’s vote on a public
referendum regarding the introduction of an earmarked nature protection tax
in Carinthia was included.
A final set of questions tried to reveal respondents’ attitudes comprising the
‘extended’ WTP model, asking for lexicographic preferences as well as the above-
discussed distinction between social role and epistemological questions (results for
the most important questions of these survey sections, including the wording of
the questions, are presented below).
Respondent as consumer
‘The Nockberge national park is not (yet) an internationally acknowledged national
park, but is protected by provincial law. Let us assume that the Nockberge would be
changed into an attractive and also reasonably priced ski and recreation resort with
numerous lifts and sports facilities, and the protection of nature would be abandoned.
Would you personally like to go skiing in the Nockberge national park or enjoy other
attractive sports and recreation facilities?’
Frequency Per cent
No 138 73.02
Yes 51 26.98
Total 189 100.00
Respondent as citizen
‘Regardless of your personal possibilities or wishes to use the Nockberge national park
for your sports and recreation enjoyment: would you as a citizen taking all arguments
into account vote in a referendum for the use of the Nockberge as an attractive ski
resort, or do you think that the Nockberge should in all cases remain a national park?’
Frequency Per cent
I think that the national park should
strictly be protected regardless of the
economic benefits of a ski resort. 178 94.18
I would vote for the ski resort because
of the economic benefits. 7 3.70
Sum 185 97.88
Missing/no answer 4 2.12
Sum 189 100.00
Note: All questions are translated from the original German questionnaire.
‘Surveys asking respondents about their personal willingness to pay for environmental
improvements are relatively rare. Answering the questions and this specific kind of
questioning is perceived by some as strange or is rejected. Which of the following
statements would you agree to?’ [Multiple answers]
Frequency Respondents (%)
This sort of questioning has neither
influenced me personally nor my opinions
about natural protection. 61 32.28
Before this survey, I had not thought about
my willingness to pay for a nature
protection programme. 124 65.61
This kind of question has influenced me.
I will in the future pay more attention to
whether nature protection yields economic
benefits compared to the costs. 26 13.76
I reject these kinds of questions on the
willingness to pay for nature protection,
because species protection cannot be valued
in monetary terms. 14 7.41
N 189
leading to an R² of 42 per cent. Finally, the inclusion of the variable ‘Con_cit’ only
marginally improves the quality of the model. About 87 per cent of all answers
can correctly be predicted.
After having explained the respondent’s principal WTP, the second part of the
‘nested’ model was estimated by OLS. Extending the valuation framework to
explain the respondent’s WTP bids by means of a simple OLS regression model
improves the model, too. As Table 2.5 shows, the usually significant socio-economic
variables in regression models are significant in this estimation as well. Most
influential are ‘Income’ and ‘Children’, and (interestingly) ‘Gender’. Some
expectedly influential variables regarding environmental behaviour and
environmental consciousness also play a significant role in explaining the
respondent’s WTP bid. Variables describing the institutional and social context in
which decisions on nature protection programmes are made are of crucial
importance. A highly significant explanatory power is revealed by those variables
that show the respondent’s decisions (behaviour) in public referenda as well as
their willingness to donate a personal contribution even if they are in a minority
position. That means that the higher the supposedly altruistic motives
operationalized with this concept, the higher the respondent’s WTP bid is.
Beginning with the usual socio-economic characteristics of respondents,
variables ‘Income’, ‘Gender’, ‘Biodiversity’, ‘Member’, and ‘Recreation’ play a
significant role. The explanatory power of the OLS model for the respondent’s
WTP bids is 19 per cent. If we include institutional context variables (respondent’s
perception of the CVM setting), about 24 per cent of the variance of the dependent
38 Michael Getzner
continued …
Table 2.4 Continued
continued …
Table 2.5 Continued
Acknowledgements
The author would like to thank all participating students of the University of
Klagenfurt as well as the instructors of those courses who ‘donated’ 25 minutes of
their course time. Particular thanks go to G. Haber, W. Hediger, S. Niemeyer, M.
Riedel and N. Wohlgemuth for commenting on earlier drafts of this chapter. The
author accepts the responsibility for all remaining errors.
Notes
1 Based on the recommendation of Arrow et al. (1993), the concept of respondents’
WTP is used. However, many results of this chapter might also be transferred to
respondents’ willingness to accept.
2 The ‘standard’ model referred to in this chapter is a sketch of the usually applied
approach to explain the respondent’s WTP (see also the theory of reasoned action by
Fishbein and Ajzen 1975).
3 The term ‘institutional’ is taken from the auction literature where the ‘institution’
refers to the elicitation instrument and auction setup.
4 The concept of biodiversity has become ‘popular’, particularly after the United Nations
Conference on Environment and Development in Rio de Janeiro (1992) which
produced a convention on the protection of biodiversity and biological resources.
This convention defines biodiversity in natural science terms as the diversity within
species (genetic diversity), species diversity (number of different species living in an
ecosystem), as well as ecosystem and landscape diversity (different ecosystems and
their protection).
5 Diversity is a characteristic of a species as well as of an ecosystem. Thus, valuation
becomes much more complicated. Respondents in CVM studies are usually asked to
assess policy-relevant proposals that often involve a number of biodiversity issues.
Thus, respondents must consider their sympathy towards single species as well as
taking into account their perceptions towards the importance of a (frequently
unknown) number of species, their interrelations to secure the stability of ecosystems
as a whole, including risk and (fundamental) uncertainties, as well as expected needs
of future generations. In the light of some ecological fundamentals including
uncertainty, the task of appropriately valuing biodiversity seems particularly difficult.
6 That means that respondents in a CVM survey might consider several elements of
an ecosystem to be substitutable or not vital for the function of that ecosystem while
ecologists and biologists would strongly oppose such beliefs.
7 For further critique, see Boulding (1994: 8): ‘It is curious that economists, almost
without exception, take human valuations for granted and assume that everybody
has a set of preferences that can be expressed as an indifference curve.’ Preferences
are often found to be imprecise, as there is ‘plentiful evidence of intervals within
which respondents found it difficult to be sure whether they would or would not pay
or accept compensation for marginal changes in risks of injury’ in road accidents
(Dubourg et al. 1994: 127).
46 Michael Getzner
8 It can be argued that preferences based on existence values are more likely to be of
lexicographic nature since existence values are often founded on rights-based beliefs
(Stevens et al. 1991). Furthermore, a WTP bid based on existence values can
additionally be founded on a number of other reasons, e.g. ‘warm glow’ giving
(Kahneman and Knetsch 1992), paying for ‘good causes’ (Stevens et al. 1994) or impure
altruism (Andreoni 1990).
9 For more examples on the ‘unwillingness to accept’, see O’Connor (1997: 463).
10 Some differences of results of these studies can, of course, be attributed to differences
in the institutional settings and contexts of these experiments.
11 Hanemann (1995: 88) argues that statements of respondents and their preferences
are ‘inherently context-laden’. Furthermore, he reviews in detail the problems involved
with CVM and voting, and offers a number of methods to overcome problems with
inadequately defined context variables.
12 Note, the diverging roles assumed here are not generally accepted in mainstream
economics ‘since citizens can be expected to contribute, in one way or another, toward
the costs of policy, the respondent trades off the prospect of an amenity improvement
against the prospect of a reduction in disposable income’ (Hoehn and Randall 1987:
229). However, the strict distinction between ‘citizen’ and ‘consumer’ is made for
illustrative purposes. Those two concepts are better thought of as two poles of a
continuum from self-interested utility maximization to purely altruistic social
behaviour. Individuals can, of course, behave according to both patterns (i.e. they are
not substitutes).
13 System theory particularly stresses the importance of differences between systems in
order to be able to strictly divide between circumstances that are inside or outside a
particular system (e.g. Luhmann 1986).
14 ‘Individual behavior is always mediated by social relations’ (Arrow 1994: 5).
15 In discussing their results, Schkade and Payne (1994: 107) explicitly mention the ‘citizen
input (values) into policy decision’. Their study shows that within CVM questionnaires
citizen’s considerations are generally excluded rather than acknowledged.
16 Public projects may run into problems not only because of project characteristics but
also due to missing or inadequate involvement of citizens. In this sense CVM as a
hypothetical market transaction is ‘blind to reasons’ (Keat 1997: 36). Nevertheless, a
CVM survey is one way of involving affected citizens.
17 The questionnaire, additional written material as well as all data, are available from
the author on request.
18 There are a number of hypotheses regarding the sign of the coefficients in the model.
For instance, WTPi / Yi 0 , with Yi being the income of respondent i as part of the
group of Sm variables in equations (3.1a) and (3.1b).
19 An actually existing nature programme of the regional government of Carinthia to
protect and enhance biodiversity was valued. The programme was described in terms
of what was to be valued (¾ page of description of the programme), how it would be
provided and at what level (regional government and nature protection fund), who
would pay (partly regional government budget and the fund), and how long it would
last (5 years).
20 Due to budgetary constraints (small sample) and lack of previous valuation estimates,
a question format including dichotomous choice was deemed infeasible. In any case,
the significance of institutional and social context variables seems independent of
the chosen question format.
21 The figures presented in this section illustrate preferences of the entire sample
consisting mainly of students. The values do not provide a basis for aggregation to
the whole population. Rather than producing WTP figures for Carinthia, the aim
was to show the influence of extended context variables on WTP bids, which can be
assumed to exist in the whole population.
A framework for valuing nature 47
22 Respondents’ problems with understanding the concept of biodiversity can also be in
part interpreted as being a semantic difficulty: In German, the word Biodiversität is
apparently poorly known. The words Artenschutz and Artenvielfalt would be understood
more clearly, but these concepts only comprise species protection and diversity, leaving
out genetic as well as ecosystems diversity.
23 Of course, the credibility of the proposed programme can also be of crucial
importance. Respondents thus might particularly value policy programmes and
environmental goods as ‘public’ goods rather than context-free discrete ‘bits’ of
environmental goods (Brouwer et al. 1999).
References
Aldred, J. (1994) ‘Existence value, welfare and altruism’, Environmental Values, 3(4): 381–
402.
Andreoni, J. (1990) ‘Impure altruism and donations to public goods: a theory of warm-
glow giving’, The Economic Journal, 100 (June): 464–77.
Arrow, K.J. (1994) ‘Methodological individualism and social knowledge’, American Economic
Review, 84(2): 1–9.
Arrow, K.J., Solow, R.M., Leamer, E., Portney, P., Randner, R. and Schuman, H. (1993)
‘Report of the NOAA panel on contingent valuation’, Federal Register, 58(10): 4602–14.
Becker, G.S. (1976) The Economic Approach to Human Behavior, Chicago: University of Chicago
Press.
Bingham, G., Bishop, R., Brody, M, Bromley, D., Clark, E., Cooper, W., Costanza, R.,
Hale, T., Hayden, G., Kellert, S., Norgaard, R., Norton, B., Payne, J., Russell, C. and
Suter, G. (1995) ‘Issues in ecosystem valuation: improving information for decision
making’, Ecological Economics, 14(2): 73–90.
Bishop, R.C. and Welsh, M.P. (1992) ‘Existence values in benefit-cost analysis and damage
assessment’, Land Economics, 68(4): 405–17.
Bishop, R.C., Heberlein, T.A. and Kealy, M.J. (1983) ‘Contingent valuation of
environmental assets: comparison with a simulated market’, Natural Resources Journal,
23(3): 619–33.
Blackburn, M., Harrison, G.W. and Rutström, E.E. (1994) ‘Statistical bias functions and
informative hypothetical surveys’, American Journal of Agricultural Economics, 76(5):
1048–88.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and the
expression of citizen values and issue options’, in V. Adamovicz (ed.) Forestry, Economics
and the Environment, Wallingford: CAB International.
Boulding, K.E. (1969) ‘Economics as a moral science’, American Economic Review, 59(1):
1–12.
Boulding, K.E. (1994) ‘Economic ethics and the environmental crisis’, in E. Hickey and
L.A. Longmire (eds) The Environment: global problems, local solutions, Westport, CT/London:
Greenwood Press.
Boyce, R., Brown, T.C., McClelland, G.H., Peterson, G.L. and Schulze, W.D. (1992) ‘An
experimental examination of intrinsic values as a source of the WTA-WTP disparity’,
American Economic Review, 82(5): 1366–73.
Brouwer, R., Powe, N., Turner, R.K., Bateman, I.J. and Langford, I.H. (1999) ‘Public
attitudes to contingent valuation and public consultation’, Environmental Values, 8:
325–47.
48 Michael Getzner
Burgess, J., Clark, J. and Harrison, C.M. (1998), ‘Respondents’ evaluations of a contingent
valuation survey: case study based on an economic valuation of the wildlife enhancement
scheme, Pevens Levels in East Sussex’, Area 30(1): 19–27.
Cameron, J.I. (1997) ‘Applying socio-ecological economics: a case study of contingent
valuation and integrated catchment management’, Ecological Economics, 23(2): 155–65.
Carson, R.T. (1998) ‘Valuation of tropical rainforests: philosophical and practical issues
in the use of contingent valuation’, Ecological Economics, 24(1): 15–29.
Carson, R.T. (2000) ‘Contingent valuation: a user’s guide’, Environmental Science and Technology,
34(8): 1413–18.
Carter, J.R. and Irons, M.D. (1991) ‘Are economists different, and if so, why?’, Journal of
Economic Perspectives, 5(2): 171–7.
Common, M. and Perrings, C. (1992) ‘Towards an ecological economics of sustainability’,
Ecological Economics, 6(1): 7–34.
Common, M., Blamey, R. and Norton, T. (1993) ‘Sustainability and environmental
valuation’, Environmental Values, 2(4): 299–334.
Common, M., Reid, I. and Blamey, R. (1997) ‘Do existence values for cost benefit analysis
exist?’, Environmental and Resource Economics, 9(2): 225–38.
Conlisk, J. (1996) ‘Why bounded rationality?’, Journal of Economic Literature, 34 (2): 669–700.
Cummings, R.G., Harrison, G.W. and Osborne, L.L. (1995a) ‘Are realistic referenda real?’,
Economics Working Paper B-95-06, Division of Research, College of Business
Administration, Columbia: University of South Carolina.
Cummings, R.G., Harrison, G.W. and Osborne, L.L. (1995b) ‘Can the bias of contingent
valuation surveys be reduced? Evidence from the laboratory’, Economics Working Paper
B-95-03, Division of Research, College of Business Administration, Columbia:
University of South Carolina.
de Dios, E.S. (1995) ‘Re-reading Menger’s table’, European Journal of Political Economy, 11(2):
317–34.
Department of the Interior (1994) ‘Natural resource damage assessment’, Federal Register,
59(85): 23098–111.
Dubourg, W.R., Jones-Lee, M.W. and Loomes, G. (1994) ‘Imprecise preferences and the
WTP-WTA disparity’, Journal of Risk and Uncertainty, 9(2): 115–33.
Edwards, S.F. (1992) ‘Rethinking existence values’, Land Economics, 68(1): 120–2.
Ehrlich, P. and Ehrlich, A. (1992) ‘The value of biodiversity’, Ambio, 21(3): 219–26.
Encarnación, J. (1990) ‘Consumer choice of qualities’, Economica, 57(1): 63–72.
Etzioni, A. (1986) ‘The case for a multiple utility conception’, Economics and Philosophy, 2(2):
159–83.
Fischhoff, B. (1991) ‘Value elicitation: is there anything in there?’, American Psychologist,
46(8), 835–47.
Fischhoff, B. and Furby, L. (1988) ‘Measuring values: a conceptual framework for
interpreting transactions with special reference to contingent valuation of visibility’,
Journal of Risk and Uncertainty, 1(1), 147–84.
Fishbein, M. and Ajzen, I. (1975) Belief, Attitude, Intention and Behavior: An Introduction to Theory
and Research, Reading: Addison-Wesley.
Frey, B.S. (1997) Not Just for the Money, Cheltenham: Edward Elgar.
Georgiou, S., Langford, I.H., Bateman, I.J. and Turner, R.K. (1998) ‘Determinants of
individuals’ willingness to pay for perceived reductions on environmental health risks:
a case study of bathing water quality’, Environment and Planning A, 30 (3): 577–94.
Getzner, M. (1999) ‘Risk, uncertainty and discounting in practical environmental decision
making’, in K. Steininger and F. Prettenthaler (eds) Proceedings of the Workshop ‘Risk and
A framework for valuing nature 49
Uncertainty in HD-Research’, Austrian Programme on Human Dimensions of Global
Environmental Change, Graz: Department of Economics of the University of Graz.
Getzner, M. (2000) ‘Hypothetical and real economic commitments, and social status in
valuing a species protection program’, Journal of Environmental Planning and Management,
43(4): 541–59.
Goodman, S., Jaffry, S. and Seabrooke, B. (1999) ‘Assessing public preferences for the
conservation quality of the British coast, in M. O’Connor and C.L. Spash (eds) Valuation
and the Environment: Theory, Method and Practice, Cheltenham: Edward Elgar: 165–82.
Gowdy, J. (1996) ‘Discounting, hierarchies, and the social aspects of biodiversity protection’,
International Journal of Social Economics, 23(4,5,6): 49–63.
Gowdy J. and Olsen P.R. (1994) ‘Further problems with neoclassical environmental
economics’, Environmental Ethics, 16(2): 161–71.
Hanemann, W.M. (1995) ‘Contingent valuation and economics’, in K.G. Willis and J.T.
Corkindale (eds) Environmental Valuation: New Perspectives, Wallingford: CAB International:
79–117.
Hanley, N., Spash, C. and Walker, L. (1995) ‘Problems in valuing the benefits of biodiversity
protection’, Environmental and Resource Economics, 5(3): 249–72.
Harrison, G.W. (1996) ‘Experimental economics and contingent valuation’, Economics
Working Paper B-96-10, Division of Research, College of Business Administration,
Columbia: University of South Carolina.
Hayden, F.G. (1993) ‘Ecosystem valuation: combining economics, philosophy, and ecology’,
Journal of Economic Issues, 27(2): 409–20.
Hodgson, G.M. (1997) ‘Economics, environmental policy and the transcendence of
utilitarianism’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment,
London/New York: Routledge: 48–66.
Hoehn, J.P. and Randall, A. (1987) ‘A satisfactory cost indicator from contingent valuation’,
Journal of Environmental Economics and Management, 14(3): 226–47.
Holland, A. (1997) ‘The foundations of environmental decision-making’, International Journal
of Environment and Pollution, 7(4): 483–96.
Holm-Müller, K., Hansen, H., Klockmann, M. and Luther, P. (1991) Die Nachfrage nach
Umweltqualität in der Bundesrepublik Deutschland, Berlin: Erich Schmidt Verlag.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision
making’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment, London/
New York: Routledge: 211–31.
Jakobsson, K.M. and Dragun, A.K. (1996) Contingent Valuation and Endangered Species,
Cheltenham: Edward Elgar.
Jeffrey, R.C. (1974) ‘Preferences among preferences’, Journal of Philosophy, 71(39): 377–91.
Johansson, P.-O. (1992) ‘Altruism in cost-benefit analysis’, Environmental and Resource Economics,
2(4): 605–13.
Kahneman, D. and Knetsch, J.L. (1992) ‘Valuing public goods: the purchase of moral
satisfaction’, Journal of Environmental Economics and Management, 22(1): 57–70.
Keat, R. (1997) ‘Values and preferences in neo-classical environmental economics’, in J.
Foster (ed.) Valuing Nature? Ethics, Economics and the Environment, London/New York:
Routledge: 32–47.
Kosz, M. (1997) ‘Probleme der monetären Bewertung von Biodiversität’, Zeitschrift für
Umweltpolitik und Umweltrecht, 20(4): 531–50.
Lazo, J.K., Schulze, W.D., McClelland, G.H. and Doyle, J.K. (1992) ‘Can contingent
valuation measure nonuse values?’, American Journal of Agricultural Economics, 74 (6): 1127–
32.
50 Michael Getzner
Lockwood, M. (1997) ‘Integrated value theory for natural areas’, Ecological Economics, 20(1):
83–93.
Lockwood, M. (1998) ‘Integrated value assessment using paired comparisons’, Ecological
Economics, 25(1): 73–87.
Luhmann, N. (1986) Ökologische Kommunikation, Opladen: Westdeutscher Verlag.
Marwell, G. and Ames, R.E. (1981) ‘Economists free-ride, does anyone else?’, Journal of
Public Economics, 15(3): 295–310.
McDaniels, T.L. and Roessler, C. (1998) ‘Multiattribute elicitation of wilderness preservation
benefits: a constructive approach’, Ecological Economics, 27 (3): 299–312.
Menger, C. (1871) Einführung in die Volkswirthschaftslehre: erster, allgemeiner Theil, Vienna:
Braumüller.
Mitchell, R.T. and Carson, R.C. (1989) Using Surveys to Value Public Goods: The contingent
valuation method, Washington DC: Resources for the Future.
O’Connor, M. (1997) ‘The internalization of environmental costs: implementing the
polluter pays principle in the European Union’, International Journal of Environment and
Pollution, 7(4): 450–82.
O’Hara, S. (1996) ‘Discursive ethics in ecosystem valuation and environmental policy’,
Ecological Economics, 16(2): 95–107.
Peukert, H. (1998) Das Handlungsparadigma in der Nationalökonomie, Marburg: Metropolis Verlag.
Sagoff, M. (1988) The Economy of the Earth, New York: Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24(2/3): 213–30.
Schkade, D.A. and Payne, J.W. (1994) ‘How people respond to contingent valuation
questions: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26(1): 88–109.
Schulze, W., McClelland, G., Waldman, D. and Lazo, J. (1996) ‘Sources of bias in contingent
valuation’, in D.J. Bjornstad and J.R. Kahn (eds) The Contingent Valuation of Environmental
Resources, Cheltenham: Edward Elgar: 97–116.
Sen, A.K. (1977) ‘Rational fools: a critique of the behavioral foundations of economic
theory’, Phil. and Public Affairs, 6(4): 317–44.
Solow, R.M. (1993) ‘Sustainability: an economist’s perspective’, in R. Dorfman and N.S.
Dorfman (eds) Economics of the Environment, 3rd edn, New York: W.W. Norton.
Spash, C. and Hanley, N. (1995) ‘Preferences, information, and biodiversity preservation’,
Ecological Economics, 12(3): 191–208.
Stevens, T.H., Echeverria, J., Glass, R.J., Hager, T. and More, T.A. (1991) ‘Measuring the
existence value of wildlife: what do CVM estimates really show?’, Land Economics, 67(4):
390–400.
Stevens, T.H., More, T.A. and Glass, R.J. (1994) ‘Interpretation and temporal stability of
CV bids for wildlife existence: a panel study’, Land Economics, 70 (3): 355–63.
Stork, P. and Viaene, J.-M. (1992) ‘Policy optimization by lexicographic preference ordering’,
Journal of Policy Modeling, 14 (5): 655–73.
Toman, M.A. (1999) ‘Sustainable decision making: the state of the art from an economics
perspective’, In M. O’Connor and C.L. Spash (eds) Valuation and the Environment: Theory,
Method and Practice, Cheltenham: Edward Elgar: 59–72.
Tversky, A., Sattath, S. and Slovic, P. (1988) ‘Contingent weighting in judgement and
choice’, Psychological Review, 95 (3): 371–84.
Vatn, A. and Bromley, D.W. (1994) ‘Choices without prices without apologies’, Journal of
Environmental Economics and Management, 26 (2): 129–48.
Non-use values and attitudes 51
Introduction
Climate change is expected to cause extensive damage in coastal areas due to a
rise in sea level and increasingly frequent violent storms. Scientists expect that, in
particular, the nature and landscape of the German Wadden Sea, a coastal wetland
ecosystem, will be negatively influenced. The adjacent salt marshes, a habitat for
endangered species, have already been adversely affected in quality and quantity
by human interference, such as by dam construction and waste disposal into the
North Sea, and are expected to deteriorate even faster. However, it is possible to
mitigate the consequences of climate change by taking suitable measures such as
removing dams (Reise 1993, 1996). To obtain information about the population’s
support for the protection of the Wadden Sea against the consequences of climate
change, the willingness to pay (WTP) of German households was determined by
using the contingent valuation method (CVM).1 Several studies have shown that
non-use values can form an important part of the total economic value of
environmental assets (Garrod and Willis 1996; Bateman and Langford 1997),
particularly for highly unique resources. Therefore, non-use values were also
thought to form an important part of the total economic value of the Wadden
Sea. This area is a very valuable natural refuge, ranking third out of thirteen
national parks in Germany.
However, despite an ongoing debate among environmental economists about
the compatibility of non-use values with neoclassical theory, it is far from obvious
whether all bids from non-users express economic values and should therefore be
used in cost-benefit analyses. As a guideline, some authors require that it should at
least be possible to explain stated non-use values partly by predictors suggested by
economic theory. ‘If it is argued that non-use values are theoretically included in
economic benefits, these values should therefore be suitable for economic models
of choice, this would imply, among other things, that the value for a good with
non-use value should be partly explained by income, and possibly by other economic
determinants of value’ (Wierstra 1996: 118; also Blamey et al. 1995: 272). One
aim of this study was to clarify the relationship between attitudes and non-use
values, testing the results by Wierstra. In his study On the Domain of Contingent Valuation,
he found that the variation of stated non-use values could be explained solely
based on attitudes.
52 Jürgen Meyerhoff
This chapter is organized in the following way. In the second section, a brief
overview of the discussion about the relationship of attitudes and WTP is given.
The third section presents the design of the contingent valuation and its results in
relation to the use and non-use of the Wadden Sea area. Section four presents
findings on the relationship between measured attitudes and WTP, while in section
five the results of multiple analyses by SPSS-CHAID and OLS regression are
discussed. Further conclusions are drawn in section six.
Table 3.2 Summary statistics of the WTP (in DM) for users and non-users
All User Non-user
n 1412 695 717
Protest votes 790 362 428
Used sample 622 333 289
Willing to pay 336 225 111
Mean WTP 7.8 11.1 3.6
Standard deviation 17.2 20.8 10.3
Median WTP 1.0 5 0
Mode WTP 0 0 0
Maximum WTP 170 170 100
Non-use values and attitudes 57
Influence of attitudes on WTP
As mentioned in section one, it was the aim of this study to test Wierstra’s results
about the relationship between attitudes and WTP. According to him, it was
hypothesized that the conditions for economic values were not met if the WTP of
non-users could be sufficiently explained by attitudinal factors; in this case the
survey would support the results by Wierstra. Also following Wierstra, attitudes
towards more general objects such as nature and climate change were gathered.
The reason for this is that only people with enough information about the good
are in a position to hold site-specific attitudes. Less informed people, such as non-
users, will possibly construct attitudes (Slaby 1996). Similar to the problem of
constructed preferences, these attitudes will supply no sound basis for any analysis.
Therefore, the gathering of attitudes was designed more generally, allowing for
uninformed non-users to be considered. Figure 3.1 shows the conceptual model
which is assumed to determine the WTP. Attitudes were mainly elicited towards
three attitude objects: climate change, nature, and valuation of nature with money.
The following subsections present the items and results regarding attitudes towards
climate change and nature, and the results regarding attitudes towards money.
• Nature
Attitudinal factors • Climate change
• Valuation with money
• User/non-user
-
WTP for protection
• Planned visits
of the Wadden Sea
Landscape-related • Knowledge about the
against
factors Wadden Sea
consequences of
• Interest in conservation
climate change
• Distance to Wadden Sea
alpha was calculated. It is the current standard statistic assessing the reliability of
a scale composed of multiple items and it is the most appropriate reliability measure
to use for Likert scales. Because it considers the degree to which items on a scale
intercorrelate, it is often referred to as a measure of internal consistency (Eagly
and Chaiken 1993: 67). The associated alpha is 0.67 for the scale ‘nature’, and
0.75 for the scale ‘climate change’. These low values of Cronbach’s alpha illustrate
the problem arising from the use of a scale which is not as well developed as the
NEP scale, for example. Compared with a generally accepted value of 0.80
(Schumann 1999: 42), the study fails to gather reliable attitudes. However, if
compared with the range of values for Cronbach’s alpha obtained in other studies,
our data prove reasonably reliable. Wierstra reported values of Cronbach’s alpha
of between 0.45 and 0.85. Diekmann (1997: 220–3) mentioned that in the General
Population Survey (Allgemeine Bevölkerungsumfrage (ZUMA-Allbus)) short scales
of four to six items were used and Cronbach’s alpha values of approximately 0.60
were obtained. Therefore, the gathered attitudes towards nature and towards
climate changes were subjected to further evaluation. Tables 3.5 and 3.6 illustrate
the relationship between the attitude score and the mean WTP.
Non-use values and attitudes 59
Table 3.5 Attitudes towards nature and WTP
Score User and non-user User Non-user
The relationship between attitudes towards nature and mean WTP is positive
for all three groups. The results confirm the expectation that people with stronger
attitudes towards nature are on average willing to pay more than people with
weaker attitudes. The mean WTP of users is for each level of the attitude score
higher than for the corresponding group of non-users. The last line of Table 3.5
presents results of the Kruskal–Wallis H-test. This non-parametric test is a test for
the null-hypothesis that the distribution of some variable (WTP in this case) is the
same in all samples defined by a certain grouping variable (attitudes towards nature
in this case). The results allow the rejection of the null-hypothesis of equal
distribution within each attitude class at a high significance level.
The figures in Table 3.6 show the same pattern as in Table 3.5. The mean WTP
is at each level of attitude score higher than at the preceding level. Again, mean
WTP of users is in each case higher than WTP of non-users. The results of the
Kruskal–Wallis H-test allow once more the rejection of the null-hypothesis of equal
distribution of WTP at a high significance level. Thus, attitudes towards climate
change also have, like attitudes towards nature, a significant influence on WTP. In
both cases, a higher score on attitude corresponds to a higher WTP on average.
In addition, the analysis of the relationship between attitudes and WTP confirms
two more findings presented by Kotchen and Reiling (2000). First, respondents
with stronger pro-environmental attitudes are more likely to participate in the
valuation procedure. People who were in-principle willing to pay show a higher
mean score on both the nature and the climate change scale. Second, people who
were defined as protest voters attain lower mean scores on both scales. The Mann–
Whitney test indicates that both differences are significant at the 0.1 per cent level.
60 Jürgen Meyerhoff
Attitude towards money as a measure of value
As another expression of attitude, the degree of agreement with the statement ‘I
reject money as a measure to express the value of nature and landscape’ was used.
Nearly 45 per cent out of the whole sample ‘completely agree’ or ‘agree’ with this
statement, while 25 per cent ‘disagree’ or ‘completely disagree’. Slightly more
than 29 per cent are in between both of these groups. The relationship between
the stated answers and mean WTP for both users and non-users is given in Table
3.7. In both cases the results of the Kruskal–Wallis H-test allow the rejection of
the null-hypothesis of equal distribution of WTP within the defined samples. At a
first glance, it is interesting that also people who ‘completely agree’ or ‘agree’ with
the presented statement were in-principle willing to pay, and, especially in the
group of users, that their mean WTP is relatively high compared with the amounts
stated by people belonging to the other categories. However, a post hoc test (Sachs
1997: 395–7) qualifies these findings. In the group of users, their mean WTP does
not differ between those who ‘completely agree or agree’ or ‘neither agree nor
disagree’ at a 5 per cent significant level. In the group of non-users, only the mean
WTP of those who ‘disagree or completely disagree’ differs from the means of the
other samples. Nevertheless, the expected result remains unchanged. People who
accept money as a measure to express the value of nature and landscape show a
significantly higher mean WTP than those who reject money as such a measure.
Table 3.7 Relationship between attitudes towards money and mean WTP
‘I reject money as a measure to User Non-user
express the value of nature
and landscape’ n % per Mean n % per Mean
column column
Completely agree 41 12.3 11.7 59 20.4 3.4
Agree 91 27.2 6.6 95 32.9 2.2
Neither agree nor disagree 101 30.2 9.7 81 28.0 3.7
Disagree or 100 30.2 17.8 54 18.7 6.0
Completely disagree
χ2 (p) 30.4; df. 3; p < 0.001 24.77; df. 3; p < 0.001
Note: Because of a low number of observations in the group ‘completely disagree’, it was merged
with the group of people who ‘disagree’.
Non-use values and attitudes 61
Unlike his study, both groups of users have to be analysed separately because only
one good with both characteristics was the subject of the CVM. Table 3.8 reports
the potential explanatory variables.
SPSS-CHAID analysis
SPSS-CHAID is a market segmentation technique that is designed for use with
categorical variables and continuous variables which have been divided into
categories (e.g. income, age). Based on the chi-square test, CHAID divides a
population at each level of the analysis into two or more segments that differ with
respect to a designated criterion. The analysis stops if the program is not able to
detect further significant relationships between the dependent and any remaining
variable or if a defined minimum of observations is reached. The first line of each
box in the tree diagrams in Figures 3.2 and 3.3 shows the corresponding variable.
The second line, except for the root of the tree, shows the category of the respective
variable by which CHAID divides the population. In the third line, the percentage
of people who are willing to pay within the subgroup is presented, while the fourth
line shows the total number of people within that group.
In both the user and the non-user group only a very limited number of variables
have influence on the dependent variable ‘in-principle WTP’. Among these
variables none is predicted by economic theory as an important predictor of WTP.
Rather, the attitudinal factors dominate. The attitude towards nature is in both
cases the most influential variable. People who show stronger attitudes towards
In-Principle
Willingness to Pay
yes: 37%
n = 289
In-Principle
Willingness to Pay
yes: 67%
n = 334
ATTMON ATTMON
tot-agree – nn disag – tot-disag
yes: 71% yes: 88%
n = 123 n = 73
nature are more likely to pay for the proposed programme to protect the Wadden
Sea. On the second level, variables differ between both models. In the case of the
non-user group, TRAVEL is the second most important variable. People who plan
a visit or wish to travel to the North Sea are more likely to pay than people who
neither plan nor wish to travel. In the case of the user group, the variable ATTMON
appears with two new segments merged out of the five-point scale. The more
non-users agree to value nature in monetary terms, the higher is the percentage of
people who are in-principle willing to pay.
Non-use values and attitudes 63
However, the third attitudinal variable, that is attitudes towards climate change,
did not show any influence on the in-principal WTP answer either in the group of
users or in the group of non-users. One explanation could be that people perceived
the offered programme to protect the Wadden Sea as not being something that is
connected to climate change. It is supported by the finding that most of the people
who were willing to pay stated that they understood that their payment was not
solely to finance measures against consequences of climate change but to finance
a programme which would also protect the landscape against other negative
interventions.
Discussion
The aim of this chapter was to test whether non-use values are only influenced by
attitudes or whether economic predictors influence these values as well. According
to Wierstra (1996), goods which have a high degree of non-use value could be out
of the domain of contingent valuation because stated WTP for these goods may
only be influenced by attitudes. His results about the WTP for the protection of
Non-use values and attitudes 65
the uninhabited island of Rottumeroog support that hypothesis. As opposed to
this, contingent valuation of public benefits of protecting the Wadden Sea against
consequences of climate change shows exactly the opposite results. While the bids
stated by non-users were influenced among other things by economic variables, in
the case of users they show no influence at all. However, attitudes and selected
landscape-related variables are significant in both models.
Confronted only with the results of the presented analysis, one could draw the
conclusion that non-use values are influenced more by economic variables than
use values. According to Randall (1998) there is no crucial experiment which allows
us to answer the question ‘Does the CVM work?’ Therefore, the findings presented
in this chapter do not allow us to draw the conclusion that non-use values are
always influenced by economic variables. They are just another contribution to
the map of performance characteristics of contingent valuation. In this sense,
they support findings of several other studies that environmental attitudes are
significant explanatory variables of both principle WTP and of the amount of
WTP. Furthermore, they also raise the question of whether it is meaningful to
draw a dividing line between use and non-use values. Instead, it seems to be more
promising to study further which factors determine people’s WTP for collective
goods (Green 1997; Green and Tunstall 1999).
It is also important to examine in which case the Fishbein–Ajzen attitude-model
could be applied to non-users. Fishbein and Ajzen’s work has recently been
recognized as being important in environmental valuation studies. One example
of a study which was similar to their approach is that conducted by Moisseinen
(1999). However, her study aimed at measuring use values. To be in a position to
apply this approach to non-users, several aspects have to be considered. For example,
the theory of reasoned action demands that some requirements such as the ‘prin-
ciple of compatibility’ be fulfilled. But in the case of non-users it is not possible to
completely fulfil this requirement. As already mentioned, non-users are, normally,
people with less information about the good in question. Therefore, it could be
misleading to use the theory of reasoned action when non-users’ WTP is of interest.
The predictability of actual behaviour which follows the behavioural intention
is much weaker if instead of the theory of reasoned action a more general attitude
approach such as the NEP, the one used by Wierstra or in this study is applied.
Therefore, Bishop and Heberlein’s (1986) point that some grounds exist for arguing
that legitimate economic values are being established, at least to a rough approx-
imation, is perhaps no longer valid. A next step in the analysis of the relationship
between attitudes and WTP could be to develop and test models which incorporate
general attitudes (e.g. environmental concern) and attitudes towards specific
behaviour. Within these models, general attitudes are seen as determinants of
attitudes towards specific behaviour (Bamberg et al. 1999; Fransson and Gärling
1999). Furthermore, the amount of knowledge people have about the environmental
goods should be incorporated in more detail.
66 Jürgen Meyerhoff
Notes
1 This was part of a study on costs of climate change under the project ‘Effects of climatic change
on the Island of Sylt: An interdisciplinary research project’. Further information is available at
http://soel.geographie.uni-kiel.de/sylt/english/Welcome.htm.
2 At this stage of the sampling procedure interviewees were not informed about the type of survey.
Therefore, people who refused to participate did not refuse because they don’t agree with
monetary valuation of environmental resources. They refused generally to participate in any
kind of survey.
3 The wording of the corresponding items is: ‘What is your attitude towards the extra payment to
protect the Saimaa seal? The idea of collecting money from the households is: good ´ bad’
(attitude towards the behaviour) and ‘Majority of people I know would participate by paying
the asked amount: likely ´ unlikely’ (subjective norm), (Moisseinen 1999: 202).
4 The definition of ‘users’ as people who have visited the North Sea coast at least once before the
interview is very broad because people who have ‘actually used’ the coast many years ago are
also defined as users. Further analysis showed no significant difference between people who
used the coast many years ago and people who have visited it only one or two years ago. Therefore,
this broader definition was applied during the whole evaluation.
References
Ajzen, I. (1996) ‘The directive influence of attitudes on behavior’ in P.M. Gollwitzer and
J.A. Bargh (eds) The Psychology of Action. Linking Cognition and Motivation to Behavior, New
York/London: The Guilford Press: 385–403.
Bamberg, S. (1996) ‘Allgemeine oder spezifische Einstellungen bei der Erklärung umwelt-
schonenden Verhaltens? Eine Erweiterung der Theorie des geplanten Handelns um
Einstellungen gegenüber Objekten?’, Zeitschrift für Sozialpsychologie, 27(1): 47–60.
Bamberg, S., Kühnel, S.M. and Schmidt, P. (1999) ‘The impact of general attitudes on
decisions’, Rationality and Society, 11(1): 5–25.
Bateman, I.J. and Langford, I.H. (1997) ‘Non-users’ willingness to pay for a national park:
an application and critique of the contingent valuation method’, Regional Studies, 31(6):
571–82.
Bishop, R.C. and Heberlein, T.A. (1986) ‘Does contingent valuation work?’, in R.C.
Cummings, D.S. Brookshire and W.D. Schulze (eds) Valuing Environmental Goods. An
Assessment of the Contingent Valuation Method, Totowa, NJ: Rowman & Allanheld: 123–47.
Blamey, R, Common, M. and Quiggin, J. (1995) ‘Respondents to contingent valuation
surveys: consumers or citizens?’, Australian Journal of Agricultural Economics, 39(3):
262–88.
Brouwer, R. and Slangen, L.H.G. (1998) ‘Contingent valuation of the public benefits of
agricultural wildlife management: the case of the Dutch peat meadow land’, European
Review of Agricultural Economics, 25(1): 53–72.
Diekmann, A. (1997) Empirische Sozialforschung, Reinbeck: Rowohlt.
Dunlap, R.E. and van Liere, K.D. (1978) ‘The “New Environmental Paradigm” ’, Journal
of Environmental Education, 9(4): 10–19.
Eagly, A.H. and Chaiken, S. (1993) The Psychology of Attitudes, Orlando: Harcourt Brace
Jovanovich College Publishers.
Fischhoff, B. (1991) ‘Value elicitation: is there anything there?’ American Psychologist, 46(8):
835–47.
Fishbein, M. and Ajzen, I. (1975) Belief, Attitude and Behaviour. An Introduction to Theory and
Research, Reading, MA: Addison-Wesley.
Non-use values and attitudes 67
Fransson, N. and Gärling, T. (1999) ‘Environmental concern: conceptual definitions,
measurement methods, and research findings’, Journal of Environmental Psychology, 19(4):
369–82.
Garrod, G.D. and Willis, K.G. (1996) ‘Estimating the benefits of environmental
enhancement: a case study of the River Darent’, Journal of Environmental Planning and
Management, 39(2): 189–203.
Georgiou, S., Langford, I.H., Bateman, I.J. and Turner, R.K. (1998) ‘Determinants of
individuals’ willingness to pay for perceived reductions in environmental health risks: a
case study of bathing water quality’, Environment and Planning A, 30: 577–94.
Goodman, S., Seabrooke, W., Daniel, H. Jaffry, S. and James, H. (1996) Determination of
Non-use Values in Respect of Environmental and Other Benefits in the Coastal Zone, Portsmouth:
Centre for Coastal Zone Management.
Green, C. (1997) ‘Are blue whales really simply very large cups of coffee?’, paper presented
at Stockholm Water Symposium, Stockholm, January 1997.
Green, C. and Tunstall, S. (1999) ‘A psychological perspective’, in I.J. Bateman and K.G.
Willis (eds) Valuing Environmental Preferences. Theory and Practice of the Contingent Valuation
Method in the US, EU, and Developing Countries, Oxford: Oxford University Press: 207–58.
Green, C., Tunstall, S., Herring, M. and Sawyer, J. (1993) ‘Customer preferences and
willingness to pay for selected water and sewerage services’, a summary report to the
Office of Water Services, Middlesex: Flood Hazard Research Centre.
Greene, W.H. (2000) Econometric Analysis, New York: Macmillan.
Hofer, D. (1987) ‘Naturschutz als Wertobjekt. Eine exemplarische Studie über Einstellungen
zu Schutzgebieten’, PhD dissertation, Munich: Ludwig-Maximilians-Universität.
Jorgensen, B.S., Syme, G.J., Bishop, B.J. and Nancarrow, B.E. (1999) ‘Protest responses in
contingent valuation’, Environmental and Resource Economics, 14(1): 131–50.
Kahneman, D., Ritov, I., Jacowitz, K.E. and Grant, P. (1993) ‘Stated willingness to pay for
public goods. A psychological perspective’, Psychological Science, 4: 310–15.
Kahneman, D., Ritov, I. and Schkade, D. (1999) ‘Economic preferences or attitude
expressions? An analysis of dollar responses to public issues’, Journal of Risk and Uncertainty,
19(1–3): 203–35.
Kaiser, G.F., Wölfing, S. and Fuhrer, U. (1999) ‘Environmental attitude and ecological
behaviour’, Journal of Environmental Psychology, 19(1): 1–19.
Kotchen, M.J. and Reiling, S.D. (2000) ‘Environmental attitudes, motivations, and
contingent valuation of nonuse values: a case study involving endangered species’,
Ecological Economics, 32(2): 93–107.
Moisseinen, E.M. (1999) ‘On behavioural intentions in the case of the Saimaa Seal.
Comparing the contingent valuation approach and the attitude-behaviour research’,
in M. O’Connor and C.L. Spash (eds) Valuation and the Environment. Theory, Method and
Practice, Cheltenham: Edward Elgar: 183–204.
Pate, J. and Loomis, J.B. (1997) ‘The effect of distance on willingness to pay values: a case
study of wetlands and salmon in California’, Ecological Economics, 20(3): 199–207.
Preisendörfer, P. (1996) ‘Umweltbewußtsein in Deutschland. Ergebnisse einer
repräsentativen Bevölkerungsbefragung 1996’, report for the Umweltbundesamt, Berlin:
Umweltbundesamt.
Randall, A. (1998) ‘Beyond the crucial experiment: mapping the performance characteristics
of contingent valuation’, Resource and Energy Economics, 20(2): 197–206.
Reise, K. (1993) ‘Verschwommene Zukunft der Nordseewatten’, in H.-J. Schellnhuber
and H. Sterr (eds) Klimaänderung und Küste, Berlin: Springer: 223–9.
68 Jürgen Meyerhoff
Reise, K. (1996) ‘Wattökologische Folgen bei Änderung von Klima und Küste’, Schriftenreihe
der Schutzgemeinschaft Deutsche Nordseeküste, 1: 31–45.
Sachs, L. (1997) Angewandte Statistik, Berlin: Springer.
Schumann, S. (1999) Repräsentative Umfrage, München: Oldenbourg.
Silberman, J. and Gerlowski, D.A. (1992) ‘Estimating existence value for users and nonusers
of New Jersey beaches’, Land Economics, 68(2): 225–36.
Slaby, M. (1996) ‘Einstellungsmessung oder Einstellungsgenerierung? Die Bedeutung der
informationellen Basis bei Befragten für die empirische Rekonstruktion von
Einstellungen zu gentechnischen Anwendungen (Measuring or generating attitudes?
Subjective knowledge base for the empirical reconstruction of attitudes towards new
applications of genetic engineering)’, Stuttgart: Institut für Sozialforschung of the
Universität Stuttgart.
Spaninks, F.A., Kuik, O. and Hoogeven, J.G.M. (1996) ‘Willingness to pay of Dutch
households for a natural Wadden Sea. An application of the contingent valuation
method’, Amsterdam: Instituut voor Milieuvraagstukken of the Free University
Amsterdam.
Spash, C.L. (1998) ‘Environmental values and wetland ecosystems: CVM, ethics and
attitudes’, in M. O’Connor (ed.) The VALSE Project: Social Processes for Environmental
Valuation, full final report to the DG-XII, European Commission, chapter 6–B.
Sutherland, R.J. and Walsh, R.G. (1993) ‘Effect of distance on the preservation value of
water quality’, Land Economics, 61: 281–91.
Wierstra, E. (1996) On the Domain of Contingent Valuation, Enschede: Twente University Press.
Modelling environmental behaviour 69
4 Modelling environmental
behaviour
Socio-psychological simulation
Hans-Joachim Mosler
Introduction
In economically highly developed countries, many of the conditions that would
allow people to behave in environmentally responsible ways are already in place.
We have a lot of knowledge; for years surveys have shown that people give top
priority to the need to act on environmental issues. The necessary technical and
economic resources are also available. But there is little sign of a real about-face
except in limited areas. We believe that the much-cited discrepancy between
cognition and behaviour, between lip service and a person’s own contribution to
conserving the environment, can be better understood if we also take people’s
perceptions of the social surround into consideration. To view human beings as
Homo economicus (Diekmann 1996; Harsanyi 1977) falls short, because the economic
man approach does not explicitly take into account inner psychological factors,
such as motives, values, norms and attitudes, nor does it consider social influences
on a person’s behaviour. The present research conceives a model of behaviour
that in addition to economic factors includes personal and social factors. Environ-
mental consciousness is determined to a significant degree by social systems, or in
other words, by people’s corresponding social representations. An individual’s
personal contribution seems insignificant in the face of massive destruction of the
environment caused by many. This perception – that there is nothing we can do
personally, that each one of us is powerless – as well as a reluctance to be the
‘sucker’, are important causal factors in behaviours that overuse environmental
resources. It does not seem rational to exercise personal restraint (for example, by
not driving), because not only will we suffer from the harm caused by the general
public’s overuse (the consequences of air pollution), but from a reduction in our
own direct return as well (time saved, comfort). However, as this state of affairs
applies equally to all individuals in a society overusing environmental resources,
people mutually trap each other in patterns of actions that harm the environment.
It is for this reason that we are particularly interested in examining the psychological
conditions that would form the basis of a collective reorientation towards environ-
mentally sustainable behaviour. Starting out from new environmentally friendly
behaviours of some ‘pioneer’ individuals, we wish to discover the social psycholog-
ical conditions that would ensure that the number of persons joining ranks with
70 Hans-Joachim Mosler
such pioneers would continue to automatically increase and result in a true, large-
scale ‘turn-around’ of previous, environmentally harmful patterns of behaviour.
The focal question of our research can thus be framed as follows: what are the
conditions that foster widespread, effective inner dynamics that change collective
environmental overuse (in thinking and action) to collective, environmentally
responsible thinking and acting? Findings generated by this research will lend
scientific support to the planning of environmental protection campaigns.
Method
Procedure
If we start from the assumption that environmental problems originate in the
overusing behaviours of very many individuals, we need to consider how new
solutions might be tested in a problem area of this magnitude. The instrument of
the questionnaire, based upon imaginary situations or conditions, seems ill-suited
(‘How expensive would gasoline have to be for you to change to public
transportation?’). Massive field experiments that translate the issue directly in real
social systems can also be eliminated; given our present state of scientific knowledge,
such experiments would be both financially and ethically irresponsible. Laboratory
experiments, which would not require intervention in existing social systems, cannot
be carried out with large groups of persons (1000 and more). Computer simulation
provides a possible solution. Simulation aims to reconstruct the relevant cause-
and-effect relationships in a problem area in the form of a model. With the aid of
empirical data, the relationships can then be validated. In this way, we can test
‘experimentally’ the most various and unconventional ideas of ways to spread
environmentally responsible thinking without incurring the risk of intervention in
real social systems.
It is important to remember, however, that simulation results are not empirical
results found in reality. Simulation results can generate fruitful hypotheses that
must be tested in reality. The great advantage is that the hypotheses are derived
consistently using a clearly defined procedure. Another important point to consider
is that findings gained through the simulation method allow only comparative
conclusions, for example that under the conditions studied, Measure A produced
better results than Measure B. The simulation method, furthermore, does not
allow statements to be made about the means to be used to implement certain
measures. For this, real experiments with real people are required. Despite all of
this, computer simulation is the only method available for investigating, by means
of many systematically varied experiments, how environmentally responsible
changes can be instituted in large populations.
Our simulation procedure consisted of the following steps.
Following a preliminary selection of the most important, empirically well-
founded theoretical approaches within the field of social influence, the theories’
most relevant and significant core statements were formulated, according to content
criteria.
Modelling environmental behaviour 71
In the ensuing modelling, the core statements of the theories were described
with the existing variables and set in relation to one another according to certain
systems-theoretical rules (see ‘Modelling of the theoretical concepts’). For lack of
space, we here dispense with discussion of programming/technical implement-
ations of the theories (see Mosler 2000 for details).
The design of the simulation model was validated through the aid of experts’
evaluations and replication of findings from the fields of environmental and social
psychology. Again, due to space considerations, the validation process is not reported
here (cf. Mosler 2000 for details).
Experiments with various strategies for the spread of behaviours were then
conducted. The following will report on the most important and meaningful
experiments.
From the findings of the simulations, conclusions were drawn pertinent to both
environmental practice and basic research. The most well known forms of inter-
vention stemming from environmental research receive a new interpretation; well-
founded recommendations for the field of intervention are derived.
The reader may wish to note that the following deals intensively with psycho-
logical concepts (to psychologists, however, the descriptions will appear somewhat
superficial). The author sees the present chapter as a contribution that takes a
perspective that differs greatly from the economic viewpoint, although some relevant
connections are pointed out. This becomes clear in the next section. The reader
interested in the exact derivation of the findings should read all sections, while
others may wish to read only the sections on the simulation experiments and results.
The most important findings are reported and discussed in the concluding section.
External and internal input variables are processed in different sub-models (see
Figure 4.1) according to the theory being applied. The output values of the sub-
models have either a direct outward effect on other persons with whom the
individual has contact (attitude, persuasiveness), or they affect variables leading
up to behaviour and decisions on the use of resources and consequent actions.
The following sub-models were designed and simulated (simulations of sub-models
II and IV will not be presented):
All of these variables are conceptualized at a general level, thus according with
Ajzen and Fishbein’s (1977) correspondence hypothesis, which states that there is
a high correlation between behaviour and the determining variables if they have
been operationalized with the same degree of specification.
As a fundamental extension of the theory of planned behaviour (Ajzen and
Madden 1986), we have added an additional factor, cost–benefit analysis. Although
the theory of planned behaviour does contain a kind of cost–benefit analysis, in
that its components are broken down in ‘expectation evaluation products,’ we
conceptualise this factor as an independent component. With this approach we
expect to achieve better explanations of the variance for intention, but more
importantly, to gain added starting points for intervention measures. Cost–benefit
analysis models people’s ‘economic’ considerations, as they determine whether it
will pay for them to act in certain ways. The model is thus the first to integrate
economic and social psychological components (see also Mosler and Tobias 2000).
These components all influence behavioural intention, which leads to behaviour
towards the environment. Additionally, we introduce volition into the model, which
76 Hans-Joachim Mosler
is closer to actual behaviour than behavioural intention. Psychological research
(Frey et al. 1993b) has shown that the act of making a decision commits a person to
the when and where of attempting to realise an intention. As the entire model is
deterministic in construction, random variables play absolutely no role in deter-
mining individual behaviour towards the environment.
INDIVIDUAL
Status of contact
persons Attractiveness of group
50
Attitude of the population
45
40
35
Not 0 5 10 15 20 25
environmentally
Runs
responsible
Intervention
A. In each group there is an isolated pioneer whose susceptibility to influence is the same in degree
as that of the members of the group majority.
B. In each group there is an isolated pioneer whose susceptibility is lower than that of the members
of the group majority.
C. In each group there is a pioneer whose susceptibiity is the same in degree to that of group
members; in addition, the pioneer is in contact with 10 other ‘pioneers’ in other groups. They
thus no longer experience themselves to be a minority.
D. As in C above, but here pioneers have a lower degree of susceptibility.
E. All pioneers are concentrated within their own core groups and show low susceptibility to
influence. Each pioneer is in contact with another person outside of the core group.
F. As in E above, whereby here pioneer group members are in contact with 10 other persons
outside of the core group.
The multiplicative relation of the three variables expresses the fact that the effect
of one of the variables is greater, the greater the value of the other variables.
There is no effect at all if one of the variables equals zero. According to the theory,
moreover, the effect of the variable N (number) is not linear, but rather is an
exponential function: I = sN t, where I is impact, s is a constant, and the exponent
t is a value less than 1. The parameters s and t are different for each situation and
have to be determined empirically. The factor ‘number’ thus has the effect that the
first person has the greatest impact and each person thereafter has ever less of an
impact. With an increasing number of influencing source persons, the social impact
on a person rapidly decreases.
Expectations of efficacy can be approximately represented by behaviour control,
because we define behaviour control as a person’s conviction that they can indeed
execute the behaviour. In accordance with the framework model, behaviour control
affects the behavioural intention and, thus, behaviour. According to Latané’s theory
of social impact (1981), behaviour arises from observation of the using behaviour
of contact persons: social influence is a multiplicative function of the number,
strength, and immediacy of contact persons. In our model (see Figure 4.4), the
variable ‘number’ is determined by the number of the individual’s contact persons
who behave in an environmentally responsible, or irresponsible, way. The ‘strength’
of contact persons’ influence is defined by their status. ‘Immediacy’ is conceived
INDIVIDUAL
Using behaviour
Using
behaviour of Behaviour control
contact persons Intention
Values Self-confirmation
Volition
Self-responsibility
Using
behaviour
A. In this experiment, 500 persons of high status are selected and caused to
adopt a behaviour that is more eco-friendly for the duration of the intervention.
They are to demonstrate this behaviour to many others (15 contacts).
B. As in A above, but here only 300 persons are selected.
C. As in A, but here the behaviour is demonstrated to only 5 contacts.
D. Here we selected 500 persons showing very environmentally sustaining
behaviour patterns, which they demonstrate to 15 contacts. In this case, their
status is raised for the duration of the intervention.
It appears to be more effective to select persons having high status as role models
and to induce them to behave in a more environmentally friendly way for the
duration of the intervention than to temporarily raise the status of persons already
showing such behaviours. An increase in the number of contacts, that is, an increase
in the visibility of the environmentally sound behaviour, achieves relevant effects.
An increased number of role models also has a strong effect. The dissemination
process continues to progress automatically for some time after the intervention,
as the entire social system must again adapt to the changes.
Environmentally
responsible 56
A. 15 CP/500 P/Act+
54
B. 15 CP/300 P/Act +
52
82 Hans-Joachim Mosler
C. 5 CP/500 P/Act +
50
D. 15 CP/500 P/Status +
48
46
44
Control situation
42
Not 40
environmentally 0 5 10 15 20 25 30 35 40 45 50 Runs
responsible
Intervention
CP: contacts per person; P: number of persons addressed (role models); Act+, Status+: more responsible acts towards environment.
Figure 4.5 Average population behaviour with differing interventions based on observational learning
Modelling environmental behaviour 83
In developing concepts for environmental campaigns, simulation can contribute
support with regards to the implementation of efficient means. It can answer
questions as to the number of role models required, how these should be selected,
and what changes they should show, the degree of visibility necessary, and so on.
In particular, simulation allows us to estimate the effects that can be achieved by
means of compensation in other values, which makes a direct contribution to an
increase in efficiency (for example, an increase in visibility at the expense of the
number of role models). However, the simulation method does not allow conclusions
about implementing the results, such as statements on how realistic it is to induce
high status people to behave in a more environmentally friendly way. As discussed
above, the simulation method permits only comparative conclusions.
Estimated proportion
Using behaviour of resource- Expectation of
of contact sustainable Sustainability
sustaining users
persons utilization motive
Values Using
behaviour
stemming from inner psychological biases or from external influences in the form
of media reports, disputes among the experts, and so on.
The expectation that one’s own way of using the environmental resource will
make a contribution towards preservation of the resource is formed from the required
proportion of resource-sustaining users in the population and the estimated proportion
of resource-sustaining users. In other words, the greater the proportion of non-
sustaining users of a resource as estimated by the person in comparison with the
required proportion of resource-sustaining users in the population, the lower the
person’s expectation that they themselves can make an effective contribution to
sustainable use of the resource. This expectation rises in the degree to which the
proportion of estimated resource-sustaining contact persons exceeds the required
proportion of resource-sustaining users. Knowledge of the number of resource-
sustaining users required in order to preserve the resource comes from resource
knowledge and the perception of the state of the resource. The model presumes
people to have – at present not a very realistic assumption! – knowledge about the
regeneration rate, regeneration characteristics, and the maximal state of the resource.
From the discrepancy between the maximal state of the resource and the current
state of the resource, the person derives the required proportion of resource-sustaining
users. The greater the discrepancy between the current and the maximal state of the
resource, the greater must be the proportion of the population that uses the resource
in a sustainable way if the resource is to recover.
The estimated proportion of resource-sustaining users in the population is based
upon the behaviour of contact persons perceived by the person. According to the
ratio of contact persons using the resource sustainably and non-sustainably, the
person makes an estimate for the entire population. This factor is at present not
Modelling environmental behaviour 85
yet weighted by socially, or environmentally, oriented values. For example, people
who are not environmentally oriented overestimate the number of like-minded
people.
If no particular interventions take place, there is a danger that dynamics such
as those shown in Figure 4.7a will develop. With an optimistic starting value with
regard to average use within a population (over 50 = environmentally friendly
use), the state of the resource briefly improves. Due to its improved state, its value
declines (only goods in short supply are valuable), and the individuals in our
simulated population resume stronger exploitation of the resource. From the fifth
step onwards, this tendency results in clear over-utilization (average use under 50).
This increases the discrepancy between actual and sustainable use patterns, and
expectations of sustainable use correspondingly decrease. The individual is less
motivated to make a personal contribution to sustainable patterns of utilization
(‘… personal restraint on my own part would not make any difference; no one else
is showing restraint, so it is better for me to help myself to the resource so long as
it is still up for grabs …’). And so the state of the resource deteriorates. Its value
rises, which does not, however, lead to a marked reduction in utilization. There is
no stopping the course of these ‘downhill’ dynamics, and the resource is destroyed.
To counteract these negative dynamics, we ran a campaign in the fifth step that
aimed to (a) lead persons, in their own use behaviour, to orient themselves less to
other persons’ patterns of use and (b) lend heavier weighting to the importance of
the resource. Through this, the common property becomes more highly esteemed.
We found that the utilization behaviour of the population becomes more
environmentally sustaining (see Figure 4.7b). If individuals use a resource in an
environmentally sustainable fashion on the average, the state of the resource
improves. This line of development continues until the resource has regenerated.
At this time, the resource is available in over-abundance so to speak, whereby its
value again declines, and intensity of use increases. Downhill dynamics develop
until that crucial point where they are again brought under control. A dynamic-
stable balance has emerged, in which utilization continually adapts to the state of
the resource by means of inner personal factors. The resource will never be
completely destroyed.
Through the interaction of the social system and the resource, the resulting
system behaviour shows large fluctuations. Following Forrester (1972: 48ff.),
fluctuating system behaviour results when in a system of interlocking feedback
loops two or more temporal delays occur. In the present simulation model, there
are delays both within the social system and between the social system and the
resource system. The average expectation that one can make an effective
contribution to sustainable use of a resource reacts with a delay to the average
resource use behaviour of the social surround. The rise and fall of the average
value of the common good reacts with a delay to the development of the resource.
Our simulation approach is not directly comparable to Forrester’s approach.
Forrester observes only one macrosystem, while the simulation builds upon
numerous microsystems (individual persons) which join to form the macrosystem
(the social system).
86 Hans-Joachim Mosler
Environmentally
`´ ýP
responsible/
resource maximal
100
80
60
40
20
0
Not 0 5 10 15 20 25 30
environmentally
responsible/ resource in % Runs
resource minimal average use
average value of common resource
average expectation of sustainable use
Environmentally
responsible/
resource maximal
100
80
60
40
20
0
Not 0 5 10 15 20 25 30
environmentally Runs
Resource in %
responsible/ Intervention
resource minimal Average use
Average value of common resource
Average expectation of sustainable use
Note: Also illustrates the course of significant inner variables. 4.7a (top): control situation; 4.7b (bottom):
with intervention.
Communicator's persuasiveness
Communicator's attitude
Communicator's status
Self-responsibility
Processing
Status
motivation
Perception of state Concern
of environment
Processing intensity Attitude
Values
Persuasiveness
Processing ability
Knowledge
Let us first examine results with individual persons selected from the population.
Person A – who is biased due to an environmentally non-friendly attitude – places
a negative value even on good arguments (curve Persuasiveness+). In contrast,
good arguments effect even more environmental consciousness in Person B, who
has environmentally responsible attitudes from the start, as information processing
is hardly biased here.
A campaign using esteemed public figures is effective with both Persons A and
B, if they tend to process information relatively superficially (curve Status+). But
if Person A’s depth of processing is increased by means of self-responsibility (curve
Self-responsibility+/Status+), they then show biased processing, and their attitude
90 Hans-Joachim Mosler
Environmentally
responsible
70 70
Person A Person B C C
60 60 B B
D D
A A
50 50
40 BC
40
DA
30 B B
C 30
C
20 20
BC
10 A A 10
DA
0 D D 0
Not 0 5 Intervention 10 15 0 5 Intervention 10 15
environmentally Runs
responsible Runs
Environmentally
responsible Campaigns with:
90
Population A Campaign: persuasiveness+
80 B Campaign status+
70 C Campaign: persuasiveness+/status+
C C
B B D
60 Campaign: self-responsibility+/status+
A A
50 D D Control situation
40
30
Not
environmentally 5 Intervention 10 15
responsible Runs
Note: Above, individuals in the population. Below, population values. The three graphs present
courses of attitude change in four different experiments and the control situation.
changes in the direction away from that advocated by the campaign. In the total
population, the average attitude becomes most changed by means of good
arguments and presentation by esteemed personalities. But as shown in Figure
4.9, we must always take into consideration that these average values within the
population hide very contrasting effects upon differing individuals.
For various experiments with multiplicators (persons who advocate their environ-
mental attitudes to others), a communication net was constructed. Each simulated
person ‘talks’ about environmental issues with five friends and one stranger during
each step, whereby the friendships are assumed to be lasting and reciprocal. The
relationships to strangers are different in each step. Initial values in the population
are the same as those described above for the information campaign experiment.
Within the framework of an environmental action campaign, 500 persons
(multiplicators) having very environmentally friendly attitudes receive training in
argumentation (increasing persuasiveness, see Gonzales et al. 1988). In a variant
of this experiment, the status of these persons is raised (as for example through
public commendation awards). The following interventions were carried out:
Modelling environmental behaviour 91
• Multiplicators with good arguments (Figure 4.10, Persuasiveness+).
• Multiplicators with raised status (Figure 4.10, Status+).
• Multiplicators with good arguments and raised status (Figure 4.10,
Persuasiveness+/Status+).
• Control situation.
In Figure 4.10 we again see two selected individuals A and B, who both stand
under the influence of the action campaign. A glance shows that these graphs are
not as smooth as the information campaign graphs. The curves would be smoother
if, after a short levelling-off phase, there were only the influence of a constant
group of friends. But the ever-changing influence of strangers, upon Persons A
and B and also upon their friends, causes fluctuations in attitudes. We can still
recognize certain tendencies in the development of attitudes, however. Person A’s
attitude becomes less environmentally friendly relatively quickly. Later changes to
a more responsible attitude are only brief and are repeatedly destroyed by the
environmentally unfriendly social surround. Person B shows initial swings in
attitude, but then develops a tendency towards a less environmentally friendly
attitude. Both Persons A and B react to the introduction of multiplicators having
high status (curve Status+) with an increasingly environmentally friendly attitude.
This effect is even stronger if multiplicators in addition show high persuasiveness
(curve Persuasiveness+/Status+). But if multiplicators demonstrate only
persuasiveness (curve Persuasiveness+), they have no effect upon Person A. There
is an initial effect in this case upon Person B, but the influence of the rather
environmentally unfriendly social surround cannot be cancelled out. The same
effects find expression at the level of the population: multiplicators with raised
status are just as convincing as those who also show increased persuasiveness.
Multiplicators who have only high persuasiveness at their disposal have a counter-
productive effect.
Environmentally
responsible
80
Population
70
A Multiplicators: persuasiveness+
60
B Multiplicators: status+
50 C C C C C
B B C Multiplicators: persuasiveness+/status+
B B B
A D D D
40 D D D
A Control situation
A A A
30
20
Not
environmentally 10 20 30
responsible Runs
Intervention
Note: Individual graphs show the course of attitude change in three different experiments and the
control situation.
Part II
Introduction
The awareness that environmental problems have a large impact on the economy
and society has risen continuously during the last three decades. Lifting sustainable
development to the political agenda nurtured the understanding that policy analysis
is characterized by conflicting economic, environmental, societal, technical, and
aesthetic objectives. Consequently numerous research projects on the inter-
dependence of environment, society and economy were conducted in this period.
With increasing knowledge about the economic and environmental processes and
their interlinkages, it became clear that decisions about environmental problems
are especially characterized by uncertainties, high stakes, urgency, and dispute
(Funtowicz and Ravetz 1990). These characteristics influence the choice of the
method.
In the field of environmental decision-making, multiple criteria decision aid
(MCDA) has become a widespread tool, able to tackle problems involving more
than one objective. A number of theoretical (e.g. Munda 1995; Castells and Munda
1999) and empirical studies (e.g. Janssen 1992; Andreoli and Tellarini 2000;
Mendoza and Prabhu 2000) showed that MCDA provides a useful tool for decision
aid in this field, as it allows for multiple objectives, for the use of different types of
data and the involvement of different stakeholders. The choice of a particular
method depends on the specific problem as well as on users’ demands. As the
number of existing methods is already quite large and is still increasing, the choice
of the ‘right’ method is not an easy task.
The aim of this chapter is the comparison of frequently used multiple criteria
methods. A number of authors have compared multiple criteria algorithms with
regard to their consistency, relative accuracy, speed, etc. (e.g. Olson 2001; Salminen
et al. 1997). Since the underlying philosophies and assumptions, and hence the
applicability, vary significantly, we chose, in contrast, a less mechanistic approach
and describe the methods with reference to a comprehensive list of quality criteria.
The criteria were chosen with particular emphasis on their relevance for questions
of sustainable development. For the evaluation we apply the list of quality criteria
to seven different MCDA methods (MAUT, AHP, Evamix, ELECTRE III, Regime,
NAIADE, and programming methods) and compare them. This list of criteria
100 Andrea De Montis et al.
provides a basis to assess the usefulness of each method for a specific application.
The chapter is structured as follows. First, the list of quality criteria are presented,
then the seven methods are characterized and evaluated by use of the list of criteria.
This evaluation leads to a comparison of the methods, which is followed by
conclusions.
Table 5.1 List of quality criteria for MCDA methods (De Montis et al. 2000)
Quality criterion Description
Operational components
Criteria
Interdependence Allowance for the interdependence of different criteria
Completeness Need for the completeness of the criteria
Non-linear preferences Possibility to express non-linear valuation patterns
Weights
Transparency of process, Type of the procedure of deriving values for the weights
type of weights
Meaning Interpretation and role of weights in the evaluation process
Solution finding Type of procedure used for the comparison of alternatives
procedure
Issues addressed by Interpretation of the results generated by the use of method
results
Applicability of MCDA methods – user context
Project constraints
Costs Implementation costs in the specific user situation
Time Implementation time in the specific user situation
Structure of problem
solving process
Stakeholder participation Possibility to include more than one person as decision
maker
Problems structuring Existence of mechanisms supporting the structuring of the
problem
Tool for learning Support of learning processes
Transparency Promotion of transparency in the decision making process
Actor communication Support of the communication between opposing parties
Applicability of MCDA methods – problem structure
Indicator characteristics
Geographical scale Applicability of different geographical scales for one case
Micro–macro link Applicability of different institutional scales for one case
Societal/technical issues Possibility for the consideration of both societal and
technical issues
Methods combination Possibility of methods combination
Data situation
Type of data Type of data supported as values for the indicators
Risk/uncertainties Possibilities for the consideration of evaluation risk and/or
uncertainties
Data processing amount Processing amount needed to compile the data required for
the method
Non-substitutability Possibility to consider sustainability standards and non-
substitutability
102 Andrea De Montis et al.
(a) methods based on the use of a single synthesizing criterion without
incomparability, (b) methods based on the synthesis by outranking with
incomparability, and (c) methods based on interactive local judgements with trial-
and-error iteration. While the first two groups embody a clear mathematical
structure, the third does not use any formalized or automatic procedure. Another
possible division refers to two groups of multiple criteria methods, namely discrete
and continuous, depending on whether the set of alternatives is finite or infinite
(Voogd 1983). In the following we will first present two types of discrete methods
(single synthesizing criterion and outranking methods) and then the continuous
methods (programming methods).
Step 1 comprises the definition of the alternatives, the objectives, the attributes
and the outcomes of each alternative in each attribute. This step is not specific to
MAUT; this is how all discrete MCDA methods start.
Step 2 concerns risk viewed as subjective probability. If there is risk concerning
the impact of the actions on the attributes, probabilities pij can be assigned to the
outcomes x1, y1, … of the alternatives on the attributes. It is assumed that the value
of the action a for attribute i is given by the expected value of ui (Vincke 1985;
Dillon and Perry 1977): ui(xi ) = Sj pijui(xij ), where pij is the probability of the j-th
outcome in the i-th attribute or in continuous terms, and ui(xi ) = Úui(xi ) ¶i (xi )dxi,
where ¶i (xi ) is the probability distribution of outcomes in the i-th attribute.
There are different methods to elicit subjective probability distributions, which
are not without difficulties. Which method to choose depends on the problem and
often on the experience of the researchers. Direct involvement methods are simple
and allow the decision-makers to see and understand the built distribution. For
these reasons, such direct involvement methods seem preferable to others, such as
lottery-type questions (Dillon and Perry 1977). The simplest method of all is asking
for a few point estimates of probability; this is crude but practicable.
Step 3 requires the elicitation of the decision-maker’s utility function ui(xi ) for
the range of outcomes on each attribute. A utility function ui(xi) is developed for
104 Andrea De Montis et al.
each attribute. This is done by asking the decision-maker a series of questions
based on the axiom of continuity (e.g. Dillon and Perry 1977: 10). Additionally
weights wi are assigned to the attributes, presenting trade-offs, by asking for the
decision-maker’s order of importance of the attributes. The assessment of the
appropriate utility function is complex and intricate. Thus it is a crucial step in
MAUT. This procedure leads us to three difficult questions (Vincke 1985): (a) what
must the properties of the decision-maker’s preferences be so that U is a certain
function of ui ?; (b) how to test the properties?; and (c) how to construct the function
U? The method can only be applied if these questions can be answered in a
meaningful way.
After the establishment of the utility functions ui (xi ) and the elicitation of the
attribute probability distributions, the expected utility is aggregated across the
attribute distributions for each alternative in Step 4. For each attribute a function
ui is built and the functions ui are aggregated in a global criterion U, such that
U(u1(x1),…, un(xn )) > U(u1(y1),…,un (yn )), if the action represented by (x1,x2,…,xn ) is
better than the action represented by (y1,y2,…,yn ), when considering all the attributes
simultaneously (Vincke 1985; Roy and Bouyssou 1985). In that way the multiple
criteria problem is reduced to a single-criterion decision problem. In order to build
the aggregate function, different aggregation procedures are possible; the additive
or the multiplicative aggregations are most widely applied. The additive form is
the simplest form of a utility function, requiring two strong assumptions, namely
utility independence and preferential independence of any subset of criteria (Fandel
and Spronk 1985).2 SMART (simple multiple attribute rating technique) presents
one method to build the additive form (Martel, in Climaco 1997). If the additive
model holds, the expected utility of each attribute for each alternative is added.
Each attribute must get a weighting factor wi to get a common scale for the utility
functions ui(xi ). The utility for a certain alternative with uncertain consequences is
then given by: U(x) = Swi [ui (xi )], where in discrete terms, ui (xi ) = Sj pij ui (xij ), pij being
the probability of the j-th outcome in the i-th attribute; or in continuous terms,
ui(xi ) = Úui (xi ) ¶i (xi )dxi , where ¶i (xi ) is the probability distribution of outcomes in the
i-th attribute. If independence does not hold, the multiplicative form is usually
applied. Another form is the quasi-additive model, which can be used if neither
marginality nor preferential independence prevails; but this procedure becomes
very complicated for cases with more than three attributes (Martel, in Climaco
1997). Step 5 consists of summarizing the results and interpreting them.
A sub-methodology of MAUT is multiple attribute value theory (MAVT). The
two approaches deal with risk in different ways. While MAUT relies upon a utility
function, which allows the comparison of risky outcomes through the computation
of an expected utility, MAVT is unable to take risk into account. It uses a value
function to represent the outcome of the alternatives, not allowing for risky
outcomes. Value functions preserve deterministic ordering, whereas utility functions
preserve the expected utility hypothesis. MAUT is suited for ex-ante evaluations
for multiple objective problems with risky outcomes as long as one accepts the
assumptions on which the method is based.
Assessing the quality of different MCDA methods 105
MAVT is based on four key assumptions. First, a complete system of preferences
pre-exists, which is an adequate representation of the axioms in the decision-maker’s
mind. Second, the preference relation of the decision-maker is of a weak order,
which means that all states are comparable, transitivity of the preferences and
transitivity of the indifferences is given. The last one is particularly difficult as it
can lead to a lack of realism; see for example the famous cup-of-coffee example by
Luce (1956).3 Third, the main hypothesis underlying MAUT is that in any decision
problem there exists a real-valued function U defined on A (the set of actions),
which will be examined. This function aggregates the utility functions ui (Vincke
1985). Fourth, in the probabilistic case, when there is risk concerning the outcomes,
the axioms of von Neumann and Morgenstern (1947) must hold. In this way the
concept of rationality comes into MAUT, which can be seen as an extension of
ordinary subjective expected utility procedures to the case of choice between multi-
attribute alternatives. They imply an existing unique set of subjective probability
distributions for outcomes in each of the attribute dimensions associated with any
risk multi-objective alternative and a utility function whose expected value gives a
measure of the decision-maker’s preference for each of the risky multi-attribute
alternatives. If the axioms are accepted, there exists both ‘a unique set of subjective
probability distributions for outcomes in each of the attribute dimensions associated
with any risky alternative and a utility function U(xi) whose expected value (in
terms of the decision-maker’s subjective probabilities) gives a measure of the
decision-maker’s preference for each of the risky multi-attribute alternatives’ (Dillon
and Perry 1977: 9).
The development of MAUT was a milestone in the history of MCDA, as it
presented a formal and transparent method able to deal with multiple objectives,
risk concerning the consequences, the pattern of consequences over time, and a
start for using qualitative data. But MAUT’s main problem is the strong assumptions
that need to be fulfilled to derive the global utility function. Basically, MAUT
relies upon the same axioms as social welfare theory like CBA does and therefore
a number of the points of critique towards CBA also apply to MAUT. A review of
case studies shows that MAUT is used frequently for decision-making for economic,
financial, and actuarial problems as well as problems concerning environmental
issues, like water management, energy management, agriculture, but not really in
the field of sustainable development, which is much broader as it covers an
economic, a social, and an environmental dimension.
Concerning the first group of quality criteria (see Tables 5.1 and 5.2) for MCDA
methods, interdependence of attributes is not allowed; on the contrary the
independence assumption is crucial. This assumption makes it hard for use in
complex decision problems, as is the case for problems in the field of sustainable
development. The meaning and assignment of weights is clear (weights stand for
the importance of criteria) and transparent. The solution is a complete ranking of
actions, according to their expected global utility. This again is clear and transparent,
but the derivation of the utility function is the most delicate part of the whole
procedure. Different issues concerning the type of decision problem are allowed,
as long as they can be addressed by attributes; a utility function is designed for
106 Andrea De Montis et al.
each attribute. Thus the more issues are addressed the more time it takes and the
higher are the costs.
For most decisions relating to sustainable development, stakeholder participation
is necessary and desirable. Stakeholders are normally involved twice during the
decision-making process. First, when developing the utility function for each attribute,
they have to elicit their preferences. Second, when calculating the expected utility,
they have to assign probabilities to certain outcomes. Preferences are directly assessed
and generally aggregated and used to develop the utility functions. Under MAUT,
transparency of the process is given as well as a clear structuring. Concerning the
indicators and data, MAUT can handle different scales as well as different issues.
The required amount of data is usually quite high. The data itself can be of qualitative
or quantitative character. Risks of the outcome are accounted for by the expected
utility theory, while uncertainties are checked with a sensitivity analysis.
In sum, MAUT ensures in risky project choices the correspondence of the choice
with the decision-maker’s preferences and it is a mechanism for doing this in cases
too complex to be handled satisfactorily by intuition, as long as the underlying
assumptions are known and always in the minds of the researchers. Information
about the preferences of the decision-maker are necessary to build the singular
utility functions for each attribute and to build the aggregate utility function based
on all attributes. These individual utility functions build the base for the decision.
No social function is constructed, thus no interlinkages or interdependencies
between the individuals are taken into account. The building of a group or social
function is a problem in every MCDA method, as we learned from Arrow’s
impossibility theorem that it is not possible to aggregate individual preference
functions to group opinions. The individual preferences and estimations can be
seen as an acceptable proxy as long as the decision projects stay on a rather micro-
related level, like the decision about an irrigation system. But if we have to deal
with macro-related problems, additional influences gain importance, such as the
impact on the different dimensions of sustainable development or on different
groups in society. This decreases the appropriateness of individual utility functions
as an acceptable proxy for social preferences. Thus MAUT, like CBA, is hardly
suited for macro-related problems.
where:
1
aii = 1 a ji = aij ≠ 0
aij
l max - n
C .I . =
n -1
Assessing the quality of different MCDA methods 109
Saaty proposes a ‘consistency ratio’ in order to appreciate the consistency of the
matrix which is calculated by dividing the C.I. by R.I., where R.I. represents an
experimental ‘random index’ which increases as the order of the matrix increases.5
A consistency ratio of 0.10 or less is considered acceptable; if this ratio is more
than 0.10, it is necessary to reformulate the judgements by means of new pairwise
comparisons.
The AHP method can take into account the interdependence existing between
the different evaluation criteria by means of the construction of suitable hierarchies.
The elements of a hierarchy (i.e. criteria, sub-criteria, alternatives) are not linked
to all the others but they are grouped into disjoint sets; if an interdependence
between some elements is recognized, they are placed in the same hierarchical
level and connected to the same element of the next higher level. If the choice of
the number of hierarchical levels is unlimited (although, as a rule of thumb, no
more than seven elements for each node is best) it is possible to insert in the hierarchy
all the criteria required by the decision-making problem. In this sense a complete
(i.e. exhaustive) list of evaluation criteria is allowed. A cardinal weight is assigned
to each element of the hierarchy by means of opportune pairwise comparisons.
The weights express the importance of one element over the others. The final
result offered by the method is a complete ranking of the evaluated alternatives.
AHP facilitates stakeholder participation. Different social actors are invited to
assign weights to the evaluation criteria; in this perspective weights reflect different
social views. In comparing elements in pairs the different judgements given by
each actor can also be integrated – by means of an arithmetic average – obtaining
only one weight for each criterion, which expresses synthetically the points of
view of all the involved stakeholders. It is also possible to promote the actors’
communication if they are called to work together in order to identify general
objectives, evaluation criteria, sub-criteria, and deduce their importance.
The method has been applied to a range of decision-making problems (e.g.
corporate policy and strategy, public policy, political strategy, environmental
planning) at different geographical scales. Sometimes also using other evaluation
methods it is possible to make a pairwise comparison between criteria as carried
out in the AHP if normalized weights are required. In this sense a combination
with other methods is possible. AHP allows use of qualitative and quantitative
data; in particular, cardinal numbers can be directly normalized without the need
of pairwise comparisons.
The set of criteria j is divided into two subsets, denoted O and C, where O is the
set of the ordinal (qualitative) criteria and C the set of cardinal (quantitative) criteria,
obtaining two distinct evaluation matrices: EO (ordinal criteria/alternatives) and
EC (cardinal criteria/alternatives). This way the differences among alternatives
can be expressed by means of two dominance measures: the first one based on
ordinal criteria and the second one based on cardinal criteria. In particular, in
order to construct the cardinal dominance score, the components ej i Œ EC are
standardized to a common unit; in this way all the quantitative criteria are expressed
on a scale from 0 to 1. All the standardized scores should have the same direction;
this means that a higher score implies a better score. In the case of criteria for
which lower scores mean better scores, the standardized scores are transformed by
subtracting them from 1.
Additionally, a qualitative weight can be assigned to each criterion constructing
a vector wj (j = 1, 2, …, m); the associated cardinal weights wj can be obtained – in
an approximated way – by looking at the ‘extreme weights sets’ which delimit the
values that metric weights may have.7 Because weights represent subjective value
judgements, it is preferable to use different priority sets so that the consequences
of different political viewpoints can be illustrated.
Step 2 requires calculation of dominance scores for all ordinal and cardinal
criteria. Differences between alternatives are expressed by means of two dominance
scores, for ordinal (aii¢) and cardinal (aii¢) criteria:
(
a ii ¢ = f e ji , e ji ¢ , w j ) ("j ŒO )
(
aii ¢ = g e ji , e ji ¢ , w j ) (" j ŒC )
where eji ( e ) represents the score of criterion j and alternative i ( i ¢ ), w j represents
ji ¢
the weight attached to criterion j, and f and g are two different functions. If each
extreme (quantitative) weight set w is substituted into the above formulae instead
of the qualitative weights w j , it is possible to define – for each extreme weight set
– a new dominance measure for ordinal ( a ii ¢ ) and cardinal ( aii ¢ ) criteria:
Assessing the quality of different MCDA methods 111
(
a ii ¢ = f e ji , e ji ¢ , w j ) ("j ŒO )
(
aii ¢ = g e ji , e ji ¢ , w j ) ( "j Œ C )
The scores expressed by means of these formulae reflect the degree in which
alternative i dominates alternative i¢ for ordinal and cardinal criteria respectively.
Step 3 calculates standardized dominance scores for all ordinal and cardinal
criteria. The dominance scores a ii ¢ and aii ¢ are standardized into the same
measurement unit in order to make them comparable. If h is a standardization
function, the ‘standardized dominance measures’ for all ordinal (d ii ¢ ) and cardinal
(dii ¢ ) criteria have the following expression:
The standardization of the scores α ii ' and a ii ' can be done in three different
ways: the ‘subtractive summation technique’, the ‘subtracted shifted interval
technique’ and the ‘additive interval technique’.8 Also in this case the standardized
dominance scores reflect the degree in which alternative i dominates alternative
i ¢ for ordinal and cardinal criteria. Different scores are obtained using the three
different techniques.
Step 4 calculates overall dominance scores. The ‘overall dominance measure’
mii ' for each pair of alternatives (i, i ¢ ) – giving the degree in which solution i
dominates solution i ¢ – is calculated by means of the following formula:
where wO represents the weight of the qualitative criterion set O and wC
represents the weight of the quantitative criterion set C:
wO = ∑ w j wC = ∑ w j
j∈O j∈C
Step 5 calculates appraisal scores. The overall dominance measure mii ¢ may also
be considered as a function k of the ‘appraisal scores’ si and si ¢ of the alternatives
i and i ¢ :
At the same time this means that the appraisal score of an alternative depends
on the overall dominance score such as calculated in Step 4. In particular, three
different appraisal scores of alternative i with respect to the n other alternatives
are obtained if the subtractive summation technique, subtracted shifted interval
technique or additive interval technique has been used in Step 3. Additionally, by
112 Andrea De Montis et al.
using the appraisal score of all three techniques (t = 1, 2, 3) a standardized ‘average
appraisal score’ ai for alternative i can be calculated:
3
Ê s - s- ˆ
ail = Â Á ti+ ti- ˜
t =1 Ë sti - sti ¯
where sti is the appraisal score of alternative i for each technique; sti- is the lowest
sti -score for each technique; sti+ is the highest sti -score for each technique. The
result is a complete ranking of alternatives and the higher ai is the better is
alternative i.
Since the starting point of Evamix is the construction of an evaluation matrix,
it is not possible with this method to take the interdependence between the different
evaluation criteria into account. But if we consider that there is no limit on the
choice of the number of criteria to insert in the evaluation matrix, it is possible to
consider all the criteria required by the decision-making problem. In this sense a
complete (i.e. exhaustive) list of evaluation criteria is allowed. An ordinal weight,
expressing the importance between all the relevant criteria, is assigned to each
criterion, and subsequently ordinal weights are transformed into cardinal ones.
Evamix supports stakeholder participation. Different social actors are invited
to assign weights to the evaluation criteria, and in this case the different weights
given by each actor are not integrated but they are used to show the different
points of view (i.e. economic, social, environmental) expressed by the stakeholders
involved. It is also possible to promote actor communication if they are called to
work together in order to identify general objectives, points of view, evaluation
criteria, sub-criteria, and deduce their importance. The method is used to cope
with different decision-making problems at different geographical scales. Most
frequently, Evamix been used in urban and regional planning. Evamix allows using
qualitative and quantitative data. In particular, quantitative data are expressed in
an ordinal scale and the cardinal numbers are standardized in a scale from 0 to 1.
Outranking methods
The development of the outranking methods was started in France in the late
1960s by Bernard Roy and his team (Roy 1985). Outranking methods were
developed in an attempt to manage with less strong assumptions (about existence
of utility function, additivity, etc.) and to require less information from decision-
makers (especially preference intensities, rates of substitution) than the methods
described above.
The outranking methods also differ fundamentally from other methods in the
approach insofar as they start from the assumption that the decision is a process
during which decision-makers may change their preferences in response to the
information provided and/or after thorough reflection about the problem. For
this reason the developers of outranking methods considered it important to allow
at the beginning for incomparability of options. This avoids a complete ranking
being identified too early; the aim is therefore to stimulate thorough consideration
Assessing the quality of different MCDA methods 113
and to avoid premature conclusions. Hence, more than other methods, these
methods encourage interaction between the model and the decision-maker/s. The
outranking algorithms also account for the political reality that options which
perform very poorly in one dimension are likely to face severe opposition from at
least some of the stakeholders and are therefore deemed unacceptable.
ELECTRE III
Roy (1985) developed one of the most widely used software packages, known as
ELECTRE, which is available in four different main releases (ELECTRE I, II,
III, IV) and two other versions (ELECTRE IS and TRI). A comprehensive
comparison of these approaches can be found in Vincke (1992). The underlying
principle of the ELECTRE approach stems from processing a system of pairwise
comparisons of the alternatives. This procedure involves the use of a system of
preference relations (SPR) that is based on the exploitation of the outranking
relation. According to this relation, an action will outrank another if the first action
is considered to be at least as good as the second. Roy (1996) defines a system like
this as a basic system of outranking relations. Recently it was pointed out that
release III is remarkable among the ELECTRE family, since it takes explicitly into
account the indifference and preference thresholds but, furthermore, it is based on
an outranking relation, which is less sensible to the variation of the input data
(Scarelli 1997). The system solves ranking problems, dividing the set A of actions
into equivalence classes (Scarelli 1997). In release III of ELECTRE, the process
can be sketched in the following steps.
A cardinal impact table is prepared for a finite number of alternatives ‘measured’
according to a set of pseudo-criteria. The latter are criteria whose SPR includes
weak preference, indifference and incomparability, and a coherent system of
thresholds has to be set. The program allows to set thresholds for indifference and
for preference, according to a linear relationship: T = a • g(a) + b, where T is the
total threshold, a is a proportional threshold, b is a fixed threshold and g is the
criterion function. This threshold is defined directly when it refers to the maximum
value of the criterion, and indirectly when it refers to the minimum value. This
allows constructing fuzzy outranking relations, given that it is possible to state how
much an action outranks another. In fact, a series of pairwise comparisons is set
through the use of concordance and discordance indices. A general framework for
the calculation of these indices is given by Voogd (1983). The degree of dominance
of alternative i over alternative i¢ is measured by the following index:
1/ a
È Â j Œy w aj ˘
cii ¢ = Í ˙
Í Â jwj ˙
a
Î ˚
where y is the set of criteria that indicate a preference for the alternative a with
respect to the alternative a ¢ , w the criterion weight and a a scaling parameter
greater than one. ‘By means of this scaling parameter the analyst is able to vary
114 Andrea De Montis et al.
the importance of the small weights and small divergences between the scores’
(Voogd 1983: 123). In many applications, the value of the scaling parameter is set
equal to one, which allows managing a simpler functional algorithm.
The degree of discordance between the same alternatives is measured by the
following index:
a 1/ a
dii ¢ =
È
Í (
 j Œj w j | e ji - e ji ¢ | ˘˙ )
Í w | e - e ji ¢ | ˙˙
a
(
ÎÍ Â j j ji )
˚
where j is the set of criteria that do not agree about the preference of alternative
i with respect to alternative i ¢ .
In the generalized concordance analysis the values of the concordance and
discordance indices individuate the subsets of alternatives that are accepted and
rejected. The first group consists of the alternatives that fulfil the following
requirements: Tc e ≥ h and Td ≥ m.
In ELECTRE III, the concordance and discordance indices are calculated with
reference to the values that the criterion functions take for the consequences of
each alternative. In the same pattern, the threshold functions are modified as linear
functions of the pseudo-criterion g. For each criterion gj a preference threshold pj
and an indifference threshold qj are set.
Given the pseudo-criterion g and two variable alternatives i and i ¢ , two
concordance indices cj (i, i ¢ ) are calculated for the comparison of alternative i with
respect to alternative i ¢ and vice versa. For the comparison between alternative i
and i ¢ , the concordance index takes values between 0 and 1.
The discordance index dj (i, i ¢ ) is calculated with reference also to the veto
threshold vj . This threshold is associated with a criterion that is considered to be
very important. If the difference between two functional values in the case of a
number of alternatives is greater than the threshold value, this condition acts as a
veto to the affirmation: the first alternative outranks the second.
The construction of the outranking relation is based upon the credibility index
s(i, i ¢ ), which ranges between 0 and 1 and is a function of the concordance and of
the discordance index. The credibility index takes the following values:
1 - d j (i , i ¢ )
s (i , i ¢ ) = c j (i , i ¢ ) ’
1 - c j (i , i ¢ )
The credibility index provides the information for the construction of two
rankings of the alternatives. First, from the worst to the best (ascending distillation)
and the second from the best to the worst (descending distillation). The intersection
of these distillations yields the final distillation graph. This graph indicates the
Assessing the quality of different MCDA methods 115
system of relationships among the alternatives, according to a system of oriented
arcs. This framework reveals incomparability relationships that the other distillations
do not consider.
Roy (1996) considers that in every decision-making situation, four problem areas
have to be taken into account: Pa (choice), Pb (sort), Pg(rank), and Pd (description).
With reference to this outline set forth by Roy (1996), the procedure does not
deliver a complete ranking of the evaluated alternatives, leaving partially unresolved
what Roy (1985) calls the problematic P. g (ranking of the alternatives). On the
other hand, the procedure of ELECTRE III is helpful for what Roy (1985) names
the problematic P.a (choosing the best action).
ELECTRE III is based on the assumption that the system of individual
preference relationships can be described by means of a sophisticated analysis of
the relationships of preference, indifference, and incomparability. As a matter of
fact, incomparability enters as a puzzling factor into the final ranking. On the one
hand, this complex analysis involving the assessment of several thresholds and
preference relationships may be judged as a remarkable effort. On the other hand,
it may lead to cumbersome work. Often it is very difficult to express credible
threshold values and the resulting ranking can end up being hardly understandable.
As a matter of fact, in the situation where results were described by a Kernel
graph, the complex meaning stemming from this kind of representation would be
hardly transmissible to a group of laypersons, such as a local community. Local
communities seem to be more keen to comprehend and discuss definitive positions
and comparable outcomes than to give personal interpretation of relationships.
The nature of the ranking, which is presented by a graph, allows also for incompar-
ability among the alternatives: such an incompletely resolved outcome challenges
the analyst who should then, if equipped with abilities of good advocacy, foster
communication among stakeholders to discuss the incomparability. Politicians often
are unwilling to accept the complexity of this structure and show a tendency to
sceptical behaviour against an incomprehensible technique for decision-making.
Often it is the expectation that the multiple criteria analysis will solve the problem
at hand completely. The obvious fact that MCDA supports decision-making, but
must not substitute decision-makers, becomes even more apparent with outranking
methods than with the previously discussed ones. These are the main reasons why
in Table 5.2 ELECTRE is rated highly with respect to criteria completeness and
interdependence, while it is rated poorly with respect to transparency and to
applicability of criteria. It should be noted that in terms of ‘meaning of weights’
ELECTRE received a zero score, since it does imply weighting procedures con-
nected to the combinatory use of the so-called ‘weights’ and to the thresholds.
Regime
The Regime method was developed by Hinloopen et al. (1983), assessed and refined
by Hinloopen (1985) and Hinloopen and Nijkamp (1990). This method can be
considered as a multiple criteria evaluation framework belonging to the family of
dominance analysis approaches, even though it has been pointed out that it is
116 Andrea De Montis et al.
quite distinct from other methods in this family, mainly because Regime is a
qualitative multiple criteria method. Qualitative multiple criteria methods aim to
provide a tool for analysing complex situations, which are impossible to model by
means of quantitative information. Hinloopen and Nijkamp (1990) list several
such methods: (a) the extreme expected value method (Kmietowicz and Pearman
1981; Rietveld 1980), (b) the permutation method (Mastenbroek and Paelink 1977),
(c) the frequency method (Van Delft and Nijkamp 1977), (d) the multidimensional
scaling method (Nijkamp and Voodg 1981), and (e) the mixed data method (Voogd
1983).
Since human activities and settlements are characterized by phenomena that
are often impossible to measure, studies have focused on the construction of
techniques able to deal with ‘soft’ and ‘fuzzy’ information. ‘These qualitative multi-
dimensional judgement methods are essentially more in agreement with “satisfying
behaviour” (based on Simon’s bounded rationality principle) than with conventional
“optimising decision making”, as imprecise information precludes the attainment
of an unambiguous maximum solution’ (Hinloopen and Nijkamp 1990: 38).
As in Evamix, an evaluation table is given and composed by eij scores of a
number n of alternative scenarios with respect to m criteria. In the case of ordinal
information the weight can be represented by means of rank orders wj in a weight
vector w: w = (w1, w2, …, wj )T. The higher the value of the weight, the better the
correspondent criterion.
The following main features characterize the multiple choice regime method.
First, with this method the evaluation table can contain cardinal as well as ordinal
data; this is accomplished by treating cardinal information as ordinal, with reference
to the ranking position of each alternative. Second, the basis of the method is the
regime vector. For two alternative choice options i and i ¢ , the difference of the
criterion scores is assessed by considering the difference between the scores sii¢j = eij
– eei¢j.
In case of ordinal information, the order of magnitude of sii¢j is not relevant,
but only its sign qii¢j . Therefore, if qii¢j = +, then alternative i is better than alternative
i¢ for criterion j; i¢ qii¢j = –, then alternative i is better than alternative i¢ for criterion
j. The extension of the pairwise comparison to all the alternatives leads to the
following J ¥ 1 regime vector: rii¢ = (sii¢1,sii¢2, sii¢3, … sii¢n ). The regime vector,
composed of + or – signs, or eventually zeros, ‘reflects a certain degree of (pairwise)
dominance of choice option i with respect to i¢ for the unweighted effects for all J
judgement criteria’ (Hinloopen and Nijkamp 1990: 41). For all the combinations
of comparisons, a regime matrix is drawn with the following pattern:
n ii ¢ = Â s ii ¢j w*j
j
1
pi = Â pii ¢
I - 1 i ¢π i
+• +•
Ú
-•
f ( x )dx = Ú g ( y )dy = 1
-•
Assessing the quality of different MCDA methods 119
The semantic distance between the two fuzzy sets (that is the distance between
all points of their membership functions) is computed. The same distance concept
can be extended to stochastic measures considering f(x) and g (y) probability density
functions of the measures.
The pairwise comparison of alternatives is based on six ‘preference relations’:
expressed for each criterion starting from the distance between alternatives. The
above preference relations are analytically defined by means of six functions that
express (for each criterion on the base of the distance between alternatives) an
index of credibility of the statements that an alternative is ‘much greater’, ‘greater’,
approximately equal’, ‘very equal’, ‘less’ or ‘much less’ than another.11 This
credibility index goes, increasing monotonically, from 0 (definitely non-credible)
to 1 (definitely credible). In particular, given a criterion j and a pair of alternatives
i and i¢, it is possible to define six membership functions (different for crisp criterion
scores and for stochastic or the fuzzy criterion scores), which are respectively
named12:
m * (i , i ¢ ) = f ( m * (i , i ¢ ) j )
Ïm >> (i , i ¢ ) H ( >> )
Ôm ( i , i ¢ ) H (>)
Ô >
Ôm @ (i , i ¢ ) H (@)
Ì
Ôm = ( i , i ¢ ) H (=)
Ôm < ( i , i ¢ ) H (<)
Ô
Óm << (i , i ¢ ) H ( << )
where m* (i , i ¢ ) is the overall evaluation of a given fuzzy relation for each pair of
actions and H(*) is the associated entropy level.14
Step 3 is the evaluation of alternatives. The evaluation of the alternatives derives
from the information provided by the above aggregate fuzzy preference relation,
defining for each action two functions f+ and f –, from which separate rankings of
alternatives are obtained. The function f+(i) is based on the ‘better’ and ‘much
better’ preference relations and with a value from 0 to 1 indicates how the alternative
i is ‘better’ then all other alternatives. The second function f–(i) is based on the
‘worse’ and ‘much worse’ preference relations, its value going from 0 to 1 which
indicates how the alternative i is ‘worse’ than all other alternatives (Menegolo and
Guimarães Pereira 1996). The final ranking of alternatives comes from the
intersection of two separate rankings obtained by means of the functions f+(i) and
f-(i), taking into account that it can also be an incomplete ranking because non-
dominated alternatives are calculated.
Additionally, NAIADE allows for another type of evaluation. It analyses conflicts
between different interest groups and the possible formation of coalitions according
to the proposed alternative options. Besides the ‘impact matrix’, also an ‘equity
matrix’ is constructed, which contains linguistic evaluations of different social groups
for each alternative. In particular, ‘equity analysis is performed by the completion
of an equity matrix from which a similarity matrix is calculated. Through a
mathematical reduction algorithm, it is possible to build a dendrogram of coalitions
which shows possible coalition formation, and a level of conflict among the interest
groups’ (Menegolo and Guimarães Pereira 1996: 1).
Because the starting point of NAIADE is the construction of an evaluation
matrix, it is not possible with this method to take into account the interdependence
existing between some different evaluation criteria. Since there is no limit on the
choice of the number of criteria that can be inserted in the evaluation matrix, it is
possible to consider all the criteria required by the decision-making problem. In
this sense a complete (i.e. exhaustive) list of evaluation criteria is allowed. No weights
are assigned to the criteria. The final result offered by the method can be an
incomplete ranking of the evaluated alternatives. In fact, the final ranking comes
from the intersection of two separate rankings which allows to calculate also the
non-dominated alternatives.
Stakeholder participation is explicitly supported in NAIADE. The method allows
the construction of an equity matrix which reflects how the different social groups
Assessing the quality of different MCDA methods 121
involved evaluate the alternatives. It is also possible to promote the actors
communication if stakeholders are called to discuss their evaluation of the
alternatives. The method is used to cope with different decision-making problems
(it has especially been used for problems of unsustainability) at different
geographical scales. NAIADE allows using qualitative and quantitative data. In
particular, qualitative data are expressed by means of linguistic evaluations and
quantitative data may be expressed either in crisp, stochastic or fuzzy numbers.
Programming methods
Multi-Objective-Programming (MOP), Goal Programming (GP) and their variants
are continuous multiple criteria methods. In contrast to the methods presented so
far, the programming methods do not rank or sort a finite number of alternatives,
but the alternatives are generated during the solution process on the basis of a
mathematical model formulation.
The roots of GP lie in the work by Charnes and Cooper in the early 1960s.
After the mid-1970s it became widely applied due to seminal work by Lee (1972)
and Ignizio (1976). A number of specific software tools were developed, but often
general programming software packages (like GAMS) are applied. From a
procedural point of view MOP consists of two steps: (1) identification of Pareto
efficient (also called non-dominated) solutions; and (2) identification of the most
preferred alternative together with the decision-maker(s).
Step 1 consists of finding the non-dominated solutions. The problem is formul-
ated as a task of simultaneous maximization/minimization of several objectives
subject to a set of constraints. In a so-called criterion space it can be defined as
follows: ‘max’ q, s.t. q Œ Q, where Q is a so-called feasible region in the criterion
space. The set Q is not specified directly, but by means of decision variables as
usually done in single optimization problems: ‘max’ q = f(x) = (f1(x),...,fk(x)), s.t. x Œ
X, where X is a feasible set. The functions fi, i = 1, 2, ..., k are objective functions
(Korhonen 2001).
Since the simultaneous optimization of all objectives is impossible – given a
certain level of conflict between them in most real problems – MOP tries to find
the set of Pareto efficient solutions. According to Ballestero and Romero (1998) a
necessary condition to guarantee the rationality of any solution to an MCDA
problem is the following: ‘A set of solutions in a MCDA problem is Pareto efficient
(also called non-dominated), if their elements are feasible solutions such that no
other feasible solution can achieve the same or better performance for all the criteria
being strictly better for at least one criterion’ (p.7). The goal is therefore to divide
the feasible set into a subset of Pareto efficient solutions and a subset of Pareto
inefficient solutions.
Step 2 requires choosing the most preferred solution. Any choice from among
the set of efficient solutions is an acceptable and ‘reasonable’ solution, as long as
we have no additional information about the decision-maker’s preference structure.
The final solution q Œ Q is called ‘the most preferred solution’. It is a solution
preferred by the decision-maker to all other solutions (Korhonen 2001). Since we
122 Andrea De Montis et al.
are always facing a number of conflicting goals in multiple criteria problems, a
solution can never be optimal (or ideal), but rather a compromise. With regards to
how much we need to know about the decision-maker’s preferences, different
approaches were developed: there are (a) those that need knowledge of the decision-
maker’s preference; (b) those in which the decision-maker’s preferences are pre-
emptively determined; and (c) those that progressively reveal the preferences of
the decision-maker through man–machine interactions (interactive methods) (Kalu
1999). The latter typically operate with decision-maker’s aspiration levels regarding
the objectives on the feasible region. The aspiration levels are projected via
minimizing so-called achievement scalarizing functions (Wierzbicki 1980). No
specific behavioural assumptions, e.g. transitivity, are necessary (Korhonen 2000).
At the beginning of MOP research the expectation was that computing all
efficient extreme points and then asking the decision-makers to examine their
criterion vectors to select the best one could solve the problem. However, after
vector-maximum codes became available, it was found that the number of efficient
extreme points generated by MOP was far larger than had been expected. As a
result, interactive procedures for exploring efficient sets for the best solution became
popular. Many different non-reference point and reference point procedures are
available now (Steuer and Gardiner 1990).
In GP the impossibility of building a reliable mathematical representation of
the decision-maker’s preferences due to conflicts of interest and incompleteness
of available information is addressed in a different way. Two major subsets of GP
models can be distinguished. In the first type it is assumed that within this kind of
decision environment, decision-makers attempt to achieve a set of relevant goals
as close as possible to the set of targets established (Ballestero and Romero 1998).
The algebraic representation is given as the sum of unwanted deviations from
target values if a number of objectives are minimized (weighted GP):
k
min z (ui ni + vi pi ), s.t. f i (x ) + ni - pi = bi , i = 1…Q , x ŒC5 ,
i =1
where fi(x) is a linear function (objective) of x, and bi is the target value for that
objective. ni and pi represent the negative and positive deviations from this target
value. ui and vi are the respective positive weights attached to these deviations in
the achievement function z (Tamiz et al. 1998: 570). The methodology rests on the
following basic scheme (Vincke 1992): (1) set the values one wishes to attain on
each criterion (the objectives); (2) assign priorities (weights) to these objectives; (3)
define (positive or negative) deviations with respect to these objectives; (4) minimize
the weighted sum of these deviations; and (5) perform a sensitivity analysis. In
practice, the weights are sometimes derived from the pairwise comparison of AHP.
Alternatively, they can be derived by use of an interactive MCDA method, like the
one by Zionts and Wallenius (1976) (as shown in Lara and Romero 1992). Variants
of this type (Chebyshev GP, MinMax GP, and Fuzzy GP) seek to minimize the
maximum deviation from amongst the set of unmet objectives instead. In the case
of another variant (Preference-Function GP) the relationship between the deviation
Assessing the quality of different MCDA methods 123
from the target and the penalty imposed is not the standard linear one, but it may
be piecewise linear, discontinuous, or even non-linear (Tamiz and Jones 1997).
In the other major subset of GP the deviational variables are assigned to a
number of priority levels and minimized in a lexicographic sense (lexicographic
GP). The assumption of non-continuity of preferences implies the impossibility
of ordering the decision-maker’s preferences by a monotonic utility function.
Lexicographic minimization means a sequential minimization of each priority
while maintaining the minimal values achieved by all higher priority level
minimizations. In mathematical terms this means:
Lex min a = (g1(n,p), g2(n,p), ... , gL(n,p)), s.t. fi(x) + ni – pi = bi, i = 1, ..., Q
The given model has L priority levels, and Q objectives. a is an ordered vector of
these L priority levels. ni and pi are deviational variables which represent the under-
and over-achievement of the i-th goal, respectively. x is the set of decision variables
to be determined. The g (within a priority level) function is given by
where u and v represent inter-priority level weights (Tamiz et al. 1998: 570). The
possibility to account for the non-continuity of preferences seems particularly
important for the analysis of environment problems (Spash 2000).
Instead of optimization, GP is based on Simon’s satisficing philosophy (Simon
1955, 1979). In comparison, GP is a more pragmatic approach. It was used in the
1960s and 1970s by practitioners in corporations and agencies and only slowly
received more attention by theorists. While the techniques applied are the same
programming techniques as for the other MOP methods, the underlying philosophy
differs. The idea of optimizing is abandoned; only satisficing is attempted.
MOP and particularly GP are the most frequently applied multiple criteria
methods. They have been applied to a wide array of problems, namely production
planning, oil refinery scheduling, health care, portfolio selection, distribution system
design, energy planning, water reservoir management, timber harvest scheduling,
problems of wildlife management, etc. Their popularity can be explained by their
flexibility; they can account for diverse variable types (continuous, integer, Boolean,
etc.) as well as constraints and objective functions (linearity, convexity,
differentiability, etc.) (Vincke 1992).
More recent additions to the field of MOP have been: (a) to allow for fuzziness in
the data which is a way to address the problem of uncertainty to some degree; and
(b) evolutionary algorithms which were developed as a response to the fact that for
models which are large in size and/or significantly non-linear, traditional solution
methods are often unable to identify the global optimum. In such cases, typically
linear approximations are applied instead in order to find a solution. Genetic
algorithms were shown to be applicable to examples of large non-linear models (e.g.
Mardle et al. 2000). They follow the concept of solution evolution, by stochastically
developing generations of solutions populations using a given fitness statistic.
124 Andrea De Montis et al.
MOP and GP allow addressing a completely different type of multiple criteria
problems. Instead of choosing from a limited number of alternatives, a continuous
set of alternatives is generated during the solution process. They are the only
methodologies in the field which provide the model-based calculation of non-
dominated alternatives. Another advantage is that restrictions which are for instance
needed for the analysis of decision-making processes based on criteria of strong
sustainability (non-substitutability) can be accounted for. These special features
turn out to be important for the decision-making situation in many case studies.
GP has often been criticized for being analytically inferior to MOP. While GP
is indeed a somewhat simpler method, it is often preferred because of its capacity
to handle large-scale operational real-life problems. MOP works efficiently for
moderate-size problems defined within a well-structured environment. When
programming methods are applied, interdependence between criteria is not
allowed. The completeness description of the phenomenon is desired, but not
explicitly mentioned. While MOP is very appealing theoretically it has the
disadvantage that for large problems and particularly if non-linear functions are
included, often no optimal solution can be found.
The development of interactive methods, which allow to reveal progressively
the preferences of the decision-maker through man–machine interactions, made
MOP much less reliant on doubtful assumptions about the decision-maker’s
preferences. A number of useful software tools were developed which allow the
decision-makers an important role in the decision process and enable ‘psychological
convergence’ (i.e. termination of the interactive process as and when the decision-
maker chooses). Hence, interactive MOP leads not only to a good structuring of
the problem but can also provide a useful tool for learning for the decision-maker
involved. A problem remaining with interactive methods is that mostly they do not
provide an objective basis upon which the decision-maker could base the decision
to terminate the interactive process (Schniederjans 1995).
The next step in perceiving the role of the decision-makers in a more democratic
way, namely accounting for stakeholder participation (group decision-making), is
however explicitly accounted for in few tools and is therefore exhausted in only
some case studies. Another drawback in the user context is that the process cannot
be made very transparent. MOP allows for the use of different geographical scales,
as well as different general scales (micro–macro levels). Concerning data, MOP is
able to deal with quantitative as well as qualitative and fuzzy data and allows for
risk via fuzziness. An interesting aspect with regard to the fuzzy version of MOP
was reported in several case studies: the large amount of time which is normally
needed to select all the data required can on some occasions be bypassed by using
the fuzzy relations (e.g. Chang et al. 1997). GP is not able to deal with a lot of
different data, such as qualitative and fuzzy sets.
Some criteria could not be considered here because they are especially designed to
help the user in the selection of a method for a concrete decision-making situation,
and not designed for a theoretical method comparison with regard to general
sustainability issues without any specific application in mind, as done in this
comparison. Reference to the whole list of criteria was made in De Montis et al.
(2000), which included selected case studies of sustainable water use.
In handling complex situations, all methods show similar performance with
respect to the aspects and scales which can be considered; but they all show
weaknesses in the characteristics of the evaluation criteria which should be used
as operational components. Only AHP allows the interdependence of evaluation
criteria, while only MAUT and NAIADE allow non-linear preferences. Evamix,
ELECTRE III, and Regime do not permit either of these characteristics. In each
case quantitative and qualitative data can be addressed. In ELECTRE III
qualitative data have to be transformed first to a quantitative scale. Additionally,
some of the methods like GP/MOP and NAIADE include features to deal with
fuzzy data and stochastic numbers, which is the way that these methods deal with
risk and uncertainties. MAUT allows to deal with risk (not uncertainty) concerning
the outcomes of the alternatives by assigning probabilities to the utility functions
(von Neumann–Morgenstern utility functions). All of the methods provide the
possibility to carry out a sensitivity analysis or to apply qualitative data to consider
uncertainties and risks. Non-substitutability aspects can only be properly analysed
by using GP/MOP and ELECTRE III which allow to set constraints and thresholds
explicitly. Within the remaining methods non-substitutability can only be considered
by assigning flags to the respective criteria.
The performance to inform stakeholders in order to increase their knowledge
and change their opinion and behaviour is satisfactory for all methods. In MOP/
GP, ELECTRE III and Regime the transparency is only of medium quality, which
hinders stakeholder participation. The other methods are satisfactory or even good
Table 5.2 Summary of method comparison
MAUT AHP Evamix ELECTRE III Regime NAIADE MOP/GP
Operational components of methods
Criteria Inter- Not allowed Allowed Not allowed Not allowed Not allowed Not allowed Not allowed
dependence
Completeness Allowed Allowed Allowed Allowed Allowed Allowed Required
Non-linear Allowed Not allowed Not allowed Not allowed Not allowed Allowed Allowed for
preferences some variants
Weights Transparency Cardinal Cardinal Ordinal Cardinal Ordinal No weights Weights
of process, weights weights weights weights weights assigned assigned for
126 Andrea De Montis et al.
ordering
128 Andrea De Montis et al.
concerning these aspects. NAIADE does not allow to assign weights explicitly but
is the only methodology which provides an explicit structure for stakeholder
participation. For MOP/GP there are tools for interactive decision-making as well
as for group decision-making. While the majority of applications did not make use
of them, these tools have been available for more than ten years now.
The methods also differ in the possibilities for involving more than one decision-
maker in the process. For MOP/GP there are tools for group decision-making,
but again, they have not been applied much. Some methods like AHP support
consideration of preferences of several decision-makers. In practice, each decision-
maker is asked individually. Concerning MAUT, if n persons are involved in the
decision, each person is asked for their single utility function for any attribute;
these single functions are then aggregated to n multi-attribute utility functions. If
they are in conflict with each other, group solutions are possible, but are not really
treated by the literature about MAUT. Concerning stakeholder information,
MAUT, AHP, and Evamix provide good transparency and allow to give weights
explicitly. In a very interesting way, NAIADE supports the involvement of more
than one person in the decision-making process, namely with reference to their
different interests, which allows for an explicit conflict analysis. In general, an
important aspect in group decision-making is how transparent the decision-making
process is, as this enhances possibilities of participation and goal-oriented discussion
within the group.
Evamix, ELECTRE III, and Regime provide only satisfactory performance
for complex situations, so they are not the first choice in this context. Non-
substitutability aspects could best be considered by using ELECTRE III or GP/
MOP. MAUT and Evamix have acceptable performance in general, however they
lack important aspects crucial for sustainability issues, namely the ability to account
for uncertainty and non-substitutability. NAIADE is a specialist to involve more
than one decision-maker, while GP/MOP and ELECTRE III allow dealing with
non-substitutability. Thus, a clear ranking of the methods concerning the different
quality criteria is impossible, as each has its own strengths and weaknesses in
different areas. The characterization shows, however, that depending on the specific
problem the choice of one method over the other is advisable and that some
methods are largely useless for the analysis of sustainability problems.
Conclusions
The goal of more sustainable development has been studied extensively during
the last fifteen years; but a key criticism has been vagueness as to recommendations
and lack of operational applications. The acknowledgement of the limits of growth
has led to the revision of the paradigms of economics regarding human and natural
resources. Besides well-known goals like efficiency and income distribution, many
other conflicting factors need to be included in the analysis. MCDA constitutes a
powerful tool that is able to take into account several concerns and to foster partici-
pation and learning processes among politicians and citizens. Thus, MCDA is
useful for decision-making processes in the context of sustainability (e.g. Munda
Assessing the quality of different MCDA methods 129
1995; Castells and Munda 1999; Janssen 1992; Andreoli and Tellarini 2000;
Mendoza and Prabhu 2000).
In contrast to earlier studies which compare MCDA methods with respect to
more technical issues (e.g. Olson 2001, Salminen et al. 1998) we used a list of
quality criteria for the assessment (De Montis et al. 2000) and concentrated on the
most relevant criteria for sustainability issues. The comparison shows that the
indicator list is applicable and useful for the evaluation of MCDA methods in the
context of sustainability. As a result, a detailed characterization of the different
MCDA methods was derived. Some rough guidelines which can be given from the
comparison of the methods are:
1 If the respective decision problem is such that relying upon social welfare
theory and its assumptions is possible and if the data to build utility functions
is available (risk and qualitative data are possible) then MAUT is a good choice.
2 If working with different conflicting interest groups is important for the case,
NAIADE and AHP provide the best performance.
3 If the decision-makers involved should primarily learn from the application
of the MCDA tool, it is advisable to use MAUT or AHP.
4 If thresholds and constraints are central for the problem under investigation,
which means that there is non-substitutability of some criteria, ELECTRE
III or GP/MOP should be chosen.
5 If the problem is a continuous one, i.e. there is no discrete number of alterna-
tives which comes out of the specific situation, GP or MOP should be chosen.
6 If a complete ranking of the given alternatives as a result of the analysis is
indispensable, MAUT, AHP, Evamix, or Regime should be applied.
Notes
1 Risk and uncertainty are sometimes used interchangeably in the MCDA literature. As the
difference is very important in the context of sustainability, we would like to distinguish between
them. We use ‘risk’ if a probability distribution can be assigned for the unknown effects; otherwise
the term ‘uncertainy’ is used (Knight 1921).
2 Utility independence: ui(xi ) is unaffected by levels of the other attributes. Preferential indepen-
dence: the preferences for some attribute levels do not depend on the levels fixed for the other
attributes.
3 No individual will sense a difference between a cup of coffee with i milligrams of sugar and one
with i + 1 milligrams of sugar, but will express a preference between a cup with a lot of sugar
and one with none; this contradicts transitivity of indifferences (Vincke 1992).
130 Andrea De Montis et al.
4 ‘In general what we mean by being consistent is that when we have a basic amount of raw data,
all other data can be logically deduced from it. In doing pairwise comparison to relate n activities
so that each one is represented in the data at least once, we need n - 1 pairwise comparison
judgements. From them all other judgements can be deduced simply by using the following kind
of relation: if activity A1 is 3 times more dominant than activity A2 and activity A1 is 6 times
more dominant than activity A3 then A1 = 3A2 and A1 = 6A3. It should follow that 3A2 = 6A3 or A2
= 2A3 or A3 = ½A2. If the numerical value of the judgement in the (2, 3) position is different
from 2 then the matrix would be inconsistent. This happens frequently and it is not a disaster’
(Saaty 1980: 18).
5 For more information on the ‘random index’ (R.I.) see Saaty (1980: § 1.4). In the following table
values of the random index for matrix of orders from 1 to 11 are reported:
Order of the 1 2 3 4 5 6 7 8 9 10 11
Random index 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49 1.51
6 Elements of qualitative criteria are expressed by means of symbols like +++, ++, +, 0, –, – –,
– – –, where . … + + + ++ + 0 - -- - - - ….
7 An example of the way to obtain cardinal weights from ordinal preferences is given by Voogd
(1983: 175): ‘if w1 £ w2 £ w3 and it is assumed that weights meet the condition
m
 wj =1
j =1
Ê1 1 ˆ Ê 1 1 1ˆ
w1 = (1, 0, 0), w2 = ÁË , , 0˜¯ , w3 = ÁË , , ˜¯ ' .
2 2 3 3 3
8 For more information about the mathematical aspects of Evamix, see Voogd (1983: ch. 10).
9 Using stochastic numbers, it is necessary to choose the probability density functions; using fuzzy
numbers, it is necessary to define their membership functions; using qualitative evaluations,
some pre-defined linguistic variables such as ‘good’, ‘moderate’, ‘bad’ and so on are considerate.
It is not possible to assign different types of measurements (i.e. crisp, stochastic, fuzzy, linguistic)
to the same criterion for different alternatives.
10 In the case of two crisp numbers, the ‘distance’ is defined as their difference; in the case of
stochastic and fuzzy numbers, the ‘semantic distance’ measures the distance between two functions
(i.e. probability density functions or fuzzy membership functions); the linguistic variables are
treated as fuzzy sets.
11 For linguistic variables,the equivalent preference relations may be used : much better than (>>);
better than (>), approximately equal to (@); very equal to (=); worse than (<); much worse than
(<<). In this case the six associated analytic functions express an index of credibility of the
statements that an alternative is ‘much better’, ‘better’, ‘approximately equal’, ‘very equal’,
‘worse’ and ‘much worse’ than another.
12 For more information about the mathematical aspects of NAIADE, see Munda (1995: ch. 7).
13 The symbol * stands for >>, >, @, < or <<.
14 ‘Entropy measure of fuzziness can be easily extended to the notion of fuzzy relations … to have
information on the diversity among the assessments of the single fuzzy relations, according to
each criterion’ (Munda 1995: 138). More precisely ‘entropy is calculated as an index varying
from 0 to 1 that gives an indication of the variance of the credibility indexes that are above the
threshold, and around the crossover value 0.5 (maximum fuzziness). An entropy value of 0
means that all criteria give an exact indication (either definitely credible or definitely non-credible),
whereas an entropy value of 1 means that all criteria give an indication biased by the maximum
fuzziness (0.5)’ (Menegolo and Pereira 1996: 5). For the mathematic formulae of entropy, see
Munda (1995: § 7.4.)
Assessing the quality of different MCDA methods 131
References
Andreoli, M. and Tellarini, V. (2000) ‘Farm sustainability evaluation: methodology and
practice’, Agriculture, Ecosystems and Environment, 77 (1–2): 43–52.
Ballestero, E. and Romero, C. (1998) Multiple Criteria Decision Making and its Applications to
Economic Problems, Boston/Dordrecht/London: Kluwer Academic.
Castells, N. and Munda, G. (1999) ‘International environmental issues: towards a new
integrated assessment approach’, in M. O’Connor and C.L. Spash (eds) Valuation and
the Environment – Theory, Method and Practice, Cheltenham, UK/Northampton, MA:
Edward Elgar: 309–27.
Chang, N.-B., Wen, C.G. and Chen, Y.L. (1997) ‘A fuzzy multi-objective programming
approach for optimal management of the reservoir watershed’, European Journal of
Operational Research, 99: 289–302.
Churchman, C.W., Ackoff, R.L. and Arnoff, E.L. (1957) Introduction to Operations Research,
New York: Wiley.
Climaco, J. (1997) Multiple Criteria Analysis. Proceedings of the XIth International Conference
on MCDM, Coimbra, Portugal: Springer, 1–6 August 1994.
De Montis, A., De Toro, P., Droste-Franke, B., Omann, I. and Stagl, S. (2000) ‘Criteria for
quality assessment of MCDA methods’, paper presented at the 3rd Biennial Conference
of the European Society for Ecological Economics, Vienna, 3–6 May 2000.
Dillon, J.L. and Perry, C. (1977) ‘Multiattribute utility theory, multiple objectives and
uncertainty in ex ante project evaluation’, Review of Marketing and Agricultural Economics,
45 (1, 2): 3–27.
Fandel, G. and Spronk, J. (eds) (1985) Multiple Criteria Decision Methods and Applications. Selected
readings of the First International Summer School Acireale, Sicily, September 1983,
Berlin: Springer.
Fishburn, P.C. (1968) ‘Utility theory’, Management Science, 14: 335–78.
Fishburn, P.C. (1970) Utility Theory for Decision Making, New York: John Wiley and Sons.
Fishburn, P.C. (1978) ‘A survey of multiattribute/multiple criteria evaluation theories’, in
S. Zionts (ed.) Multiple Criteria Problem Solving, Berlin: Springer: 181–224.
Funtowicz, S.O. and Ravetz, J. (1990) Uncertainty and Quality in Science for Policy, Dordrecht:
Kluwer.
Hewitt, N. (1995) European Local Agenda 21 Planning Guide – How to Engage in Long Term Environ-
mental Action Planning Towards Sustainability? Brussels: European Sustainable Cities and
Towns Campaign.
Hinloopen, E. (1985) De Regime Methode, MA thesis, Interfaculty Actuariat and Econometrics,
Free University Amsterdam.
Hinloopen, E. and Nijkamp, P. (1990) ‘Qualitative multiple criteria choice analysis’, Quality
and Quantity, 24: 37–56.
Hinloopen, E., Nijkamp, P. and Rietveld, P. (1983) ‘Qualitative discrete multiple criteria
choice models in regional planning’, Regional Science and Urban Economics, 13: 73–102.
Hwang, C. and Yoon, K. (1981) Multiple Attribute Decision Making, Berlin: Springer.
Ignizio, J.P. (1976) Goal Programming and Extensions, Lexington: Lexington Books.
Janssen, R. (1992) Multiobjective Decision Support for Environmental Management, Dortrecht:
Kluwer Academic.
Kalu, T.C.U. (1999) ‘An algorithm for systems welfare interactive goal programming
modelling’, European Journal of Operational Research, 116: 508–29.
Keeney, R.L. and Raiffa, H. (1976) Decisions with Multiple Objectives, New York: John Wiley.
Kmietowicz, Z.W. and Pearman, A.D. (1981) Decision Theory and Incomplete Knowledge,
Aldershot: Gower.
132 Andrea De Montis et al.
Knight, F. (1921) Risk, Uncertainty, and Profit, Boston: Houghton Mifflin.
Korhonen, P. (2001) ‘Multiple objective programming support’, in C.A. Floudas and P.M.
Pardalos (eds) Encyclopedia of Optimization, Dortrecht: Kluwer: Vol. 3, 566–74.
Lara, P. and Romero, C. (1992) ‘An interactive multigoal programming for determining
livestock rations: an application to dairy cows in Andalusia (Spain)’, Journal of the
Operational Research Society, 43: 945–53.
Lee, S.M. (1972) Goal Programming for Decision Analysis, Philadelphia, PA: Auerbach.
Luce, R.D. (1956) ‘Semiorders and a theory of utility discrimination’, Econometrica, 24:
178–91.
Mardle, S.J., Pascoe, S. and Tamiz, M. (2000) ‘An investigation of genetic algorithms for
the optimization of multi-objective fisheries bioeconomic models’, International
Transactions in Operational Research, 7(1): 33–49.
Mastenbroek, P. and Paelinck, J.H.P. (1977) ‘Qualitative multiple criteria analysis –
applications to airport location’, Environment and Planning A, 9 (8): 883–95.
Mendoza, G.A. and Prabhu, R. (2000) ‘Multiple criteria decision making approaches to
assessing forest sustainability using criteria and indicators: a case study’, Forest Ecology
and Management, 131(1–3): 107–26.
Menegolo, L. and Guimarães Pereira, A. (1996) NAIADE Manual, Joint Research Centre –
EC, Institute for System, Informatics and safety, Ispra (VA), Italy.
Munda, G. (1995) Multiple Criteria Evaluation in a Fuzzy Environment – Theory and Applications
in Ecological Economics, Heidelberg: Physika Verlag.
Munda, G., Nijkamp, P. and Rietveld, P. (1994) ‘Fuzzy multigroup conflict resolution for
environmental management”, in J. Weiss (ed.) The Economics of Project Appraisal and the
Environment, Aldershot: Edward Elgar.
Nijkamp, P. and Voogd, H. (1981) ‘New multiple criteria methods for physical planning by
means of multidimensional scaling techniques’, in Y. Haimes and J. Kindler (eds) Water
and Related Land Resource System, Oxford: Pergamon Press: 19–30.
Olson, D.L. (2001) ‘Comparison of three multicriteria methods to predict known outcomes’,
European Journal of Operational Research, 130(3): 576–87.
Rietveld, P. (1980) Multi Objective Decision Making and Regional Planning, Amsterdam: North
Holland.
Roy, B. (1985) Méthodologie Multicritère d’Aide à la Décision, Parigi: Economica.
Roy, B. (1996) Multiple Criteria Methodology for Decision Aiding, Dordrecht: Kluwer Academic.
Roy, B. and Bouyssou, D. (1985) ‘An example of comparison of two decision-aid models’,
in G. Fandel and J. Spronk (eds) Multiple Criteria-Decision Methods and Applications. Selected
readings of the First International Summer School Acireale, Sicily, September 1983,
Berlin: Springer: 361–81.
Saaty, T.L. (1980) The Analytic Hierarchy Process for Decision in a Complex World, Pittsburgh, PA:
RWS Publications.
Saaty, T.L (1988) Decision Making for Leaders, Pittsburgh, PA: RWS Publications.
Saaty, T.L. (1992) Multicriteria Decision Making – The Analytic Hierarchy Process, Pittsburgh, PA:
RWS Publications.
Saaty, T.L. (1994) Fundamentals of Decision Making and Priority Theory with the Analytical Hierarchy
Process, Pittsburgh, PA: RWS Publications.
Saaty, T.L. and Alexander, J.M. (1989) Conflict Resolution – The Analytic Hierarchy Process,
New York: Praeger.
Saaty, T.L. and Forman, E. (1993) The Hierarchon, Pittsburgh, PA: RWS Publications.
Saaty, T.L. and Vargas, L.G. (1991) Prediction, Projection and Forecasting, Dordrecht: Kluwer
Academic Publishers.
Assessing the quality of different MCDA methods 133
Salminen, P., Hokkanen, J. and Lahdelma, R. (1998) ‘Comparing multicriteria methods
in the context of environmental problems’, European Journal of Operational Research, 104:
485–96.
Scarelli, A. (1997) Modelli matematici nell’analisi multicriterio, Viterbo: Edizioni Sette Città.
Schniederjans, M.J. (1995) Goal Programming Methodology and Applications, Boston: Kluwer
Publishers.
Simon, H. (1955) ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69:
99–118.
Simon, H. (1979) ‘Rational decision making in business organizations’, American Economic
Review, 69: 493–513.
Spash, C.L. (2000) ‘Multiple value expression in contingent valuation: economics and
ethics’, Environmental Science & Technology, 34(8): 1433–8.
Steuer, R.E. and Gardiner, L.R. (1990) ‘Interactive multiple objective programming:
concepts, current status, and future directions’, in C.A. Bana e Costa (ed.) Readings in
Multiple Criteria Decision Aid, Berlin/Heidelberg/New York: Springer: 413–44.
Taha, H.A. (1976) Operations Research, New York: Macmillan.
Tamiz, M. and Jones, D. (1997) ‘An example of good modelling practice in goal pro-
gramming: means to overcoming incommensurability’, in R. Caballero, F. Ruiz and
R.E. Steuer (eds) Advances in Multiple Objective and Goal Programming, Berlin/Heidelberg/
New York: Springer: 29–37.
Tamiz, M., Jones, D. and Romero, C. (1998) ‘Goal programming for decision making: an
overview of the current state-of-the-art’, European Journal of Operational Research, 111:
569–81.
Van Delft, A. and Nijkamp, P. (1977) Multiple Criteria Analysis and Regional Decision Making,
The Hague/Boston: Martinus Nijhoff.
Vincke, P. (1985) ‘Multiattribute utility theory as a basic approach’, in G. Fandel and J.
Spronk (eds) Multiple Criteria-Decision Methods and Applications. Selected readings of the
First International Summer School Acireale, Sicily, September 1983, Berlin: Springer:
27–40.
Vincke, P. (1992) Multiple Criteria Decision-Aid, New York: John Wiley & Sons.
von Neumann, J. and Morgenstern, O. (1947) Theory of Games and Economic Behaviour,
Princeton, NJ: Princeton University Press.
Voogd, H. (1981) ‘Multicriteria analysis with mixed qualitative–quantitative data’, Delft
University of Technology, Department of Urban and Regional Planning, Planologisch
Memorandom, 81: 86.
Voogd, H. (1983) Multiple Criteria Evaluation for Urban and Regional Planning, London: Pion.
Wierzbicki, A. (1980) ‘The use of reference objectives in multiobjective optimization’, in
G. Fandel and T. Gal (eds) Multiple Objective Decision Making, Theory and Applications, New
York: Springer.
Zionts, S. and Wallenius, J. (1976) ‘An interactive programming method for solving the
multiple objective programming problem’, Management Science, 22(6): 652–63.
134 Wendy Proctor
Introduction
Making decisions about the environment often involves balancing the conflicting,
incommensurate and incompatible values of many users and uses of a resource.
One of the most fundamental and difficult tasks involved, therefore, is the effective
integration or synthesis of all values related to the resource issue in question.
Consideration and effective integration of all resource values, whether they are
environmental, economic or social, is a necessary first step to achieving and main-
taining ecologically sustainable development.
In 1992, after decades of environmentalist versus forest industry conflict, the
Australian Government embarked on the largest and most expensive environmental
planning exercise ever undertaken in Australia. This was the programme of Com-
prehensive Regional Assessment (CRA) of Australia’s forests that would result in
the signing of Regional Forest Agreements (RFAs) and ensure the sustainable
management of forests in Australia. One of the most difficult tasks for planners
and policy-makers was that of integrating all of the different forest values upon
which a final decision about reserved and unreserved areas could be made. Without
an adequate process of integration, important information could have been lost in
the volumes of complex assessment documents that were created. As well, the
process would fail to adequately take account of the preferences of all stakeholders.
This chapter outlines a practical application of the decision support tool multiple
criteria decision analysis, also known as multiple criteria analysis (MCA). The
case study concerns the integration of various forest values using a stakeholder
forum from one of the assessment regions. In recent years, MCA, as a structured
procedure for aiding complex environmental, social and economic decisions, has
gained popularity. However, MCA has had limited application in environmental
policy-making in Australia. Although not part of the official process, the case study
utilizes the results of the assessments themselves and the decision-making criteria
and priorities placed on different forest values provided by a representative stake-
holder group that was part of the CRA of the Southern Forest Region of New
South Wales, Australia.1
MCDA and stakeholder participation 135
Forest assessments
Of the 157 million hectares of forest in Australia, around 20 per cent comprise
rainforest and open eucalypt forests, both of which are major sources of timber
production. Around 30 per cent of this rainforest and open eucalypt forest area is
situated on state government timber reserves and 18 per cent in nature conservation
reserves. The day-to-day management of these areas rests largely under the control
of the relevant state government with the Commonwealth (i.e. Federal) Government
having jurisdiction over the issuing of woodchip export licences and obligations to
ensure sustainable forest management under various international protocols, such
as the Montreal Process (Commonwealth of Australia 1997a). However, the
management of Australia’s public forests has been characterized by intense public
debate for several decades. As a result, Australian governments are now carrying
out a series of CRAs of these forests that are designed to determine the major part
of Australian forest policy for the next twenty years.
In 1992, the Commonwealth and State Governments of Australia reached
agreement on a National Forest Policy that would provide a long-term management
strategy for Australia’s forests (Commonwealth of Australia 1992). As part of this
strategy, the governments agreed to: undertake CRAs of environment, heritage,
social and economic values of forests; set up Comprehensive, Adequate and
Representative (CAR) conservation reserves on the basis of these assessments; sign
RFAs about the long-term management and use of Australia’s forests
(Commonwealth of Australia 1997b). The CAR Reserve system is intended to
safeguard biodiversity, old growth, wilderness and other natural and cultural values
of forests. Forests outside of these reserves will be available for timber or other
types of suitable ecologically sustainable production. The conservation criteria set
by a panel of experts and agreed to nationally by governments (known as the
JANIS criteria) aim to reserve where practicable: 15 per cent of the estimated
extent of each forest ecosystem prior to European arrival; at least 60 per cent of
old growth forest; and 90 per cent or more of high quality wilderness. Other
criteria have been similarly set for flora and fauna species. The assessments have
resulted in volumes of documents for each region that need to be considered
carefully before the drawing up of a Regional Forest Agreement between the
Commonwealth government and the relevant state government. The RFAs will
be maintained for a period of twenty years. The long-term nature of these agree-
ments has been designed to provide a secure basis on which the timber industry
can make long-term investment decisions. These agreements are, however, to be
reviewed every five years, to ensure that the objectives are being satisfactorily met.
In 1994, the first of a series of extensive assessments of 25 million hectares of
forests began. Relevant government departments provide assessments on
environment and heritage, National Estate, wilderness, resource, economic and
social values of the forests under review. An important part of the decision-making
phase leading up to these agreements is the ‘integration’ process. That is, the process
that ideally allows a rigorous comparison of economic, environmental and other
forest values in order to allow decisions to be made about the extent, in these
forests, of reserved, logged and other use areas. The following is an overview of
136 Wendy Proctor
how these assessments are undertaken and how the official ‘integration’ process
takes place (for more information, see for example Commonwealth of Australia
1998).
The ‘Environment and Heritage Assessment’ covers biodiversity, endangered
species, old growth and world heritage areas. The biodiversity assessment includes
a compilation of flora and fauna data and an assessment of species and forest
ecosystems. The species assessment includes information on distribution, habitat,
life history and risk of extinction factors for each terrestrial and aquatic forest
flora and fauna species including those that are rare and threatened. The forest
ecosystems assessment attempts to estimate pre-1750 distributions of forest
ecosystems and compare these with present-day distributions. The old growth
assessment lists and maps areas classified under certain scientific principles. The
framework for the world heritage assessment is provided by a panel of experts
who identify themes of outstanding universal value relevant to Australia and related
to Australian forests. The analysis then seeks to identify if any of these themes can
be applied to the forest region under review.
National Estate Assessment uses a broad range of technical expertise and public
input. The assessment seeks to identify areas of both natural and cultural value. It
draws on a wide range of existing flora, fauna, historical, ecological and pre-logging
surveys as well as some new data collected especially for the purpose. Also included
is an ongoing assessment of indigenous cultural and heritage values.
The Wilderness Assessment has been largely based on either existing surveys or
on expert opinion or a combination of the two. Data on variation in wilderness
quality across the natural areas of the region are mapped. Measures of wilderness
quality are essentially ‘remoteness’ from access, structures and settlement and
‘naturalness’, both apparent and biophysical. Other guidelines agreed upon by
technical experts (such as a minimum of 25,000 hectares for an area to be classified
as wilderness) are also used.
Resource and Economic Assessment details the existing resources and uses of
the forest area and provides an economic evaluation of the current situation and
potential future scenarios. The usage categories may include: native forests
(production of sawlogs and residual logs), plantations (native and exotic species),
other forest produce (e.g. honey), recreation and tourism, water and mineral
resources. The net value of the timber industry is assessed using survey data
collected from existing sawmillers. Future returns to the industry are estimated
using a GAMS (Generalized Algebraic Modelling System) based constrained linear
optimization model of the local timber industry called FORUM (Forest Resource
Use Model), and estimates of sustainable yields provided by independent forecasting
models of forest growth and production for the particular region (Dann and Clarke
1996). Data on the values of other forest uses is provided by existing surveys. The
potential for mineral resources is assessed using standard geological survey tech-
niques that place varying probabilities on the existence of valuable minerals
according to the basic geological structure of the area.
Social Assessment uses survey, interview and workshop techniques. This analysis
provides a socio-demographic profile of the CRA area as well as details of the
MCDA and stakeholder participation 137
current community infrastructure and an outline of community attitudes and
perceptions with regard to the use of forest resources. Various techniques based
on social assessment theory are employed in the analyses of the data.
‘Integration’ describes the official government process by which the various
assessments are brought together and used as the basis of coming to a decision
about reserved and other use areas in a region. The integration process differs
from state to state depending on the preferences of particular governments and on
the availability of data. In general, though, after the assessments have been
completed, a series of meetings are held and attended by representatives from the
Commonwealth, the state government and various stakeholder groups. The
objective of these meetings is to develop feasible scenarios, and subsequently, a
negotiated resource use option or set of options for the region, that is released for
public comment before a final decision is made.
A Geographic Information System (GIS) based reserve selection model called
‘C Plan’ is used extensively as the basis for developing resource use options. The
C Plan model maps areas of differing ‘irreplaceability’ and finds the smallest
possible set of areas that will achieve a nominated conservation goal (Pressy et al.
1995). Irreplaceability measures the extent to which a particular conservation
goal can be achieved. Two definitions of irreplaceability are used: one is the
likelihood that any of the areas in a region will be needed to achieve an explicit
conservation goal, the other is the extent to which the options for achieving an
explicit conservation goal are narrowed if any of the areas in a region are
destroyed or made unavailable for conservation. The conservation goals can
include the preservation of such items as particular forest ecosystems, species of
flora and fauna, etc. Also, the model maps areas of old growth forest, wilderness,
potential mineral deposits, high cultural and heritage values, timber production,
private and leasehold land as well as areas which have been targeted for indigenous
land rights claims.
A process of visual interpretation of the irreplaceability indices is followed and
each area of high irreplaceability investigated and selected for inclusion as a possible
reserve area. A visual interpretation of the areas of wilderness, old growth forests
and areas where threatened species are reported to exist is also carried out and
these areas included in the possible reserve system. Reserve design is also taken
into account. Despite the technical and structured nature of the C Plan method,
the reserve design process also includes a degree of ad hoc selection. For example,
this is displayed in the choice of criteria being used to select areas for reserve and
sometimes on the heavy reliance on local knowledge of conservation groups and
government officers. After several selections have been made, the extent to which
conservation targets have been met for the various flora and fauna species and
ecosystems are assessed. Next, a timber volumes model, the Forest and Resource
Management System (FRAMES) is run. This model estimates the amount of
commercial timber that could be harvested from the remaining area of forest that
had been excluded from reserve and is not on private or leasehold land (Turner
1998). The results are then fed into the FORUM economic model to determine
the likely consequences on the value of timber production in the area as a result of
138 Wendy Proctor
the change in reserve selection. FORUM provides an analysis of the commercial
value of the forest harvest, net returns to the regional industry and employment.
Further details on ‘integration’ can be found in Proctor (1999), however, it should
be noted that at no stage in the official integration process for the Southern New
South Wales (NSW) Region were the data obtained from the various assessments
comprehensively and methodically incorporated into the decision-making. It should
also be noted that, unlike previous assessments, for this region, the stakeholder
groups were largely excluded from the integration process. The Southern Region
integration process resulted in the description of five options, along with various
data associated with each option, being released for public comment. These options
represented five different reserve designs (with implications for economic indicators
as well as various levels of conservation and National Estate targets being met)
associated with levels of high quality sawlog production of 32,000 cubic metres
per annum, 35,000 cubic metres per annum, 45,000 cubic metres per annum,
55,000 cubic metres per annum and 65,000 cubic metres per annum. Details of
each option are provided later.
The following sections outline an MCA to utilize the forest assessment data in
a much more rigorous, objective and structured way than has been undertaken in
the RFA integration process. This method also allows for a much more participatory
approach to the decision-making process than has so far occurred for any of the
forest agreements. The guidelines for the CRA process (see Commonwealth of
Australia 1992) outline the need for public participation covering all interest groups
and accounting for all stakeholder preferences in a comprehensive and transparent
manner. In addition, stakeholder participation is a fundamental principle of
ecologically sustainable development. It will be shown that MCA is an effective
means of achieving stakeholder participation in the decision-making process.
Method
MCA is a means of simplifying complex decision-making tasks which may involve
many decision-makers, a diversity of possible outcomes and many and sometimes
intangible criteria by which to assess the outcomes. It is one means by which
structure and transparency can be imposed upon the decision-making process. Its
origins lie in the fields of public policy, mathematics and operations research and
it has had a great deal of practical usage by public planners in such areas as the
location of health facilities, motorways and nuclear reactors (Massam 1988).
An MCA seeks to make explicit the logical thought process that is implicitly
carried out by an individual when coming to a decision. In complex decision-
making tasks, which sometimes involve many objectives and many decision-
makers, this structured process of logic may be lost in the complexity of the
issues. In general, an MCA seeks to identify the alternatives or options that are
to be investigated and decided upon, a set of criteria by which to rank these
alternatives and the method by which the alternatives are to be ranked and
preferences aggregated. Finally, a sensitivity analysis is carried out on the results.
The final outcome is a preferred option or set of options that is based upon a
MCDA and stakeholder participation 139
rigorous definition of priorities and preferences decided upon by the decision-
maker.2
In general, the following steps are carried out by the decision-maker in an MCA.
The first is to identify the feasible options or preferred outcomes involved in the
decision problem. Next the decision-maker seeks to identify the overall objective
that is to be achieved in the process and then identifies the criteria by which to
judge the selected options. An important part of the process is then to apply
appropriate weights on each of the criteria that reflect the particular preferences
of the decision-maker in how important each criterion is in relation to the overall
objective of the decision process.
The analyst then chooses an appropriate form of MCA with which to analyse
the data and aggregate the preferences. An assessment is then made of each of the
options based on the chosen MCA technique, the data on how each option performs
under the different criteria and the weightings supplied by the decision-maker. An
investigation of the sensitivity of the ranking of options with respect to, for example,
the chosen weights and method of aggregation can then be carried out. An iterative
process can then be pursued where the analyst reports back to the decision-maker
the findings of the MCA and the sensitivities. ‘Fine tuning’ can then be performed
by the decision-maker to aspects related to their input. Also this part allows for the
identification of trade-offs, if any, that may be made in the overall decision process.
A common criticism in public policy-making and in particular environmental
policy-making in Australia is the high incidence of ad hoc decision-making that
takes place under such processes (see, for example, Dovers 2000; Doyle and Kellow
1995; Ferguson 1996). The need for a more structured and objective approach has
been addressed in the public policy literature, with some authors providing general
guidelines on how this might be achieved (for example Hogwood and Gunn 1984).
Similarly, an important aspect of environmental policy-making in recent years has
been the increasing practice of incorporating public and stakeholder participation
into some part of the policy-making process. The need for a participatory frame-
work to environmental policy formulation in Australia has grown out of increasing
interest by individuals in environmental issues (Lothian 1994), a grass-roots desire
by the public to become involved in the policy-making process over regional issues
that affect them, and a recognition by governments that involving the general
community early in the process can bring benefits by avoiding disagreements and
conflicts in later stages and increasing the knowledge base that is available to policy-
makers (Howlett and Ramesh 1995).
Whereas much of the literature on both prescriptive methods for policy-making
and the incorporation of public participation into such policy-making has addressed
general requirements and concerns, the steps outlined under the MCA process
are more explicit and provide a more detailed and structured guide for practical
approaches to participatory and other public policy-making. In the context of this
research, MCA is primarily regarded as an aid in the process of decision-making
and not necessarily as a means of coming to a singular optimal solution. As such,
the MCA process is valued for the enlightenment and unravelling of issues that it
can provide in the decision-making problem. The process adds to the knowledge
140 Wendy Proctor
of the decision-maker and is greatly aided by including the decision-maker in
each step of the analysis. This is one reason why some forms of MCA are regarded
as superior to other decision-aiding techniques (for example, mathematical
programming methods) which are usually carried out in isolation from the decision-
maker. Such techniques are designed to provide an ‘optimal’ result at the end of
the analysis, and so they do not necessarily increase the understanding of the
important elements of the process, especially to those who are not familiar with
such models. Other techniques are not designed to unravel the decision-making
problem or provide a process that will enlighten the decision-maker and add to the
decision-maker’s knowledge of the problem. The non-transparency and singular
solution nature of such techniques may only result in increased mistrust in the
process by those who are not familiar with how a decision was finally derived
(Prato 1999). The approach followed in this current research is therefore similar to
the ‘heuristic’ approach emphasized by Stirling and Mayer (1999) in their ‘multi-
criteria mapping’ technique (see also Chapter 7 in this volume).
a j1 a j1
Aj A1 a j1 a jI
a jI a jI
AI a j1 a jI
MCDA and stakeholder participation 141
Each cell entry of G reflects a ratio scale of the underlying priority weights assigned
to each of the criteria, i.e. gjj ¢ = wj/wj ¢. The weightings (wj ) have to be derived from
the cell entries gjj ¢. A nine-point Intensity of Importance scale was developed by
Saaty to determine these. It is claimed that the scale is based on psychological
experiments and is designed to allow for an accurate reflection of priorities in
comparisons between two items whilst minimizing the difficulties involved in doing
so (Table 6.1).
In G, each cell entry is positive and the diagonal elements (gjj ) are a series of 1’s.
If it is assumed that transitivity of preferences prevails (i.e. that if G1 is preferred
by a scale of 5 to G2, then G2 is preferred by a scale of 1/5 to G1) then the reciprocal
property gjj’ = 1/gjj’ is satisfied and estimates need only be provided for those cells
which lie above the diagonal. Saaty proves that, if G displays ‘cardinal’ consistency,
in that gjj’ gj’j” = gjj”, then by normalizing the positive reciprocal matrix G so that the
columns sum to unity, a solution to w, the vector of overall priority weights, can be
obtained by reading any column of the matrix, as each column in this normalized
matrix will be identical (Saaty and Vargas 1982). Imposing cardinal consistency
on the matrix also means that only one row of the matrix needs to be entered and
all other values can be derived. If cardinal consistency were not imposed, then
each column vector may be different and it would then be necessary to average
across the rows to determine the overall priority weights. In this instance, Saaty
provides a ‘check for consistency’ measure and acceptable bounds within which
this measure should fall.
To gauge how each plan performs with respect to each of the criteria, another
series of pairwise comparisons are carried out. For j = 1, …, J criteria and i = 1,
…, I different plans, a pairwise comparison matrix comparing all plans under the
criteria Gj would be:
A1 A1
a j1 a j1
Aj A1 a j1 a jI
a jI a jI
AI a j1 a jI
The following example clearly describes the above process. Assume that three
broad criteria groups (which could incorporate a number of sub-criteria) are
thought to be important in a decision concerning various natural resource use
options. These represent environmental, social and economic considerations. A
pairwise comparison for each of the criteria could then be as shown in Table 6.2.
The comparison of a criterion with itself always denotes equal importance and
so the central diagonal is a series of 1’s. In this example, environmental criteria are
considered to be weakly dominant over social criteria and so a 3 is recorded in the
appropriate cell. Its reciprocal is placed in the corresponding cell that compares
social with environment. This matrix is consistent in the sense that a value of 5/3
is recorded in the space when social criteria are compared with economic criteria.
This value was derived by imposing the cardinal consistency property on the matrix
although cases could exist where a different value is recorded here thus denoting
an inconsistency of preferences of the decision-maker. The matrix is then
normalized (by dividing each cell entry by its column total) so that meaningful
comparisons can be made amongst the components, and the relative priorities for
each criteria with respect to the overall objective can be read from any column.
This is shown in Table 6.3.
The vector of overall priorities shows the relative importance of each of the
three criteria in affecting the final decision. In this example, environmental criteria
are considered most important by a large margin and then follow social and finally
economic considerations. Using the criteria and priority weights estimated above,
three different plans are compared by first carrying out a pairwise comparison of
the performance of each plan against the others under the three different criteria
identified. An example is shown for the environmental criteria (Table 6.4). The
same procedure would then be carried out for the social and the economic criteria.
In this example, Plan a performs quite strongly under the social and economic
criteria and less so under the environmental criteria. Plan b performs poorly under
the environmental criteria and slightly better for the other two, whilst Plan c is
strongly preferred under environmental criteria but is least preferred when only
social or economic criteria are considered. To obtain an overall ranking of the
plans under all of the criteria combined, a weighted sum of each plan’s performance
under each of the criteria is calculated as in Table 6.5. From this analysis, Plan c is
MCDA and stakeholder participation 143
chosen as the preferred plan because of its strong performance with respect to the
environmental criteria and the associated weighting placed on the environmental
issues in achieving the overall objective.
Case study
The NSW Southern Region Regional Forest Forum was formed as part of the
RFA process. The Forum provided an opportunity for regional stakeholders to
share information with governments and offer recommendations. The Forum was
not, however, given an official decision-making role in the process. Members of
the Forum represented local state and federal government natural resource
management organizations, forest industry, conservation, farmer and beekeeping
lobby groups as well as academics and indigenous tribal elders and representatives.
The Forum met regularly for almost three years – around once every six weeks
between 1997 and 2000. Although not part of the ‘official’ process, the Forum
members agreed to take part in the MCA as part of their regular meetings, with
144 Wendy Proctor
the time allocated to collect information for the analysis forming only a small part.
Much of the information was gathered outside the meetings by mail survey, phone
or face-to-face interview.
The first stage of the analysis was to identify a complete set of criteria by which
to assess each alternative plan. The criteria identified relate directly to the stated
objectives of the RFAs as outlined in the National Forest Policy Statement. Figure
6.1 shows a hierarchy of objectives and criteria, developed in association with the
Forum members. Forum members were then asked to carry out the AHP pairwise
comparison exercise outlined previously. A survey asking for details about their
pairwise preferences of the identified criteria and sub-criteria was distributed and
information on how to fill out the survey was provided at one of the meetings.
Follow-up phone calls also provided assistance. A pairwise comparison of the three
broad criteria (Environment, Economic and Social) was requested as well as three
sub-criteria group comparisons. Information was only required on the matrix cell
entries which were above the diagonal. Reciprocals were later entered below the
matrix diagonals. Cardinal consistency was not imposed on the comparisons. Time
and other constraints did not allow for further interaction with the Forum for
‘fine-tuning’ of their preference data.
Results
Apiary
Timber
Minerals
Old growth
Indigenous
Wilderness
Biodiversity
National estate Indicators
Figure 6.1 Hierarchy of objectives, criteria and indicators for the Southern Region RFA
145
146 Wendy Proctor
0.90
0.80
Normalized priorities
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Environment Economic Social
the overall RFA objective with a figure of 0.41. Next followed the Economic criteria
(0.34) and finally the Social criteria (0.25).
The normalized priorities of the sub-criteria were then calculated. Within the
Environmental sub-criterion, the most important criteria were identified as Water
and Soil Resources, the Productive Capacity of the Forest and Biodiversity (Figure
6.3). The normalized priorities of the Economic criteria show less dispersion than
those of the Environmental criteria. Disparities in preferences exist in the Productive
Capacity, Timber Values and Minerals Values categories although the range of
priorities for Productive Capacity is not spread evenly with one substantial outlier
influencing the results (Figure 6.4). Disregarding this outlier, the highest priorities
were given to Timber Values and Employment and Community Needs.
In the category of Social Criteria, the highest priorities were given to
Employment and Community Needs and Indigenous Values. The largest range of
priorities for the social criteria (ignoring outliers) appeared to be in the Employment
and Community Needs category (Figure 6.5). More detailed analysis of the
preference structures of the Forum can be found in Proctor (1999) and Proctor
(2001).
0.60
0.50
Normalized priorities
0.40
0.30
0.20
0.10
0.00
Carbon
Hazard
Prod cap
Biodiversity
Wilderness
Water and
Old growth
soil
0.60
0.50
Normalized priorities
0.40
0.30
0.20
0.10
0.00
Productive Timber Minerals Apiary Other Employment Recreation and
capacity values values values and community tourism
Resources and Indigenous Values) were not provided for the different proposed
options.
The individual Forum members’ preference matrices of the sub-criteria were
re-normalized, after excluding those sub-criteria for which data were not available.
Table 6.7 shows the estimated pairwise comparisons of options under the
148 Wendy Proctor
0.70
0.60
0.50
Normalized priorities
0.40
0.30
0.20
0.10
0.00
Employment and Recreation and National Cultural and Indigenous
community tourism estate heritage
haulage employment, change in other employment); National Estate (National Estate areas reserved, National Estate values).
3 These data address the broad criteria under Maintenance of Long-Term Economic Benefits – Timber (gross value of timber output); Other Products (total gross value of
output) Employment and Community Needs (direct mill employment, harvesting and haulage employment, change in other employment).
149
150 Wendy Proctor
Table 6.7 Pairwise comparison of options under environmental criteria
Biodiversity* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 3 3 4 4
Option 2 0.33 1 2 2 3
Option 3 0.33 0.5 1 2 2
Option 4 0.25 0.5 0.5 1 2
Option 5 0.25 0.33 0.5 0.5 1
analyses are also important parts of the MCA to help to unravel the complexities
of the decision-making problem.
Sensitivity analysis
A sensitivity analysis was carried out to test the outcome of the overall ranking of
options when the priorities placed on the broad Environmental and Economic
criteria were both placed at their maximum values in turn.3 The outcomes are
depicted in Figure 6.9 and Figure 6.10, with an environmental preference leading
to an even higher ranking of Option 1 and an economic preference leading to
Option 5 being ranked first.
MCDA and stakeholder participation 151
Table 6.9 Pairwise comparison of options under economic criteria
Timber Option 1 Option 2 Option 3 Option 4 Option 5
employment*
Option 1 1 0.5 0.33 0.25 0.2
Option 2 2 1 0.33 0.25 0.2
Option 3 3 3 1 0.5 0.33
Option 4 4 4 2 1 0.5
Option 5 5 5 3 2 1
*Includes direct mill and harvesting and haulage employment.
Timber values* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 0.25 0.25 0.17
Option 2 1 1 0.33 0.25 0.2
Option 3 4 3 1 0.5 0.33
Option 4 4 4 2 1 0.5
Option 5 6 5 3 2 1
*Gross value of timber output.
Other values* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 1 0.5 0.5
Option 2 1 1 1 0.5 0.5
Option 3 1 1 1 0.5 0.5
Option 4 2 2 2 1 0.5
Option 5 2 2 2 2 1
*Total gross value of output.
Sensitivity analysis was also carried out to assess the order of rankings under
different assumptions about the sub-criteria. As the Employment and Community
Needs criterion displayed the greatest difference in the priorities placed on it by
Forum members, two scenarios were tested. The first used the maximum value
obtained for this criterion from the individual pairwise comparisons of both the
Economic and the Social sub-criteria (with other relative priorities maintained).
The effect was to change the ranking to Option 5 (0.29), Option 1 (0.22), Option
4 (0.21), Option 3 (0.16) and Option 2 (0.13). The second scenario used the
minimum value obtained from the individual pairwise comparisons for the
Employment and Community Needs criterion. Minimizing this criterion’s
152 Wendy Proctor
Table 6.11 Pairwise comparison of options under social criteria
National Estate* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 2 3 4 5
Option 2 0.5 1 2 3 4
Option 3 0.33 0.5 1 2 3
Option 4 0.25 0.33 0.5 1 2
Option 5 0.2 0.25 0.33 0.5 1
*The National Estate criterion has been given an arbitrary weighting because only
qualitative data were available.
Emp. & comm. Op ion 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 0.25 0.2 0.17
Option 2 1 1 0.25 0.2 0.17
Option 3 4 4 1 0.25 0.25
Option 4 5 5 4 1 0.5
Option 5 6 6 4 2 1
preference weighting only strengthened the lead of Option 1 further and only
slightly changed the lower rankings with an ordering of Option 1 (0.30), Option 5
(0.21), Option 2 (0.17), Option 4 (0.17), Option 3 (0.15).
A final sensitivity analysis was carried out on the relatively arbitrary values of
the National Estate impacts for the different options. Changing these values from
very small differences in impacts to very large differences in impacts had the effect
on the final rankings of swapping the placements of Options 1 and 5 between first
and second in each case. For small differences in National Estate impacts, the final
ranking was Option 5 (0.25), Option 1 (0.24), Option 4 (0.19), Option 3 (0.17),
Option 2 (0.15). For large differences in National Estate impacts, the final ranking
was Option 1 (0.27), Option 5 (0.24), Option 4 (0.18), Option 3 (0.15), Option 2
(0.15).
Conclusion
The outcome of the MCA based on the data released for the Southern Region
reveals that the preferred option is Option 1. Contrary to what may be thought
MCDA and stakeholder participation 153
0.45
0.40
0.35
0.30
Priority weight
Environmental
0.25
Economic
0.20
Social
0.15
0.10
0.05
0.00
1 2 3 4 5
Option
0.30
0.25
Priority weight
0.20
0.15
0.10
0.05
0.00
1 2 3 4 5
Option
from initial observations of the data, however, is that the next preferred option is
Option 5. On closer inspection of the data, this outcome is seen to be the result of
the high performance of Option 1 under the environmental criteria when compared
with all the other options and, similarly, the much better performance of Option 5
under the employment criteria when compared with all the other options. This
result accurately depicts the polarized nature of the priorities of the Forum members
and shows the nature of the trade-off in the forest management debate as being
one between conservation and employment. The outcome of the analysis shows
154 Wendy Proctor
0.5
Max
Priority weight
0.25 Min
Av
0
1 2 3 4 5
Option
0.40
0.30
Priority
0.20
0.10
0.00
1 2 3 4 5
Option
Figure 6.9 Environmental preference
0.40
0.30
Priority
0.20
0.10
0.00
1 2 3 4 5
Option
References
Commonwealth of Australia (1992) National Forest Policy Statement: A New Focus for Australia’s
Forests, Online. http://www.rfa.gov.au/nfps/contents.html (accessed April 2000).
Commonwealth of Australia (1997a) Australia’s First Approximation Report for the Montreal Process,
Canberra: Department of Primary Industries and Energy, Montreal Process Imple-
mentation Group.
Commonwealth of Australia (1997b) East Gippsland Resource and Economics Report, Joint
Commonwealth and Victorian Regional Forest Agreement (RFA) Steering Committee,
Canberra and Melbourne.
Commonwealth of Australia (1998) Central Highlands: Comprehensive Regional Assessment Report,
Canberra.
Dann, T. and Clarke, J. (1996) ‘An economic model for comprehensive regional forest
assessments, a case study – some issues and considerations’, paper presented at 40th
Annual Conference of the Australian Agricultural and Resource Economics Society,
Melbourne, 13–15 February 1996.
Dovers, S. (ed.) (2000) Environmental History and Policy: Still Settling Australia, Melbourne:
Oxford University Press.
Doyle, T. and Kellow, A. (1995) Environmental Politics and Policy Making in Australia, South
Melbourne: Macmillan.
Ferguson, I. (1996) Sustainable Forest Management, Melbourne: Oxford University Press.
Hogwood, B. and Gunn, L. (1984) Policy Making for the Real World, Oxford: Oxford University
Press.
Howlett, M. and Ramesh, M. (1995) Studying Public Policy: policy cycles and policy subsystems,
Oxford: Oxford University Press.
Lothian, J. (1994) ‘Attitudes of Australians towards the environment: 1975–1994’, Australian
Journal of Environmental Management, 1: 78–99.
Massam, B. (1988) Multi-Criteria Decision Making (MCDM) Techniques in Planning, Oxford:
Pergamon Press.
Prato, T. (1999) ‘Multiple attribute decision analysis for ecosystem management’, Ecological
Economics, 30: 207–22.
Pressy, R., Ferrier, S., Hutchinson, C., Sivertsen, D. and Manion, G. (1995) ‘Planning for
negotiation: using an interactive geographic information system to explore alternative
protected area networks’, in D. Saunders and J. Craig (eds) Nature Conservation 4: The
Role of Network, Sydney: Surry Beaty and Sons.
Proctor, W. (1999) ‘A practical application of MCA to forest planning in Australia’, paper
presented at IUFRO International Symposium: From Theory to Practice – Gaps and
158 Wendy Proctor
Solutions in Managerial Economics and Accounting in Forestry, Czech University of
Agriculture, Prague, 13–15 May 1999.
Proctor, W. (2000) ‘Towards sustainable forest management – an application of MCA to
Australian forest policy’, paper presented at the 3rd International Conference of the
European Society for Ecological Economics, Vienna, Austria, 3–6 May 2000.
Proctor, W. (2001) ‘Multi-criteria analysis and environmental decision-making: a case study
of Australia’s forests’, unpublished PhD thesis, Australian National University, Canberra.
Saaty, T. (1980) The Analytic Hierarchy Process, New York: McGraw-Hill.
Saaty, T. (1982) Decision-making for Leaders, California: Wadsworth.
Saaty, T. and Vargas, L. (1982) The Logic of Priorities, Boston: Kluwer-Nijhoff.
Stirling, A. and Mayer, S. (1999) ‘Rethinking risk: a pilot multi-criteria mapping of a
genetically modified crop in agricultural systems in the UK’, report for the UK
Roundtable on Genetic Modification, SPRU, University of Sussex, August 1999.
Turner, B. (1998) ‘Review of FRAMES data for the upper north east and lower north east
RFA regions of NSW’, unpublished consultant’s report for the State Forests of NSW
and the Bureau of Resource Sciences, Commonwealth DPIE.
Confronting risk with precaution 159
Introduction
The ‘precautionary principle’ is becoming an increasingly prominent theme in
the debate over technological risk. Many questions are raised over the implications
for policy making. In particular, concerns have been expressed over the relationship
between ‘precautionary’ and more traditional ‘science-based’ approaches to
decision making such as cost–benefit and risk analysis. Fears are sometimes raised
that – unlike risk assessment – a ‘precautionary approach’ is too ambiguous and
impractical to serve as a basis for real decision making, that it is somehow antag-
onistic to science and even that it threatens to stifle technological innovation and
economic growth.
The first part of this chapter takes a close look at some of the key practical and
theoretical issues bearing on the relationship between ‘science’ and ‘precaution’ in
the management of technological risk. It is found that – far from being in tension
– these two concepts might actually be seen as entirely consistent and even mutually
reinforcing. The real distinction is found to lie between a narrow ‘risk-based’ concept
of regulatory appraisal and broader precautionary approaches. A series of eight
criteria are developed against which appraisal techniques might be evaluated in
terms of both their scientific and precautionary validity. The second part of the
chapter then reports on a recent pilot exercise that seeks to apply these criteria to
the specific case of the regulatory appraisal of a genetically modified crop. The
main features of a novel ‘multi-criteria mapping’ methodology are outlined. Key
findings are reported, with particular attention paid to the extent to which the
method conforms to the scientific and precautionary quality criteria. Conclusions
are drawn concerning the degree to which the quality criteria can be made
operational in the practical business of regulatory appraisal.
ABOUT
continuum of set of discrete outcomes
LIKELIHOODS outcomes outcomes poorly defined
INCERTITUDE
RISK
Apply:
UNCERTAINTY IGNORANCE
no basis
for Apply: Apply:
probabilities scenario analysis diversity
sensitivity analysis flexibility
precaution precaution
not only is it impossible definitively to rank the different options, but even their full
characterization is difficult. Under a state of ignorance (in this strict sense), it is
always possible that there are effects (outcomes) which have been entirely excluded
from consideration.
Figure 7.1 provides a schematic summary of the relationships between these
formal definitions for the concepts of risk, uncertainty, ambiguity and ignorance.
It is quite normal, even in specialist discussion, for the full breadth and depth of
these issues to be rolled into the simple concept of ‘risk’ (and sometimes ‘uncert-
ainty’), thus seriously understating the difficulties involved. In order to avoid
confusion between the strict definitions of the terms ‘risk’ and ‘uncertainty’ as
used here, and the looser colloquial usages, the term ‘incertitude’ can be used in a
broad overarching sense to subsume all four subordinate conditions. Either way, it
is not difficult to see that it is the formal concepts of ignorance, ambiguity and
uncertainty – rather than mere risk – which best describe the salient features of
regulatory decision making in areas such as energy technologies, toxic chemicals
164 Andy Stirling and Sue Mayer
and genetically modified organisms. Indeed, many of the most high profile
technologically-induced ‘risks’ of recent years – such as stratospheric ozone
depletion, endocrine disrupting chemicals and BSE, for instance – are all cases
where the problem lay not so much in the determination of likelihoods, but in the
anticipation of the very possibilities. They were surprises!
The crucial point is that intractable uncertainties, ambiguities and ignorance
are routinely treated in the regulatory appraisal of technology simply by using the
probabilistic techniques of risk assessment. This treatment of uncertainty and
ignorance as if they were mere risk effectively amounts to what the economist
Hayek dubbed (in his Nobel acceptance speech) ‘pretence at knowledge’ (Hayek
1978). Far from displaying a respect for science in regulatory appraisal, the effect
of such scientistic oversimplification is actually to ignore and undermine the
scientific principles on which risk assessment itself purports to be based. Given the
manifest inapplicability – in their own terms – of probabilistic techniques under
uncertainty and ignorance, this is a serious and remarkable error. The self-
contradictions in aspirations to a ‘science-based’ approach reliant solely on
quantitative risk assessment, already noted in the last section, are thus further
underscored and reinforced.
Why is it that pursuit of (and claims to) the definitive authority of ‘science-
based’ approaches continues to be so prominent in regulatory appraisal? It seems
that the elegance and facility of probabilistic calculus has had a seductive effect on
many risk analysts and their sponsors. This may be understandable, yet it is also
curious. Despite the intractability of the condition of ignorance, there is no shortage
of operational tactical and strategic ‘precautionary’ responses. Some specific
features of these approaches will be reviewed in some detail later. For the moment,
in the specific context of ignorance, the point is simply that there do exist practical
alternatives to the use of probabilistic methods. For instance, there exists a variety
of institutional procedures for including in the regulatory appraisal process a range
of different scientific disciplines and people with pertinent professional and local
knowledge and relevant socio-economic perspectives. By providing for the
identification of a wider range of possibilities, this effectively helps to convert
some part of the domain of ignorance into the more tractable condition of
uncertainty. Here, techniques such as scenario and sensitivity analysis can also
help systematically to characterize these neglected possibilities and explore their
implications under different perspectives.
Beyond this, there are a series of broader strategies that may be employed and
which will be returned to later. In particular, rather than focusing entirely on efforts
to characterize the ‘problem’ (ignorance) attention can also be devoted directly at
aspects of the ‘solution’. Although the manifestations of ignorance are, by definition,
not characterizable in advance, certain dynamic properties of the different options
themselves can offer valuable ways of hedging against ignorance. Here properties
such as flexibility, reversibility, resilience, robustness and adaptability are all
potentially valuable (Stirling 1999; Salo 1999). Perhaps even more important, are
the possible merits of deliberate diversification across a range of options. After all,
it is a well-established matter of common sense that, when we don’t know what we
Confronting risk with precaution 165
don’t know, we don’t put all our eggs in one basket! It may be that a persistent
preoccupation with probabilistic methods has left these kinds of strategies unduly
neglected in the regulatory appraisal of technological risk.
ONSHORE WIND
PHOTOVOLTAICS
BIOMASS
HYDROELECTRICITY
NUCLEAR FISSION
GAS
OIL
COAL
hour of electricity production. The lowest values of the bottom range in Figure
7.2 are less than four hundredths of a cent per kilowatt-hour. The difference is
more than four orders of magnitude – a factor of more than fifty thousand! Detailed
analysis of the reasons for these discrepancies show that they do not arise as a
result of any single factor. It is not a simple matter of some studies being more
‘accurate’ or ‘reasonable’ than others in any definitive sense. Instead, the variability
is the cumulative consequence of the adoption of divergent assumptions and
priorities concerning the whole range of the different ‘dimensions of appraisal’
identified in the preceding sections (Stirling 1997).
The second crucial feature that is illustrated in Figure 7.2 concerns the ambig-
uities in the ordering of the different options under appraisal. The lowest values
obtained for the worst ranking option (coal) are lower than the highest values
obtained for the apparently best ranking options (wind). Since the effect of the
particular assumptions adopted in individual studies is to produce results at the
high end of the overall range for some options but lower in the distributions for
others, the overall picture yielded by the literature as a whole would accommodate
virtually any conceivable ranking order for these eight options! By the judicious
choice of framing assumptions, then, radically different conclusions can be justified
for regulation.
This evident disjuncture between precision and accuracy in supposedly ‘science-
based’ risk assessment paints a rather negative picture. One of the first and most
basic tasks in the management of risk is to construct some robust overall notion of
the relative merits of the different options under consideration from the point of
view of society as a whole. This then serves as a basis for regulatory intervention,
Confronting risk with precaution 167
market-based measures or investment initiatives. Where this cannot be achieved
in any absolute (or even relatively robust) sense, then the value of appraisal lies in
systematic exploration of the relationships between different assumptions in analysis
and the associated pictures of the relative importance of different options. Where
aspirations to the ‘science-based’ appraisal of risk lead to the assertion of the
intrinsic authority of narrow risk assessment procedures, then these crucial
exogenous factors typically remain unacknowledged and unexplored. In this event,
the problem is not simply one of a lack of rigour concerning the theoretical
contradictions noted in the previous sections. The difficulties are also very concrete
and pragmatic. For, without a robust appreciation of the assumptions under which
appraisal yields differing pictures of performance, serious questions must be raised
over whether the associated results – no matter how confidently and precisely
expressed – are of any practical policy use at all.
SCIENTIFIC Precautionary
Risk-based approach
DISCIPLINE
approach
systematic
sceptical
peer-reviewed
independent
accountable
learning
Cornucopian Apocalyptic
BREADTH OF SCOPE
Figure 7.3 A model of the relationships between risk, science and precaution
Confronting risk with precaution 169
affected. It requires scrutiny of claims to benefits and justifications as well as risks
and costs, with full account given to the available alternatives. Finally, a
precautionary approach involves the adoption of long-term, holistic and inclusive
perspectives in regulatory appraisal (Stirling 1999).
In many ways, these attributes of a ‘precautionary’ approach can be seen to
concern different aspects of the breadth of the regulatory appraisal process. A
‘broad’ regime is one that takes account of a wide range of different types of
impact, including qualitative as well as quantitative issues and including indirect
as well as direct effects. Likewise, a ‘broad’ framework accommodates a diverse
array of different points of view (including, importantly, those of potential ‘victims’)
and anticipates a wide range of possibilities in the face of uncertainty and ignorance.
It extends consideration to the benefits and justifications associated with the
introduction of the technology in question and examines a variety of alternative
ways in which the benefits of a regulated technology might be realized at lower
levels of risk. Taken together, these features constitute a more ‘precautionary’
approach because they increase the number and intensity of the constraints that
any technological option must satisfy in order to be approved by the regulatory
process, thus making it more difficult for certain innovations to pass through the
regulatory ‘filter’. At the same time, however, such measures might equally serve
to encourage other technological innovations that might otherwise remain
neglected.
What is interesting about this characterization of ‘precaution’ in terms of the
‘breadth’ of the associated regulatory regime, is that it reveals an inherently con-
sistent – and in many respects complementary – relationship between ‘precaution’
and ‘science’ in the management of technological risk. Accordingly, Figure 7.3
distinguishes between different approaches to risk management based on the degree
to which each embodies the respective characteristics of ‘scientific discipline’ and
‘breadth of framing’ identified here. Of course, both the ‘broad’/‘narrow’ and
the ‘scientific’/‘unscientific’ dichotomies drawn here are highly stylized and
simplified. However, the general picture revealed in Figure 7.3 is at least richer
and more realistic than the prevailing one-dimensional dichotomy between ‘science’
and ‘precaution’. Taken together, the combination of these two dichotomies
generates a fourfold array of idealized permutations. The adoption of a ‘narrow’
regime without reference to scientific understandings or disciplines in appraisal
might be described as a permissive position. Taken to an extreme, this would amount
to an entirely uncritical ‘anything goes’ approach to the regulation of technology
of the kind associated with caricature ‘cornucopian’ visions of technological
progress. Likewise, a broad-based regime might be similarly unscientific. The
resulting restrictive position might be associated with a caricature ‘apocalyptic’
vision of technology. In the extreme, it would lead to a situation of paralysis under
which no new technological innovation that offends in the slightest respect would
ever be approved for deployment. The crucial point is that neither the ‘permissive’
(cornucopian) nor the ‘restrictive’ (apocalyptic) positions as defined here would be
subject to challenge or reversal by the disciplines of scientific discourse associated
with the vertical axis.
170 Andy Stirling and Sue Mayer
Clearly, neither the established procedures of risk regulation (based on relatively
narrowly framed risk assessment methods) nor the emerging precautionary
approach (based on broader perspectives and considerations) actually resemble
these stylized ‘permissive’ or ‘restrictive’ caricatures. Existing risk-assessment-based
regulation includes a host of effective checks and balances. It certainly does not
necessarily provide for the uncritical approval of any new technology that may be
developed. Likewise, even the most progressive formulations of a ‘precautionary
principle’ are circumscribed in their scope, admit an incremental series of instru-
ments and allow for regulatory approval under a host of favourable conditions.
Both approaches are compatible – at least in principle – with the requirements of
systematic methodology, scepticism, transparency, accountability, quality control
by peer-review, professional independence and an emphasis on learning which are
held for the purposes of this discussion to be among the key aspirations of a ‘science-
based’ approach.
Now let us return to the earlier discussion of the profound importance of the
conditions of uncertainty, ignorance and multidimensionality in risk assessment.
Questions over the scope of appraisal, the plurality of different value positions
and framing assumptions, the diversity of different anticipated possibilities and
the degree of confidence placed in the available knowledge are all matters that are
central to the ‘scientific’ status of the appraisal process. That probabilistic
approaches are inapplicable under strict uncertainty and ignorance flows directly
from the theoretical foundations of risk assessment, and CBA (and, indeed, all
‘rational choice’ approaches to decision-making on risk). Following equally directly
from these fundamental theoretical principles, different priorities, framing
assumptions and value systems cannot be definitively aggregated across different
groups. Thus, for both these reasons, there can be no analytical fix for the definitive
ranking of different technology or policy options in the social appraisal of risk. All
that can be done to maximize scientific rigour in appraisal is to ensure that the
process is as broadly based as possible in terms of the value systems and framing
assumptions that are included and the options and possibilities that are addressed.
Seen in this way, then, key elements of the ‘breadth’ of the regulatory regime
themselves become issues of ‘sound science’ in the management of environmental
risk, as well as institutional features of the wider regulatory regime. Precaution, in
this sense, is not just entirely consistent with science – it is a necessary pre-requisite
for a truly scientific approach to the regulatory appraisal of risk.
These are the considerations that have informed the development of the ‘multi-
criteria mapping’ technique employed in the present case study (Stirling 1997;
Stirling and Mayer 1999). In short, the motivation behind this approach is to seek
to combine the openness and qualitative flexibility of participatory deliberation
with the clarity and focus of quantitative assessment. The specific case study with
which this method has been piloted concerns the hotly contested debate over the
use of genetically modified crops in UK agriculture.
172 Andy Stirling and Sue Mayer
Applying the multi-criteria mapping technique
Table 7.1 The definitions of the ‘basic options’ appraised by all participants
Option Definition
Organic agriculture All farming and food production conducted under present-
day organic standards.
Integrated pest All farming and food production conducted using systems
management designed to limit but not exclude chemical inputs and with
greater emphasis on biological control systems than
conventional systems.
Conventional All farming and food production conducted under present-
agriculture day intensive systems.
GM oilseed rape with Labelling based on the presence of foreign DNA or protein in
segregation and present the final product.
systems of labelling
GM oilseed rape with Monitoring for effects (mainly environmental) conducted on
post-release monitoring an on-going basis after commercialization.
GM oilseed rape with Areas of growing of GM oilseed rape restricted on a
voluntary controls on voluntary basis to avoid unwanted effects such as gene-flow
areas of cultivation and cross-fertilization of non-GM crops.
Up to six additional Any option of participant’s choice including combinations of
options to be specified the above if desired.
by participant
174 Andy Stirling and Sue Mayer
Table 7.2 The participants
Area Code
Agriculture and food industry B, L, H, K
Academic scientists C, J
Government safety advisors E, F
Religious and public interest groups A, D, G, I
OPTIONS
CRITERIA WEIGHTS A B C D E
environment
2 health 5
DEFINE
CRITERIA economics
SCORES CALCULATE
RANKS
RANKS
3
ASSESS
4
SCORES
ASSIGN
WEIGHTS
EXAMPLE
GM 0.14
pessimistic assumptions 0.14 1 optimistic assumptions
GM0.572 0.5717 1
GM0.566 0.5658 1
con0.252 0.2521 0
organic IPM0.142 0.1423 1
org 0.21 0.21 1
integrated pest management
conventional
GM - current
GM - monitoring
GM - voluntary controls
Findings
The procedure summarized above generated a rich body of empirical material,
reflecting a very wide range of issues bearing on the regulatory appraisal of GM
176 Andy Stirling and Sue Mayer
technologies in agriculture. This is reported on in more detail elsewhere (Stirling
and Mayer 1999, 2000) but the key points can be summarized here. In addition to
the six ‘basic options’ a further 18 alternative agricultural strategies were identified
and evaluated by participants. These included many different labelling and control
regimes and the application of a series of different regulatory assessment procedures
and criteria. They also involved a variety of different agricultural strategies,
including a focus on hypothetical strategies using GM technologies under an organic
farming regime. Interestingly these tended to perform relatively well under the
perspectives of participants from both sides of the GM debate (Table 7.3). A total
of 117 appraisal criteria were defined by different participants. Although some of
these are apparently closely related, all embody important differences of emphasis
or framing. It is interesting that criteria typically reflect issues not only of consumer
choice but also of citizenship and wider questions of participation and agency.
Collectively they cover a very wide range of considerations, including environ-
mental, agricultural, health, economic, social and ethical issues. Most of these
issues are very remote from the factors included in orthodox risk assessment. Indeed,
for no participant was it true that the full range of their concerns is fully addressed
by existing regulatory appraisal procedures in the UK (Table 7.4).
Public interest
A D
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Industry
I B
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
K L
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Each chart shows one perspective on the relative performance of six agricultural
options for producing oilseed rape. Individual perspectives are grouped and
colour coded according to their affiliation with four different constituencies. In
descending sequence on the vertical axis, the options are organic farming,
integrated pest management, conventional intensive agriculture and three
different strategies for the implementation of genetically modified crops. The
horizontal axis shows option performance and uncertainties on a subjective
interval scale (low to left, high to right).
Conclusions
This chapter has sought to cover quite a bit of ground. It has been shown that
there are significant theoretical, methodological and practical grounds for concern
over the utility of orthodox ‘narrow risk assessment’ techniques in the regulatory
appraisal of risk. The frequently cited dichotomy between ‘science-based’ and
‘precautionary’ approaches to regulatory appraisal has been shown to be unfounded
Confronting risk with precaution 181
Notes:
Participants were asked their opinion concerning diversity among options.
The results for each participant are shown as a pie chart displaying the
proportional importance of each option in their favoured mix of options.
The six basic options 1–6 are colour coded. The basic options are: 1
organic, 2 IPM, 3 conventional, 4 GM with segregation and current
labelling, 5 GM with monitoring, 6 GM with voluntary controls. The
additional options are: I7 GM with segregation, labelling, monitoring
and other restrictions; I8 GM with full assessment of impacts and need;
B7 GM with quality; K7 conventional and organic without GM; K8 mix
of GM, conventional and organic; J7 GM in regulated organic system,
J8 GM in IPM system; F7 GM with segregation, labelling and monitoring.
References
Arrow, K. (1963) Social Choice and Individual Values, New Haven: Yale University Press.
Bezembinder, T. (1989) ‘Social choice theory and practice’, in C. Vlek and G. Cvetkovitch
(eds) Social Decision Methodology for Technological Projects, Dordrecht: Kluwer.
EPCAG/Biotechnology and the European Public Concerted Action Group (1997) ‘Europe
ambivalent on biotechnology’, Nature, 387: 845–7.
Fisher, E. and Harding, R. (eds) (1999) Perspectives on the Precautionary Principle, Sydney:
Federation Press.
Funtowicz, S. and Ravetz, J. (1990) Uncertainty and Quality in Science for Policy, Amsterdam:
Kluwer.
Grove-White, R., Macnaghton, P., Mayer, S. and Wynne, B. (1997) ‘Uncertain world.
Genetically modified organisms, food and public attitudes in Britain’, Lancaster:
Lancaster University, Centre for the Study of Environmental Change.
Hayek, F. (1978) New Studies in Philosophy, Politics, Economics and the History of Ideas, Chicago:
Chicago University Press.
Loasby, B. (1976) Choice, Complexity and Ignorance: An Inquiry into Economic Theory and the Practice
of Decision Making, Cambridge: Cambridge University Press.
Mayer, S., Hill, J., Grove-White, R. and Wynne, B. (1996) ‘Uncertainty, precaution and
decision making: the release of genetically modified organisms into the environment’,
briefings no 8, Sussex: University of Sussex, Global Environmental Change Programme.
O’Riordan, T. and Cameron, J. (eds) (1994) Interpreting the Precautionary Principle, London:
Earthscan.
O’Riordan, T. and Jordan, A. (2001) Interpreting the Precautionary Principle, 2nd edn, London:
Cameron May.
Raffensberger, C. and Tickner, J. (eds) (1999) Protecting Public Health and the Environment:
Implementing the Precautionary Principle, Washington: Island Press.
Renn, O. and Klinke, A. (1999) ‘Risk evaluation and risk management for institutional
and regulatory policy’, field study for ESTO project on ‘Technological risk and the
management of uncertainty’, Sevilla: Institute for Prospective Technological Studies.
Salo, A. (1999) ‘Technological risk and the management of uncertainty: the role of decision
analytic modeling’, study for ESTO project on ‘Technological risk and the management
of uncertainty’, Sevilla: Institute for Prospective Technological Studies.
Smithson, M. (1989) Ignorance and Uncertainty: Emerging Paradigms, New York: Springer.
Stirling, A. (1994) ‘Diversity and ignorance in electricity supply investment: addressing the
solution rather than the problem’, Energy Policy, 22(3): 195–215.
Stirling, A. (1997) ‘Multi-criteria mapping: mitigating the problems of environmental
valuation’, in J. Foster (ed.) Valuing Nature, London: Routledge.
Stirling, A. (1999) ‘On science and precaution in the management of technological risk –
final report’, Sevilla: Institute for Prospective Technological Studies.
Stirling, A. and Mayer, S. (1999) ‘Rethinking risk: a pilot multi-criteria mapping of a
genetically modified crop in agricultural systems in the UK’, report no 21, Sussex:
University of Sussex, SPRU.
184 Andy Stirling and Sue Mayer
Stirling, A. and Mayer, S. (2000) ‘Precautionary approaches to the appraisal of risk: a case
study of a GM crop’, International Journal of Occupational and Environmental Health, 6(4):
296–311.
Von Schomberg, R. (1998) ‘An appraisal of the working in practice of directive 90/220/
EEC on the deliberate release of genetically modified organisms’, Luxembourg:
European Parliament, Scientific and Technological Options Assessment Unit (STOA).
Wynne, B. (1992) ‘Uncertainty and environmental learning: reconceiving science and policy
in the preventive paradigm’, Global Environmental Change, 2(2): 111–27.
Wynne, B. and Mayer, S. (1999) ‘The release of genetically modified organisms to the
environment’, in EU 98, Copenhagen: European Environment Agency.
Consumer valuation and citizen deliberation 185
Part III
Deliberation,
participation and value
expression
186 Jonathan Aldred
Consumer valuation and citizen deliberation 187
Introduction
There is now a relatively well-developed critique of the contingent valuation method
(CVM) for adding decisions about environmental problems. Theories of
deliberative democracy have been invoked which question the individualistic,
preference-based calculus of contingent valuation. A particular deliberative institu-
tion which has recently received much attention is the citizens’ jury (CJ). This is
taken here to involve around 16 ordinary members of the public, selected to
represent a cross-section of the local community. The jury comes together over a
period of usually four days to consider an important public policy question. The
members of the jury are briefed by and question expert ‘witnesses’, and discuss
the issues with each other in small and large groups, chaired by an independent
moderator. However, while the original impetus for the adoption of deliberative
institutions in environmental policy-making was undoubtedly dissatisfaction with
the CVM and the search for an alternative, a perhaps unexpected development is
that advocates of deliberative institutions no longer see them as direct alternatives
to the CVM. For instance, as the CJ methodology has been elaborated in more
detail, researchers have increasingly realized that the CJ ‘valuation exercise’ serves
different purposes, answers different questions, and makes different assumptions
about the individual participants than the CVM.
Although much of interest must be set aside in order to pursue any sort of
comparison between the CVM and CJ, it is still possible to explore the relative
merits of public discussion about the environment, vis-à-vis private reporting of
individual judgements. What does the presence of discussion per se add to the
environmental decision-making process? A number of putative advantages of
discussion are explored in detail, and in particular the critique of these deliberative
virtues, coming from both neoclassical environmental economists and the ‘rational
choice’ school of political theorists, is assessed. An aim has been to address not
just existing criticisms of deliberative institutions from this perspective, but to
anticipate and respond to criticisms which are likely to be made in the same spirit.
Rather than repudiate this perspective altogether (although I do not share it), I
show that the criticisms do not even succeed on their own terms. That is, I hope to
show that the arguments against discussion fail, in terms which the critics of
deliberative democracy must acknowledge.
188 Jonathan Aldred
The chapter begins by exploring the extent to which the CVM and CJ can be
meaningfully compared. It is argued that this is possible to a limited extent, by
treating both as explicit decision-making processes, with only CJs involving
discussions before decision, and restricting the comparison to an assessment of
epistemic advantages and disadvantages of that discussion. The assessment in this
chapter is made from the perspective of rational choice theory, because that
approach has been influential amongst critics of deliberative democracy. A final
section examines the desirability of using CJs to generative explicit decisions,
including a brief discussion of methodologies which attempt to combine the CVM
and CJ.
It is not immediately obvious that the comparison specified by this question maps
neatly and helpfully on to that between the CVM and CJ. A number of features of
the comparison must be clarified.
Consumer valuation and citizen deliberation 189
Decision-making processes?
(Q) implies the end point of the process is a decision. There are two separate issues
at stake here. Roughly speaking, the first may be regarded as the question of whether
the input to the process is open or closed; the second issue is whether the output is
open or closed. In both cases, there is no dichotomy but a continuum of varying
degrees of openness. The process input is said to be closed if the participants’
have no control over the inputs to the process. These inputs are principally the
task the participants are set, and the information they are given at the beginning
of the process, or are able to obtain during it. The more open the process input,
the more control participants have over their role, and in particular the subject
matter of their thinking, during the process. The process output is said to be closed
if the outcome of the process is narrow and determinate. The more open the
output, the more it admits diverse interpretations, and the less structure it will
have. While these notions of openness and closedness of outputs and inputs seem
intuitive, they are hard to define clearly in isolation. A few examples should help to
clarify them.
The following process outputs are (arguably) progressively more open: decision,
valuation, recommendation, competing recommendations, process report
suggesting consensus but making no recommendations, process report suggesting
dissent. Clearly the openness of the output will be influenced by the openness of
the input – for instance, if the participants are set a relatively closed task in being
asked to reach a decision. This example might suggest that a relatively closed
input entails a closed output. However this is mistaken for two reasons. First, in
the example, the participants may be asked to reach a decision, but the process
input may be very open in other respects, implying a relatively open input overall.
Second, the participants may simply ignore the exhortation to reach a decision.
The openness of inputs and outputs are in general independent of one another,
and only the former concerns the degree of control the participants possess. A
very strict requirement to reach a decision is likely to be followed by a closed
output; but it is the closed input rather than the closed output which is responsible
for the participants’ lack of control; in contrast, a decision may emerge from a
process in which the participants were given an extremely open, sketchy task, but
chose to narrow this down into a particular decision problem. Another example
of confusion between open inputs and outputs is captured in the dubious claim
that if a process requires only a report of deliberation rather than a decision, then
the participants have more control. Certainly a decision is a less open output, but
this does not entail a less open input. In other words, there is no a priori reason why
the requirement to produce a report on a specific task gives participants any more
control than the requirement to produce a decision. It does, however, demand less
agreement among them.
While the process input is principally classified according to control, the focus
in classifying the process output is agreement – a closed output will demonstrate
more agreement among the participants.4 In a later section, this feature will be
used to distinguish different types of CJ, but the more immediate concern is the
190 Jonathan Aldred
comparison between the CVM and CJ. The CVM methodology presumes and
necessitates a closed input in which the set of policy alternatives, among which a
single choice is to be made, is exogenously determined. CJ methodology makes no
such assumption. With an open input, the set of alternatives is not well defined,
perhaps because there is no fact of the matter, or established understanding,
regarding the nature of the problem. The public authority may present the CJ
with an open input because the authority’s understanding of the problem does not
permit closure. Part of the resolution of the problem is in framing the decision to
be made – ‘asking the right questions’. While the open input approach may be the
richer and more realistic characterization, it remains incompatible with the CVM
methodology. If a comparison of the CVM and CJ is to be pursued, a closed input
must be assumed.
Another difficulty for the comparison between the CVM and CJ is that they
typically have different outcomes. A CJ provides a report, while the CVM gives a
valuation. Since comparing a report and a valuation directly is impossible, it is
helpful to construe the comparison between the CVM and CJ as one between
different decision-making procedures. So, in the absence of unanimity, deliberation
in the CJ is terminated with a voting or aggregation procedure; it may then be
compared with ‘CV-in-CBA’, which also yields a decision, rather than the CVM
alone.5 Once both the CVM and CJ are defined to stretch as far as a final decision,
a direct comparison is possible. However adopting this approach is unnecessary in
order to compare the CVM and CJ indirectly. Referring again to (Q), it is still
possible to examine whether, and in what respect, discussion affects participants’
views. But the comparison is clearly much sharper if the effect of discussion can
be traced through to a decision.
To sum up, the CVM and CJ are regarded here as decision-making processes
in two respects: participants in both are set the task of deciding between exogenously
specified options (closed input), and both processes reach a decision (closed output).
These features are clearly exhibited by some, but not all, CVs and CJs in practice.
For the purpose of tractable comparison, they will be assumed in what follows,
although I shall return to the possibility of open input and output processes at the
end of the chapter. This approach also has the virtue of consistency with the more
abstract comparison between deliberative and aggregative institutions. Insofar as
they have been compared, both are assumed to terminate with a decision (Elster
1998a). The comparison between deliberative and aggregative institutions is then
a rather strange one – between aggregation preceded by discussion, and aggregation
alone. Most advocates of deliberative democracy would of course reject the claim
that the only difference between deliberative and aggregative institutions is that in
the latter, voting is not preceded by discussion. But this approach reflects the ‘thin’
view of deliberation, as discussion, adopted here.
Epistemic merits
The approach taken here might appear to bias the debate heavily in favour of
deliberative methods; the comparison between aggregation preceded by discussion,
and aggregation alone, is, to repeat, a strange one. But the debate is far from over,
since advocates of deliberative democracy must claim:
decisions reached after discussion will be better decisions than those reached
without it (B)
Strategic reasons
A speaker (S) might advance justifications which appeal to the public interest for
what have been termed ‘strategic’ reasons (Elster 1995). The basic idea is
straightforward: if S invites others to choose proposal P merely because P serves S’s
private interest, few will be persuaded. Johnson (1998) is characteristic of the
deliberative democracy critics in seeing the presence of self-interested motives behind
apparently impartial arguments as crucially undermining the normative weight, which
deliberative democracy advocates attach to the argumentative content of deliberation.
The epistemic justification of deliberative democracy, which appeals to ‘the force of
the better argument’, is subverted by the self-interested motives of participants to
the deliberative process. The objection runs that listeners are swayed by arguments
which have no objective merit but have been proposed simply because they serve the
speaker’s interests. Johnson concedes that ‘psychological pressures’ generated by the
process of discussion might over time make speakers sincerely come to embrace a
principle which they earlier invoked for self-interested reasons alone – one example
of Elster’s ‘civilizing force of hypocrisy’ (Elster 1995). But this possibility does not
nullify the objection – Johnson (1998: 172) argues that such a mechanism which
‘first induce[s] actors to engage in self-censorship and eventually cause[s] them to
embrace the resultant views … does not so much generate a “reasoned agreement”
as induce a conformity that is at once rather shallow and normatively suspect’. The
reasons for suspicion are not stated, although in a note Johnson (1998: 172 n. 47)
worries about the autonomy of actors subject to such mechanisms and quotes Elster
(1986a: 113) approvingly: ‘Dissonance reduction does not tend to generate
autonomous preferences’. He summarizes his critique by implying that deliberative
democracy advocates must demonstrate both ‘(a) that impartial claims are, even
tacitly, quite widespread in public debate and (b) such claims are animated by
mechanisms other than … strategic calculation and attendant psychological pressures’
(Johnson 1998: 172). It is argued here that the case for deliberative institutions such
as CJs need not satisfy (b) and that there are reasons to expect (a) will be satisfied, at
least in the ideal case.
Regarding (b), a claim can be impartial, in the sense of appealing to widely
shared principles or interests, but still be ‘animated by strategic calculation’.
194 Jonathan Aldred
Arguments that appeal to the public interest, to widely shared principles, have
normative force not because they purportedly reveal the noble character of the
speaker, but because they are much more likely – in the ideal case, the only way –
to secure the consent of the listener. The motives of the speaker are not in general
relevant to winning this consent. For example, in a debating contest S may be
acting as a ‘devil’s advocate’, but this does not in itself undermine S’s arguments,
even if the listeners know of S’s rejection of S’s arguments. Of course motives
matter: if the listener (L) is aware that S’s self-interest is best served by the action
urged by S, then deliberative democracy recommends (and expects) that L will
examine S’s claim that the public interest is served by the action all the more
carefully. L will also be sceptical of emotional appeals to the public interest, or
uncorroborated factual claims, in this context. But L should in general be able to
establish, independently of S’s motives, whether the action urged by S serves the
public interest – and in particular, whether it does so for the reason(s) advanced in
S’s argument.
Johnson also misrepresents Elster’s ‘civilizing force of hypocrisy’. While
advocates of deliberative democracy are not always clear, the essential point is
that agents, through discussion, become aware of new ideas and the positions of
others. In particular, as S advances arguments in which S has no conviction in
order to persuade others, S may come to realize their merits over opposing argu-
ments S privately held at first – a situation I shall term devil’s advocacy conversion.
It occurs directly through the recapitulation of arguments to others in order to
persuade them: finding new perspectives which enjoy common support, realizing
how the position can be strengthened further, or simply coming to understand it
more fully.
Returning to Johnson’s claim (a): as already suggested, whenever S seeks to win
consent, whether for self-interested reasons or otherwise, S will seek to advance
impartial arguments – arguments showing how the public interest is served – because
by showing how diverse private interests are simultaneously served, an argument
will command widespread support. This motivation for appeals to the public interest
would not seem to apply to majority groups: a subset of deliberators, if they form
a majority of the whole group, will be free to make blatant appeal to their collective
self-interest, that is the collective interest shared by the members of the subset but
not shared by all. Yet there is anecdotal evidence that even majorities appeal to a
broader public interest than their own constituents. This suggests there must be
other reasons besides strategic ones for justifications in terms of the public interest.
That credibility has both epistemological dimensions – the speaker must have
good sense and be reliable in the formation of judgements – and ethical
dimensions – the speaker must have the moral character that allows us to trust
their utterance and there must be grounds for believing that they are not
inclined to impart falsehoods to their audience.
(O’Neill 1997)
Beginning with the first question, the problem is not the emergence of a closed
output per se, but the extent to which this is forced. The requirement that a closed
output be reached, is a form of closed input which reduces the control the
participants have over the process. In contrast, theories of deliberative democracy
generally encourage as much openness as possible. Deliberative democracy
advocates argue that, by demanding a decision on a specific ‘charge’ or question,
the body which commissions a CJ may exclude important issues, information and
perspectives. At best, the CJ may be biased, while if the exclusion is intentional
then the CJ is manipulated. At the other extreme:
Experiments with complete juror control of the process have found that
participants, in the initial stages, do not have enough of an overview on a
subject to deal competently with setting the charge, agenda organisation or
witness selection.
(Smith and Wales 1998: 12)
Consumer valuation and citizen deliberation 201
Thus there is a trade-off between the openness of inputs and deliberative
competence. How this trade-off should be resolved depends on the role attributed
to the CJ. One view holds that the role of a CJ is simply to ‘read off ’ the given
opinions or tastes of the jurors. On this view, it may be plausible to separate entirely
the framing of the question from opinions about its answer. In contrast, supporters
of CJs usually interpret their role as centrally involving matters of ethical concern,
where the neoclassical economists’ mantra de gustibus non est disputandum is irrelevant.
Jurors must debate their judgements of ‘right’ and ‘wrong’ rather than simply
express their tastes of ‘like’ and ‘dislike’, and a central part of this debate is to
explore the issues at stake. It seem indefensible that this exploration be subject to
an exogenous constraint.
Part of the process of reaching the ‘right answer’ in environmental problems is
framing the question correctly, just as it is for, say, abortion or capital punishment
issues. For such problems, it seems that any defence of imposing a specific question
on a CJ must rely on roughly the claim that ‘the commissioning body knows best’.
Whether this claim is based on the view that the commissioning body is more
experienced at making complex problem-framing decisions, or better informed,
or more public-interested than the jury, it faces many difficulties. An obvious one
is that a commitment by the commissioning body to a closed input reflects a lack
of trust in participants. This may trigger a corresponding mistrust among
participants – in particular a suspicion that the commissioning body will ‘cherry-
pick’ recommendations from the report of the CJ as it suits them (Aldred 1998;
Aldred and Jacobs 2000). It is difficult to overemphasize the need for jurors to
maintain trust in the process if a CJ is to be successful (Armour 1996; Crosby
1996, 1999; Coote and Lenaghan 1997; Smith and Wales 1999; Ward 1999).
A less obvious difficulty is an apparent tension between the claim that ‘the
commissioning body knows best’ and rational choice theory. For on the broader –
and arguably more consistent – account of rational choice theory which embraces
public choice theory, it is not just ordinary market agents who act as if homo
economicus, but politicians and bureaucrats too. Politicians are neither better
informed, nor more public-interested, than anyone else. Within this tradition, it is
very difficult to defend any claim that ‘the commissioning body knows best’, yet it
is from within this tradition that environmental economists emerge who most
strongly argue in favour of requiring closed outputs to the CVM and CJ.
In summary, reconciling the demand for a closed output with a view of CJs as
calling forth ethical judgements of right and wrong, proves difficult. Suppose
however that such a reconciliation can be achieved, or that this view of CJs is
rejected, and that in answer to the first question, a closed output is deemed desirable.
Moving to the second question, should this output be quantitative or qualitative?
A common early view of the role of CJs was that they could offer a useful space
for discussion before proceeding to a standard CVM exercise: the CJ would finish
with the now well-informed participants each privately filling in a CVM
questionnaire, a close variant of the ‘focus group CV’ (Brown et al. 1995; Aldred
1996; Jacobs 1997b; Brouwer et al. 1999). Both Ward (1999) and James and Blamey
(2001) are more radical in seeking an average valuation from the jury not as a
202 Jonathan Aldred
statistical average of individual CVM responses, but as the direct response to asking
the jurors collectively for their valuation. They both hope to avoid some of the
problems of imposing forced agreement: they hope that an average valuation will
emerge naturally as a kind of quantitative consensus from group discussion, rather
than being enforced by means of a vote. Nevertheless, they both limit jurors’ control
in the sense discussed above; they require a closed output, to be achieved by some
kind of voting procedure, if consensus does not emerge. This new approach, termed
‘deliberative monetary valuation’ (DMV) (see the introduction to this volume),
seeks to combine the advantages of discussion with the merits of a valuation output.
Since DMV aims to be more readily comparable with the CVM than conventional
CJs, with their qualitative output, it will undoubtedly attract much attention.
In practice, probably the main advantage of the DMV approach is the appeal
of quantitative measures in contemporary policy discourse in market economies;
they have overtones of scientific rigour and accuracy. However, at a theoretical
level, the DMV approach appears to have two main virtues over a conventional
CJ. They mimic two of the arguments made for favouring the CVM over CJ. First,
DMV captures preference intensities. If, as seems likely, the deliberation in both
CJ and DMV processes does not generate unanimity, then some kind of aggregation
mechanism must be employed to generate a decision. If the aggregation procedure
used in DMV at this point measures preference intensities, then it is argued to be
superior because it captures more information about individual preferences. Second,
this additional information enables DMV to avoid standard Arrovian results in
social choice theory under which there exists no mechanism to aggregate
preferences which also satisfies certain weak and plausible properties (Sen 1970).
The threat to post-deliberation preference aggregation posed by Arrovian
impossibility is arguably a weak one – indeed, as noted above, the language and
concepts of social choice theory may help make clear the extent to which
deliberation can generate consensus (Miller 1992; Aldred 2000). But the argument
that individual valuations provide extra information seems more robust, and can
be generalized to the claim that DMV does more than simply decide amongst a
given set of policy options, because it provides a willingness to pay which is
applicable in broader resource allocation contexts. For instance, the willingness to
pay value could help a public authority make budgetary decisions beyond the
options facing the jury – say between health care and environmental preservation
expenditure.
The fundamental difficulty with this argument, and the DMV method, can be
easily stated: ‘is the collective valuation question meaningful?’ Ward (1999: 78)
writes: ‘The question posed should be “what is a particular level of provision of
environmental quality worth to the average citizen?”’ (1). Later he adds (p.79):
‘Jurors are not asked to express their personal evaluations but their judgements
about what environmental quality is worth to society as a whole’ (2). In contrast,
James and Blamey’s (2000, emphasis added) stylized valuation question takes the
form: ‘What is the maximum amount the citizens of City X would be prepared to
pay to ensure the proposed harbour development proceeds, if they knew what you
now know about it?’ (3). In the actual DMV exercise they conducted, the format was
Consumer valuation and citizen deliberation 203
different: ‘How high would a park management levy have to be, before the jury
would recommend Option 1 rather than Option 4 below? In other words, how
high would the levy have to be before the New South Wales public would be no
better off under Option 4 than Option 1?’ (5).
It seems possible that these five different question formats may elicit five
substantively different kinds of response. The variety of question formats may
reflect an underlying ambiguity in the variable they seek to measure. Although
both Ward and James and Blamey suggest they seek to capture ‘citizen’ rather
than ‘consumer’ responses (Sagoff 1988), it is not obvious that all the formats
would achieve this, particularly formats (1) and (3). Formats (1) and (3) might both
be interpreted by jurors as asking them to make some kind of prediction regarding
what others (non-jurors) would be willing to pay. Format (5) also appears to ask for
a prediction, but this time relative to welfare rather than willingness to pay. In
contrast, formats (2) and (4) seem to ask the jurors for their valuations, although
particularly in (2) it remains unclear how they are to interpret the question, and,
correspondingly, the meaning to be attached to an answer. One conclusion to
draw is simply that questions in DMV exercises need very careful formulation,
another is that DMV exercises reflect similar problems of meaning as do the CVM
exercises, if not more so. There is no space to go into this very large issue,9 but the
kinds of questions which both the CVM and DMV exercises raise are easily stated.
In the DMV context, do the jurors assume all members of the affected community
have to pay? And an equal amount? Or just those who express a hypothetical
willingness to do so? Or are the jurors being asked how much they would voluntarily
contribute to a joint effort to secure the specified outcome? It is difficult to see how
the DMV format, or any decision procedure which seeks a valuation, will ever be
as readily comprehensible to participants as the underlying ‘what is best for society?’
in a CJ.
Conclusions
This chapter has discussed how a comparison between the CVM and CJ might
proceed, focusing on the epistemic merits of discussion. It was argued that the
rational choice theory critique of these merits largely fails on its own terms. For
the CJ/CVM comparison, the focus in this chapter on the feasibility of competent
deliberation seems more informative than the early emphasis in the deliberative
democracy literature on an ideal of deliberation. However this focus is subject to
a number of caveats. The social facts which constrain and determine the deliberative
setting are not fixed but contingent. There is already some evidence of a dynamic
feedback effect from deliberative institutions to the social conditions, which
determine their feasibility. The evidence is generally positive – increasing the
presence of deliberative institutions may improve the confidence, capacity and
willingness of ordinary citizens to make decisions about public policy questions
(Smith and Wales 1998, 1999). Thus it is important to set the detailed discussion
of feasibility constraints in this chapter in context: they will be subject to – and
may be nullified by – the deliberative institutions, which they presently constrain.
204 Jonathan Aldred
One important feasibility constraint which has not been mentioned here (it receives
little attention from the rational choice theory perspective), but which deserves detailed
treatment, is inequalities of voice in any community – what might be termed the
problem of ‘willingness to say’ (O’Neill 1997). It is arguable that income and wealth
inequalities may generate the same unequal distribution of power in influencing the
decision, whether they work through constraining willingness to pay in CVM or
constraining willingness to say in CJs. More worrying still: the very requirement that
discussion be reasoned and serious to qualify as deliberation may be enough to
disenfranchise many members of society (Sanders 1997).
The question (Q) invokes a comparison both narrower and more general than
that between the CVM and CJ: it captures only one element of any overall com-
parative assessment (although it is not clear in what other respects the methodologies
can be helpfully compared), but it applies to the more general comparison of
deliberative versus aggregative institutions. It will by now be apparent that the
advantages of discussion canvassed here – and the objections raised – are highly
interdependent. Together they generate a complex web of forces (see in particular
Bohman 1996: 57–66), which even on the account of rationality advanced by
rational choice theory, create some significant benefits of discussion.
With these caveats firmly in mind, the discussion here has a number of policy
implications. Very broadly, where bounded rationality is likely to be severe, public-
interested principles are called forth, or important private information is present,
the epistemic merits of discussion are likely to be more significant. CJ appears
more attractive in these contexts. This conclusion aims to stimulate further research
rather than general judgements about the superiority of one method in all cases.
In a particular policy context, there is no substitute for a careful, empirical, case-
by-case comparison. While both the CVM and CJ have drawbacks, the final section
suggests that appeal to a mixed method such as DMV may not be the best way to
overcome them.
Acknowledgements
The author would like to thank Alan Holland, Michael Jacobs, Rosemary James,
Jürgen Meyerhoff, Simon Niemeyer, John O’Neill, Graham Smith, Andrew Stirling
and Hugh Ward for useful discussions on these issues.
Notes
1 See Bohman (1998) for a survey. The leading early statements of deliberative democracy are all
in Bohman and Rehg (1997), which also contains important papers. Another key collection is
Elster (1998a).
2 I shall ignore the distinction between judgements and preferences in what follows. Much of the
literature to be discussed uses ‘preference’ in contexts where ‘judgement’ seems more appropriate
– where the content of the ‘preference’ can be assessed for reasonableness. I shall follow the
usage of this literature while acknowledging its shortcomings.
3 There is substantial work discussing CJs in an environmental context (Armour 1996; Crosby
1996; Coote and Lenaghan 1997; Jacobs 1997b; Kuper 1997; Aldred 1998; Crosby 1999; Rippe
Consumer valuation and citizen deliberation 205
and Schaber 1999; Smith and Wales 1999; Ward 1999; Aldred and Jacobs 2000), in addition to
the contributions elsewhere in this volume.
4 Of course absence of agreement does not ensure an open output. There may be no agreement
among the CVM participants, yet an apparently closed output – at least if the CVM is regarded
as having only one output, an aggregate valuation, and data on the distribution of individual
valuations or protest bids is ignored.
5 The distinction (Jacobs 1997b) between an institution with decision-making power, and one
with power only to make a recommendation to a decision-making authority, is ignored in this
chapter. A recommendation is termed a decision regardless of whether those making the recom-
mendation have power to enforce it. (Of course the possession or absence of such power will
typically affect the recommendation.)
6 Although the term deliberative institution will be retained instead of discursive institution.
7 In this chapter, rational choice theory is understood in a strong sense as comprising both the
formal axioms which make up decision theory (Anand 1993 offers a detailed critical treatment),
and a substantive assumption of self-interest – in short, the agent is homo economicus. See Pettit
(1998).
8 The arguments in the next three sections are developed in more detail in Aldred (2001).
9 There is a large literature on what the CVM question means in theory (as opposed to
methodological difficulties in the practical application of the method), most of it written by
non-economists. For three different, and powerful, critiques by economists which cast doubt on
the coherence of the CVM question, see Sen (1995a), Sugden (1999) and Vatn and Bromley
(1994).
References
Aldred, J. (1996) ‘Value conflicts in environmental decision-making’, PhD thesis, Faculty
of Economics and Politics, Cambridge: University of Cambridge.
Aldred, J. (1998) ‘Land use in the fens: lessons from the Ely citizens’ jury’, Ecos, 19(2):
31–7.
Aldred, J. (2000) ‘Deliberation and social choice theory’, paper presented at ISUS
Conference, Winston-Salem/North Carolina, March 2000.
Aldred, J. (2001) ‘Citizens’ juries: discussion and rationality’, Risk, Decision and Policy, 6(2):
71–90.
Aldred, J. and Jacobs, M. (2000) ‘Citizens and wetlands: evaluating the Ely citizens’ jury’,
Ecological Economics, 34(2): 217–32.
Anand, P. (1993) Foundations of Rational Choice under Risk, Oxford: Clarendon.
Armour, A. (1996) ‘The citizens’ jury model of public participation’, in O. Renn, T. Webler
and P. Wiedemann (eds) Fairness and Competence in Citizen Participation: Evaluating Models
for Environmental Discourse, Dordrecht: Kluwer.
Austen-Smith, D. (1990a) ‘Credible debate equilibria’, Social Choice and Welfare, 7: 75–93.
Austen-Smith, D. (1990b) ‘Information transmission in debate’, American Journal of Political
Science, 34(1): 124–52.
Austen-Smith, D. and Riker, W. (1987) ‘Asymmetic information and the coherence of
legislation’, American Political Science Review, 81 (3): 897–918.
Bohman, J. (1996) Public Deliberation, Cambridge, MA: MIT.
Bohman, J. (1998) ‘The coming of age of deliberative democracy’, Journal of Political
Philosophy, 6 (4): 400–25.
Bohman, J. and Rehg, W. (eds) (1997) Deliberative Democracy, Cambridge, MA: MIT.
Brouwer, R., Powe, N., Turner, R.K., Bateman, I.J. and Langford, I.H. (1999) ‘Public
attitudes to contingent valuation and public consultation’, Environmental Values, 8(2):
325–47.
206 Jonathan Aldred
Brown, T., Peterson, G. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71(2): 250–60.
Buchanan, J. (1954a) ‘Individual choice in voting and the market’, Journal of Political Economy,
62: 334–43.
Buchanan, J. (1954b) ‘Social choice, democracy and free markets’, Journal of Political Economy,
62: 114–23.
Coote, A. and Lenaghan, J. (1997) Citizens’ Juries, London: Institute of Public Policy
Research.
Crosby, N. (1996) ‘Citizens’ juries: one solution for difficult environmental questions’, in
O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer.
Crosby, N. (1999) ‘The citizens jury process and environmental decisions’, in K. Sexton
and T. Burkhardt (eds) Better Environmental Decisions: Strategies for Governments, Businesses
and Communities, Washington, DC: Island Press.
Elster, J. (1986a) ‘The market and the forum: three varieties of political theory’, in J. Elster
and A. Hylland (eds) Foundations of Social Choice Theory, Cambridge: Cambridge University
Press.
Elster, J. (1986b) The Multiple Self, Cambridge: Cambridge University Press.
Elster, J. (1995) ‘Strategic uses of argument’, in K. Arrow (ed.) Barriers to Conflict Resolution,
New York: Norton.
Elster, J. (ed.) (1998a) Deliberative Democracy, Cambridge: Cambridge University Press.
Elster, J. (1998b) ‘Introduction’, in J. Elster (ed.) Deliberative Democracy, Cambridge: Cambridge
University Press.
Farrell, J. and Gibbons, R. (1989) ‘Cheap talk with two audiences’, American Economic Review
79(3): 1214–23.
Fearon, J. (1998) ‘Deliberation as discussion’, in J. Elster (ed.) Deliberative Democracy,
Cambridge: Cambridge University Press.
Gambetta, D. (1998) ‘Claro! An essay on discursive machismo’, in J. Elster (ed.) Deliberative
Democracy, Cambridge: Cambridge University Press.
Goodin, R. (1986) ‘Laundering preferences’, in J. Elster and A. Hylland (eds) Foundations of
Social Choice Theory, Cambridge: Cambridge University Press.
Gowans, C. (1987) Moral Dilemmas, Oxford: Oxford University Press.
Gutmann, A. and Thompson, D. (1996) Democracy and Disagreement, Cambridge, MA:
Harvard University Press.
Hylland, A. (1986) ‘The purpose and significance of social choice theory’, in J. Elster and
A. Hylland (eds) Foundations of Social Choice Theory, Cambridge: Cambridge University
Press.
Jacobs, M. (1997a) ‘Alternatives to contingent valuation’, paper presented at CSEC/Green
Alliance Conference, London, 23 January 1997.
Jacobs, M. (1997b) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature?, London: Routledge.
James, R. and Blamey, R. (2000) ‘Deliberative non-market valuation’, paper presented at
ISEE 2000 Conference: People and Nature: Operationalising Ecological Economics,
5–8 July.
Johnson, J. (1998) ‘Arguing for deliberation: some skeptical considerations’, in J. Elster
(ed.) Deliberative Democracy, Cambridge: Cambridge University Press.
Johnson, S. (1963) ‘The adventurer, Saturday 28th April 1753’, in W. Bate (ed.) ‘The Idler’
and ‘The Adventurer’, New Haven: Yale University Press.
Krehbiel, K. (1991) Information and Legislative Organization, Michigan: Ann Arbor.
Consumer valuation and citizen deliberation 207
Kuper, R. (1997) ‘Deliberating waste: the Hertfordshire citizens’ jury’, Local Environment,
2(2): 139–53.
Lane, R. (1991) The Market Experience, Cambridge: Cambridge University Press.
Liebrand, W., Messick, D. and Wilke, H. (eds) (1992) Social Dilemmas: Theoretical Issues and
Research Findings, Oxford: Pergamon Press.
Macedo, S., (ed.) (1999) Deliberative Politics, Oxford: Oxford University Press.
Mackie, G. (1998) ‘All men are liars’, in J. Elster (ed.) Deliberative Democracy, Cambridge:
Cambridge University Press.
Miller, D. (1992) ‘Deliberative democracy and social choice’, Political Studies, 40: 54–67.
Nagel, T. (1979) ‘The fragmentation of value’, in T. Nagel (ed.) Mortal Questions, Cambridge:
Cambridge University Press.
O’Connor, M. (2000) ‘Pathways for environmental evaluation: a walk in the (Hanging)
Gardens of Babylon’, Ecological Economics, 34(2): 175–93.
O’Neill, J. (1993) Ecology, Policy and Politics, London: Routledge.
O’Neill, J. (1997) ‘Deliberation and its discontents’, working paper presented at the
conference Environmental Justice, Global Ethics for the 21st Century, October 1997,
Melbourne: University of Melbourne.
Orbell, J., van der Kragt, A. and Dawes, R. (1988) ‘Explaining discussion-induced co-
operation’, Journal of Personality and Social Psychology, 54(5): 811–19.
Orbell, J., Dawes, R. and van der Kragt, A. (1990) ‘The limits of multilateral promising’,
Ethics, 100: 616–27.
Pettit, P. (1998) ‘Institutional design and rational choice’, in R. Goodin (ed.) The Theory of
Institutional Design, Cambridge: Cambridge University Press.
Riker, W. (1982) Liberalism against Populism: A Confrontation between the Theory of Democracy and
the Theory of Social Choice, San Francisco: Freeman.
Rippe, K. and Schaber, P. (1999) ‘Democracy and environmental decision-making’,
Environmental Values, 8(1): 75–88.
Sagoff, M. (1988) The Economy of the Earth, Cambridge: Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24(2–3): 213–30.
Sally, D. (1995) ‘Conversation and cooperation in social dilemmas’, Rationality and Society,
7: 58–92.
Sanders, L. (1997) ‘Against deliberation’, Political Theory, 25(3): 347–76.
Sen, A. (1970) Collective Choice and Social Welfare, Amsterdam: North Holland.
Sen, A. (1995a) ‘Environmental evaluation and social choice: contingent valuation and
the market analogy’, Japanese Economic Review, 46(1): 23–37.
Sen, A. (1995b) ‘Rationality and social choice’, American Economic Review, 85(1): 1–24.
Sen, A. (1997) ‘Individual preference as the basis of social choice’, in K. Arrow, A. Sen
and K. Suzumara (eds) Social Choice Re-examined, London: Macmillan.
Simon, H. (1978) ‘Rationality as process and as product of thought’, American Economic
Review, 68 : 1–16.
Smith, G. and Wales, C. (1998) ‘Citizens’ juries and deliberative democracy’, mimeo,
Glasgow: University of Strathclyde, Department of Government.
Smith, G. and Wales, C. (1999) ‘The theory and practice of citizens’ juries’, Policy and
Politics, 27(3): 295–308.
Sugden, R. (1999) ‘Alternatives to the neo-classical theory of choice’, in I. Bateman and
K. Willis (eds) Valuing Environmental Preferences, Oxford: Oxford University Press.
Vatn, A. and Bromley, D. (1994) ‘Choices without prices without apologies’, Journal of
Environmental Economics and Management, 26: 129–48.
208 Jonathan Aldred
Ward, H. (1999) ‘Citizens’ juries and valuing the environment: a proposal’, Environmental
Politics, 8(2): 75–96.
Williams, B. (1979) ‘Conflicts of values’, in A. Ryan (ed.) The Idea of Freedom, Oxford:
Oxford University Press.
Three approaches to valuing nature 209
9 Three approaches to
valuing nature
Forest floodplain restoration
Wendy Kenyon and Nick Hanley
Introduction
Given the extensive literature documenting problems related to economic valuation
techniques such as the contingent valuation method (CVM), interest has developed
in approaches which provide an alternative source of information on public value
judgements for natural resource decision-making (Brown et al. 1995). In particular,
some authors have suggested that deliberative methods might have a role to play
as alternatives or complements to more traditional project appraisal techniques
(Jacobs 1997; Sagoff 1998). A number of researchers have assessed the role of the
Citizens’ Jury (CJ) as discussed in the preceding chapter (see also, Stewart et al.
1994; Aldred and Jacobs 1997; James and Blamey 1999). Here we take the CJ to
consist of a small group of people, selected to represent the general public rather
than any particular interest group or sector, who meet to deliberate upon a policy
question.
The CJ offers a number of advantages over economic appraisal methods, and
in particular over the CVM. First, respondents to CVM surveys must understand
exactly what it is they are to value if useful information about preferences is to be
elicited (Arrow et al. 1993). However, recent literature highlights the fact that many
respondents do not appear to be well informed about the issues or the good to be
valued (Brown et al. 1995; Jacobs 1997). As Munro and Hanley (1999) show,
changing people’s information sets can be expected to change their willingness to
pay (WTP). CJs tackle this ignorance problem by combining information, time,
scrutiny and deliberation in the preference elicitation process (Coote and Lenaghan
1997). They allow participants to question witnesses, discuss witnesses’ evidence
with other jurors, and thereby gradually learn about and reach a richer
understanding of the issue (Sagoff 1998). CJs therefore address the information
problem better than the CVM.
Second, economists and others have suggested that a CVM questionnaire asks
respondents the wrong question, assuming that consumers think about
environmental goods in the same way they do about private goods (Blamey 1996;
Jacobs 1997; Sagoff 1988). Blamey (1996) suggests that respondents should not be
treated as consumers of environmental goods, but rather as citizens who think of
the welfare of the community when responding to environmental issues. In other
210 Wendy Kenyon and Nick Hanley
words, individuals approach decision-making relating to environmental goods as
citizens rather than consumers. Whilst the validity of this as a universal description
fitting all cases of environmental management can be disputed, the use of CJs as
a method of preference revelation allows consumers to be asked what Sagoff and
Jacobs call ‘the right question’, as it allows deliberation on the environmental issue
in terms of what is best for society. Indeed, while the question for the jury can be
framed in the context of individual consumer values and preferences if necessary,
CJs were developed specifically to determine opinions that represent the general
public, rather than any individual interest (Coote and Lenaghan 1997; Brown et
al. 1995).
Third, CJs may also be useful in dealing with the equity and distributional
issues which the CVM has attracted criticism over. Economic value is effectively
determined by preferences, underpinned by ability to pay. Therefore, in the CVM,
any value that a consumer places upon a good is not registered unless she is able to
pay for it. CJs however allow participants’ opinions and preferences to be expressed
and registered regardless of their ability to pay (Crosby 1995).1
Fourth, the United Nations Conference on Environment and Development has
suggested that environmental decisions will not be sustainable unless local
communities participate fully in the decision-making process (United Nations 1993).
The CVM does not include the participation of the community in a central way,
and therefore may not encourage sustainable actions. The use of CJs is, in contrast,
a means by which public participation can be more fully incorporated into the
environmental decision-making process.
Finally, the notion of value construction suggests that respondents do not have
well-defined preferences for many complex environmental goods prior to the elicit-
ation process, but that these preferences are constructed during this elicitation
process itself (Gregory et al. 1997; Payne and Bettman 1999). The way in which
people construct their preferences is important as it is argued that decision-makers
should attach more weight to the preferences of someone who knows both sides of
the argument, than to someone whose knowledge of the problem is more limited
(Elster 1983; Payne and Bettman 1999). According to Elster (1983) however,
methods which explicitly try to determine how people construct their preferences
(for example by aiding value construction, such as Gregory et al.’s (1997) decision
pathways, or by asking respondents to ‘think aloud’ whilst determining their value
as Schkade and Payne (1994) did) are likely to be contradictory – akin to telling
someone to ‘be spontaneous’! CJs, however, provide information about the process
of preference construction as a by-product of the process rather than as a central
role. Therefore, using CJs may offer a means of circumventing the contradiction
inherent in helping respondents construct their values, and provide information to
decision-makers on the weight which should be given to those preferences expressed.
Elster (1983) argues that because the CVM is based on a ‘thin theory of
rationality’, which requires only consistency in the expression of preferences, poor
policy choices may be made on the basis of CVM results. CJs on the other hand
appear more consistent with a ‘broad theory of rationality’ which examines not
only the consistency of expressed preferences, but also the beliefs and desires behind
Three approaches to valuing nature 211
decisions. He argues that decisions should be made by the ‘public and rational
discussion about the common good, not the isolated act of voting to private
preferences’ (Elster 1983: 35). CJs facilitate this type of rational discussion more
than the use of economic appraisal methods.
Despite the advantages that CJs appear to have over the CVM, they do not
provide an economic estimate of the value of any particular project, nor whether
it constitutes an efficient use of resources. These are important weaknesses. A
third approach therefore seems potentially desirable, which allows deliberation,
but also the estimation of the economic value of a project. Such a method might
aim to combine the strengths of the CVM and CJ. This chapter reports on an
application of three approaches to one decision-making context, a proposed
floodplain restoration project in the south of Scotland. A CVM, a CJ and a
combined approach (the ‘valuation workshop’) were conducted to assess the value
of the project to local communities. The valuation workshop was an attempt to
combine the quantitative outputs of the CVM with the participatory, deliberative
and preference construction attributes of the CJ.
The rest of the chapter is organised as follows. We briefly describe the context
of the three approaches. Then we report on the design and the results of the
CVM. The next section describes the design and results of the CJ, and the following
section relates to the workshop approach. Then we assess the methods, and look at
the implications of the CJ and valuation workshop for dealing with problems related
to the CVM. The chapter concludes with a discussion of the complementarity of
the three approaches.
Table 9.1 Descriptive statistics for WTP for the Ettrick project
Mean Std dev Median Range 95% Confidence intervals N
(£ per person)
13.18 69.71 1 0–1,000 4.15–22.1 232
Three approaches to valuing nature 215
the success of such projects be determined? Clearly, ‘economic efficiency’ could
be one answer to either of these questions, but may not be a priority for participants.
To provide information with which the jury could assess the project, ten
witnesses attended sessions to provide ‘evidence’ and to answer questions. The
witnesses were selected in consultation with the Borders Forest Trust, and in
discussions with stakeholders from all sides of the debate to ensure balanced
representation. The witnesses came from a variety of backgrounds including
Scottish Natural Heritage, the local council, the Forestry Commission, the Scottish
Tourist Board, and environmental consultants. Witnesses made short
presentations to the jury of 10 to 15 minutes followed by a discussion session
with the jury of about 30 to 40 minutes. In addition to sessions involving witnesses,
the process included a number of jury-only sessions, where the jurors discussed
particular issues as a whole unit, or in smaller groups using participatory appraisal
techniques. Participatory tasks were designed to including listening, hearing and
watching. According to Pretty et al. (1995: 24) participants are likely to remember
‘10 per cent of what they read, 20 per cent of what they are told, 30 per cent of
what they see and 50 per cent of what they hear and see’. Although participants
did not need to remember long term what went on in the session, ensuring they
were engaged in the tasks was vital. The use of brainstorming techniques,
organising comments written on cards, and using large flip-chart paper served a
number of purposes: to provide a focus for attention while the discussion ensued;
to stimulate discussion between people with different levels of numeracy and
literacy; to re-enforce spoken issues; to provide a means of cross-checking what
was said, and to record the proceedings. The final recommendations were
achieved entirely by discussion and consensus, and approved by all of the jurors.
A report on the process and outcome of the jury was written and sent to the
jurors for approval, before being sent to the Borders Forest Trust and other
interested parties (Kenyon 1999).
The jury made recommendations about how the Ettrick project should be
managed and co-ordinated in order to achieve the environmental and social
objectives which the jurors felt were desirable in southern Scotland. The jury also
noted a number of aspects related to the project they felt were valuable (Table
9.2). Interestingly, not only environmental issues were identified, but also aspects
which might be classed as social values, such as community involvement, and the
role of the project for education.
In addition to the aspects of value identified in Table 9.2, the jury noted two
areas of concern, namely access and future management. Jurors felt that visitors
should be allowed access to the site, but were concerned about the damage they
might inflict. Specific recommendations were made regarding arrangements to
promote benign access, and the provision of information to visitors to encourage
careful use of the site. The second area of concern related to future management
of the site and the availability of funding to secure it. The jurors felt that it may be
possible to start a trust fund dedicated to the Ettrick project to make sure that
money was still available for future management. However, after speaking with a
member of the local community, the jurors became less concerned as they were
216 Wendy Kenyon and Nick Hanley
Table 9.2 Positive issues identified with the forest floodplain project
Environmental issues Social issues
Preservation of a natural ecosystem – Good demonstration scheme for copying
a world resource
Flood control Education of the young and encouraging
educational studies
Balance of different habitats and getting Decrease the number of sheep and fencing
back to nature of sensitive areas
Encouraging wildlife and preservation of Getting rid of ugly blanket plantation
indigenous life forms forestry
Monitoring of species Community involvement
assured that the local community was fully involved in the project and that they
would ensure that the site was well managed into the future.
The jury were able to look at the Ettrick project in the wider context of the
south of Scotland, and made a number of recommendations regarding the
management of individual environmental projects in the region. They felt that a
variety of projects were needed in the region, which met a number of different
needs, but that all of these projects should be co-ordinated in an integrated way.
For example, some environmental projects might aim to attract tourism to the
area, but such projects must be situated in less environmentally sensitive areas.
Others might aim to increase biodiversity, but not aim to attract tourists. Finally,
the jury considered how the success of environmental projects such as the Ettrick
Forest Floodplain Project, might be assessed. Table 9.3 shows the criteria offered
by the jury, who also suggested that these measures could be considered as
elements of future project design and planning. It is interesting to note that the
jury considered environmental and social elements important in judging the
success of the project, but that they did not seem to consider economic criteria
important.
Each group was asked to provide a ranked list of good points related to the Ettrick
project. The second task was to consider the problems of the project, and the
concerns they had with it. Each group were asked to discuss problems, rank their
relative importance and offer suggestions of how each problem might be mitigated
or solved. The final task in this section of the workshop asked the participants to
make recommendations on how the success of the project might be judged. This
involved creating a list of means by which the project could be assessed, with
discussion surrounding each suggestion. In the third part of the valuation workshop
participants were asked to complete some further questions individually in a survey.
These questions asked whether the participants would change the WTP they stated
at the beginning of the workshop, and explain why. Participants had now had a
chance to discuss the project, assimilate information more fully and consider the
project from different perspectives, given the interaction with other participants.
The valuation workshops were conducted in December 1999 and took around
three hours each. Two workshops were carried out in each of two towns in the
Borders, giving a total of four workshops and 44 participants. In order to select
participants, 500 letters were sent out to a random selection of addresses in or
near each of the towns. The letter invited the addressee to one of the workshops,
explained that the Scottish Agricultural College was carrying out research on the
local environment and were interested in their views, and asked them to return an
enclosed form if they wished to participate. A pool of possible participants was
then drawn up from those who replied, and participants were chosen from this
pool to be as representative of the local population as possible. Each participant
was paid £20 for their attendance.
The workshop provided a range of qualitative and quantitative data on the
Ettrick project. Of the 44 participants attending the workshops, 27 (61 per cent)
made genuine CVM bids initially, the rest being either protest bids (5 per cent) or
‘don’t know’ responses (34 per cent) to the valuation question. After the discussion,
in the second part of the questionnaire, the number of valid bids increased as 2 of
the ‘don’t know’ respondents offered a new valid bid. Of the 44 participants, only
6 (14 per cent) changed their bid after the discussion (Table 9.4). Participants were
clearly considering their budget constraints carefully as the most common reason
given for not changing their bid was that their financial circumstances had not
changed and they could not therefore afford a higher amount even if they now
wanted to. Interestingly none of the participants revised their bid downwards.
Although the preceding part of the workshop comprised discussion of both the
good and bad aspects of the project, the discussion of the good points clearly
carried more weight.
218 Wendy Kenyon and Nick Hanley
Table 9.4 Original and revised bids with reasons for change
Original bid Revised bid Reason
25 50 Wish it would happen
10 20 It would have a beneficial effect on the local community,
Borders community, and is a valuable recreational and
educational asset
25 50 Awareness of the problem
20 ?? Would be willing to sponsor the project on a monthly
basis
DK 10 More knowledge
DK 25 Because of points for the project brought up in the
discussion
Table 9.5 and Table 9.6 display the descriptive statistics for pre-discussion and
post-discussion bids. Although the mean has increased, a two-tailed paired T-test
showed no significant difference between the two workshop mean WTPs at 10 per
cent significance (t = –0.57, p = 0.57).6 Although the discussion did have an impact
on individual bids, it did not therefore have a statistically significant impact on
mean WTP. A Mann–Whitney U-test gives the same results, the hypothesis that the
pre-discussion median is equal to the post-discussion median cannot be rejected (W
= 747.0, p = 0.6735 adjusted for ties).
In addition to this quantitative output, the valuation workshop provided
information about what the participants felt was good and bad about the project,
and how the project’s success might be measured. Environmental aspects of the
project were in all but one of the 8 groups (2 per workshop) considered the most
important. The one group who dissented felt that the promotion of education and
tourism in the area was most important. Table 9.7 shows the results from one
discussion group which is typical of the responses from other groups.
One strength of the valuation workshop approach relative to the CVM is that it
provides additional information on participants’ views of the project – both positive
and negative. Participants were asked not only to discuss issues which they thought
were problematic, but also to offer solutions to these problems. Table 9.8 shows the
typical concerns and the solutions offered by one of the workshop groups. Every
group considered the cost of the project to be a problem and many groups
differentiated between start-up costs and running costs. One of the other most
discussed problems was that of access. Roads to the site were single track and not
suitable for a large number of cars, and there was virtually no public transport
available to the site. Despite having been provided with very little information,
participants were able to offer solutions to these problems in most cases, clearly
some more practical than others (Table 9.8).
Finally, the participants were asked to suggest how the success of the project
might be measured. A range of ideas was suggested including monitoring flora and
Three approaches to valuing nature 219
Table 9.5 WTP descriptive statistics for initial valuation
Mean Median Std dev Range 95% CI N
11.07 10 12.59 0–50 6.09–16.05 27
fauna, evaluating visitor levels, jobs, the level of community involvement, and interest
in the site by educational establishments. As with the CJ, no direct financial criteria
were suggested here, although many of the criteria have an economic element,
such as jobs and visitor levels.
Conclusions
Both the CVM and CJ provide policy-relevant information, the information
provided by each being useful in different ways. The results of the CJ identify the
project needs and how it should develop. Table 9.2 lists those issues that the jurors
felt to be positive, and that might be used as objectives by the managers of the
project. Similarly, by identifying concerns relating to the project, jurors provide
direction to managers and policymakers as to the development of the project. The
information provided by the jury therefore plays a practical role in directing the
management of the project.
The results of the CVM provided different, but equally policy-relevant
information. The CVM is able to measure both the intensity and direction of
220 Wendy Kenyon and Nick Hanley
Table 9.8 Project problems, solutions and the importance of bad aspects
Problems Rank Solutions
Costs – who pays 1 Start-up cost: central funding, lottery, EU
Running cost: 3 for admission fee, 3 against
Prevent damage by visitors
Rural depopulation – 2 Change/expansion of jobs related to the
reduction of number exercise, e.g. catering, maintenance, paths,
of viable farms bridges, etc.
Rangers
Tourism facilities
Farm shop/equipment hire shop
Access roads 3 Straighten roads
Bridges
Improved public transport, i.e. minibuses
Agreement with 4 No solutions offered
landowners
Damage by visitors 5 Designated picnic area and toilets
Designated play area
Car parking 6 No solutions offered
preferences, and provides data that is consistent with welfare economic analysis
which can feed into cost-benefit analysis. The benefit of the project estimated at
£568,677 could be compared with the costs of the project at £350,000. The project
passes the cost-benefit test, and thus offers a potential Pareto improvement.
The workshop approach is an attempt to amalgamate aspects of the two methods
and provide both practical guidelines to the project managers and cost-benefit
estimates, whilst maintaining the participation of the local community in the project
evaluation. The results of the workshop are comparable on the one hand with the
CVM results and on the other with CJ results. A T-test shows that neither the pre-
discussion workshop mean nor the post-discussion workshop mean WTP are
significantly different from the CVM mean bid at 10 per cent level (t = –0.41, p =
0.68 and t = 0.01, p = 0.99 respectively). However, a Mann–Whitney U-test does
shows a significant difference between the medians (W = 4228.5, p = 0.0350 and
W = 4265.5, p = 0.0268 respectively). The results of the discussion of the workshop
correspond quite closely to the recommendations provided by the jurors.
Workshop participants were also asked to evaluate the final aggregate CVM
result at the end of the session. The calculation was explained to the group, and
the final figure presented to them. The comments from all the groups can be split
into four categories. First, that it is not possible to put a value on such a project.
Second, comments surrounded whether or not this aggregate figure would be
obtained in reality, as due to the poor economic situation in the Borders there
were more important things to spend money on. Third, the clear feeling that money
for such projects should come from government, lottery or EU funds was expressed.
Finally, there was discussion that information on costs (which was not made
Three approaches to valuing nature 221
available) was required before any pronouncement was made on its value. This
additional feature provides interesting information on the CVM figures, and on
the authenticity of the economic estimates. Overall the valuation workshop
approach goes some way to developing a method which provides economic
estimates of environmental projects whilst adding value to traditional CVM
exercises, and offering qualitative recommendations. However, the workshop results
suffer in this study from a small sample size relative to the CVM survey.
The research also related to problems identified with the CVM in the literature,
and the role that more deliberative methods may have in mitigating these concerns.
One relates to the provision of information to CVM respondents. The respondents
to the CVM survey were given visual and verbal information about the project
site, and asked, ‘Do you prefer the site with or without the project?’ Thirteen per
cent of respondents did not know whether they preferred it with or without. This
may indicate that the information was not sufficient for them to be able to determine
their preference. This did not appear to be a problem in the CJ. Jurors were all
able to determine their own preferences regarding the Ettrick project, and further
were able to identify those aspects they preferred most (Table 9.2). The workshop
participants were given the same information as the CVM respondents, but were
allowed to discuss it in a social setting. The fact that two participants changed their
WTP bid from a ‘don’t know’ response to a positive response implies that the
discussion did allow the participants to better understand the project they were
asked to pay for.
A number of researchers suggest that CVM questionnaires ask the wrong
question and assume that respondents act as consumers and not citizens when
responding. Our results indicate that this assumption may be valid (Sagoff 1998;
Blamey 1996). Analysis of the CVM results showed that the most significant variable
in influencing whether a respondent would be willing to pay anything was whether
she was likely to visit the site if the project went ahead (Kenyon and Nevin 2001).
This may indicate that the respondents acted as consumers and not citizens when
responding to the questionnaire.7 If this is the case, the CJ can be seen as a comple-
ment to the CVM by evaluating the project from a citizen’s point of view. The
valuation workshop incorporates elements of both a citizen and consumer
approach. The initial CVM responses were from a consumer standpoint, but the
output from the discussion was clearly from a citizen standpoint, since community
issues were considered an important aspect of the Ettrick project. However, this
result also suggests some relationship between a ‘consumer’ standpoint and the
respondent accounting mainly for use and option values (and not other components
of total economic value). It is clear that further research is required on the relation-
ship between these two issues.
The issue of value construction is also a matter of debate within the environ-
mental valuation literature. The CVM results provide no evidence regarding the
detailed construction of final values. However, comments from jury participants
in the evaluation discussion and questionnaire at the end of the process suggested
that breaking down the decision-making process into tasks made the whole task
less daunting, and more manageable. The by-product of these tasks is to provide
222 Wendy Kenyon and Nick Hanley
information that may indicate how jurors constructed their final response. The
evidence of the valuation workshop is mixed on this point. Whilst the pre-discussion
WTP was constructed in the normal CVM way, the post-discussion response had
the benefit of discussion in a social setting, and the outcome of the discussion may
be an indicator of the reasons for a (limited) changed bid.
Finally, in a climate where both national and international agreements exist
which seek to enhance public participation in environmental decision-making, it
seems clear that the process should rely on more than just economic estimates of
value as provided by traditional CVM surveys. Whilst such estimates are still useful,
policy-makers are increasingly required to incorporate decision-making, planning
and management into one process, and include both expert and lay opinion. New
methods which are able to provide a more holistic approach to environmental
policy are needed so that these national and international targets can be met. The
CJ is able to offer such an integrated approach. Evidence from the CJ suggests
that the jurors were able to think holistically, offering comment on benefits and
pitfalls of the project and a wide range of management recommendations. One
of the recommendations was that the Ettrick project should not be considered in
isolation, but as part of a suite of projects, to ensure that the Borders environment
developed in an integrated way. The valuation workshop also provided a more
integrated approach, in which participants were able to complement economic
estimates with wider indicators of values and preferences.
In many instances where environmental project evaluation is required,
conducting both a CVM and CJ would be prohibitive in terms of time and money.
The challenge for the future may therefore lie in developing complementarities.
Building on the valuation workshop approach, and increasing sample sizes may
go further in offering a more appropriate combined approach, so that decision-
makers can benefit from both a wide range of policy-relevant information when
evaluating environmental projects.
Acknowledgements
This work was funded by the Economic and Social Research Council under the
Global Environmental Change Programme, and the Scottish Executive Rural
Affairs Division. The authors would also like to thank the Borders Forest Trust,
and the people of the Borders for their input and enthusiasm.
Notes
1 Although self-selection of the participants in a CJ could mean that those with higher education
and income, for example, may be more likely to participate and the preferences of the ‘excluded’
may not be registered.
2 How much money (on a once-only basis) would your household be willing to donate to the fund
in order to ensure this project went ahead? Please bear in mind that this money would only go
towards paying for the management of the natural environment in the Ettrick Valley. You might
also like to think about the spending on other items that you might have to give up if you did
make this payment.
Three approaches to valuing nature 223
3 More detailed statistical analysis of the CVM results can be found in Kenyon et al. (2000).
4 Figure obtained from the Borders Forest Trust.
5 The Borders Jury was selected in this way to minimise recruitment costs. Despite the final CJ
being a reasonable representation of the Borders population (when compared by socio-economic
characteristics), this is probably not the best means of jury selection. Selection processes utilising
a random draw from the electoral register are considered to provide a better sample of jurors.
6 The two respondents who changed from a ‘don’t know’ to a positive bid were taken out of the
sample for the purpose of the test.
7 It could be argued that the most important feature of the project was therefore its use or option
use value.
References
Aldred J., and Jacobs, M. (1997) ‘Citizens and wetlands. What priority, if any, should be
given to the creation of wetlands in the fens?’, report of the Ely citizens’ jury, Centre
for the Study of Environmental Change (CSEC), Lancaster: Lancaster University.
Arrow K., Solow R., Portney P.R., Leamer E.E., Radner R. and Schumann H. (1993)
‘Report of the NOAA panel on contingent valuation’, Federal Register, 58: 4601–14.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and the
expression of citizen values and issue-options’, in W.L. Adamowicz, P.C. Boxhall, M.K.
Luckert, W.E. Phillips and W.A. White (eds) Forestry Economics and the Environment, Oxford:
CAB International.
Boyle, K.J. and Bergstrom, J.C. (1999) ‘Doubt, doubts and doubters’, in I. Bateman and
K.G. Willis (eds) Valuing Environmental Preferences. Theory and Practice of the Contingent Valuation
Method in the US, EU and Developing Countries, Oxford: Oxford University Press.
Boyle, K.J., Johnson, F.R., McCollum, D.W., Desvouges, W.H., Dunford, R.W. and Hudson,
S.P. (1996) ‘Valuing public goods: discrete versus continuous valuation responses’, Land
Economics, 72: 381–96.
Brown, T.C., Peterson, G.L. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71: 250–60.
Carson, R.T., Groves, T. and Machina, M.J. (1999)‘Incentive and informational properties
of preference questions’, paper presented at the European Association of Resource
and Environmental Economists Conference, Oslo, June 1999.
Coote, A. and Lenaghan, J. (1997) Citizens’ Juries: Theory into Practice, London: Institute for
Public Policy Research.
Crosby, N. (1995) ‘Citizens juries: one solution for difficult environmental questions’, in O.
Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation,
Dordrecht: Kluwer.
Dechamps, H., Joachim, J. and Lauga, J. (1987) ‘The importance for birds of the riparian
woodlands within the alluvial corridor of the River Garonne, SW France’, Regulatory
Rivers Restoration Management, 1: 301–16.
Elster, J. (1983) Sour Grapes. Studies in the Subversion of Rationality, Cambridge: Cambridge
University Press.
Gregory, R., Flynn, J., Johnson, S.M., Sattefield, T.A., Slovic, P. and Wagner, R. (1997)
‘Decision-pathways surveys: a tool for resource managers’, Land Economics, 73: 240–54.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision
making institutions’, in J. Foster (ed.) Valuing Nature? Economics Ethics and Environment,
London: Routledge.
James, R.F. and Blamey, R.K. (1999) ‘Public participation in environmental decision-making
– rhetoric to reality?’, paper presented to the International Symposium on Society and
Resource Management, Brisbane, Australia, July 1999.
224 Wendy Kenyon and Nick Hanley
Kenyon, W. (1999) Report of the Galasheils Citizens’ Jury, Edinburgh: Scottish Agricultural
College.
Kenyon, W. and Nevin, C. (2001) ‘The use of economic and participatory approaches to
assess forest development: a case study in the Ettrick Valley’, Forest Policy and Economics,
3(1–2): 69–80.
Macmillian, D., Smart, T.S. and Thornburn, A.P. (1998) ‘The importance of realism to
experiments comparing cash and CV charitable donations: the case of the isle of Eigg
Trust’, paper presented at the Agricultural Economics Society Conference, Reading,
March 1998.
McGhee Smith Associates (1995) ‘Wild rivers. Aspects of the location and ecology of the
floodplain woodlands in Scotland’, report to World Wildlife Fund.
Mitchell, R.C. and Carson, R.T. (1989) Using Surveys to Value Public Goods: The Contingent
Valuation Method, Washington, DC: Resources for the Future.
Munro, A. and Hanley, N. (1999) ‘Information, uncertainty and contingent valuation’, in
I. Bateman and K.G. Willis (eds) Valuing Environmental Preferences. Theory and Practice of the
Contingent Valuation Method in the US, EU and Developing Countries, Oxford: Oxford University
Press.
Payne, J.W. and Bettman, J.R. (1999) ‘Measuring constructed preferences: towards a building
code’, Journal of Risk and Uncertainty, 19: 243–70.
Peterken, G.F. and Hughes, F.M.R. (1995) ‘Restoration of floodplain forests in Britain’,
Forestry, 68: 187–202.
Petts, G. (1990) ‘Forested river corridors: a lost resource’, in D. Cosgrove and G. Petts,
(eds) Water, Engineering and Landscape: Water Control and Landscape Transformation in the Modern
Period, London: Belhaven Press: 12–34.
Pretty, J., Guijit, I., Thompson, J. and Scoones, I. (1995) Participatory Learning and Action: A
Trainers Guide, London: International Institute for Environment and Development.
Ready, R., Buzby, J.C. and Hu, D. (1996) ‘Differences between continuous and discrete
contingent valuation estimates’, Land Economics, 72: 397–411.
Sagoff, M. (1988) The Economy of the Earth: Philosophy, Law, and the Environment, Cambridge:
Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24: 213–30.
Schkade, D.A. and Payne, J.W. (1994) ‘How people respond to contingent valuation
questions: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26(1): 88–109.
Stewart, J., Kendall, E. and Coote, A. (1994) Citizens’ Juries, London: Institute of Public
Policy Research.
United Nations (1993) The Global Partnership for the Environment and Development. A Guide to
Agenda 21, post-Rio edition, New York: United Nations.
Yon, D. and Tendron, G. (1981) Alluvial Forests of Europe, Nature and Environment Series
200, Strasbourg: Council of Europe.
Deliberation and economic valuation 225
Introduction
Among the wide variety of methods available for incorporating community attitudes
and values into environmental decision-making are public hearings, calls for sub-
missions, referenda and opinion polls. Each of these approaches has its limitations.
Concerns also surround the use of stated preference techniques such as the
contingent valuation method (CVM) and choice modelling as inputs to environ-
mental decision-making, in part because responses are based on limited levels of
information and deliberation. Techniques based on the theory of deliberative
democracy have the potential to correct these adverse conditions. Citizens’ juries
(CJs), deliberative polls, consensus conferences and related techniques offer
participants time, information and opportunities for social interaction and
deliberation, which may result in the development and articulation of robust
preferences.1
Under the tenets of participatory democratic theory, public participation is
functionally, logically and ethically central to the operation of democracy (Webler
and Renn 1995). As Thomas-Slatyer (1995: 11) noted (in relation to community
participation in project development in developing countries): ‘Central to these
approaches is the belief that ordinary people are capable of critical reflection and
analysis and that their knowledge is relevant and necessary.’ Public participation is
required in order to ensure that the aspirations and needs of citizens are heard by
decision-makers and is considered to produce many benefits including definition
and communication of the public will (Rousseau 1968), popular sovereignty
(Rosenbaum 1978), political equality (Rosenbaum 1978, Van Valey and Petersen
1987), protection of citizens’ interests (Rosenbaum 1978), definition of the public
will (Dienel 1989) and enhancement of personal (Daneke et al. 1983) and social
(Rosenbaum 1979) development. An experimental application of the CJ method
to economic valuation is described in this chapter. The theoretical basis of the
method is first described, followed by a detailed description of this application of
the method. Some of the methodological and theoretical issues which arose from
this work are discussed.
226 Rosemary F. James and Russell K. Blamey
Public participation through cost–benefit analysis
There are numerous methods commonly used in developed countries to allow
public input into environmental decision-making, including methods in which
respondents are wholly self-selected (such as calls for submissions) and those in
which potential respondents are identified by some other agency (such as opinion
polls). Another form of public participation is provided by the market, where
people’s preferences are reflected in their market transactions. In the absence of a
market, surrogate markets and economic valuation surveys may provide information
regarding the preferences of citizens. An assessment of the value (benefits or costs)
accruing to the community from an actual or proposed course of action is an
important component of environmental cost–benefit analysis (CBA), of which the
objective is to inform decisions on the allocation of resources so as to maximize
the excess of benefits over costs to society as a whole.
The basis of the practice of environmental CBA lies in a form of preference
utilitarianism in which it is assumed that preferences reflect utility functions, and
willingness to pay (WTP) or willingness to accept (WTA) is taken as an indicator
of strength of preference. The democratic basis of environmental CBA and
associated estimates of community value can thus be summarized as ‘one dollar,
one vote’. Environmental CBA involves an assessment of the implications for
community environmental preferences, typically in WTP terms, of proposals under
investigation, which may have impacts on the natural environment. Stated
preference methods alone are potentially capable of estimating the full range of
environmental values, while revealed preference methods only provide information
on direct use values (see Chapter 1).
Stated preference surveys involve eliciting information pertaining to strength
of preference through the use of community surveys in which respondents are
requested to state what they would do under prescribed circumstances. Respondents
are questioned in such a manner as to elicit their WTP for a programme involving
some potential environmental improvement (or prevention of some potential
environmental deterioration) or WTA compensation for a programme involving
some environmental deterioration or lack of environmental improvement. The
two most commonly used stated preference methods are the CVM and choice
modelling. The CVM, and to a lesser extent choice modelling,2 have been criticized
on several grounds.
One argument is that ‘stated preference surveys represent a highly constrained
environment for preference construction’. Whilst some respondents to stated
preference questionnaires may reveal their pre-existing preferences for the attributes
and/or goods and services under examination, it is probably more common for
respondents to construct their preferences at the time of completion of the
questionnaire. Environmental goods and services are complex and often unfamiliar,
and the method chosen to elicit WTP should allow respondents the opportunity to
articulate their values for the goods and services under consideration and then to
construct their preferences. Contingent valuation in particular is not seen as a
method which permits this to occur (Gregory et al. 1993; Gregory and Slovic 1997;
Deliberation and economic valuation 227
Sagoff 1998). CVM respondents are required to make choices in the face of
necessarily limited time and information and in the absence of the capacity to
seek clarification of many of the issues of concern, which may severely constrain
the construction of informed and deliberated preferences.
Stated preference surveys, in common with other types of surveys, are subject
to coverage, sampling, non-response and measurement error. Much, if not most,
of the controversy surrounding techniques such as CVM and choice modelling
focuses on measurement errors, particularly those attributable to respondents and
their interactions with the questionnaire. Sources of such error include yea-saying
and nay-saying (when respondents answer positively or negatively to valuation
questions without regard to the amount involved in the scenario), the non-
identification of protest responses (objections to the use of a tax or levy to fund
environmental improvements), strategic responses and amenity misspecification
biases. Methodological criticisms of the CVM include the production of
unacceptably large estimates for total WTP or WTA and the insensitivity of WTP
or WTA estimates to price and scope variations (Perman et al. 1999).
There are a range of philosophical concerns associated with environmental
valuation surveys. These relate both to the use of the methodology itself and to
respondent motivations and decision processes (Blamey and Common 1999). The
‘one dollar, one vote’ assumption of environmental CBA raises the question as to
whether the wealthy should have more say than others. Ethical concerns are raised
regarding the valuation of the non-monetary components of environmental goods
in monetary terms (Blamey and Common 1992; O’Neill 1997; Gregory and Slovic
1997). The potentially reductionist and atomistic nature of stated preference methods
is seen as problematic. Alternative methods which are more pluralistic in terms of
community perspectives and values may offer a more appropriate and legitimate
input into environmental decision-making. Environmental issues may then best be
addressed by individuals acting as citizens with a societal perspective, rather than as
consumers acting on the basis of individual preferences (Sagoff 1998).
The potential for use of the fourth approach, deliberative valuation, has been
discussed under various names by Brown et al. (1995), Jacobs (1997) and Sagoff
(1998), but no examples of the practice appear to be reported in the literature to
date (although see other chapters in this volume). The conduct of a stated preference
method in the context of a CJ has the potential to reduce some of the difficulties
associated with stated preference methods. The use of a discursive rather than
questionnaire-based stated preference valuation method in a CJ may provide
additional benefit.
Deliberation and economic valuation 229
Design and conduct of this application
The CJ reported in this chapter was conducted as a research application of the
technique, rather than commissioned by a government agency in response to an
existing environmental issue. The specific topic examined in the CJ was the
management of national parks5 in the most populous Australian state, New South
Wales. This CJ formed part of an examination of the potential for use of the
technique in environmental decision-making in Australia.
There are several ways in which the CJ technique can be applied to non-market
valuation, i.e. used for deliberative valuation. These are illustrated in Figure 10.1.
The suite of work undertaken in this project is shown in italics. The CJ reported
herein included the following elements: a conventional option choice charge which
did not involve deliberative valuation; two approaches to deliberative valuation,
involving a further charge in which the jurors were asked to determine societal
WTP for a specified programme involving environmental improvement; and a
stated preference (choice modelling) survey, completed by the individual jurors
before and after the CJ proceedings. The principal focus of this chapter is on the
first of these deliberative valuations, in which a process of group discussion and
decision-making was used to arrive at a valuation (WTP) for a posited programme
involving environmental improvement.
A scenario based on consideration of five major national park management
activities was developed and then refined using focus groups. The scenario specified
options differing in terms of the levels of effort allocated to the following key
national park management activities: fire management, weed control, feral animal
control, maintenance of visitor facilities and management of historic sites.6 The
relationships between park management activities and biophysical attributes such
as wildlife and native plants were detailed in an explanatory document provided
to the jurors. The same scenario format was used for both the conventional and
deliberative valuation charges and the choice modelling questionnaire. Further,
the same base allocation of effort (Option 1), representing the current situation,
was used in all three approaches. The charges are shown in Table 10.1 and
Table 10.2.
The jury was given a set of information and asked which of three options they
preferred before the introduction of a fourth option. They were told that the
National Parks and Wildlife Service manages about 5 million hectares of parks in
New South Wales (6.3 per cent of the state) in 479 parks and reserves. This includes
national parks, nature reserves, historic sites, state recreation areas and regional
parks. They were told that the term ‘national park’ included all these different
types of parks and reserves. They were told that budgets for national park
management are always tight and that there would never be enough money to do
all the things which might be required to meet the expectations of the public
about national park management. A possibility was described as to continue
spending the existing budget as shown in Figure 10.2 and in Option 1. Further
information about these management activities was provided in a pamphlet.
Another possibility described was to keep the total amount of money spent on
230 Rosemary F. James and Russell K. Blamey
Citizens’ Jury
Framing of charge
Jury to estimate WTP for a
specified option
Jury to select preferred
option from several given;
each option includes a cost
per potential payer
national park management the same, but change how it is spent. For example,
some money could be taken away from maintaining facilities and spent on feral
animal control. This is what is proposed in Options 2 and 3 of the charge.
The jurors were told that another possibility was to raise more money to spend
on national parks. State governments could introduce a park management levy to
help pay for increased levels of current activities. The levy would be paid by all
Deliberation and economic valuation 231
Table 10.1 Charge 1: CJ group preference among three options
Outcomes of national park management Option 1 Option 2 Option 3
current
situation
Number of national parks with good fire 100 40 160
management
Area of feral animal control each year 50,000 ha 100,000 ha 30,000 ha
(hectares)
Area of weed control each year (hectares) 3,000 ha 1,000 ha 10,000 ha
Proportion of facilities that are well 35% 45% 25%
maintained
Number of well-protected historic sites 7,000 6,000 7,500
those who live in NSW and pay income tax, and would be a set amount per year,
paid as part of the income tax return system. They were then told: ‘Another way
of raising extra funds would be to introduce entrance fees for more national parks
or to increase those fees that already exist. Unfortunately, opportunities to
significantly increase the funds raised from these fees are very limited. The costs
of collecting entrance fees are often higher than the money collected. For this
reason, entrance fees are not considered in this charge. We have put together a
new option, Option 4, that involves a substantial increase in expenditure for all
five park management activities.’ The results of the straw preference poll were
used in establishing the priorities for this option. The jury was told that what
Option 4 would cost to implement is unknown. However, it would cost a lot more
money than the NSW National Parks and Wildlife Service currently has to spend
on these activities. For the purposes of this scenario, jurors were asked to assume
that the only way that Option 4 could be achieved is via the introduction of a park
management levy. They were then asked ‘How high would a park management
levy have to be before the jury would recommend Option 1 rather than Option 4?
232 Rosemary F. James and Russell K. Blamey
Protection of historic sites
and items (13%)
479 parks covering
5 million hectares
Feral animal control (6%)
In other words, how high would the levy have to be before the NSW public would
be no better off under Option 4 than Option 1?’
The population relevant to the scenario was identified as the 3.8m population
of the Australian state of New South Wales between 20 and 69 years of age. The
jurors were selected using a two-phase telephone survey, in which various
demographic and attitudinal data were collected. A sample was selected to be as
representative as possible of the population on the basis of the latest available
Australian Bureau of Statistics data on the following criteria: gender, age, place of
residence, ranking of the environment in relation to other social issues, occupation,
income, income source and education. One selected juror failed to attend.
Jurors were provided with written instructions regarding both their roles and
the roles of witnesses and with training in note-taking and strategic questioning.
All travel, meal and accommodation costs were met and the jurors were paid a
small sum (A$100 per day) to compensate them for out-of-pocket expenses incurred.
Five witnesses with expertise in particular aspects of the charges were selected (i.e.
fire, weeds, feral animals, facilities and historic sites), in addition to two general
witnesses on national park management. Witnesses were provided with written
instructions regarding the content of their presentations, their roles and the roles
of the jurors in the jury process.
The jury met over the period 5–7 October 1999. The first day of the process
was devoted to preparation of the jurors and the jury for their roles over the
following two days. The second day was largely occupied by presentations from
and questioning of the witnesses; on the third day, the jury consulted two further
witnesses and reached decisions on the two charges. The second charge was
presented to the jury only after a decision had been reached on the first charge.
Option 4 in the second charge was developed from the results of an initial
confidential identification by the jurors of their individually preferred option from
the first charge. The jurors were aware that Option 1 in each choice set represented
the current situation. They were also informed that the results of this experimental
application of the CJ method would be relayed to the NSW National Park Service
for use in budgeting and in the design of public participation programmes.
Deliberation and economic valuation 233
Little has been documented about appropriate decision-making methods for
use in citizens’ juries; there does not appear to be a method consistently applied as
part of the CJ process. It was necessary, therefore, to develop a consistent, robust
method for facilitating group consensus, which would prevent false or forced
consensus and would allow for alternative strategies if a true consensus could not
be reached. In this work, different decision-making methods and decision rules
were used for the two charges. The processes followed are outlined at Figure 10.3
and Figure 10.4. For the first charge, jurors were asked to choose their preferred
option privately and to provide written reasons for this choice. The results of this
work were then displayed without citation of the source of the individual comments.
This confidential approach was adopted in order to ensure the jurors first
determined their own preferences and then considered the reasons cited by others
for their choices without consideration of their authorship. This was done in order
to avoid contamination of the ideas by jurors’ perceptions of the status of and
personal relationships with the author. As anticipated, this was effective for a short
time, after which the individual jurors in the minority (those who did not select
Option 1) declared their choices and the discursive process commenced. Assessment
of support for each option was then discussed openly by the group and consensus
reached. When the second charge was presented, the approach adopted for the
first charge was modified in anticipation of the need for the jurors to express their
views regarding both the proposed introduction of some type of payment and the
payment type proposed. Following considerable discussion by the jurors, public
polling on the options was conducted in several stages.
Results
After their initial and independent consideration of the options, nine jurors chose
Option 1, one chose Option 2 and three chose Option 3 as their preferred option.
After consideration of the reasons for the choice and discussion amongst the jurors,
the jury reached consensus and selected Option 1 (the status quo) as the preferred
option. The outcome was stated by the jury as follows:
1 that the feral animal, weed and fire experts be told that we can see benefits,
based on their separate presentations to us, in their getting together. Some
of the programs have some common benefits, and we think they could all
be more efficient as a result of that liaison.
234 Rosemary F. James and Russell K. Blamey
Consensus reached
by jury
2 that we are all very concerned about the adequacy of the programs in
Option 1, and that the level of funding which permits that type of com-
bination of programs is inadequate. We would like to send a strong message
to the politicians that, as we are a representative jury in the community,
there ought to be more funding, which could be put to better use in
developing these programs.
3 A sub-issue under that (relates) to the National Parks and Wildlife Service –
it is possibly a little unfair because we have not heard from anyone from
the National Parks and Wildlife Service – is that we can see from what the
experts told us that better management would possibly enhance the way in
which these programs work.
Deliberation and economic valuation 235
Framing statements
by the facilitator and project team members
Discussion re tax –
occurrence
methods of collection
by jury and facilitator
The second charge required the jury to undertake a deliberative valuation. The
jury entered into considerable debate, in stages as follows: (i) discussion of the
charge and its interpretation – some jurors had initial difficulty with the idea of
maximum willingness to pay; (ii) clarification regarding the timing of collection of
the proposed levy (weekly or annual) and the eventual use of the funds collected;
(iii) discussion about the way in which the levy should be determined; one juror
put a contribution model7 in the jurors’ minds by stating ‘If we don’t know how
much it is going to cost for each one of those (activities), how do we know the
amount of the levy to impose to cover the costs?’;8 (iv) discussion about the
acceptability of the levy to jurors, the type of levy (flat or scaled rate) and other
potential sources of money to fund Option 4; and (v) discussion amongst pro-levy
jurors about an appropriate type and amount of levy. At this point, in a public
poll, the jury voted nine to four in favour of Option 4 (the ‘with levy’ option).
Further discussion followed regarding the nature of the levy, during the course of
which one additional juror came to support a levy. A lengthy discussion then
followed about appropriate fixed levy amounts. The jurors who now supported a
levy eventually voted ten to zero in favour of a proportional rather than a fixed
price levy. The levy was to be calculated as a percentage of gross income. Following
a discussion of appropriate percentage figures, a vote was taken to determine
support for the two figures which had been proposed by the jurors, 0.1 per cent
236 Rosemary F. James and Russell K. Blamey
and 0.25 per cent of gross income. A majority rule preference of a 0.1 per cent
levy on income tax resulted, by a vote of eight to two. Three jurors maintained
their objection to the use of a levy. Using the proposed levy of 0.1 per cent, an
additional A$109.7m (1997/8 income data), equivalent to €70.2m, would be
collected annually for use on national park management. Total annual expenditure
on the five programmes was A$15.2m in 1999–2000.
Conclusions
The conduct of this deliberative valuation exercise has raised a number of
significant methodological and theoretical issues which require resolution in order
that the method may be rendered more rigorous. These include methodological
issues covering consensus formation, decision rules for polling or voting, equality
of juror impact, and provision of information. Theoretical issues which arose
concern framing, representation, non-consensual outcomes, and the economic
interpretation of the results.
The development of a consensual decision is considered the ideal outcome of
a CJ, although alternative approaches can be accommodated. The selection of an
appropriate concluding process is necessarily strongly influenced by the wording
of the charge; as is illustrated by the case studies discussed by Coote and Lenaghan
(1997) it may involve voting, listing of information or some other method. The
importance accorded to consensus in the CJ method suggests a need for clarity
and exactitude in identifying consensus. When is a consensus a consensus?
Consensus requires that the eventual decision is agreed to by all members of the
group (Center for Conflict Resolution 1981). As this jury considered the first charge,
it was apparent that jurors differed markedly in their degree of participation in
the discussion of the options. This is hardly surprising; the same situation pertains
to most circumstances where a group is given a complex task to complete. However,
the task in a CJ is to reach a consensual decision; thus there should be clear evidence
that the group decision was willingly accepted by all and that apparently consensual
outcomes are not false. Factors which may lead to false consensus include: compliant
behaviour by jurors whose initial views are in the minority, driven by, for example,
a need to accord with a group norm or with the views of particular jurors
(conformity bias); and failure to participate in the process by individual jurors (due
to, for example, lack of communication or interpersonal skills, or to irritation or
boredom with the process). A protocol to reach consensus and robust methods of
testing the validity of apparently consensual outcomes are required in order to
prevent acceptance of false consensus. Such testing should necessarily be capable
of distinguishing between the ‘workable agreements’ referred to by Dryzek (2000:
48), arrived at by participants agreeing to a common outcome for different reasons
(i.e. content-related) and an apparent but false consensus developed in response to
some failure within the process of a particular CJ in response to factors such as
juror boredom (i.e. process-related). One possible approach would be to examine
the decision-making processes of individual jurors by means of a post-decision
questionnaire. Further, juries may fail to arrive at a consensus. At what point should
Deliberation and economic valuation 237
the possibility of a non-consensual decision be revealed to the jury? If this is done
too early in the deliberation phase, abandonment of the consensus-seeking process
may result. If done too late, a false consensus may have been reached. In addition,
at what point in the deliberation and under what conditions should the jury be
encouraged to implement other decision rules? What alternative decision rules
should be used?
As noted above, if a jury fails to reach a consensus, alternative decision rules
are necessary. If these involve voting or polling, care is required to establish an
appropriate procedure, as the polling process may affect the outcomes of a CJ
significantly. There are four major methodological aspects of polling which
require consideration: the form of the question posed, the polling method used,
the structure of the deliberative and polling processes and the decision rule
employed. The question to be considered may be framed in one of several forms,
dependent in part on the construction of the charge. Possible approaches include:
a choice between ‘yes’/‘no’/‘abstention’; forced choice ballots, in which all
answers but ‘yes’ are taken as ‘no’; and various forms of rating and ranking
(Gastil 1993: 56–60). The polling method used – confidential or public polling –
may also influence poll results. The use of confidential straw or preliminary
polls (as used in initial consideration of the first charge in this CJ) would seem to
offer advantages, by allowing jurors to make their initial choice uninfluenced by
the choices of others. The use of a subsequent anonymous tabling of reasons for
these preliminary choices permits analysis of those reasons unclouded by social
dynamics associated with the status of and personal feelings towards the source.
The structure of the deliberative and polling processes may also influence the
poll outcome. For example, multiple polls separated by deliberative phases are
more likely to lead to consensus on controversial issues than a single poll, by
allowing those holding differing views to consider and reflect on both their own
views and those of other jurors. The decision rule to be applied also requires
consideration. Should a simple majority vote, some proportional threshold or
some other measure be used? The decision on this aspect of polling will clearly
exert a strong influence on the poll results.
Bohman noted that deliberation may be considered to have been successful
when ‘participants in the joint activity recognize that they have contributed to and
influenced the outcome, even when they disagree with it’. He considered that
equality, at a minimum, required ‘equal standing and effective voice’ for every
citizen, both in terms of access to deliberative processes and participation within
specific processes (Bohman 1996: 33, 35). A lack of within-process equality was
evident in this CJ. Preliminary analysis shows that, of the thirteen jurors, one
generally spoke only when directly questioned and several others made minimal
contributions to the discussions. Whilst it might be that these jurors were not
distressed by their low profile in the jury and their minimal impact on the outcome
(and may have in fact agreed with the decisions reached), their lack of contribution
suggests a less than ideal deliberative process. The sources of this within-process
inequality lie in the wide range of social and cognitive skills and capacities, and of
prior knowledge, found in a jury derived from effective sampling of a diverse
238 Rosemary F. James and Russell K. Blamey
population. To what extent it is desirable and possible to reduce this inequality is
a fundamental issue affecting the legitimacy of both deliberative and non-
deliberative public participation methods.
Two key aspects of the provision of information to jurors are: the appropriate
level of juror access to witnesses; and how best to deal with inaccurate or misleading
information. Witnesses will vary both in their personal appeal to the jury and the
level of interest engendered by their subject matter. Hence some witnesses may be
sought after by the jury to provide additional information. Equality of access
appears reasonable on procedural grounds. However, the purpose of the witness
presentations is to provide the jurors with the information they consider necessary
to address the charge. Hence, one could argue for unlimited jury access to each
witness. In adversarial citizens’ juries, however, interested parties who consider
themselves disadvantaged by relative lack of access to the jury could claim the
outcome was biased. On the other hand, refusing the jury access to additional
information from witnesses could be considered to subvert the jury process, by
forcing jurors to reach a decision based on potentially inadequate information.
Further, witnesses will vary in their capacity to provide information to the jury, in
terms of accuracy, relevance and pedagogic effectiveness. Means are required to
deal with these potential distortions in information delivery. The provision of a
neutral expert to assist the jury in assessing information provided by witnesses and
the use of some type of discovery process (as used in the legal system) have been
proposed as possible mechanisms for dealing with inaccurate, misleading or
irrelevant information (Crosby 1995). The effectiveness of these approaches has
yet to be assessed in practice.
A citizen rather than a consumer frame is considered appropriate for most
citizens’ juries, although it may produce results differing from the estimates of
Hicksian measures of welfare change sought for environmental CBA. A citizen
frame, for example, encourages participants to think about costs and benefits in
aggregate societal terms which may not fit well with the notion of individual WTP.
Further, to the extent that individuals do consider costs and benefits at the individual
level, they may adopt a contribution model rather than the purchase model typically
assumed in environmental economics. This is problematical for environmental
CBA, as are several other features of citizen responses to valuation questionnaires
(Blamey 1996, 1998). In the study reported in this chapter, the charges were framed
so as to have jurors provide WTP estimates according to a purchase model and
hence potentially of use to environmental CBA, whilst maintaining a citizen rather
than consumer perspective.
A CJ engaged in deliberative valuation may provide either a maximum WTP
or a minimum WTA figure for some course of action or programme involving
altered levels of environmental goods and services. How should that figure be
extended to reflect the societal valuation? The most obvious approach to this issue
is to extrapolate the figure across the relevant population, as was done for this CJ.
However, the question of the degree of representation of the sample arises in this
context. If the jurors are acting fully as citizens – behind a Rawlsian veil of ignorance
– and have perfect knowledge of the full range of relevant societal circumstances
Deliberation and economic valuation 239
and views, then representation is of limited importance in terms of determination
of the societal view on the issue under consideration.9 That is, if the construction
of the citizen frame of reference has been successful and the jurors are clearly
deliberating and deciding in a citizen rather than an individual consumer context,
the influence of juror characteristics may be low, if sufficient information about
the relevant circumstances of other citizens is available to them. Relevant knowledge
would include details of the costs of substitute and complementary goods and
services and the amount of discretionary expenditure possible for all those potent-
ially liable to pay for the posited environmental change. This is of particular
importance where the jury is not representative of the population in terms of
factors such as income which may exert a strong influence on WTP or WTA.10 If
this is not the case, however, and a jury is selected which is not representative of
the wider population in terms of income and other characteristics and has less
than perfect knowledge of societal circumstances, biased WTP or WTA estimates
may result. These estimates may not reflect the WTP/WTA which would apply to
the population under the same conditions. The importance of this effect will be
determined by several factors, including: the degree to which the jurors are acting
as citizens; their knowledge of the relevant circumstances of the population; and
the impact of any ignorance of relevant circumstances on the resultant collective
estimate of WTP or WTA.
Decisions reached via voting are unlikely to satisfy the tenets of neoclassical
economics because the strength of each juror’s preference is not taken into account
in reaching an outcome. An argument can perhaps be made for the use of some
kind of averaging process under these circumstances. In this CJ, three jurors were
not in favour of any payment and ten were in favour of a payment of 0.1 per cent
of taxable income. Should an average WTP be determined in those circum-
stances?11 Or should the majority view of 0.1 per cent be accepted and used in
estimation of the societal WTP?
The interpretation of WTP recommendations reached by consensus raises some
issues also associated with results from applications of stated preference methods.
These include: the transferability of results; the consensual WTP determined by a
jury represents the maximum levy any individual would agree to after having
participated in a CJ on the topic under the same conditions of information,
deliberation, decisiveness and responsibility; miscounting of benefits and costs;
altruistic motivations and inclusion of bequest values by jurors may result in biased
estimates of WTP (see Quiggin 1998 and Milgrom 1993). For reasons such as
these, the results of deliberative valuation exercises may be better interpreted in
terms of social welfare functions and social optimality than individual utility
functions and Pareto optimality. Using this approach, the results of a deliberative
valuation WTP exercise should be interpreted as the maximum amount of money
which could be charged to members of the public using a given payment vehicle
in order to obtain the posited environmental improvement programme without
leaving social welfare as a whole worse off than prior to the change. This is an
advantage of deliberative valuation over stand-alone applications of stated
preference methods such as the CVM.
240 Rosemary F. James and Russell K. Blamey
The application of the CJ method to environmental decision-making and
specifically to deliberative valuation of environmental goods and services has been
considered in this chapter. The use of deliberative valuation clearly has potential
to overcome some of the deficiencies evident in valuation methods such as the
CVM. The provision of information, and of time and opportunity for deliberation,
allow for the development and expression of more robustly-held preferences than
is considered to be the case in both CVM and choice modelling. However, some
of the difficulties associated with the CVM and, to a lesser extent, choice modelling,
still remain with deliberative valuation. Further, the use of the CJ method in deliber-
ative valuation raises some additional methodological and theoretical issues. Aspects
including representation, management of the jury process to ensure equity of
participation, robust and effective decision rules and minimization of various forms
of bias, all require the development of appropriate protocols. It is likely that research
into the management and behaviour of legal juries will prove to be a rich source
of procedural guidance in regard to these matters, particularly in relation to the
conduct of adversarial citizens’ juries. There is also no doubt that potential future
litigation surrounding the results of adversarial citizens’ juries will provide much
impetus to the development and refinement of the method.
Acknowledgements
Mick Common of the University of Strathclyde originated this project whilst at
the Australian National University and has provided valuable advice during its
course. Michael Lockwood of Charles Stuart University and Rebecca Smith of
the Australian National University assisted in the conduct of the CJ. Mick Common,
David Stern and Jonathan Aldred provided useful comments on this material.
Notes
1 Writing with reference to aid projects, Slocum and Thomas-Slatyer (1995: 3) defined public or
popular participation to be ‘… active involvement of people in making decisions about the
implementation of processes, programmes and projects which affect them’. The definition applies
equally well to projects in developed countries.
2 Whilst choice modelling may provide a means to address some of the flaws cited with the
CVM, the extent to which this is so remains to be established. To date, choice modelling in
environmental applications has not been subject to the same degree of rigorous examination
and litigious critique as CVM.
3 Whilst these terms tend to be used interchangeably, Dryzek’s (1990) discursive theory has stronger
roots in the critical theory of Habermas (Dryzek pers. com.; Blaug 1999). Dryzek (2000) discusses
the distinction in some detail.
4 Constructions (i), (iii) and (iv) may result in adversarial proceedings, in which the witnesses and
perhaps the jurors perceive advantage from strategic behaviour and act accordingly. Hence
careful management of the jury process is necessary in order to minimize bias.
5 The term was used to describe several categories of land managed by the NSW National Parks
and Wildlife Service including national parks, nature reserves, state recreation areas and regional
parks.
6 Fire management in this context referred to the preparation for wildfires; this included training
of staff, controlled burning and the maintenance of fire trails and fire breaks. Weed and feral
Deliberation and economic valuation 241
(or pest) animal control are necessary in order to protect the natural biophysical resources of
the national parks from damage.
7 Repeated framing was necessary to ensure the jurors were operating on a purchase rather than
a contribution model. The distinction, as enunciated by Kahneman et al. (1993), is critical;
respondents must focus on the amount they are WTP or WTA when purchasing the posited
environmental change, in order to enable estimation of societal WTP or WTA. If respondents
instead adopt a contribution model, under which they nominate their willingness to contribute
to achieve environmental change, the resultant amounts cannot be aggregated to provide
estimates of the usual Hicksian measures of welfare change.
8 The jurors were unaware of the current levels of expenditure on the five management activities
during their consideration of the two charges.
9 It may, nonetheless, be of political importance in gaining widespread acceptance of the results
of the jury’s deliberations.
10 For example, this jury was under-representative in the lowest age (20–29 years) and income (less
than A$10,400 per annum) categories. Can the WTP figure from this jury be validly extrapolated
to the population? In this case, it is perhaps considered legitimate to do so, as the amount to be
paid by each individual was set at a percentage of taxable income. In situations where juries opt
for absolute payments rather than payments scaled in relation to income, however, the
development of societal WTP or WTA by extrapolation from a jury value is more questionable.
11 Using this approach, a per-taxpayer contribution of 0.077 per cent of taxable income resulted.
The resultant WTP for the improvements posited under Option 4 would be A$84.5m.
References
Aldred, J. and Jacobs, M. (2000) ‘Citizens and wetlands: evaluating the Ely CJ’, Ecological
Economics, 34 (2): 217–32.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and
expression of citizen values and issue opinions’, in W. Adamowicz, P. Boxall, M.K.
Luckert, W.E. Phillips and W.A. White (eds) Forestry, Economics and the Environment,
Wallingford, UK: CAB International: 103–33.
Blamey, R.K. (1998) ‘Contingent valuation and the activation of environmental norms’,
Ecological Economics, 24: 47–72.
Blamey, R.K. and Common, M. (1992) ‘Sustainability and the limits to pseudo-market
valuation’, in M. Lockwood and T. De Lacy (eds) Valuing Natural Areas: Applications and
Problems of the Contingent Valuation Method, Albury, Australia: Johnstone Centre, Charles
Stuart University: 117–46.
Blamey, R.K. and Common, M. (1999) ‘Valuation and ethics in environmental economics’,
in J. van den Bergh (ed.) Handbook of Environmental and Resource Economics, Northhampton:
Edward Elgar: 809–23.
Blaug, R. (1999) Democracy, Real and Ideal: Discourse Ethics and Radical Politics, Albany, NY:
State University of New York.
Bohman, J. (1996) Public Deliberation, Cambridge, MA: MIT Press.
Brown, T.C., Peterson, G.L. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71(2): 250–60.
Center for Conflict Resolution (1981) ‘Building united judgement’, Madison, WI: Center
for Conflict Resolution.
Coote, A. and Lenaghan, J. (1997) ‘Citizens’ juries: theory into practice’, London: Institute
for Public Policy Research.
Crosby, N. (1991) ‘Citizens’ juries as a basic democratic reform’, Minneapolis, MN: Jefferson
Center.
242 Rosemary F. James and Russell K. Blamey
Crosby, N. (1995) ‘Citizen’s juries: one solution for difficult environmental questions’, in
O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer: 157–74.
Crosby, N. (1996) ‘Creating an authentic voice of the people’, paper presented to the
Panel on Deliberation on Democratic Theory and Practice, Chicago, IL: Midwest
Political Science Association.
Daneke, G.A., Garcia, M.W. and Delli Priscoli, J. (eds) (1983) Public Involvement and Social
Impact Assessment, Boulder, CO: Westview Press.
Dienel, P.C. (1989) ‘Contributing to social decision methodology: citizen reports on
technological projects’, in C. Vlek and G. Cvetkovich (eds) Social Decision Methodology for
Technological Projects, Dordrecht: Kluwer: 133–51.
Dryzek, J.S. (1990) Discursive Democracy: Politics, Policy and Political Science, Cambridge:
Cambridge University Press.
Dryzek, J.S. (2000) Deliberative Democracy and Beyond: Liberals, Critics, Contestations, Oxford:
Oxford University Press.
Gastil, J. (1993) Democracy in Small Groups: Participation, Decision Making and Communication,
Philadelphia, PA: New Society Publishers.
Gregory, R. and Slovic, P. (1997) ‘A constructive approach to environmental valuation’,
Ecological Economics, 21: 175–81.
Gregory, R., Lichenstein, S. and Slovic, P. (1993) ‘Valuing environmental resources: a
constructive approach’, Journal of Risk and Uncertainty, 7: 177–97.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment,
London, UK: Routledge: 211–31.
Jefferson Center (1995) Citizens’ Jury on Hog Farming: Final Report, Minneapolis, MN: Jefferson
Center.
Jefferson Center (1996) Citizens’ Jury on Comparing Environmental Risks, Minneapolis, MN:
Jefferson Center.
Jefferson Center (1997) Citizens’ Jury on Dakota County’s Comprehensive Land Use Plan,
Minneapolis, MN: Jefferson Center.
Kahneman, D., Ritov, I., Jacowitz, K.E. and Grant, P. (1993) ‘Stated willingness to pay for
public goods: a psychological perspective’, Psychological Science, 4–5: 310–5.
Milgrom, P. (1993) ‘Is sympathy an economic value? Philosophy, economics and the
contingent valuation method’, in J.A. Hausman (ed.) Contingent Valuation: A Critical
Assessment, Amsterdam: North-Holland: 417–40.
O’Neill, J. (1997) ‘Value pluralism, incommensurability and institutions’, in J. Foster (ed.)
Valuing Nature? Ethics, Economics and the Environment, London: Routledge: 75–88.
Perman, R., Ma, Y., McGilvray, J. and Common, M. (1999) Natural Resource and Environmental
Economics, New York: Pearson Education.
Quiggin, J. (1998) ‘Existence value and the contingent valuation method’, Australian Economic
Papers, 37(3): 312–29.
Rosenbaum, N. (1978) ‘Citizen participation and democratic theory’, in S. Langton (ed.)
Citizen Participation in America, Lexington, KY: Lexington Books: 43–54.
Rosenbaum, W. (1979) ‘Elitism and social participation’, in Lincoln Filene Center for
Citizenship and Public Affairs Citizen Participation Perspectives, proceedings of the National
Conference on Citizen Participation, Tufts University, Washington DC, 28 Sept.–1
Oct. 1978.
Rousseau, J. (1968) The Social Contract, Harmondsworth: Penguin.
Deliberation and economic valuation 243
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent valuation’, Ecological Economics, 24: 213–30.
Slocum, R. and Thomas-Slatyer, B. (1995) ‘Participation, empowerment and sustainable
development’, in R. Slocum, L. Wichhart, D. Rocheleau and B. Thomas-Slatyer (eds)
Power, Process and Participation – Tools for Change, London: Intermediate Technology
Publications: 3–8.
Thomas-Slatyer, B. (1995) ‘A brief history of participatory methodologies’, in R. Slocum,
L. Wichhart, D. Rocheleau and B. Thomas-Slatyer (eds) Power, Process and Participation –
Tools for Change, London: Intermediate Technology Publications: 9–16.
Van Valey, T.L. and Petersen, J.C. (1987) ‘Public service science centers: the Michigan
experience’, in J. DeSario and S. Langton (eds) Citizen Participation in Public Decision Making,
Westport: Greenwood Press: 39–63.
Webler, T. and Renn, O. (1995) ‘A brief primer on participation: philosophy and practice’,
in O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer: 17–33.
244 Katja Arzt
11 The challenges of
stakeholder participation
Agri-environmental policy
Katja Arzt
Introduction
Some environmental economics approaches based on neo-classical theory attempt
to reveal people’s values related to the natural environment by aggregating single
statements, e.g. the contingent valuation method (CVM). Such approaches are
based on the assumption that a person has complete information about the natural
good at stake and that they already have formed preferences. However, people
tend to form values and attitudes over time and may change their value perception
of the natural environment as their knowledge and experiences increase (e.g. Jiggins
and Röling 1998; O’Hara 2000). The case study presented in this chapter is based
on the hypothesis that values for nature emerge from interactions between people
(see O’Connor 1998) and could therefore be linked to Sahlins’s (1976) remark that
valuation approaches should be recognized as a subjective, cultural and contextual
phenomenon. This study is based on the observations that ecosystems are dynamic
and that many of their changes are unpredictable (see Hanna et al. 1996; Berkes
and Folke 1998; Holling et al. 1998). Therefore an optimal solution for the manage-
ment of different ecosystems cannot be presumed beforehand (see Munda 2000).
Complexity and unpredictability in physical and social systems has thus to be
considered when valuing environments.
An ‘interactive valuation process’ is here defined as different stakeholders
concerned with a particular environmental issue coming together to exchange
their views and values in order to decide together about future programmes, plans
or targets for managing the environment. This kind of decision making pays
attention to ‘plural rationality’ (Söderbaum 1999) and thus could be termed a
participatory valuation process as different orientations and attitudes are to be
considered beside monetary assumptions. A discursive process will enable indivi-
duals, in groups, to construct values rather than express prior preferences and is
hence a deliberative approach which leaves it to the respondents to clarify among
themselves explicitly what they are valuing and why. The deliberative group can
define precisely which aspect(s) of the public good it values and by how much
(Sagoff 1998: 224). According to Sagoff (1998: 227): ‘If individuals … do not
come with predetermined preferences but must construct them, then the process
of construction may legitimately involve social learning, since this is precisely what
The challenges of stakeholder participation 245
occurs in other contexts in which people work out their values’. Hence, it is
hypothesized in this chapter that a mutual learning process is central for a successful
decision making process. If different stakeholders learn about each others’
perceptions and sets of values it is more likely that their actions can be co-ordinated
towards mutual agreements. As Ostrom (1998a: 71) has already explained for
common-pool resources: ‘When all appropriators of a common-pool resource share
a common set of values and interact within a complex set of arrangements, there
is a much greater probability that they will develop adequate rules and norms to
manage resources’.
This study supports that learning refers to the accumulation of insights into
system causes (see Meppem and Gill 1998). The intention is to promote the develop-
ment of a more unified and shared impression of the underlying processes of
cause and effect that describe the issues and problems under investigation. ‘Experts’
and citizens alike can be instrumental in the general learning process. This general
consultative process is proposed as an alternative to the more exclusive, funda-
mentally ‘expert’ driven, and certainly less learning-oriented, decision-making
framework of neo-classical economics (Meppem and Gill 1998; Jiggins and Röling
1998). Learning to communicate about values needs a process which goes beyond
the supply of information. DeBono (1993, quoted in Meppem and Gill 1998: 118)
states: ‘There are people who believe that if you get enough information then the
information will do your thinking for you, however, it is the concept through which
we perceive the information that gives it any value’.
The interactive valuation process brings different orientations together; some
participants may join the process because of individual interests while others may
do so for altruistic reasons. Hence, it is the aim of the ‘learning’ institution that the
limited number of participants in that deliberative approach act as citizens as
opposed to individuals who try to choose an outcome most benefiting their
individual interests. In other words, consumer preferences reflect conceptions of
the good life that individuals seek for themselves, while citizens’ preferences reflect
conceptions of the good society offered for the consideration and agreement of
others (Sagoff 1998: 215). A well-performed learning process is one where the
citizens’ preferences are emphasized; decision in the learning process may depend
more on the force of the better argument about public interest (Habermas 1982).
Discourse processes attempt to bring such different orientation together, and
constitute a conscious decision process, in which individuals may reconsider their
way of thinking (O’Hara 1999, 2000). A decentralized institution for valuation
may therefore have the following positive impacts compared with those valuation
methods which are based on neo-classical theory: it is a ‘learning institution’ in
which new experiences and knowledge as well as local knowledge can be integrated
continuously; decision results of such an institution are based on norms and rules
which correspond to the local, social and natural environment; results can therefore
be better accepted and transformed into action.
An interactive valuation process which aims for sustainable solutions and
sustainable actions needs an institutional design, capable of achieving a mutual
learning process. At present different institutional settings for interactive decision
246 Katja Arzt
making are being discussed. Different tools – such as citizens’ juries, consensus
conferences, roundtable groups, and focus groups – are used for different purposes;
some exist only for a short time to solve specific problems within communities
while others persist over time in order to be established as a new democratic
institution. The aim of this study is therefore to find the most significant circum-
stances aiding the creation of an institutional arrangement for decision making at
the local level. The first section draws some theoretical assumptions for a successful
institutional design. The second section presents a research project in northeast
Germany, which established a ‘roundtable group’ in order to decentralize decisions
about agri-environmental policies and schemes. In the conclusion the interactive
valuation process in this project is argued to have been only partially successful
due to an inappropriate institutional setting.
Fair and competent discourse must mean giving particular attention to the
differences in verbal skill among discourse participants. The distribution of
verbal skill can become just as much a source of power and distortion as the
distribution of wealth in price based valuation. The broader rationality of
discursive reason may not at all be supported by featuring culturally rewarded
skills like articulated speech, deductive logic, or abstract thinking to the
exclusion of less articulate, empathetic, or traditional thinking.
The more concrete the objects of decisions and the more obvious the consequences
of action, the better is a mutual learning process initiated. Environmental problems
which are worked out in a learning process should be obvious for all participants.
Only then will these instruments be better than existing institutional arrangements.
The source of information is also important for the learning and trust-building
process. Some sociological studies in the UK (see Clark and Murdoch 1997) showed
that in the past farmers lost trust in scientific results and ideas. However, those
findings cannot be generalized; in other areas it may be observed that farmers
depend on and use scientific research. In any case, while implementing a valuation
process, researchers should be aware of the information used, as Bernabo (1998b)
observed: ‘When researchers apply assessment models as mechanistic predictive
tools, they risk producing results that are not sufficiently relevant and are prone to
misuse in decision making (Bernabo 1998a). Policy-relevant assessment is a more
organic process involving researchers and stakeholders in an interactive exercise
that builds understanding rather than trying to define “truth”. The emphasis must
be on providing practical information in an easily usable form.’
We may find that participants decide easily on one issue, but fail on another.
Property rights theory certainly gives useful interpretations as to why people value
items the way they do, as the owner of rights of a special natural good will have
the pleasure of the corresponding utility, or, on the contrary, if the corresponding
right is a duty, the owner has to carry the negative effects and costs (Hagedorn
2000; O’Neill and Spash 2000). For example, property rights on natural goods
such as soil or water do not have homogeneous property right structures, but are
differentiated by their different ecological characteristics; correspondingly the cost
and benefits will vary, leaving room for different institutional designs of duties and
rights (see Hagedorn 2000; Bromley 1997, 1998; Ostrom 1998a). The learning
process should clarify through discourse which property right structures are
connected to which participants and how much room is left to change the
institutions. Lawrence (2000: 16) suggested to distinguish between ownership,
control and use of property resources in environmental policy definition and
applications. Soil erosion, for example, is different for farmers who actually are
250 Katja Arzt
the owners of land as opposed to being tenants. Owners may have an interest in
keeping erosion as small as possible to maintain the fertility of the land for their
children or for future sales, whereas tenants may tend to reduce the fertility in the
short run for the time of the contract. However, if arable land should be converted
into meadow for ecological reasons, the owner of the land may not agree to convert
the land, because the rent received will be less for meadows. In this case the property
right structures are difficult to overcome and a decision will not be easily obtained.
In order to build reciprocity within the group it seems necessary that the property
right structures are clarified. This also corresponds with the control mechanisms.
Only if participants are convinced that breaking the rules about the decisions
made by the group (e.g. to reduce nitrogen fertilizer) can be controlled, is there
then a chance that the group formulates rules and norms which consolidate trust,
reciprocity and reputation.
The urgency of a problem can also be an important factor for participants
reaching consensus. For example, if farmers and environmentalists of a certain
region have problems with the water regime, the farmers would like to reduce the
water level in order to keep on farming, whereas the environmentalists would like
to have a higher water level to increase biodiversity. For both, the situation is urgent
and needs a solution in order to plan future activities. In this case the participants
may invest more time and money to reach a consensus compared with a situation
where nobody’s activities are considerably harmed.
Apart from the learning process the decision procedure itself strongly influences
the outcome. In the case of a roundtable group, decisions are made by consensus,
and not by voting or other mechanisms. Consensus means that all participants have
to agree upon the results by accepting the arguments addressed during discussion.
In this case verbal skills and power structures may have an important impact. Also
analytical methods such as multi-criteria analysis which recognize the plurality of
values that inform environmental choices can be combined with more deliberative
methods that enable citizens to articulate values and which are sensitive to concerns
about fairness in procedures and outcomes (O’Neill and Spash 2000).
In Table 11.1 the above-mentioned variables and their characteristics are
summarized. The next section will analyse a project that was carried out in Germany
to see how far it meets the idealized model presented here. It will do so by evaluating
the variables described above. The data for the criteria were collected from
qualitative interviews before and during the process and from meeting observations.
only partially in agreement with the practice of farming or were argued not to
include enough ecological benefits (Müller et al. 2000). A roundtable group operates
on the basis of seeking consensus between people with different viewpoints on the
issues discussed. In this approach each participant has an equal opportunity to
voice their option. Those involved should speak and listen to others with a view to
seeking areas of consensus. Agreements come out of the discussion process. This
approach works mainly over a series of meetings. In the case of the Agri-
Environmental Forum (AEF) the discussion for each meeting was prepared by a
project team and led by a professional moderator. For the project team it was
important to find an institutional arrangement that is quite simple to handle and is
time- and cost-efficient, because the idea of the interactive valuation institution
should be transferable to other regions and should be adopted by organizations
within the region.
On the basis of the common regulation of the European Union (EU), Germany
developed different agri-environmental schemes at the Länder (county) level.
Compared with other EU countries, where schemes are developed by central
governments (e.g. the Countryside Stewardship Scheme in the UK), Germany
already has a more decentralized approach. However, the weight given to different
environmental measures and the types of environmental consideration that are
important are clearly influenced by political priorities, policy objectives and cultural
attitudes to the farmed environment (see also Curry and Winter 2000). Local
farmers’ or local conservationists’ opinions about the different measures for the
schemes are not taken into consideration when establishing new programmes, yet
different experts are consulted by the local governments when setting up schemes.
In Brandenburg, two environmental schemes for farmers exist: one is called the
KULAP (cultural environmental landscape programme), the other is the Contract
Nature Conservation Programme. In KULAP farmers can choose among different
measures and apply for them. However, due to limited funds, the scheme operates
on a ‘first come, first served’ basis. KULAP is mainly financed by the EU and at a
national level (see EWG VO 2078/92 [Agri-environment Regulation (EEC) No.
2078/92] and VO 2052/88 [Council Regulation (EEC) No. 2052/88]). The
Contract Nature Conservation Programme is financed completely at the Länder
252 Katja Arzt
level. Farmers enter into individual contracts for their farms based on discussions
with environmental officers. Only farmers whose land lies within a protected area
can apply for the scheme (Hampicke et al. 1998). However, decisions about the
measures are not taken in consensus with different stakeholders, and the targets
for the scheme are fixed by experts.
For the implementation of the project, a region in the northeast of the
Uckermark (Brandenburg county, 180 km north of Berlin) was chosen. The
selection was made in arrangement with the local department for agriculture and
environment. The region contains 11,000 ha of agricultural land, is one of the
most thinly populated districts of Germany (53 inhabitants/km) and has a high
unemployment rate. An important factor for the selection of this particular region
was its strong agricultural character, the variety of landscapes and the low density
of protected areas.
The project team is interdisciplinary and consists of several natural scientists
and socio-economic scientists as well as a moderator. These professionals come
from different universities and institutions. The different backgrounds of the project
team made it difficult to find a general understanding of the definition of ‘value’
and ‘learning processes’ beforehand. The team could not create a common
understanding about the institutional arrangement as discussed in this chapter.
However, they already had agreed upon a phase- and process-oriented process.
The strategy was to choose relevant topics with the stakeholders and subsequently
each problem was to be discussed with regard to (a) reasons for the appearance of
the problem and the expected long-term consequences; (b) solutions that should
be achieved, including justifications; (c) measures to achieve these solutions,
including a monitoring-process; (d) financing of the measures; and (e) action
planning. However, this procedure could only be partially realized, because most
stakeholders considered it to be too abstract and too long. It became increasingly
evident by the third meeting that the stakeholders’ main interest was to talk chiefly
about measures and their financing. Therefore the project team had to plan each
meeting separately.
After the region was chosen the project team contacted different stakeholders
for initial interviews, in order to get to know the region and its problems, and also
to voice the intention of the project. The results of those interviews were presented
at a community meeting, to which over 100 different stakeholders had been invited
(selected using telephone books and contacts), but only 43 stakeholders attended.
During that meeting the project team asked for volunteers from different stake-
holder groups who would like to participate in the roundtable group. At the end
23 stakeholders wanted to attend the AEF. From October 1999 to April 2000 six
roundtable meetings took place (Figure 11.1), one each month, except in the summer
months when farmers are too busy on the farm to engage in other activities. The
process continued until December 2001, during which time 13 meetings took place.
The project team found another regional organization who organizes the meetings
until the present time but with a different focus. Todays meetings are more seen as
‘information meetings’, to inform farmers and environmentalists about agro-
political subjects.
The challenges of stakeholder participation 253
35
30 Agricultural institutions
Farmers
25
Environmentalists
20
Conservation association
15 Tourism
10 Politics
5 Administration
Trade
0
Others
1
6
Scientists
g
g
t in
t in
t in
t in
t in
t in
ee
ee
ee
ee
ee
ee
M
The actual proceedings are outlined in Table 11.2. During the first meeting
much time was spent on gettting to know each other and clarifying expectations.
The different stakeholders revealed their interest in starting mutual discussions in
order to find solutions for regional ecological problems, but most of them were
unsure about the process itself and about the functioning of a roundtable discussion.
At the first meeting, the stakeholders identified the topics that should be discussed
over the following months. In order to have a better understanding of the current
environmental problems in the region, the scientists were asked to prepare a general
lecture about the environmental situation and a more detailed lecture on landscape
characteristics for the second meeting. However, it became clear in the following
meeting that the different participants including the project team had different
intentions in mind or understood the topics differently.
Accordingly, during the second meeting – which started with a scientific lecture
about the reasons for structuring the landscape (e.g. biodiversity, biotope network,
and other ecological functions) – several farmers complained that the financial
aspects were of major concern, but were not included in the lecture. They said for
instance: ‘Well as a human I find all those aims important, but as a farmer, I
cannot accept all those targets, because I must work on my fields in order to survive’,
and ‘I would do anything if I got enough cash’. The participants wanted to
concretize the topic (‘We should not talk so scientifically, otherwise this evening
does not mean anything to me.’) and decided to talk about Sölle (water bodies
developed during the ice age, which are a typical element of the regional landscape)
during the next meeting. As a result of this ‘unsuccessful’ meeting in terms of a
mutual learning process about the topic, the project team had to decide how the
topic of Sölle could be introduced and how stakeholders could come to a consensus
about a regional programme. Prior to the next meeting the organization team sent
information materials about Sölle to the stakeholders.
Table 11.2 Proceedings at the agri-environmental forum
What have been the What methods were used during the What came out of the discussion process?
expectations of the meetings?
organizers?
1. AEF To become acquainted with Questioning of expectations and Participants want a priority catalogue for the most
254 Katja Arzt
Conclusions
Certainly, six meetings are insufficient to acknowledge if the AEF will be
institutionalized as an interactive valuation process in which learning processes
are initiated. Within four meetings the participants agreed upon a list of different
measures for Sölle. The list was transformed into an agri-environmental programme
for Sölle and presented by the organization team to the local government of agri-
culture, but it has not yet been accepted as part of a new agri-environmental
scheme in the future. For the second subject, soil erosion, no consensus could be
reached after three meetings within the group. Overall, it was the first time that
local stakeholders together with scientists decided and reached consensus about
agricultural measures, and on this basis the process can be called a success. They
exchanged their value perceptions on different issues. However, the decline in
numbers of participants over the series of meetings signalled that not all participants
were satisfied with the process itself.
This study hypothesized that a good learning process is a result of well-
implemented variables (e.g. selection of stakeholders, issues discussed, flow of
information, etc.). In the case of Sölle it can be concluded that the participants
learned about them during the meetings and were helped to create a value
perception about the issue. It also became clear that they needed different
information sources. The third meeting, for instance, was successful in stimulating
learning about the ecology of the subject under consideration, but participants
failed to decide about a local programme because more information was needed.
After the presentation of practical experiences from farmers a decision procedure
was possible. In the case of soil erosion, however, a decision process could not be
stimulated. The issue may not be of any relevance to the participants or the infor-
mation given to them did not correspond to their needs. However, it was shown
that the presentation of scientific lectures alone neither stimulated a dialogue among
the participants nor helped to create an atmosphere of trust.
What could be improved in future settings in order to stimulate a mutual learning
process? With regard to stakeholder selection it should be ensured that the
representatives of different organizations back-check with and can speak on behalf
of the constituency group, because it happened that two representatives from the
same organization had opposite opinions. This fact did not help to build reciprocity
among the participants. The group tended to exclude some stakeholder groups,
like farmers from family businesses, women, or people who do not belong to any
organization. More thoughts should be given how those groups could be integrated
The challenges of stakeholder participation 259
or how the selection of stakeholders should be organized in order to have an
adequate representation of the county community. The organization team should
seek to be more independent of the other stakeholder groups. Scientists therefore
should be restricted to an advisory function, and should only make a contribution
when requested.
The chosen topics did not correspond to the interests, knowledge or urgent
concerns of the participants. Therefore a better method for choosing topics seems
to be necessary. However, the method still has to be quite simple so that it can be
transferred to other regions. This requirement precluded a careful analysis of needs
of the participants through social scientific methods (e.g. interviews) prior to the
process. Conversely, to start with conflicting problems may impede a constructive
discussion and may not help to build trust among the participants in the beginning.
Therefore it might be best to start with an issue which is relatively uncontroversial.
Property right structures (ownership, control and use) have not been exclusively
mentioned during the discussion but this should be changed so that action can
easily follow the decision process and reciprocity among the participants can emerge.
The information flow during the process was strongly influenced by scientists.
There is unmistakably a big gap between a scientist’s way of thinking and that of
practitioners. Scientific training emphasizes the hierarchical organization of an
issue: analysis of the situation, targets, instruments, implementation and testing. A
practitioner, on the other hand, tends to be more interested in concrete decisions
about actions. The organization of information flow is central for the learning
process. It helps people if a discussion can be followed easily in order to contribute
fruitfully to the debate. Scientific lectures should only be held on urgent request.
Further it was observed that participants with good verbal skills dominated the
discussion process; a more balanced level of participation by all stakeholders could
be achieved by using better moderation techniques or a different procedure.2
The decision procedure about the Sölle programme was inefficient. It lacked the
participation of all stakeholder groups and did not consider enough different aspects.
A well-performed MCA might improve the procedure and also stimulate a learning
process. However, institutional difficulties within the decision procedure may be
overcome by better institutional design or by learning about the process itself. Certainly
more information and perceptions are exchanged in such deliberative institutions
compared with monetary evaluations, but at the same time it is costly and time-
consuming. Yet, getting the institution right is a time-consuming process. The
stakeholders are dissatisfied with the present situation of agri-environmental policy
decisions but at the same time sceptical about the success of such new valuation
institution(s). So far none of the stakeholder groups or the organization team know
the direction that a change of the existing institutional setting in agricultural policy
may be taking. The decision procedure may improve over time and with more
experience, but the acceptance of those procedures from existing institutions can
still threaten the whole process. Therefore it is important to fit such new valuation
institutions in with existing institutions right from the beginning.
The framework used in this study certainly helped to analyse the process with
its different aspects. However, at this early stage in the deliberative process of the
260 Katja Arzt
case study it is difficult to draw any conclusions on the relevance of each element
to the process and whether there are other important elements which might lead
to a successful performance of interactive valuation processes. Only the comparison
of different case studies will clarify this question.3 Certainly more information is
needed about the reasons why people are involved in the process and their
expectations in order to understand their needs. This information might shed more
light on the adequate design of the interactive valuation process. Another question
is whether all environmental problems of agriculture can be solved with these
kinds of participatory methods. More research has to be done on which co-
ordination structures correspond to which kinds of transaction and property right
structures. In this respect a more hierarchical decision procedure might address
some problems better when transaction costs, and also fairness, are considered.
Finally, more knowledge is needed about the underlying assumptions of the core
elements of trust, reciprocity and reputation and how these stimulate a learning
process among different stakeholder groups. Further investigations are needed as
to how institutions enhance or restrict the building of mutual trust, reciprocity,
and reputation. In this respect future investigations should also be interdisciplinary
in their design.
Acknowledgements
This chapter arose from work commissioned by the German Federal Ministry of
Research and Education (Bundesministerium für Bildung und Forschung); it is
part of a scientific co-operative project called GRANO. The author is grateful to
the steering group for helpful discussions, and also to the Chair of Resource
Economics and Prof Dr K. Hagedorn, the scientists at the Countryside and
Community Research Unit in Cheltenham, UK, and especially to John Powell.
The author would like to thank Simon Andrews, Claudia Carter and the colleagues
who joined the discussions during the ESEE Conference 2000 in Vienna. The
author alone is responsible for any errors and omissions.
Notes
1 GRANO project stands for ‘Approaches for a sustainable agricultural production: Application
for North-eastern Germany’. GRANO is a scientific co-operative project supported by the Federal
Research Ministry. GRANO chose the districts of Uckermark, Barnim, and Elbe-Elster to search
for, develop, and test new ideas and solution strategies. The ‘round table’ project is one among
other projects carried out by GRANO.
2 The organization team discussed that problem, together with some participants, and decided
that bilateral interviews about the programme measures (in this case measures against soil erosion)
should precede the next meeting. Consequently the organization team prepared a questionnaire
and asked all participants separately during the summer months when no meeting took place,
and presented the results during the next meeting. This procedure was very time and resource
consuming.
3 At the moment a PhD student is working on that question.
The challenges of stakeholder participation 261
References
Berkes, F. and Folke, C. (1998) Linking Social and Ecological Systems. Management Practices and
Social Mechanisms for Building Resilience, Cambridge: Cambridge University Press.
Bernabo, J.C. (1998a) ‘Improving integrated assessments for application to decision making’,
in S. Lee and T. Schneider (eds) Risk Assessment: Linking Science to Policy, Amsterdam:
Elsevier Science.
Bernabo, J.C. (1998b) ‘Perspectives and principles from other assessments’. Online.
Available at: http://www.agci.org/publications/EOC97/eoc97session2/Bernabo.html
(accessed 16 June 2004).
Bromley, D.W. (1991) Environment and Economy: Property Rights and Public Policy, Cambridge,
MA: Basil Blackwell.
Bromley, D.W. (1997) ‘Property rights in environmental economics’, in H. Folmer and T.
Tietenberg (eds) The International Yearbook of Environmental and Resource Economics 1997/
1998: A Survey of Current Issues, Cheltenham: Edward Elgar.
Clark, J. and Murdoch, J. (1997) ‘Local knowledge and the precarious extension of scientific
networks: a reflection on three case studies’, Sociologia Ruralis, 37(1): 37–60.
Curry, N. and Winter, M. (2000) ‘The transition to environmental agriculture in Europe:
learning processes and knowledge networks’, European Planning Studies, 8(1): 107–21.
DeBono, E. (1993) Parallel Thinking, Ringwood: Penguin Books.
Habermas, J. (1982) Theorie des kommunikativen Handelns, Frankfurt/Main: Suhrkamp.
Hagedorn, K. (2000) ‘Umweltgenossenschaften aus institutionenökonomischer Sicht’, in
M. Kirk, W. Jost and S. Kramer (eds) Genossenschaften und Kooperation in einer sich wandelnden
Welt. Festschrift zum 65. Geburtstag von Prof. Dr. Hans-H. Münkner, Münster: LIT Verlag:
267–91.
Hampicke, U., Müller, K., Meyer-Aurich, A. and Kachele, K.U. (1998) ‘Agrarumweltpolitik
nach dem Subsidiaritätsprinzip. Teil III Fallstudie Brandenburg/Uckermark’, working
paper of the Universität Greifswald und Zentrum für Agrarlandschafts- und
Landnutzungsforschung e.V.
Hanna, S., Folke, C. and Mäler, K.-G. (eds) (1996) Rights to Nature, Washington, DC: Island
Press.
Holling, C.S., Berkes, F. and Folke C. (1998) ‘Science, sustainability and resource
management’, in N.G. Röling and E. Wagemakers (eds) Facilitating Sustainable Agriculture,
Cambridge: Cambridge University Press.
Jiggins, J. and Röling, N. (1998) ‘Interactive valuation’, paper presented at the Conference
of the European Society for Ecological Economics, University of Geneva, Switzerland,
5 and 6 March 1998.
Kolloge, S. (2000) ‘Die Wirkung der Agenda 21. Eine institutionentheoretische Wirkungs-
analyse am Beispiel von sechs Kommunen im ländlichen Raum Großbritaniens und
Deutschlands’, unpublished PhD thesis, Landwirtschaftliche-Gärtnerische Fakultaät
der Humboldt-Universität zu Berlin, Germany.
Lawrence, R. (2000) ‘Property, rights and fairness’, in C.L. Spash and C. Carter (eds) EVE
Policy Research Brief Series, no. 6, Cambridge: Cambridge Research for the Environment.
Meppem, T. and Gill, R. (1998) ‘Planning for sustainability as a learning concept, survey’,
Ecological Economics, 26: 121–37.
Müller, K., Bork, H.-R., Dosch, A., Hagedorn, K., Kern, J., Peters, J., Petersen, H.-G.,
Nagel, U.J., Schatz, T., Schmidt, R., Toussaint, V., Weith, T., Werner, A. and Wotke, A.
(eds) (2000) Nachhaltige Landnutzung im Konsens: Ansätze für eine dauerhaft-umweltgerechte Nutzung
der Agrarlandschaften in Nordostdeutschland, Giessen: Focus Verlag.
262 Katja Arzt
Munda, G. (2000) ‘Conceptualising and responding to complexity’, in C.L. Spash and C.
Carter (eds) EVE Policy Research Brief Series, no. 2, Cambridge: Cambridge Research for
the Environment.
O’Conner, M. (1998) ‘Walking in the Garden(s) of Babylon: an overview of the VALES
project’, Paris: C3ED Rapport de Recherche.
O´Hara, S. (2000) ‘The challenges of valuation: ecological economics between matter
and meaning’, in C. Cleveland, R. Costanza and D. Stern (eds) The Nature of Economics
and the Economics of Nature, London: Edward Elgar.
O’Hara, S., Shandas, V. and Vazquez, J. (1999) ‘Communicating sustainable development
options – who evaluates the trade-offs?’, in I. Ring, B. Klauer, F. Waetzold and B.
Mansson (eds) Regional Sustainability. Applied Ecological Economics Bridging the Gap between
Natural and Social Sciences, Heidelberg: Physica Verlag.
O’Neill, J. and Spash, C.L. (2000) ‘Conceptions of value in environmental decision-making’,
in C.L. Spash and C. Carter (eds) EVE Policy Research Brief Series, no. 4, Cambridge:
Cambridge Research for the Environment.
Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions for Collective Action,
Cambridge: Cambridge University Press.
Ostrom, E. (1998a) ‘The institutional analysis and development approach’, in E.T. Loehman
and D.M. Kilgour (eds) Designing Institutions for Environmental and Resource Management,
Cheltenham: Edward Elgar: 68–90.
Ostrom, E. (1998b) ‘A behavioral approach to the rational choice theory of collective
action’, presidential address, American Political Science Association, American Political
Science Review, 92(1): 1–22.
Ostrom, E., Gardener, R. and Walker, J. (1994) Rules, Games, and Common-Pool Resources,
Ann Arbor: University of Michigan Press.
Röling, N. and Wagemakers, E. (eds) (1998) Facilitating Sustainable Agriculture, Cambridge:
Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24(2–3): 213–30.
Sahlins, M.D. (1976) Culture and Practical Reason, Chicago, IL: University of Chicago Press.
Simon, H.A. (1983) Reason in Human Affairs, Stanford, CT: Stanford University Press.
Söderbaum, P. (1999) ‘Values, ideology and politics in ecological economics’, Ecological
Economics, 28: 161–70.
Preference transformation through deliberation 263
12 Preference transformation
through deliberation
Protecting world heritage
Simon Niemeyer
Introduction
Preferences are prone to change. Yet much effort in environmental economics has
been applied to the measurement of static preferences for use in policy. This
‘endogenous preference’ orthodoxy has recently been challenged on at least two
fronts. One criticism stems from findings that support the idea of ‘constructed’
preferences in view of exogenous influences (for example Slovic 1995; van den
Bergh et al. 2000). A second criticism, with implications for environmental policy,
is the emerging tension between the normative goal of sustainability and
aggregation of sovereign preferences (for example Common and Perrings 1992).
These have led to arguments within ecological economics for policies that will
lead to so-called ‘sustainable’ preferences (Norton et al. 1998), although it is uncertain
what is needed to achieve such a normative goal.
The fact that preferences can change invokes notions of preference quality –
that some preferences might be ‘better’ than others. Most economists do not make
such distinctions: a stated preference is a legitimate preference where consumers
are sovereign and free to choose as they wish. Others, such as O’Neill (1997) and
Elster (1983) are less sanguine. For them, the process of producing ‘better’ (or
reasonable, in the sense that they have been through a process such that chosen
means match justified ends) preferences is more important than simply realising
given preferences – the reason for a preference is as important as the preference
itself.
Deliberation is one (if not the) process for reasonable preference formation. It
is the process of deliberation, be it interactive deliberation with others, or
deliberation within oneself (e.g. Goodin 2000) whereby one recursively constructs
those means that suit desirable ends. In the case of environmental policy, the most
vaunted end is sustainability although it is unclear and possibly indefinable. Many
authors are turning to deliberation in the search for mechanisms capable of
elevating environmental considerations in policy making (for example Sagoff 1998).
An important element of this ‘deliberative turn’ in environmental policy is the
tendency for deliberation to transform preferences in the direction of ‘nature-
loving’, or sustainability. The relationship between preference for nature and
sustainability is partially accounted for by the simple premise of ‘more nature
264 Simon Niemeyer
preserved is more nature sustained’; but sustainability is inherently more complex
and infused with a specific meaning depending on context.
One danger in attempting to produce ‘sustainable’ preferences is to pre-empt
outcomes that are nominally designated sustainable when that goal itself is not
well defined. Just as Elster (1983: 9) states that utility maximisation as a single
endeavour is futile because happiness eludes those who actively strive for it, the
same may also be true when actively seeking to achieve sustainability. Instead of
defining a sustainable end, a better approach is to focus on those processes that
tend to produce ‘sustainable’ outcomes; better still if those same processes help
clarify the debate concerning definitions thereof. Fortuitously, deliberation is also
identified as such a process through which sustainable outcomes are defined and
achieved as part of the same process (Rydin 1999). In other words, ‘deliberators’
formulate those ends deemed sustainable and form judgements about which policy
means are best suited.
This chapter seeks to explore the ‘deliberative alternative’ and its ability to
transform preferences to suit so-called sustainable ends. The case study concerns
a controversial road constructed through World Heritage Listed rainforest in Far
North Queensland, Australia, on which a Citizens’ Jury (CJ) was held in January
2000. CJs are seen here as an example of a formal deliberative process that has
some similarity to a legal jury where decisions are made collectively by a randomly
selected group of citizens based on evidence presented (see other chapters in this
volume and also Smith and Wales 2000). The focus of this chapter is on the margins
of ‘preference shift’, or those mechanisms through which preferences are
transformed by the deliberative process. The aim is to measure both changes to
(rank ordered) preferences and the underlying subjectivity for those preferences.
Subjectivity is measured using Q method, a form of ‘quantitative discourse analysis’,
which is also compared with actual jury discourse and interview data.
The following discussion begins by considering tentative mechanisms whereby
preferences are transformed through group deliberation in an information-rich
environment. This is followed by a description of the case study, its history and
politics, considered important because, as will be shown, issue context is a major
factor in preference dynamics. The impact of the deliberative processes on both
preference and subjectivity is then discussed with particular mechanisms
explaining the processes that contribute to preference shifts. Conclusions are
then drawn about the role of deliberative processes in improving environmental
policy outcomes.
Results
In the following discussion the results from the CJ are briefly outlined. This is
followed by more substantial analysis and discussion to establish the implications
of deliberation for preferences and attitudes, and mechanisms for change thereof.
Attitudes
Measurable changes to attitudes did occur during the deliberative process. Yet
there were a number of constant aspects – in terms of their relative weighting.
Most significant of these was the tendency for jurors to place considerable emphasis
on long-term solutions to the issue, and its impact on future generations. For
example, jurors consistently agreed that a short-term perspective was inappropriate
for deciding the issue (see statement 30 in Table 12.1, x1 x3 = –2.08, s1 = 1.44, x3 =
–2.25, s3 = 1.22); and that inaction would damage the opportunities of future
Preference transformation through deliberation 271
generations (statement 25, x1 = 3.00, s1 = 0.74, x3 = 2.83, s3 = 0.94). There was
also a strong, though variable, commitment to conservation in the Daintree
irrespective of cost (statement 20, x1 = 2.67, s1 = 1.97, x3 = 2.42, s3 = 2.27). All
jurors tended to agree that the Bloomfield Track issue was an important one
(statement 12, x1 = 1.58, s1 = 1.00, x3 = 1.67, s3 = 0.78). One of the most intractable
issues among the jurors throughout the process was the level of permissible
development within the World Heritage area (statement 10, x1 = 0.92, s1 = 2.39,
x3 = 1.25, s3 = 3.39).12 There was also widespread disagreement regarding whether
the track should stay, given that it had already been constructed, although there
tended to be a net shift away from this position (statement 39, x1 = 0.75, s1 = 1.54,
x3 = –0.83, s3 = 2.66). As will be seen in the preference analysis below, this was
reflected in a division among jurors over the preferred policy option. Using
conventional Q analysis, described above, it is possible to establish an overview of
the attitudinal changes within the jury as a result of deliberation. By tracking
changes to juror’s loadings on each factor – representing the degree to which they
‘agreed’ or ‘disagreed’ with that subjective position – we can draw insights into
changes to attitudes during the deliberative process. The combined data for all
three Q sorts by the jurors produced four distinct factors: Preservation, Optimism,
Pragmatism and Symbolism. Statement scores for each factor are provided in Table
12.1. The following four paragraphs briefly describe the contours of each factor
and the degree to which their role in the attitudes of jurors changed during the
deliberative process. The description of each factor takes the combined form of
narrative regarding its defining features and includes a summary of factor scores
for relevant statements.
The first and most dominant factor – accounting for 40 per cent of variation
between sorts – is Preservation. Based on the statement scores for the factor given in
Table 12.1, Preservationists feel a personal commitment to protection of the Daintree
area. Statements with high scores for this factor tended to reflect an anti-development
sentiment towards the area (such as responses to statement numbers 10 and 32).
There is also a perception that the Bloomfield Track is a contributing factor to future
development (statement 2). (This is significant given, as previously mentioned, that
there was an increasing sense among the jury during deliberation that the Bloomfield
Track was part of an interconnecting web of regional impacts, which cannot be
properly considered in isolation.) Also evident is the belief that the use of vehicles on
the Bloomfield Track is detrimental to both rainforest (statement 21) and, to a lesser
extent, the reef (statement 7). Preservationists also feel they would be better off, or at
least no worse off, because of protection of the area (statement 11). Figure 12.2
shows how jurors changed their loadings on Factor 1 during deliberation. Each of
the bars represents the loading of each juror on the factor before, during and at the
conclusion of the CJ. (The names in the figure are the abbreviated codenames chosen
by jurors.13) The left-most bar represents the initial loading at the beginning of
deliberation. The middle (thinner) bar shows loadings midway and the darker right-
hand bar shows the loading of jurors at the end of the process. From Figure 12.2,
most jurors’ loadings on Preservation have changed substantially.14 Eight of the 12
jurors are significantly loaded at the beginning of deliberation (based on a level of
Table 12.1 Statement scores for Q attitudinal factors
Factor Score
No. Statement 1 2 3 4
1 Laying bitumen on the Bloomfield Track would be beneficial for the environment. It may even help reduce fuel usage –3 0 –3 1
and the greenhouse effect.
2 I don’t know if improving the Bloomfield Track would lead to a rapid acceleration of development in the area to the –2 3 –1 0
272 Simon Niemeyer
43 Everyone in Queensland is better off for having a road like the Bloomfield Track. –1 –2 –3 –2
274 Simon Niemeyer
100
90
80
Factor loading (x100)
70
60
50
40
30
20
10
0
ADV ASW BOA JAN JUL KEI KOD MAT PEA RAS SNO TAM
Juror
significance of 0.4).15 During deliberation there has been a net increase in factor
loadings ( x1 = 0.57, x3 = 0.61), with the exception of Adventure, Boat, Julie and
Keith, who have all decreased their loadings.
The Optimist archetype is characterised by someone who seeks an outcome where
everybody wins – environment and humans (exemplified by statement 18). Though
optimistic, they are very uncertain about a number of issues, including damage to
the area because of development (statement 2). They feel that the continued existence
of the Track will probably not lead to more development (statement 6), or destroy
the fringing reef (statement 7). Possibly in view of this, they are more predisposed to
development in the region than others (statement 10). Optimists are not inured to
future consequences. Indeed, they are less concerned about the present than others
(statement 42), but they are uncertain about whether now or the future is the more
important (statement 38). This could be a result of the level of uncertainty attached
to future outcomes, consequently greater significance is attached to the present – a
crude kind of discounting of future benefits.16 The Optimist is anthropocentric in
orientation.17 They feel that assessment of environmental problems should be based
on a perception of the efficacy of technological solutions.18 Greatest emphasis is on
how to achieve the best outcome. When questioned after deliberation, jurors most
strongly loaded on Optimism felt that closure of the Bloomfield Track should only
be undertaken once the costs are known to outweigh the benefits. These benefits
should be given proper account (statement 19). They feel that the Daintree could in
fact be under threat, but its conservation may not require the closure of the Bloomfield
Track (statements 20 and 27). Optimists tended to prefer technologically based ‘win–
win’ solutions. Optimism accounts for 9 per cent of variation between sorts. Net
loadings on the factor have steadily declined ( x1 = 0.26, x2 = 0.21, x3 = 0.17). As
Preference transformation through deliberation 275
90
80
70
60
Factor loading (x100)
50
40
30
20
10
0
-10
ADV AS W BO A J AN J UL KEI KO D MAT P EA RAS S NO TAM
-20
J uror
First sort Second sort Third sort
shown in Figure 12.3, only a few of the jurors subscribed to Optimism at any stage
of deliberation. Notably, Julie and Keith – the main dissenters at the end of
deliberation – have increased their loadings on the factor.
Pragmatism has many similarities to Optimism (with a correlation of 0.59).
Dominant features that distinguish Pragmatists from Optimists include scepticism
about anything but firm facts; confidence in their held beliefs; and less emphasis
on ‘win–win’ solutions. In contrast to the ecologically cautious approach of the
Preservationist – where the burden is to prove that activities do not damage the
area – Pragmatism focuses on the available facts. For example, because existing
research could not sustain the contention of reef damage, it is dismissed almost
entirely (statements 33, 7 and to a lesser extent 24). The pragmatic view of science
reflects this. Science is considered as the definitive input into decisions, rather than
a conservative and heuristic process that yields probabilistic outcomes. At its extreme
the logic is simply ‘no evidence, therefore no damage’ (and therefore, no road-
closure). As discussed below, there is a strong correlation between this position and
those jurors who prefer the Status quo option, provided that there is strict regulation
of Track use. The Pragmatist attaches more weight to present consequences than
future ones (statement 42), although they believe prudent decision making should
also benefit the future (statement 30). They are less ecocentric than Optimists
(statement 34); but they are not against preservation per se, believing that develop-
ment in World Heritage areas should be curtailed (statement 10). Because of the
conservatism of the Pragmatist they tend to place the burden of proof onto those
who wish to change the Status quo. Like Optimism, because damage to the Daintree
area does not appear significant (statement 27), conservation should be weighed
against benefits that development of the area may bring (statement 20). (However,
276 Simon Niemeyer
unlike the Optimist, Pragmatists are prepared to accept winners and losers
(statement 18).) Now that the Track is built and no definitive evidence says that it
should go, it may as well stay (statement 39); after all, it provides a benefit to the
community north of the Bloomfield River (statement 35). Pragmatists tend not to
be in favour of upgrading the Track (statements 6 and 9). As can be seen in Figure
12.4, pragmatism saw a marginal overall increase in factor loadings, although
these were far from uniform. Most of the increase can be attributed to three jurors
(Adventure 11 to 37, Janine –6 to 36 and Snoopy 7 to 33).
The last factor, Symbolic, resonates with the symbolic preferences discussed
earlier. Symbolists emphasise the impact of the road on fringing reefs, as well as its
benefit to the local community. There are only four statements that define the
Symbolic factor. Two of these concern strong belief in damage to the reef caused
by the Bloomfield Track (statements 7 and 33). There is a good deal of confidence
among Symbolists in ascribing benefits of protection (statement 40) as well as the
feelings of locals in the Bloomfield area (statement 5). Apart from these statements,
there is a high degree of affinity with factor 1, the Preservationist – with a high
level of correlation between the two (r = 0.63).19 Figure 12.5 shows changes to
loadings for Factor 4, Symbolism. Of all factors, it suffered the biggest decline
during deliberation. Beginning with five jurors loaded at the 0.15 level, by the end
of deliberation not one juror was significantly loaded – except for Keith and
Tamarra who were negatively loaded.
Preference shift
Attention is now turned to how preferences changed during deliberation. Table
12.2 shows the rankings given for each of the policy options before and after the
CJ process. To facilitate comparison of attitudes to preference, preference rakings
were subjected to factorial analysis similar to the attitude Q sort. This resulted in
four preference factors: Close Road, Stabilise, Bitumenise and Status quo.
Archetypal preference rankings for each preference factor position are provided in
Table 12.3. Figure 12.6 shows the loadings of jurors on each preference factor
before and after the deliberation. It can be clearly seen from both Table 12.3 and
Figure 12.6 that preferences changed substantially during deliberation. Two jurors
in particular – Snoopy and Aswad – experienced a near complete reversal of their
original orderings. During deliberation, preferences among jurors converged,
although they did not reach consensus.20 At the beginning there was a small
preference for minor road works to stabilise the Track and reduce run-off, or to
maintain the status quo. By the end of deliberation, road closure was clearly the
most preferred option.
70
60
Factor loading (x100)
50
40
30
20
10
-10
ADV AS W BOA J AN J UL KEI KOD MAT P EA RAS S NO TAM
-20
J uror
First sort Second sort Third sort
70
60
50
Factor loading (x100)
40
30
20
10
0
-10
-20
-30
ADV ASW BOA JAN JUL KEI KOD MAT PEA RAS SNO TAM
-40
Juror
First sort Second sort Third sort
questionnaires, how they felt they had changed their opinions and the emphasis of
issues they felt were most important. Although this supports the hypothesis that
preferences change with attitudes, alone it does not reveal a causal relationship.
One way of detecting possible relationships between preference and attitude is to
compare whether those individuals correlated to a preference factor also tended
to load to a corresponding attitude factor. These overall correlations before and
Table 12.2 Rank ordering of preferences for management options
Rank before CJ Rank after CJ
278 Simon Niemeyer
Codename Bitumenise Upgrade Stabilise Status quo Close Bitumenise Upgrade Stabilise Status quo Close
Adventure 5 4 3 2 1 5 4 3 1 2
Aswad 4 3 1 2 5 5 4 3 1 2
Boat 5 3 1 2 4 5 4 2 3 1
Janine 1 4 3 2 5 5 4 3 2 1
Julie 2 1 4 4 5 5 3 1 2 4
Keith 4 3 2 1 5 5 3 1 2 4
Koda 2 5 1 4 3 4 5 2 3 1
Matilda 5 4 2 1 3 5 4 3 2 1
Pearl 4 3 2 1 5 5 4 3 2 1
Rastus 4 3 1 2 5 5 4 3 2 1
Snoopy 1 2 3 4 5 5 4 3 1 2
Tamarra 2 5 3 4 1 5 4 3 2 1
Overall rank 3 4 1 2 5 5 4 3 2 1
Average rank 3.25 3.33 2.17 2.42 3.92 4.92 3.92 2.50 1.92 1.75
St dev 1.54 1.15 1.03 1.24 1.56 0.29 0.51 0.80 0.67 1.14
Preference transformation through deliberation 279
Table 12.3 Factor scores for preference ranking
Factor Description Rank preference for factor
Bitumen Upgrade Stabilise Maintain Close
1 Close road 5 4 3 2 1
2 Status quo 5 3 2 1 4
3 Bitumenise 1 4 3 2 5
4 Stabilise 2 5 1 4 3
after deliberation are shown in Table 12.4. The following discussion explores these
relationships and mechanisms that might explain such correlations. It does so by
considering each preference factor in turn and those attitude factors that are
significantly correlated to it.
From Table 12.4 it can be seen that loading on the Close preference factor is
strongly correlated with the Preservation attitude. It is also negatively correlated
with both Optimism (decreasingly so) and Pragmatism (increasingly so). The nature
of the changes in correlation between attitude and preference can be illustrated by
plotting each juror’s loading on both factors at the stages before and after
deliberation. Figure 12.7 shows loadings on the Preservation and Close Road factor
spectrum. The asterisks represent jurors’ initial loadings (before deliberation), the
solid dots represent the comparative loading of jurors on each factor at the
conclusion of the deliberative process, and the slender arrows indicate the
‘migration’ of preference and attitude during deliberation. The more significant
shifts (those that are subject to further scrutiny herein) are highlighted using bold
arrows. The two heavily dashed regression lines (the lighter line indicates the
relationship at the beginning of deliberation and the darker after deliberation)
indicate the relationship between preference and attitude before and after
deliberation respectively. From Figure 12.7, most jurors have migrated north-east,
increasing loadings on both factors indicated by the broad arrow. Others (Keith,
Julie and Boat) have moved north-west, increasing loading on Close, but decreasing
on Preservation. Adventure’s shift is particularly notable: beginning the process at
the top of the Preservation–Close spectrum, but retreating somewhat from this
position as deliberation has progressed.
That Preservation and Close should be so strongly correlated is intuitively
appealing: given the way they feel that the track impacts on the region in concert
with multiple pressures, a Preservationist would feel it to be the only solution to
280 Simon Niemeyer
Before deliberation
100
80
60
40
20
-20
-40
-60
-80
-100
ADV1 ASW1 BOA1 JAN1 JUL1 KEI1 KOD1 MAT1 PEA1 RAS1 SNO1 TAM1
After deliberation
100
80
60
40
20
-20
-40
ADV3 ASW3 BOA3 JAN3 JUL3 KEI3 KOD3 MAT3 PEA3 RAS3 SNO3 TAM3
Figure 12.6 Changes to preference factor loadings before and after deliberation
Preference transformation through deliberation 281
JAN3 TAM MAT PEAADV1RAS
100
BOA3 ASW SNO3 KOD3
First Sort Final Sort
ADV3
80
60
Factor loading (close road) × 100
40
KEI3 JUL3 TAM1
20 BOA1MAT1
0
KEI1 PEA1 KODA1
-20
RAS1ASW1
-40
JAN1
-60
-80
SNO1 JUL1
-100
0 10 20 30 40 50 60 70 80 90 100
Factor loading (preservation) ×100
protecting the area. However, while Preservation is the most significant attitude
factor throughout, track closure only reached prominence by the end of the process.
Why did Preservationists not vote more strongly for track closure from the outset?
This anomaly might be partially accounted for by the possibility that jurors also
initially loaded on other attitude factors which ‘polluted’ their preferences. The
role of these other factors is discussed below. Another related possibility is that the
preferences were not fully ‘constructed’ at the outset, the correspondent preference
position crystallising during deliberation. There was some evidence to support the
claim that jurors did not initially vote in a manner true to their convictions. One
juror in particular expressed that she would have liked to rank track closure as her
first preferences, but did not think that it was a realistic outcome – her opinion,
and conviction, on this matter changed dramatically during deliberation.
As revealed in Table 12.4, preference for the Status quo is positively correlated
with Pragmatism throughout deliberation. During the process it has also become
correlated with Optimism and strongly negatively correlated with Preservation. The
position of jurors on the Pragmatism–Status quo spectrum is shown in Figure 12.8.
From the figure it appears that only Adventure and Snoopy were influenced by
Pragmatism such that they increased preference for the Status quo – the rest having
more or less vertical or horizontal shifts. Keith and Julie did not increase their loadings
on Pragmatism, but remained the most significantly loaded of the jurors within this
spectrum. There is evidence here regarding how different attitudinal factors interact
in forming preferences. Those individuals who increased their loadings in this
spectrum were also significantly loaded on the Preservationist. However, while they
clearly sympathised with the Preservationist attitude, they hesitated in whole-heartedly
advocating road closure because they felt more proof was needed beyond that
presented to prove the case beyond doubt.21 In Adventure’s case in particular, this is
282 Simon Niemeyer
consistent with a decline on loading on both Preservation and Close (see Figure
12.7), as well as the jury discourse. Like Adventure, the impact of Pragmatism also
marginally increased Snoopy’s preference for the Status quo, but migrating from an
entirely different overall attitudinal position. From Figure 12.7 it can be seen that
Snoopy experienced the most dramatic shift of all jurors, having begun the process
at the low end of the Preservation–Close Road spectrum. Figure 12.9 shows the
relationships between Optimism and Status quo. Snoopy began deliberation as an
Optimist, decreasing loading significantly by the end.
Based on the changes observed in Figure 12.9, increases in loading on the Optimist
attitudinal factor appears to have affected the preferences of Keith and Julie at the
end of deliberation. Both felt very strongly that the track should remain open to
permit access to the area by anyone who wished to visit. Like other jurors, they also
agreed with the substance of the preservationist position that the area is important
and under threat, but sought to find outcomes whereby everyone could benefit. Keith
in particular was keen to explore other solutions outside the mainstream of discourse
among the jurors – advocating the use of alternative approaches such as hovercraft,
or even a monorail. One plausible explanation for their position is that Keith and
Julie were the elder members of the jury. They were concerned that Closure would
effectively exclude access to those who, like them, were not able to use the walking
track that would replace it. They were optimistic that a solution could be found to
the satisfaction of all users. Due to restriction of time and evidence, the merits of
these alternative options were not fully deliberated – with view to their cost and
feasibility. Optimism alone does not account for the attitude of those jurors who
loaded significantly on the Status quo. It appears instead to act in concert with
Pragmatism. As can be seen in Figure 12.8, Keith and Julie were also highly loaded
PE A
80 KE I1
60
Factor loading (Status quo)
MAT
40
SNO3ADV3
AS W
20
BOA3
TAM3 ADV1 KODA1 J AN3 J UL1
RAS3 MAT3 PE A3
0
J AN1
SNO1 KOD3
–20
–40
–60
TAM1
–80
–20 –10 0 10 20 30 40 50 60 70 80
Factor loading (Pragmatist)
50
MAT1
Factor loading (Status Quo)
ADV3
SNO
30 ASW3
BOA3
10 JAN3
JUL1 TAM3 KOD1
PEA3 ADV1 MAT RAS
JAN1
–10
SNO1
KOD3
–30
–50
TAM1
–70
–10 0 10 20 30 40 50 60 70 80
Factor loading (Optimist)
on the Pragmatism factor. Optimism and pragmatism shared many features but
were distinguished by emphasis on different facets of the issue.
Stabilise and Bituminise are considered together here because of both their
similarity in preference structure and interaction with attitudes. Most strongly
associated with either stabilising the track, or bituminising is Symbolism (Table
12.4). The correlation between Symbolic and Stabilise is significant at the outset,
but not by the end of deliberation. The explanation lies in the ability stabilisation
or bituminisation (which is second preference for the factor, see Table 12.2) to
address symbolic issues of reef damage and community access. For a Symbolist,
choosing other options would place the competing goals of preservation and
community access in tension with one another. As deliberation progressed and
symbolic concerns dissipated, Symbolism was supplanted by emphasis on indirect
regional and long-term impacts, epitomised by the Preservationist factor.
Subsequently the loadings on Symbolism declined, along with a statistically
detectible relationship with the Stabilise preference factor.
Conclusions
Overall there appears to be a significant shift in attitudes resulting from discourse
in the CJ. Jurors adopted a ‘big picture’ perspective of the issue and generally
increased their predisposition toward preservation of the Daintree area. When
questioned about these changes, many jurors cited that improved knowledge helped
them better understand the implications of different actions (information effect).
The decline of symbolic attitudes, with concomitant changes to preference ranking
was one of the most striking outcomes of the process. This appears to reflect the
ability of deliberation to decouple attitudes from extreme positions transmitted to
284 Simon Niemeyer
the public domain through the political discourse of interest groups and issue
polarisation.
The evidence presented to the jury was based on a very small amount of research
directly relevant to the road. Damage to the fringing reefs is one of the few impacts
of the Bloomfield Track that has been studied to date. However, there was enough
information to address many of the myths and symbols that had dominated the
issue to that date. Most notable was the effect of evidence on the impact of the
road on the fringing reefs. Findings from this research, as presented to the jury, do
not support this contention. Rather than focus on these issues, jurors considered
the broader and longer-term implications of the future of the Bloomfield Track.
The fact that jurors were encouraged by the process to look at the issue from
the perspective of the citizen, rather than what benefited them individually (as
consumer) was also cited by jurors as a significant feature in changes to attitudes
and preferences. Based on observations and interviews with jurors, this perspective
change was most important insofar as it served as a pretext for the manner in
which information was weighted and used in the judgement process. There is
evidence in the Q sort for an attitudinal change that impacted on preferences.
There is also evidence of ‘better construction’ of preferences, to the extent that
individuals are better informed about what policy means best suit their desired
ends and a concomitant increase in correlations between attitudes and preferences.
Another lesser factor is increased fortitude of jurors to vote for those options they
most preferred.
Preferences tended to converge during deliberation, although discernible differ-
ences remained. Contours of difference emerged around different interpretations
of evidence. These reflected individual predispositions, such as whether they
believed science could provide certain answers, or whether technology could solve
otherwise intractable problems. All jurors agreed on the importance of the area.
However for some, the fact that science could not prove beyond doubt that the
track caused irreparable damage led some jurors to adopt a slightly more conser-
vative and pragmatic approach, deferring closure until more evidence becomes
available. Two other jurors saw opportunities for ‘win–win’ in the evidence
presented: that technically sophisticated options might satisfy competing ends.
Differences also reflected a form ‘starting point bias’, where jurors who began
loaded on one extreme of an attitude–preference spectrum, maintained a residual
affinity with that position. The contours of difference, though still significant,
became more clearly identifiable and therefore arguably easier to address. The
question remains what role further deliberation directly addressing these difference
may have. It is possible that directly addressing these issues might have the same
impact as evidence by dissipating symbolic concerns, leading to acceptance by the
rest of the jury. Based on the evidence from the CJ on the harmonising capacity of
deliberation, it is less likely to lead to increasing polarisation, but this is dependent
on a number of factors, not least being willingness of jurors to tolerate scrutiny of
their formed position by others. In either case it seems that deliberation is never
complete. Despite emergent differences among the jury, the overwhelming trend
was toward increased concern for the longer-term and interconnected issues with
Preference transformation through deliberation 285
the Bloomfield Track, and a concomitant preference for track closure. This outcome
is very different to preferences and attitudes elicited at the beginning of the process,
which varied widely but tended to support maintaining the status quo.
This chapter is part of a broader project of ecological economics to organise
human institutions to elevate the cause of ecological sustainability. It is specifically
concerned with processes that shape preferences to be consistent with this goal.
Within this aim, formal deliberative processes such as CJs are an arguably better
foundation for public input into environmental policy than aggregation exercises
such as referenda or the contingent valuation method (see Niemeyer and Spash
2001, and also Aldred this volume, Chapter 8). The deliberative process helps to
improve judgement as far as preferences tend to match means to desired ends
better than when elicited without subjecting individuals to the deliberative process.
The matching of preferred means to ends suits the goal of sustainability if, because
of deliberation, each individual desires that collective outcome. Indeed, it appears
that the attitudes of participants changed in a manner conducive to the provision
of the common good.
Deliberative approaches may also help to avoid a potential trap for ecological
economics as the ‘science of sustainability’ where predefining outcomes subverts
both democratic process and the achievement of sustainability. Although the trend
favouring preservation was a strong feature of the CJ, the detail of jury discourse
has revealed more subtle differences on the margins of debate about what is the
most preferred outcome. The main driver of these differences was perceptions
among the jurors about the nature of the scientific evidence and the ability of
technical solutions to produce optimal outcomes. If unanimity is the benchmark
for deliberative success, that the CJ produced multiple options could be considered
a less than optimal result. An alternative view is that democratic process should
help to identify substantive differences (for example Benhabib 1996). Within this
view, the CJ marks an important step in defining substantive issue contours to be
addressed by policy. These results are arguably more ‘democratic’ because the
information and deliberative process provided individuals with the necessary tools
to make their own judgements, rather than defer to information they might normally
receive through conventional channels of political discourse. Where conventional
democratic processes might normally provide individuals with predefined policies
deemed by experts or elites to be ‘sustainable’, the deliberative process has produced
positive contributions by identifying a number of possible paths. The result places
the onus back on elites to establish better information from research, or better
educate the public about the use of science in policy. Even if this was the only
outcome of formal deliberative processes, they provide a surer footing for the
development of sustainable policies.
There are important caveats to the conclusions drawn here. First, the results
are only from one deliberative process. Others have conducted similar processes,
with mixed results (for example see Pelletier et al. 1999 and also the previous chapter
in this volume). If anything, this proves that the outcome of deliberative forums is
highly sensitive to context and design. Therefore, they are amenable to problems
of political manipulation and legitimacy. An important question also concerns
286 Simon Niemeyer
whether changes to preferences observed within the CJ could be replicated in the
broader public if the political environment was more ‘discursive’ (in the sense of
Dryzek 2000). If so, rather than rely exclusively on manufacturing formal
deliberative process, a better approach is to facilitate critical deliberation in the
public sphere.
It is possible that adopting formal deliberative approaches in the medium term
will facilitate this transformation, over the long term empowering citizens in the
democratic process through improved and relevant information. However, with
scope for manipulation, the adoption of deliberative approaches will depend on a
genuine desire on the part of organisers and elites to replicate discursive ideals.
An important conclusion is that the ability of deliberation to transform preferences
cannot be divorced from an appreciation of political context. This provides even
greater impetus for ecological economics to embrace the political dimension of
the science of sustainability.
Notes
1 For example, in the case of voting, Downs (1957) observed that voters had little incentive to
acquire information about alternatives because of the small marginal gain and individual impact
on election outcome.
2 There is a possibility that deliberation may not always lead to better preferences. Symbolic
preferences may not dissolve in the process of acquiring information during the deliberative
process, but become entrenched. Depending on an individual’s ability to assimilate information,
it is entirely possible that too much detail can lead to information overload and reduce capacity
for judgement, for that is what contributes to symbolic preference in the first place. Within the
deliberative process there is an ongoing tension between the acquisition and assimilation of
information, which if not properly balanced can lead to dramatically different results (e.g. Pelletier
et al. 1999).
3 There was a statutory obligation under state law that an environmental impact assessment be
prepared for major public works, often treated as no more than a symbolic requirement in an
administrative environment geared for development. To satisfy this requirement, the shire
engineer who oversaw the construction of the road prepared a short impact assessment, which
failed to find against road development.
4 A combination of factors has meant that the Bloomfield Track has effectively been in policy
stasis since construction. This was partially broken in 1998 when, due to prohibitive cost of
maintenance of a dirt track following steep contours in a heavy rainfall region, the Douglas
Shire Council applied to the Wet Tropics Management Authority to upgrade the steepest section
and lay gravel along major sections. They granted an approval for the works on the condition
that the local council prepared a comprehensive environmental impact assessment with
recommendations for providing for long-term management. To date the impact assessment has
yet to be commissioned.
5 Jurors were accompanied on this trip by an engineer from the Douglas Shire who was on hand
to answer questions about construction of the road, and a well-known local and tour operator
in the Daintree area who gave the issue a historical and local context as well as pointing out
major features.
6 Representation of indigenous interests was mainly achieved by the fortuitous recruitment of a
Kuku-Ylanji (traditional owners of the Daintree area) woman as part of the jury.
7 Preferences were not surveyed at this midway point because of concern that, because of the
simpler nature of the questionnaire, preferences surveyed too close together might be biased by
jurors feeling that they should be consistent with previous preference orderings.
Preference transformation through deliberation 287
8 Although for this research there is a mixing of methodologies by assessing changes within
individuals over time.
9 This approach, referred to as ‘forced distribution’ is commonly used in Q method to facilitate
analysis by keeping means equivalent, but does not affect the actual result so long as the rest of
the study is well designed (Brown 1980: 198). Some of the jurors found it difficult to conform to
this imposed regime and were given the freedom to depart from the template where they felt
doing so better represented their position.
10 Statements were grouped into whether they pertained to beliefs about the consequences or the
fundamental beliefs about the importance of those consequences. Statements were also grouped
according to whether they involved a level of knowledge about the issue, the way in which the
issue was framed or conceived – a citizen or a consumer, for example – and the time frame over
which the implications of decisions about the issue are considered. Statements within each
category were phrased either as a positive statement, a negative statement, or a statement
concerning the level of confidence in a particular belief. Expression of certainty is an important
element of the concourse and including it in the survey assists the researcher to detect changes
in confidence as a result of the deliberative process.
11 Importantly, when making recommendations at the end of the process the jury was not limited
to these options. The ranking of preferences was simply used to follow the way in which their
ordering of these particular options had changed.
12 The loading on statement 12 tended to reflect the juror’s predisposition to finding technical
solutions (of which one was the construction of a mono, or sky rail) or regulatory solutions to
the Bloomfield Track issue.
13 Because of the controversial nature of the issue, jurors were given the option of choosing
codenames to protect their identity in publications arising from this research.
14 Significantly, most revealed markedly higher shifts during the first half of the process, in some
cases ‘settling back’ toward their original position, though still substantially changed. Due to
space restrictions, the implications of this cannot be discussed here.
15 Based on a level of significance of 1/ n 0.15 all jurors are significantly loaded on the factor at
the outset, with only Keith defecting. However, although statistical significance is the commonly
accepted test for difference, Q method is based on a measure of ‘psychological significance’,
which is a phenomenologically rather than statistically determined level (Brown 1980: 198).
Using the statistically accepted level for significance suggests a higher level of agreement among
jurors than was actually observed during the CJ proceedings. For the purposes here, because of
the high level of correlation between factors and significant loading of jurors on more than one
factor, a level of 0.4 is adopted.
16 This type of response is consistent with the ‘pioneering’ ethic ascribed to remote areas of
Northern Australia. Many authors portray early pioneers in terms of rapacious developers with
little concern for the environment and future consequences (e.g. Lines 1992). Passmore (1974)
illuminates the issue from a different perspective whereby actions were driven precisely because
of concern for future generations, but unconditioned by awareness, or belief in, negative environ-
mental consequences. If there is certainty that exploitation of environmental resources has
negative consequences, then there is a case for changing course.
17 There are a number of similarities with the Individualist position identified in Steg and Sievers
(2000).
18 This is also consistent with the pioneering ethic, as was epitomised by the Premier of Queensland
at the time the Bloomfield Track was constructed (Passmore 1974: 77; Fitzgerald 1984).
19 Symbolic contrasts most strongly with the Pragmatic (r = 0.38), with disagreement on the question
of the reef (7, 24, 33) and impact of the road on the rainforest (27).
20 One participant expressed pleasure that consensus was not reached because: ‘the fact that we
did not come to a definite solution, I think, shows how successful the jury process is at tackling
any problem … this problem does not have a definite solution’.
21 Many of the technical witnesses themselves pointed out the poor quality of the evidence directly
relevant to the road gathered to that date.
288 Simon Niemeyer
References
Barry, J. and Proops, J. (1999) ‘Seeking sustainability discourses with Q methodology’,
Ecological Economics, 28: 337–46.
Benhabib, S. (1996) ‘Toward a deliberative model of democratic legitimacy’, in S. Benhabib
(ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton, NJ: Princeton
University Press.
Brown, S.R. (1980) Political Subjectivity: Applications of Q Methodology in Political Science, New
Haven: Yale University Press.
Brown, S.R. (1993) ‘A primer on Q methodology’, Operant Subjectivity, 16: 91–138.
Callicot, J.B. (1982) ‘Hume’s is–ought dichotomy and the relationship of ecology to
Leopold’s land ethic’, Environmental Ethics, 4: 163–73.
Common, M. and Perrings, C. (1992) ‘Toward an ecological economics of sustainability’,
Ecological Economics, 6: 7–34.
Downs, A. (1957) An Economic Theory of Democracy, New York: Harper & Row.
Dryzek, J.S. (2000) Deliberative Democracy and Beyond: Liberals, Critics, Contestations, Oxford:
Oxford University Press.
Elster, J. (1983) Sour Grapes: Studies in the Subversion of Rationality, Cambridge: Cambridge
University Press.
Etzioni, A. (1986) ‘The case for a multiple utility conception’, Economics and Philosophy, 2:
159–83.
Fitzgerald, R. (1984) From 1915 to the Early 1980s: A History of Queensland, St Lucia, Australia:
University of Queensland Press.
Goodin, R.E. (1996) ‘Enfranchising the earth and its alternatives’, Political Studies, 44:
835–49.
Goodin, R.E. (2000) ‘Democratic deliberation within’, Philosophy & Public Affairs, 29: 81–
109.
Harrison, C.M., Burgess, J. and Clark, J. (1999) ‘Capturing values for nature: ecological,
economic and cultural perspectives’, in J. Holder and D. McGillivray (eds) Locality and
Identity: Environmental Issues in Law and Society, vol. 3, Dartmouth: Ashgate: 85–110.
Keat, R. (1994) ‘Citizens, consumers and the environment: reflections on “The economy
of the earth” ’, Environmental Values, 2: 333–49.
Kellert, S.R. and Wilson, E.O. (1993) The Biophilia Hypothesis, Washington, DC: Island
Press.
Kellow, A. and Niemeyer, S. (1999) ‘The development of environmental administration in
Queensland and Western Australia: why are they different?’, Australian Journal of Political
Science, 34: 205–22.
Lines, W.J. (1992) Taming the Great South Land: A History of the Conquest of Nature in Australia,
North Sydney: Allen & Unwin.
Niemeyer, S.J. and Spash, C.L. (2001) ‘Environmental valuation analysis, public delibera-
tion, and their pragmatic syntheses: a critical appraisal’, Environment and Planning C, 19:
567–86.
Norton, B.G., Costanza, R. and Bishop, R.C. (1998) ‘The evolution of preferences: why
‘sovereign’ preferences may not lead to sustainable policies and what to do about it’,
Ecological Economics, 24: 193–211.
O’Neill, J. (1997) ‘Managing without prices: the monetary valuation of biodiversity’, Ambio,
26: 546–50.
Papadakis, E. (1993) Politics and the Environment: The Australian Experience, St Leonards, NSW:
Allen & Unwin.
Preference transformation through deliberation 289
Passmore, J. (1974) Man’s Responsibility For Nature: Ecological Problems and Western Traditions,
London: Duckworth.
Pelletier, D., Kraak, V., McCullum, C., Uusitalo, U. and Rich, R. (1999) ‘The shaping of
collective values through deliberative democracy: an empirical study from New York’s
North Country’, Policy Sciences, 32: 103–31.
Rydin, Y. (1999) ‘Can we talk ourselves into sustainability? The role of discourse in the
environmental policy process’, Environmental Values, 8: 467–84.
Sagoff, M. (1988) The Economy of the Earth: Philosophy, Law and the Environment, Cambridge:
Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24: 213–30.
Schkade, D.A. and Payne, J.W. (1994) ‘How people respond to contingent valuation
question: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26: 88–109.
Sears, D.O. (1993) ‘Symbolic politics: a socio-psychological theory’, in S. Iyengar and W.J.
McGuire (eds) Explorations in Political Psychology, vol. 5, Durham: Duke University Press.
Slovic, P. (1995) ‘The construction of preference’, American Psychologist, 50: 364–71.
Smith, G. and Wales, C. (2000) ‘Citizens’ juries and deliberative democracy’, Political Studies,
48: 51–65.
Spash, C.L. and Hanley, N. (1995) ‘Preferences, information and biological diversity’,
Ecological Economics, 12: 191–208.
Steg, L. and Sievers, I. (2000) ‘Cultural theory and individual perceptions of environmental
risks’, Environment and Behaviour, 32: 250–69.
van den Bergh, J., Ferrer-Carbonell, A. and Munda, G. (2000) ‘Alternative models of
individual behaviour and implications for environmental policy’, Ecological Economics,
32: 43–61.
Ward, H. (1999) ‘Citizens’ juries and valuing the environment: a proposal’, Environmental
Politics, 8: 75–96.
Wilderness Action Group (1983) The Trials of Tribulation, Mossman: Wilderness Action
Group.
290 Index
Index