. 1
( 11)



>>

Applied Environmental Economics

The complex real-world interactions between the economy and the environment
form both the focus of and the main barrier to applied research within the ¬eld
of environmental economics. However, geographical information systems (GIS)
allow economists to tackle such complexity head on by directly incorporating di-
verse datasets into applied research rather than resorting to simplifying and often
unrealistic assumptions. This innovative book applies GIS techniques to spatial
cost-bene¬t analysis of a complex and topical land use change problem “ the con-
version of agricultural land to multipurpose woodland “ looking in detail at issues
such as opportunity costs, timber yield, recreation, carbon storage, etc., and em-
bracing cross-cutting themes such as the evaluation of environmental preferences
and the spatial transfer of bene¬t functions.

ian j. bateman is Professor of Environmental Economics at the School of
Environmental Sciences, University of East Anglia, and Senior Research Fellow
at both the Centre for Social and Economic Research on the Global Environment
(CSERGE) and the Centre for the Economic and Behavioural Analysis of Risk
and Decision (CEBARD), University of East Anglia. His previous publications in-
clude Economic Valuation with Stated Preference Techniques (2002, with Richard
Carson et al.), Valuing Environmental Preferences (1999, edited with Ken Willis),
and Environmental Economics (1993, with R. Kerry Turner and David Pearce). He
is Executive Editor of the journal Environmental and Resource Economics.

a n d r e w a . l ov e t t is Senior Lecturer at the School of Environmental Sciences,
University of East Anglia. His research focuses on the application of geographical
information systems, and he has previously published articles in Risk Analysis,
Social Science & Medicine, the Journal of Environmental Management, and the
International Journal of GIS. He is currently chair of the Geography of Health Re-
search Group of the Royal Geographical Society“Institute of British Geographers.

julii s. brainard is Senior Research Associate at CSERGE, University of East
Anglia. Her research background includes GIS, bene¬t transfer, outdoor recreation
and environmental equity.
A P P L I E D E N V I RO N M E N TA L
ECONOMICS
A GIS Approach to Cost-Bene¬t Analysis


I A N J . BAT E M A N
A N D R E W A . L OV E T T
J U L I I S . B R A I NA R D
©¤§ µ®©© °
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press
The Edinburgh Building, Cambridge  µ, United Kingdom
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521809566

© Ian J. Bateman, Andrew A. Lovett and Julii S. Brainard 2003


This book is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.

First published in print format 2003

©®-± 
isbn-13 978-0-511-06409-8 eBook (NetLibrary)
©®-±° 0-511-06409-8 eBook (NetLibrary)

isbn-10
©®-± 978-0-521-80956-6 hardback

isbn-13
©®-±° 
isbn-10 0-521-80956-8 hardback




Cambridge University Press has no responsibility for the persistence or accuracy of
µ¬s for external or third-party internet websites referred to in this book, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
For Fiona, Ben, Freya and Natasha: my world. With love, Ian.
For Mum and Dad. With love and many thanks, Andrew.
For Isabel, Dan and John. Con cari˜ o, Julii.
n
Contents




List of plates page ix
List of ¬gures x
List of tables xii
Foreword by David W. Pearce xv
Preface xix
Acknowledgements xxi
1 Introduction 1
2 Recreation: valuation methods 15
3 Recreation: predicting values 43
4 Recreation: predicting visits 91
5 Timber valuation 111
6 Modelling and mapping timber yield and its value 158
7 Modelling and valuing carbon sequestration in trees,
timber products and forest soils 184
8 Modelling opportunity cost: agricultural output values 219
9 Cost-bene¬t analysis using GIS 250
10 Conclusions and future directions 285
References 293
Index 332




vii
Plates




between pages 266 and 267

1 Predicted timber yield class (YC): (a) Sitka spruce; (b) beech
2 (a) Predicted farm-gate income for sheep farms; (b) Predicted shadow value for
sheep farms; (c) Predicted farm-gate income for milk farms; (d) Predicted
shadow value for milk farms
3a The farm-gate net bene¬t of retaining sheep farming as opposed to conversion
to conifer woodland (de¬ned as timber plus grants only, i.e. present
situation): 6% discount rate
3b The social net bene¬t of retaining sheep farming as opposed to conversion to
conifer woodland (de¬ned as timber, carbon storage and recreation, the
latter measured using contingent valuation): 6% discount rate
3c The farm-gate net bene¬t of retaining milk farming as opposed to conversion to
conifer woodland (de¬ned as timber plus grants only, i.e. present situation):
6% discount rate
3d The social net bene¬t of retaining milk farming as opposed to conversion to
conifer woodland (de¬ned as timber, carbon storage and recreation, the
latter measured using contingent valuation): 6% discount rate
3e The farm-gate net bene¬t value of retaining sheep farming as opposed to
conversion to broadleaf woodland (de¬ned as timber, carbon storage and
recreation, the latter valued using the ITC measure): 6% discount rate
3f The social net bene¬t of retaining sheep farming as opposed to conversion to
broadleaf woodland (de¬ned as timber, carbon storage and recreation, the
latter valued using the ITC measure): 6% discount rate
3g The farm-gate net bene¬t of retaining sheep farming as opposed to conversion
to conifer woodland (de¬ned as timber plus grants only, i.e. present
situation): 3% discount rate




ix
Figures




1.1 The total economic value of woodland page 2
1.2 Representing real-world phenomena as raster or vector data layers 6
1.3 Costs and bene¬ts of woodland 8
2.1 Methods for the monetary assessment of non-market and environmental
goods 16
2.2 The value formation process 21
3.1 Graph of the ratio of stated to GIS-calculated distance against calculated
distance 79
3.2 Comparison of 1 km grid reference with county centroid trip origins 86
4.1 Travel time zones for the Thetford Forest study 93
4.2 Digital road network for Wales and the English Midlands 98
4.3 Population density surface for Wales and the English Midlands 99
4.4 5 km grid points used to generate the predicted woodland visitors
surface 101
4.5 Woodland recreation demand in Wales: predicted annual total party
visits per site 103
4.6 Woodland recreation demand in north-western Wales: predicted annual
total party visits per site 103
4.7 Predicted value of total annual woodland recreation demand per site using
two valuation estimates: (a) lower-bound values based on cross-study
analysis of CV values; (b) upper-bound values based on ITC study 104
5.1 Forestry Commission, private sector and total annual forestry planting,
Great Britain 1946“2000 114
5.2 Price“size curve for conifers in England and Wales 131
5.3 Discount factor curves 133
5.4 Price“size curves for beech in Great Britain 135
5.5 Farmers™ private timber values for Sitka spruce (annualised equivalents
of a perpetual series of optimal rotations: r = 3%). Various yield
classes and subsidy types 149
5.6 Farmers™ private timber values for beech (annualised equivalents of a
perpetual series of optimal rotations: r = 3%). Various yield classes
and subsidy types 150


x
List of ¬gures xi

5.7 Social value for Sitka spruce (annualised equivalent of a perpetual series
of optimal rotations). Various yield classes and discount rates 156
5.8 Social value for beech (annualised equivalent of a perpetual series of
optimal rotations). Various yield classes and discount rates 156
6.1 Aspect effects for Sitka spruce and beech in differing locations 173
6.2 Predicted timber social NPV sums for perpetually replanted Sitka spruce:
3% discount rate 180
7.1 Total carbon storage curves for unthinned and thinned Sitka spruce: 5%
discount rate 190
7.2 Longevity of Sitka spruce timber when put to different uses 192
7.3 Thinning factor for beech 205
7.4 Annual carbon liberation distributions for products and waste expressed as
a proportion of total carbon sequestration in wood from one rotation of
Sitka spruce 206
7.5 Annual carbon liberation distributions for products and waste expressed
as a proportion of total carbon sequestration in wood from one rotation
of beech 207
7.6 NPV of net carbon storage in live wood, products and waste from an
optimal ¬rst rotation of Sitka spruce: 3% discount rate 213
7.7 NPV of net carbon storage in live wood, products and waste from an
optimal ¬rst rotation of beech: 3% discount rate 214
7.8 NPV of net carbon ¬‚ux (live wood, products, waste and soils), Sitka
spruce: 3% discount rate 217
8.1 Model of a typical CAP price support system 221
8.2 Sheep stocking intensity in Wales, 1972 to 1997 227
9.1 Location of Forestry Commission sub-compartments of Sitka spruce in
Wales (superimposed upon elevation) 283
Tables




1.1 Typical questions that a GIS can be used to answer page 6
2.1 Welfare change measures obtained from expressed preference
measures 18
2.2 WTP for preservation of the Norfolk Broads using various elicitation
methods 27
2.3 Payment vehicle analysis results 28
2.4 ZTC/ITC consumer surplus estimates for six UK forests 32
3.1 Forest users™ per person per visit recreation values from ZTC studies 45
3.2 Forest users™ per person per visit recreation values from CV studies 46
3.3 Woodland recreation values from a cross-study analysis of CV
estimates 51
3.4 Summary WTP responses for the Thetford 1 CV study 53
3.5 Thetford 1 TC study: consumer surplus estimates for three functional
forms 56
3.6 Summary WTP results: per annum (WTPpa) and per visit (WTPfee)
formats 58
3.7 Stepwise regression of lnWTPpa on signi¬cant predictors 60
3.8 Farm characteristics and farmers™ willingness to accept compensation for
transferring from present output to woodland 62
3.9 Mean WTP (tax) per annum and 95 per cent con¬dence intervals for each
subsample (including payment principle refusals as zeros) 68
3.10 Mean WTP (fee) per visit and 95 per cent con¬dence intervals for each
subsample (including payment principle refusals as zeros) 69
3.11 Average road speed estimates 77
3.12 Sensitivity analysis: ML models (best-¬tting model shown in italics) 82
3.13 Sensitivity analysis: OLS models (best-¬tting model shown in italics) 84
3.14 Sensitivity analysis: effects of varying outset origin on TC bene¬t
estimates 87
3.15 Valuing recreational visits to woodland: a synthesis of studies 88
4.1 Observed and predicted visitor rates 94
4.2 Of¬cial recreational visit numbers, predictions of arrivals and consumer
surplus estimates for twenty-seven English woodlands 106


xii
List of tables xiii

5.1 Forestry Commission holdings: Great Britain 1978“2000 (™000 ha) 116
5.2 High forest by general species: Forestry Commission and private
woodland in Great Britain 1947“2000 (™000 ha) 121
5.3 Woodland Grant Scheme payments (£/ha) 125
5.4 Woodland Management Grants 125
5.5 Payments under the Farm Woodland Premium Scheme (£/ha per
annum) 127
5.6 Optimal felling age for various discount rates: Sitka spruce, YC6“24 134
5.7 Optimal felling age for various discount rates: beech, YC4“10 136
5.8 Agricultural nominal rate of return (RoR) on tenants™ capital: Wales
1987/88“1991/92 140
5.9 Farmers™ private timber values for high-output Sitka spruce and beech
across various discount rates (annualised equivalents of a perpetual
series of optimal rotations) 150
6.1 Variables obtained from the SCDB 164
6.2 Variables obtained from LandIS 166
6.3 Comparing actual with predicted YC for Sitka spruce (cell contents are
counts) 171
6.4 Comparing actual with predicted YC for beech (cell contents are
counts) 173
6.5 Predicted Sitka spruce YC under three scenarios 176
6.6 Predicted beech YC under two scenarios 177
6.7 NPV sums for perpetually replanted Sitka spruce timber across various
discount rates 181
6.8 Annuity values for perpetually replanted Sitka spruce timber across
various discount rates 182
6.9 NPV sums for perpetually replanted beech timber across various
discount rates 182
6.10 Annuity values for perpetually replanted beech timber across various
discount rates 183
7.1 The social costs of CO2 emissions ($/tC): comparison across studies 188
7.2 Softwood and hardwood end uses for UK domestic production 1991/92 193
7.3 Post-afforestation changes in equilibrium soil carbon storage levels
for various soils previously under grass (tC/ha): upland and
lowland sites 196
7.4 Date of ¬rst thinning (TD1) for Sitka spruce yield models (r = 0.05
throughout) 201
7.5 Thinning factor for Sitka spruce (TF SS,t ): YC12 202
7.6 Date of ¬rst thinning (TD1) for beech yield models (r = 0.05
throughout) 204
7.7 NPV of net carbon ¬‚ux (sequestration in live wood and liberation from
products and waste) for an optimal rotation of Sitka spruce: various
yield classes and discount rates (£, 1990) 210
7.8 NPV of net carbon ¬‚ux (sequestration in live wood and liberation from
products and waste) for an optimal rotation of beech: various yield
classes and discount rates (£, 1990) 212
xiv List of tables

7.9 NPV of carbon in live wood, waste and products from an optimal rotation
of Sitka spruce and beech: linear predictive equations with yield class
as the single explanatory variable: various discount rates 212
7.10 NPV of Sitka spruce and beech carbon ¬‚ux for live wood, waste and
products: various discount rates (r) 215
7.11 NPV perpetuity sums for soil carbon ¬‚ux: all tree species (£/ha) 216
7.12 Number of 1 km land cells at differing levels of NPV for net carbon ¬‚ux
(live wood, waste, products and soils): Sitka spruce, various discount
rates (r) 216
8.1 Change in Welsh agriculture 1990 to 2000 226
8.2 FBSW annual farm account data: example of a typical farm record 232
8.3 Agroclimatic variables obtained from LandIS 234
8.4 Farm cluster characteristics: average income and mean percentage of
total revenue from speci¬ed activities in each cluster of farms 236
8.5 Best-¬tting stage 1 models of farm surplus/ha on sheep (cluster 1)
and milk (cluster 2) farms 240
8.6 Best-¬tting stage 2 models for sheep farms 242
8.7 Best-¬tting stage 2 models for milk farms 244
8.8 Predicted farm surplus values for sheep and milk farms 247
8.9 Predicted farm-gate income and shadow values for sheep and milk
farms 248
9.1 Distribution of the net bene¬ts of retaining sheep farming in Wales as
opposed to conversion to conifer (Sitka spruce) woodland: 6%
discount rate 254
9.2 Distribution of the net bene¬ts of retaining milk farming in Wales as
opposed to conversion to conifer (Sitka spruce) woodland: 6%
discount rate 260
9.3 Distribution of the net bene¬ts of retaining sheep farming in Wales as
opposed to conversion to broadleaf (beech) woodland: 6% discount
rate 264
9.4 Distribution of the net bene¬ts of retaining milk farming in Wales as
opposed to conversion to broadleaf (beech) woodland: 6% discount
rate 268
9.5 Distribution of the net bene¬ts of retaining sheep farming in Wales as
opposed to conversion to conifer (Sitka spruce) woodland: 3%
discount rate 270
9.6 Distribution of the net bene¬ts of retaining milk farming in Wales as
opposed to conversion to conifer (Sitka spruce) woodland: 3%
discount rate 274
9.7 Distribution of the net bene¬ts of retaining sheep farming in Wales as
opposed to conversion to broadleaf (beech) woodland: 3% discount
rate 276
9.8 Distribution of the net bene¬ts of retaining milk farming in Wales as
opposed to conversion to broadleaf (beech) woodland: 3% discount
rate 278
Foreword




Much of environmental change is driven by land use change. To some, the whole
history of economic and social development re¬‚ects the exchange of one form
of asset “ ˜natural™ landscape “ for another form of asset “ man-made capital.
Certainly, viewed from a global perspective, there is a one-to-one relationship
between the decline of forested land and the increase in land devoted to crops
and pasture. The factors giving rise to land use change are many and varied. But
one of the most powerful is the comparative economic returns to ˜converted™ land
relative to the economic returns to ˜natural™ land. In short, the issue is conservation
versus conversion, and this is a con¬‚ict that is invariably resolved in the favour
of conversion. This systematic erosion of the natural capital base is what worries
environmentalists, a term I take to embrace anyone with the slightest modicum
of concern about what humankind is doing to its own environment and its fellow
species. Acting on that concern takes several forms, as everyone knows. Some want
to lie down in front of the bulldozers, protest to their Members of Parliament, write
to the newspapers, appeal to some moral principle or other. For the most part quietly,
environmental economists have sought a different route. First, they observe that the
bias towards conversion arises from all kinds of incentive systems, including, for
example, subsidies to agriculture or monocultural forestry. Second, some of those
incentive systems are far more subtle, and arise from the fact that many of the
functions and services provided by natural systems have no market. At the end of
the day, and like it or not, the ¬nancial balance sheet drives land conversion. It pays
to convert land because the ¬nancial returns from conversion exceed those from
conservation. The same bias works in reverse: existing land is not converted back
to, say, woodland because some of the woodland bene¬ts have no market.
But this is a result that derives from a perversion of economics “ markets ˜fail™
to allocate resources properly because many of those resources have no price, even
though they have potentially substantial economic value. Markets are the medium
through which prices materialise. If there is no market in the carbon stored in forest

xv
xvi Foreword

biomass, then markets will ignore the fact that the carbon has an economic value.
In turn, that value derives from carbon dioxide being ˜¬xed™ by growing biomass
or from the fact that it is stored rather than released as carbon dioxide, the main
greenhouse gas.
These observations de¬ne the ¬rst stage of the economic argument for correcting
the economic system™s biases. This stage consists of ˜demonstrating™ that economic
value resides in natural systems and estimating how much it is. The second stage is
partly addressed in this volume, but it involves the redesign of institutions so that the
˜missing™ economic value is captured and represented as a ¬nancial ¬‚ow. There are
many examples of such capture mechanisms “ environmental taxes, tradable pol-
lution and resource permits, payments for ecological services, and so on. If there is
an encouraging trend in the environmental world it is that, gradually, these capture
mechanisms are expanding. Sometimes aided by policy initiatives, and sometimes
spontaneous, they help shift the bias of conversion back towards more conservation
than would otherwise be the case. In terms of this volume, Ian Bateman and his
colleagues look at how farm incomes would change if only the non-market value
of land (e.g. stored carbon, recreation) was ˜monetised™ and added to some of the
market values from changed land use (e.g. timber).
Determining economic values has become ˜big business™ for environmental
economists, and few can match the authors of this volume for ingenuity and ap-
plication of the various techniques that have evolved for ¬nding these values. But
˜valuation™ is expensive, or, at least, that™s how policy-makers like to see it. Millions
may be spent on engineering design and legal fees in the context of policy or in-
vestment projects. A few tens of thousands of pounds on a valuation study often
produces the cry that it is ˜too expensive™. In the absence of a saner approach,
environmental economists have to live with the very limited resources allocated
to valuation. That means that short-cuts are unavoidable. Results from one study
have to be ˜borrowed™ and applied to another study area. But a much understud-
ied issue is the reliability of making these ˜transfers™. Transferability requires that
the conditions at the ˜new™ site should at least be similar to the conditions at the
previously studied site. Often they are not. A few attempts have been made in the
past to adapt transferred values to account for different site characteristics. With
hindsight, it seems almost obvious that the logical way to handle variability in site
characteristics is through geographical information systems (GIS). But it wasn™t
done, and the dominant attraction of this volume is that it shows how to do it in the
context of a detailed case study. The ¬nal analysis is a mix of ˜transfer™ estimates,
modulated by the GIS, and validation of those transfers against actual data for their
geographical focus, Wales.
Ian Bateman and his colleagues have successfully pushed back the frontiers in
several ways. First, they have ˜married™ economic valuation with GIS. Second,
Foreword xvii

they have taken a very broad area for their application “ the whole of Wales. Third,
they have hypothetically recon¬gured land use in Wales under the assumption that
currently non-market land services and changed market values are integrated into
farm incomes. This amounts to a cost-bene¬t analysis because they compare the
costs of this change with its bene¬ts. They are far more modest than I would be
about the power and importance of cost-bene¬t analysis. It is fashionable to criticise
the economic approach for all kinds of supposed ethical aberrations, but it has an
ethical force of its own. It is democratic in that it allows individuals™ preferences to
rule rather than those of unelected ˜stakeholders™ and experts. It reminds us all the
time that all decisions involve costs as well as bene¬ts. While these may seem small
claims, the reality is that actual decision-making all too often reduces to choices
by an elite with little reference to cost. It is worth remembering that cost always
reduces to a taxpayer™s burden: there is no such thing as ˜government money™.
Finally, cost-bene¬t analysis is itself changing. Recent work on valuing the long
distant future and on allowing for irreversibility and uncertainty (effectively making
rigorous sense of the otherwise ill-de¬ned ˜precautionary principle™) means that it
is time to rewrite the cost-bene¬t textbooks. In so doing, we would overcome many
of the criticisms advanced against it.
So, I would make greater claims for the approach adopted in this book than the
authors make for it themselves! But what cannot be disputed is that we have a ¬ne
example here of economic valuation being put to an imaginative and unique use by
some of the most exciting practitioners of the art of economic valuation.

David W. Pearce
Preface




This book concerns the application of environmental economic analysis to real-
world decision-making. In particular it seeks to demonstrate a number of ways
in which geographical information systems (GIS) can be employed to enhance
such analyses. We have written it because, in our opinion, GIS techniques can
considerably improve the way in which the complexities of the real world can be
brought into economic cost-bene¬t analyses (CBA)1 , so reducing the reliance upon
simplifying assumptions for which economists are infamous.
As we are primarily interested in demonstrating the ¬‚exibility and applicability
of GIS techniques to a diversity of situations, we assume no prior knowledge of
such techniques and avoid unnecessary technicalities wherever possible by referring
the interested reader to related academic papers throughout. In so doing it is our
objective to appeal to students, researchers, academics and, in particular, decision-
makers and analysts across a broad spectrum of disciplines including economics
(especially environmental, agricultural and resource economics), geography, land
use planning and management, environmental science and related policy studies.
The application of GIS to environmental economic analyses is introduced grad-
ually through the use of a diverse land use change case study. This concerns the po-
tential for converting surplus agricultural land to multipurpose woodland in Wales.
However, neither the speci¬cs of this case study nor its location need be of par-
ticular interest to the reader as the study is designed primarily to demonstrate the
¬‚exibility of the underlying approach. The book opens by reviewing some basic
economic ideas concerning value and CBA (Chapter 1), focusing in particular
upon methods for valuing individuals™ preferences for non-market goods such as
those provided by the environment (Chapter 2). Previous studies of the recreational
value of open-access woodland are reviewed and some new applications presented
(Chapter 3) through which we ¬rst introduce the use of GIS techniques as a means
1 Or bene¬t-cost analysis, depending upon which side of the Atlantic/Paci¬c you reside.


xix
xx Preface

of enhancing valuation methods. This approach is then extended to the estima-
tion of the numbers of visitors arriving at existing or potential future woodland
recreation sites (Chapter 4). We then turn to consider certain other forest bene¬ts
starting with the value of timber (Chapter 5). Again GIS techniques are used to
bring together a host of diverse datasets to permit modelling of timber yield and its
net value (Chapter 6). These techniques are then extended to conduct an analysis
of the carbon sequestration value of woodland, combining models of carbon ¬‚ux
in live trees, timber products and forest soils (Chapter 7). The opportunity cost of
converting agricultural land to woodland is then examined, with GIS providing the
medium for undertaking assessments of the principal farming sectors in the case
study area (Chapter 8). All of these sub-analyses are synthesised through our GIS to
undertake a spatial CBA considering, for each location across our entire study area,
what the consequences of land use change from agriculture to woodland would be
(Chapter 9). Finally we summarise the strengths and weaknesses of our particular
application and consider the wider conclusions to be drawn from the approach set
out in this volume (Chapter 10).
We hope that readers will ¬nd this book interesting and enjoyable and that it
might contribute to what we believe would be a timely infusion of realism into
economic analyses.
Acknowledgements




The inherently interdisciplinary nature of this project involved a lot of help from a
lot of people. In particular we wish to thank Stavros Georgiou, Phil Judge, the late
(and much missed) Ian Langford, Frances Randell, Gilla S¨ nnenberg and Kerry
u
Turner at the University of East Anglia and Chris Ennew and Tony Rayner at the
University of Nottingham.
We are also tremendously grateful to the Farm Business Survey of Wales (in
particular to Nigel Chapman, Tim Jenkins and the surveyors at FBSW, Aberyst-
wyth), to the Soil Survey and Land Research Centre (in particular to Ian Bradley
and Arthur Thomasson) and to the Forestry Commission (in particular Chris Quine
and Adrian Whiteman at the Commission™s Northern Research Station, Roslin) for
provision of, and advice concerning, the data used in this analysis. Quite simply
this work could not have been undertaken without their support.
The research contained in this volume was funded in part by the Economic and
Social Research Council (ESRC) as part of the Centre for Social and Economic
Research on the Global Environment (CSERGE) Programme in Environmental
Decision Making.




The publisher has used its best endeavours to ensure that the URLs for external
websites referred to in this book are correct and active at the time of going to press.
However, the publisher has no responsibility for the websites and can make no
guarantee that a site will remain live or that the content is or will remain appropriate.

xxi
1
Introduction




The nature of value: differing paradigms
Perhaps the most often quoted de¬nition of an economist is of someone who knows
the price of everything and the value of nothing.1 However, it is an awareness of
the distinction between value and price which separates out the true economist
from the glori¬ed book-keepers and accountants who so often masquerade under
such a title. Recent years have seen a growth of badge-engineering in which so-
called new disciplines such as environmental or ecological economics have risen to
prominence. However, whilst these are appealing titles, in essence they represent
not a radical departure but rather a very welcome return to the basic principles and
domain of economics “ the analysis of true value.
It is one of these basic principles which underpins this study: namely the assump-
tion that values can be measured by the preferences of individuals.2 The interaction
of preferences with the various services provided by a commodity generates a va-
riety of values. Many economists have studied the nature of these values; however,
a useful starting point is the concept of aggregate or total economic value (TEV)
(Pearce and Turner, 1990; Turner, 1999; Fromm, 2000).
Figure 1.1 shows how TEV can be broken down into its constituent parts and
illustrates these with reference to some of the values generated by the principal
commodity under consideration in this study; woodland.
The bulk of economic analyses concentrate upon the instrumental or use values
of a commodity. Most prominent amongst these are the direct use values generated
by private and quasi-private goods (Bateman and Turner, 1993) which are often
partly re¬‚ected by market prices, and those indirect use values associated with pure

1 This is an appropriation of Oscar Wilde™s de¬nition of a cynic in Lady Windermere™s Fan (Act III). However,
given the perceived similarity between the two groups, it is easy to see how such a confusion may have arisen
(with thanks to Olvar Bergland, Colin Price and others regarding this.)
2 Speculations upon this issue and, in particular, about whether individuals have de¬nite preferences are presented
by Sugden (1999a).


1
2 Applied Environmental Economics




Figure 1.1. The total economic value of woodland. (Source: Adapted from Bateman,
1995.)


and quasi-public goods (ibid.) which generally have no market price description. A
unifying characteristic of these values is that they are all generated via the present
use of the commodity by the valuing individual. An extension of the temporal frame
allows for the possibility of individuals valuing the option of future use (Weisbrod,
1964; Cicchetti and Freeman, 1971; Krutilla and Fisher, 1975; Kristr¨ m, 1990).
o
Related to this is the notion of bequest value wherein the valuing individual gains
utility from the provision of use or non-use values to present and/or future others.
Pure non-use values are most commonly identi¬ed with the notion of valuing the
continued existence of entities, such as certain species of ¬‚ora and fauna or even
whole ecosystems. As before, this is generally both an intra- and intergenerational
value and because of the lack of an instrumental element has proved problematic
to measure. Nevertheless, the theoretical case for the ˜existence of existence value™
is widely supported (e.g. Young, 1992).
Wider de¬nitions of value have been argued for. An important issue concerns
the extent of the ˜moral reference class™ (Turner et al., 1994) for decision-making.
One question here involves the treatment of other humans (both present elsewhere
and future) while another is whether animal, plant and ecosystem interests should
be placed on an equal footing with human preferences. The modern origins of such
a view can be traced to O™Riordan (1976), Goodpaster (1978) and Watson (1979)
who take the Kantian notion of universal laws of respect for other persons and
extend this to apply to non-human others. Watson feels that those higher animals
such as chimpanzees (which he argues are capable of reciprocal behaviour) should
be accorded equal rights with humans. Hunt (in Perman et al., 1996) and Rollston
(1988) build upon the land ethic of Leopold (1949) to extend this de¬nition of
moral reference even further to include all extant entities, an approach which
Introduction 3

Singer (1993) de¬nes as the ˜deep ecology™ ethic. Such a paradigm argues that
these entities possess an ˜intrinsic™ value separate from anthropocentric existence
values. A further departure from conventional utilitarianism is proposed by Turner
(1992, 1999) who argues that all the elements of TEV can be seen as secondary
to a primary environmental quality value which is a necessary prerequisite for
the generation of all subsequent values. Side-stepping the theoretical case for
such philosophical extensions, a practical problem with these non-TEV values
is that they are essentially beyond the scope of conventional, anthropocentric,
preference-based economic valuation. If, as in this study, we constrain the moral
reference class to present humans alone, TEV is the appropriate extent of value
de¬nition. However, this still leaves the problem of how such values should be
measured.
One solution to the problem of valuation might be to abandon conventional
neoclassical economic analysis in favour of modi¬ed or alternative appraisal and
decision-making strategies. One such alternative is to base decisions upon expert
judgement and restrict the role of economics to the identi¬cation of least cost
methods for achieving stated aims (see, for example, Organisation for Economic
Cooperation and Development, 1991). Such a cost-effectiveness approach may be
optimal for a resource-rich risk-averse society faced with high risk, high uncer-
tainty, irreversible problems such as the treatment of highly persistent pollutants
(Opschoor and Pearce, 1991). Here a useful decision guide is provided by the
precautionary principle advocated by ˜ecological economics™ (see, for example,
Costanza and Daly, 1992; Toman, 1992; Turner et al., 1995). However, in other,
arguably more general, situations where the precautionary principle does not apply,
a cost-effectiveness approach may entail avoidable and, in some cases, major net
welfare losses compared to a solution based upon cost-bene¬t analysis (CBA).
Such a position is adopted by those who argue for an ˜environmental economics™
paradigm (see, for example, Pearce et al., 1989; Department of the Environment,
1991; Price, 1997a; Pearce, 1998; Grif¬n, 1998; Pearce and Barbier, 2000). Support-
ers of this view accept preference-based values as the basis of decision-making but
argue for full assessment of TEV as opposed to the concentration upon market-based
measures which appears to dominate much present practical decision-making.
This choice between ecological and environmental economics could be char-
acterised as one between principle and pragmatism. The argument for an eco-
logical economics approach is that nothing less will preserve the environmental
integrity which is vital if the present, resource-exploitative, ˜cowboy economy™
(Boulding, 1966) is to attain a state of sustainable development. The environmental
economic critique is that such a rigid approach fails to recognise the mechanisms
through which present-day decision-making operates and thereby risks being ig-
nored by those in power. In the absence of hindsight it is impossible to know which
4 Applied Environmental Economics

strategy is most likely to in¬‚uence the presently unsustainable course of economic
growth.
Our own position is that the two paradigms need not be in con¬‚ict and that
a modi¬ed precautionary principle can be used to assess the most appropriate
approach for any given decision situation. Furthermore, we see a role for public
preferences within this process. In cases where expert assessment and/or informed
public opinion identi¬es high potential risks or uncertainties from a given strategy
or decision then a precautionary, ecological economics approach would appear
justi¬able. For situations where this is not the case then an environmental economics
analysis seems likely to be optimal. From a sustainability perspective, both are
signi¬cantly superior to simple market-based appraisals.


The theoretical and methodological basis of the study
We therefore need to select the appraisal paradigm which is most appropriate for the
subject under analysis. This study examines the economic potential for conversion
of land from conventional agriculture to multipurpose woodland in Wales. Two
points are immediately important here. First, we are interested in the full range
of economic values generated by such a change in land use. Second, following
initial review (Bateman, 1991a,b, 1992), it has become apparent that large-scale
unquanti¬able risks or uncertainties are not a major factor in such an analysis. Given
this, the adoption of a CBA paradigm appears defensible.
CBA is generally thought of as an appraisal of the worth of a project from a
social perspective. That does not mean that CBA tells us about what is good or bad.
Rather, it provides information, going beyond simple market-based assessments to
a more complete analysis of value, which, if correctly employed, should improve
decision-making (Adler and Posner, 1999). In our consideration of the social value
of woodland we have attempted to be reasonably comprehensive although our main
foci of interest are timber production, open-access informal recreation, and the value
of carbon sequestration (i.e. global warming abatement). This is compared to the
social value of agriculture. In both cases we consider items such as the differing
subsidies currently paid by society to those who produce agricultural and forest
products. However, while such a CBA assessment is of use in informing decision-
makers and shaping optimal policy change, it cannot alone predict land-owners™ and
farmers™ responses to that change unless the impacts upon farm incomes are also
known. Consequently, the study also examines farm-gate incomes under present
and future policy scenarios.
The ultimate objective of this study, therefore, is to provide a policy analysis tool.
However, whilst the theoretical CBA framework of the research is conventional,
the extent of application and the methodology employed is innovative and unique.
Introduction 5


The role of geographical information systems
One distinctive feature of our research is the extensive use of geographical in-
formation systems (GIS) throughout our study. A GIS is commonly de¬ned as ˜a
system for capturing, storing, checking, integrating, manipulating, analysing and
displaying data which are spatially referenced to the earth™ (Department of the
Environment, 1987: p. 132). From an organisational perspective, a GIS typically
involves computer hardware, software, data and operating personnel. The origins
of what we now regard as a GIS can be found in the 1960s, but use has only become
widespread in the past ten years (Burrough and McDonnell, 1998; Longley et al.,
1999, 2001). Technologies such as computer-aided design (CAD), image process-
ing, database management systems and automated mapping have all contributed to
the development of GIS, but the last of these represents a distinct advance in terms
of the capacity to integrate data from different sources (e.g. relate point measures of
timber yield to environmental characteristics of areas) and undertake a wide range
of analytical operations. Examples of the types of questions that can be investigated
using a GIS are given in Table 1.1.
The use of GIS in environmental economics is a relatively recent innovation3
and in many ways their application could not be more overdue. The unrealistic as-
sumptions, implicit or otherwise, made by economists in order to implement their
analyses have often attracted critical comment, but GIS provide a means of avoid-
ing many of the worst simpli¬cations (Lovett and Bateman, 2001). For instance,
studies using travel cost techniques to estimate the recreational value associated
with open-access countryside locations have often assumed that all trips take place
in straight lines between origins and destinations, and ignored much of the spatial
heterogeneity within study areas (see discussions in Bateman et al., 1996a, 1999a).
With a GIS, travel costs can be calculated in a manner which is far more sensi-
tive to the nature of the available road network and much greater account can be
taken of spatial variations in the socio-economic characteristics of populations or
the availability of substitute destinations (Brainard et al., 1999). Another example
where the application of GIS has already proved bene¬cial involves hedonic pricing
techniques which aim to isolate the in¬‚uence of environmental characteristics on
property prices. In the past, efforts to examine factors such as views of parks, water
features or industrial areas from properties have required considerable ¬eldwork
(and involved appreciable subjectivity). The combination of high-resolution digital
map databases and GIS, however, now makes it feasible to determine the compo-
sition of viewsheds from far larger numbers of properties in a more objective and
cost-effective manner (Lake et al., 1998, 2000a,b; Bateman et al., 2001a).
3 Among the few studies to date, not otherwise mentioned, to combine GIS and environmental monetary valuation
are Eade and Moran (1996), Bhat and Bergstrom (1997), Geoghegan et al. (1997) and Powe et al. (1997).
6 Applied Environmental Economics

Table 1.1. Typical questions that a GIS can be used to answer

Type of question Example
Identi¬cation What is at a particular location?
Location Where does a certain type of feature occur?
Trend Which features have changed over time?
Routing What is the best way to travel between two points?
Pattern Is there a spatial association between two types of feature?
What if What will happen if a particular change takes place?

Source: Based on Rhind, 1990; Kraak and Ormeling, 1996.




Figure 1.2. Representing real-world phenomena as raster or vector data layers. (Source:
Based on Lovett, 2000.)


It needs to be emphasised that GIS are no universal panacea for improving data
analysis. The quality of the results obtained depends on factors such as the accuracy
of the input information and the appropriateness of the data structures used to store
digital representations of real-world phenomena. Different types of features are
most commonly held in a GIS as separate layers, usually in the form of either raster
grids (where values are assigned to cells) or vector structures (where the positions of
entities such as points, lines or areas are de¬ned by sets of co-ordinates). Figure 1.2
illustrates these two main approaches. Other methods of data storage are possible
(Laurini and Thompson, 1992), but the key principle is that data structures should
be selected to minimise distortion when creating a digital representation of reality
and maximise analytical or presentational options given the intended use of the data
(Berry, 1993; Martin, 1996a; Burrough and McDonnell, 1998).
Introduction 7

Notwithstanding the above caveats, by using GIS in this research we hoped to
overcome many of the limitations in data handling and modelling which have re-
stricted previous research. With such computing facilities we were able to combine
environmental and other spatial data in the form of digital maps and satellite im-
agery with more conventional variables to enhance the stochastic economic models
which are central to this study. As we demonstrate in the contexts of modelling
timber yield, carbon sequestration, recreational demand and agricultural produc-
tivity, the ability to integrate diverse datasets substantially improves our capacity
to understand and predict such variables. However, equally important is the scope
for querying and visualising model output (e.g. in the form of maps), so permitting
the decision-maker readily to comprehend the impact of alternative policy choices.
It is this dual capability to improve modelling and display which we feel allows
GIS signi¬cantly to enhance many aspects of economic analysis. (For a parallel
example in the context of land use, see O™Callaghan, 1996.)


Costs and bene¬ts of woodland: limitations of the study
Forestry has long struggled to compete ¬nancially with other land uses (Green,
1996) but has also been a consistent focus of attention regarding its non-market
attributes (Hodge, 1995; Mather, 1998). Figure 1.3 illustrates the complexity of
internal and external costs and bene¬ts which are generated by woodland. In this
diagram the internal costs and bene¬ts are shown in shaded boxes. These items all
have market prices from which shadow values, de¬ning the value to society of these
goods,4 may be derived. Certain external items also have related market prices from
which values may again be estimated; these are shown in the dotted line boxes of
Figure 1.3. However, the remaining externalities do not have related market prices,
thereby making valuation problematic; indeed such items are typically excluded
from appraisals (Pearce, 1998; Hanley, 2001).
Our study sets out to provide a relatively comprehensive assessment of the values
associated with the proposed conversion of agricultural land into woodland. How-
ever, we have to recognise certain limitations in the research. First, methods for
the monetary evaluation of preferences for non-market goods and services are not
uniformly developed for all types of value. In particular, methods for the evaluation
of non-use bene¬ts, such as existence values, have been the subject of sustained
criticism during recent years (see Chapter 2). Our study re¬‚ects these reservations
by concentrating upon use values. Second, time constraints and data availability

4 Shadow values adjust market prices (which may be zero for unpriced goods) to provide estimates of the value to
society of such goods. Typically this involves adjustments to allow for market failures, such as non-competitive
markets, and transfer payments such as grants and subsidies which are funded by society. Chapter 8 provides
an example of how shadow values may be derived from market prices.
8 Applied Environmental Economics




Figure 1.3. Costs and bene¬ts of woodland. (Source: Bateman, 1992.)

problems mean that even our treatment of all use values is somewhat uneven. Third,
we are only considering conversions from agricultural land to woodland and not to
any other alternative use. Strictly speaking, this contravenes the principles of CBA,
which state that the appraisal of opportunity costs should include the assessment
Introduction 9

of a wide range of feasible alternative resource uses (Pearce, 1986; Bateman et al.,
1993a; Price, 2000; Hanley, 2001). A fourth issue is that of equity “ and its root:
ethics.

Ethical questions5
Ethics and economics have often been presented as strange bedfellows. Indeed,
many proponents of the ˜positive economics™ which has dominated so much of
twentieth-century economic analysis argue that the two concepts cannot be related
˜in any form but mere juxtaposition™ (Robbins, 1935: p. 148). However, this has not
always been a widely held belief. Indeed the early great economists were explicitly
concerned with morality and ethics.6,7
Two ethical positions which have had a major impact upon the development of
economic thought are the libertarian and utilitarian schools of thought. The libertar-
ian view, which may be traced from John Locke and Adam Smith to Robert Nozick
(1974), emphasises respect for the rights of individuals. A fundamental concept
here concerns the just acquisition of property. This has been interpreted as empha-
sising both the rights of ownership and also the requirement of appropriate payment
or transfer in return for acquisition. However, libertarianism makes no prescriptions
concerning the outcome of any trade or transfer. In particular, such a view would
almost always condemn any redistributive policy, whether between present-day
populations or to future populations (intra- and intergenerational transfers) unless
they are freely entered into by all groups including donors.8 This focus upon pro-
cesses rather than outcomes differs from the utilitarian view (which derives from
the writings of David Hume, Jeremy Bentham and, most notably, John Stuart Mill
(1863)), which explicitly highlights the ethical consequences of actions. Classi-
cal utilitarianism judges actions according to whether they are ˜good™ for society,
with ˜good™ being de¬ned (by Mill) in terms of happiness or utility. Actions which
promote utility are therefore good and should be judged by the amount of utility
created. However, for utility to be cardinally measurable, individuals must be able
to express it in terms of a numeric value. Furthermore, in order to assess the social
utility of an action we have to assume that we can compare and add utilities across
individuals.
These strong assumptions make classical utilitarianism of little use for the prac-
tical economic analysis of projects. The neoclassical utilitarianism (Kneese and
Schulze, 1985) which underpins modern welfare economics involves rather weaker

5 This discussion relies heavily on Perman et al. (1996), Kneese and Schulze (1985) and Pearce and Turner (1990).
Relevant discussions are also presented in Beauchamp and Bowie (1988) and Sen (1987).
6 Interestingly Adam Smith™s post at the University of Glasgow was Professor of Moral Philosophy.
7 Reviews of the work of Marx, Marshall, Pareto, Keynes and others are presented in Schumpeter (1952).
8 This would conventionally rule out any governmental action towards the enforced provision of such transfers.
10 Applied Environmental Economics

assumptions (Layard and Walters, 1978; Varian, 1987). In particular, a common as-
sumption underpinning CBA is that the marginal utility of consumption is equal
across all individuals. If this is so we can ignore distributive issues (which are vital
under classical analysis) since any action which creates net bene¬ts unambigu-
ously raises social welfare. However, in reality, such an assumption seems unlikely
to hold, prompting some users of CBA to consider explicitly the equity implications
of their analyses (e.g. Squire and van der Tak, 1975). For many years such views
were held by an inconspicuous minority within the profession of economics. How-
ever, since the 1960s, concerns regarding the effects of environmental degradation
on present and future generations, together with the issue of North/South inequality,
have meant that discussions regarding the ethical basis of economics have grown.
These arguments over the need to consider equity as well as economic ef¬ciency
have recently coalesced within what has been termed the sustainable development
(SD) debate (WCED, 1987; Pearce et al., 1990; Perman et al., 1999).
Both intra- and intergenerational equity issues are central to the SD debate which
has, in essence, proposed an alternative to utilitarianism as a new ethical basis for
economics. Pivotal to this has been the work of Page (1977) and, in particular, Rawls
(1972). Rawls™ theory of justice can be seen as a direct development of Kant™s
universal laws. Here the individual enjoys common liberties compatible with equal
rights for others, while valid inequalities result only from personal qualities which
are attainable by all (e.g. inequalities arising from diligent work or learning as
opposed to those based upon sex or creed). This latter prescription has important
consequences for equity, as Rawls argues that under such a system the optimal allo-
cation of resources is one that is made behind a ˜veil of ignorance™ as to their intra-
and intergenerational distribution. This can be seen as being in direct con¬‚ict with
the individual maximisation principle of utilitarianism.9 Such a contrast is perhaps
most clearly demonstrated in the recent literature regarding sustainability. Turner
and Pearce (1993) identify four alternative positions ranging from ˜very weak™ to
˜very strong™ sustainability. Each de¬nition moves further from a conventional util-
itarian towards a Rawlsian position on equity, steadily imposing more constraints
upon resource use (most notably, natural capital).

The ethical position adopted in this study
There are a number of ethical positions which could be adopted in this re-
search. Despite some considerable personal sympathy with the Rawlsian/˜strong

9 The economic implications of classical and neoclassical utilitarian and Rawlsian ethical positions can be ex-
pressed through consequent social welfare functions (SWF). Classical utilitarianism implies an additive SWF of
the form: W = β 1 U A + β 2 U B where W = social welfare; U A , U B = the total utility enjoyed by individuals A
and B respectively; β 1 , β 2 = weights used to calculate W. Neoclassical utilitarianism relaxes the assumption of
additivity such that W = W(U A , U B ). Finally, following Solow (1974a), the Rawlsian position can be expressed
as the maxi-min function in which we maximise W = min (U A , U B ). Note that Perman et al. (1996) suggest
that Rawls may have strongly objected to the latter utilitarian reformulation of his work.
Introduction 11

sustainability™ view, our self-assessment is that this study is essentially neoclassi-
cally utilitarian in its ethical basis. The de¬nition of values inherent in the TEV
concept remains anthropocentric and is therefore consistent with the extended util-
itarian view discussed by Perman et al. (1996, 1999). The most non-Rawlsian
characteristic of this study is the absence of an explicit incorporation of any pre-
cautionary principle or equity constraint. It might be argued that the sensitivity
analysis across various discount rates (discussed in Chapter 6), which we include
in our CBA, effectively addresses the issue of intergenerational equity. However,
as Hanley and Spash (1993) highlight, such an approach will not ensure equality of
well-being across generations. Similarly, we do not include explicit considerations
of distributional effects nor do we include any analysis which could be construed
as compatible with a Rawlsian maxi-min criterion. Our approach is therefore con-
ventional in terms of both theory and the ethical basis of such theory. It is only
in the practical implementation of our analysis that we have attempted to improve
upon convention.
This theoretical standpoint should not be taken as implying a wholesale rejec-
tion of the Rawlsian or ˜strong sustainability™ positions. Rather it is a pragmatic
extension of accepted decision-analysis practice.


Selection of the case study and data sources
While the fundamental objective of this study is the comparison of woodland
with agricultural values, a supplementary goal is to see how such differences vary
across areas of differing environmental character. The country of Wales consti-
tutes one of the most diverse areas of the UK with altitudes ranging from sea
level to heights above those found in neighbouring England. While smaller than
its neighbour,10 the entirety of Wales represents a very much larger area than has
been considered in virtually any CBA to date.11 Furthermore, from the perspective
of land use change, Wales provides a more interesting case study in that its di-
verse and relatively more adverse environment means that agricultural production
is limited to sectors such as sheep-breeding which have been in long-term decline
(see Chapter 9) and are therefore potentially more likely to be suitable for conver-
sion to woodland (which has expanded throughout the past century; see Chapter 5).
Wales is also interesting from an environmental point of view. While other areas
10 The ¬nal CBA results presented in Chapter 9 are given in terms of 1 km square cells. The land area of Wales
comprises some 20,563 such cells.
11 Consideration was also given to extending the analysis to include England, which is considerably more populous
than Wales. However, at the time our research commenced, agricultural census data for England were only
available down to the parish level. Such resolution fails to identify individual farm locations thus rendering
accurate production modelling infeasible. More recently the parish data have been interpolated to a grid cell basis
that is available from the University of Edinburgh Data Library (see http://datalib.ed.ac.uk/EUDL/agriculture/).
However, even these data do not report certain key pro¬tability variables vital to our analysis of the opportunity
costs of converting land from agriculture to woodland.
12 Applied Environmental Economics

of the European Union (EU) have responded to falls in the real price of sheep
by diversifying into other sectors, Welsh agriculture has seen an intensi¬cation of
sheep-rearing with steadily increasing stocking densities (Fuller, 1996; Woodhouse,
2002). This in turn has raised concerns regarding overgrazing and its impacts upon
wildlife (ibid.). A number of economic and environmental factors therefore single
out Wales as a particularly suitable subject for our case study.


Data sources
Our research draws upon data from a number of sources. All data were provided
free or for a reasonable handling charge. We are very grateful to a number of people
and organisations for this co-operation without which the research could not have
been undertaken (see Acknowledgements to this volume). Detailed descriptions of
the various datasets are provided in subsequent chapters, but a brief summary is
given here.
Data on farm-level agricultural activities, costs and revenues were obtained from
the Farm Business Survey in Wales (FBSW). We are indebted to the enlightened
attitude of the FBSW which, by being prepared to enter into a con¬dentiality
agreement whereby no farm-level results were reported, allowed us to use grid-
referenced farm data which could be linked to local environmental characteristics, so
facilitating a substantial improvement in the ability to model agricultural production
and its value.
Environmental data were provided in the form of the LandIS database, kindly
loaned by the Soil Survey and Land Research Centre (SSLRC), Cran¬eld. This is the
premier repository of land information data for England and Wales. When used in
conjunction with the FBSW data, LandIS provided the highest-quality combination
of information possible for modelling agriculture in the study area.
Our other principal data source was the Forestry Commission™s (FC) Sub-
Compartment Database (SCDB). This is the most extensive and comprehensive
source of woodland data in the UK and is again geographically referenced to a high
degree of accuracy, permitting integration with the environmental data contained
in the LandIS database.
A number of other sources were employed to provide speci¬c variables. These
included Bartholomew™s 1:250,000 digital map database made available to UK
universities under a CHEST agreement, 1991 Census data purchased for academic
research use by ESRC/JISC, details of windiness provided by the Forestry Com-
mission and digital maps of Environmentally Sensitive Area boundaries supplied
by the Ministry of Agriculture, Fisheries and Food. The project also involved sur-
veys and interviews which are described later in this book, the structure of which
we now consider.
Introduction 13


Context and structure of the book
The majority of the research presented in this book was undertaken for a Ph.D. thesis
(Bateman, 1996).12 A number of journal articles discussing individual aspects of
the research have since been published and are referenced in appropriate chapters,
but this book brings these elements together allowing the integrated results to be
considered in detail.
Many of the data used refer to the early 1990s and, given, in particular, the
constantly changing context of agriculture and to a lesser extent forestry, we are wary
of asserting that all ¬ndings are directly transferable to the present day. However,
it remains our strong contention that the methodology adopted is still relevant and
capable of wider application. At appropriate points in the text we have sought to
provide some updating of the economic and policy context and to comment on the
applicability of the substantive ¬ndings in the light of this.
As discussed above, the book considers the application of environmental eco-
nomics using GIS through a case study concerning woodland, agriculture and a
CBA comparison of land use change between the two. We begin with a consid-
eration of the recreation value of woodland. This is subdivided into an appraisal
of methods for the monetary evaluation of woodland recreation (Chapter 2), a re-
view of previous valuation studies and presentation of our own studies (Chapter 3)
and GIS-based analysis transferring results from these various evaluations to the
case study area through predictions of the latent demand for visits (Chapter 4). The
focus of attention then shifts to timber and a further evaluation model is constructed
(Chapter 5) and applied to newly estimated timber yield models (Chapter 6). Our
analysis of woodland values is concluded by extending the de¬nition of values to
include the net bene¬ts of carbon sequestration (i.e. counteracting the greenhouse
effect of global warming) provided by forests (Chapter 7).
We then turn to consider the opportunity cost of converting land to woodland,
which in the case study area of Wales involves losses of agricultural production.
Models of both the farm-gate and social values of such production are presented
for the dominant farming types of the area (Chapter 8).
The preceding analyses are ¬nally synthesised through cost-bene¬t analyses of
potential conversions of land from agricultural to woodland use (Chapter 9).13 Both
market and social-perspective assessments are presented and the results clearly
demonstrate the sensitivity of ¬ndings to whether analyses are restricted to consid-
eration of market prices alone or extended to include non-market values. Further
12 Further details of this thesis can be obtained by navigating from the CSERGE website at
http://www.uea.ac.uk/env/cserge/ or by going directly to Ian Bateman™s personal home page at
http://www.uea.ac.uk/∼e089/.
13 The analysis also indicates, by default, whether conversions from existing woodland to new agriculture are
justi¬ed, although our results indicate that this is rarely the case.
14 Applied Environmental Economics

sensitivity is found regarding which agricultural sector is considered for conversion,
the choice of discount rate, choice of woodland tree species and many other pol-
icy variables. Perhaps most markedly, our GIS-based methodology highlights the
spatial dimension of CBA decisions, showing that the same policy decisions yield
social and market gains or losses depending upon the location chosen for policy
application (Chapter 9). This analysis therefore identi¬es a number of interesting
results from which policy implications and conclusions are drawn and presented
(Chapter 10) along with an assessment of the methodology adopted in the research
and consideration of the scope for further extensions, certain of which are ongoing.
2
Recreation: valuation methods




Introduction
At the heart of cost-bene¬t analysis (CBA) theory lie two basic principles (Pearce,
1986; Hanley and Spash, 1993): ¬rst that, as far as possible, all the costs and
bene¬ts arising from a project should be assessed; and, second, that they should
be measured using the common unit of money. While these seem common-sense
precepts, in application both principles raise highly complex problems. The issue
of complete appraisal is, when taken to the extreme, ultimately insoluble in a world
ruled by the laws of thermodynamics where, as noted by commentators such as
Price (1987a, 2000) and Young (1992), everything affects everything else. For real-
world decision-making, practical rules regarding the limits of appraisal are needed.
Such rules are the stuff of numerous project appraisal guidelines, for example the
Treasury™s ˜Green Book™ (H.M. Treasury, 1991), whereas the research described
here focuses on the second principle “ of monetary valuation.
In discussing approaches to the monetary evaluation of environmental prefer-
ences we can ¬rst identify a wider global family of monetary assessment methods
(see Figure 2.1). This comprises both the formal ˜valuation™ (or demand curve)
methods discussed below and a quite separate group of ad hoc environmental
˜pricing™ techniques (see the review in Bateman, 1999). In theoretical terms valua-
tion and pricing approaches are quite distinct. Whereas the former are based upon
individuals™ preferences and yield conventional, neoclassical, welfare measures
(hence the term ˜valuation methods™), the pricing techniques are much more akin to
market-price observations. For example, the shadow project pricing approach uses
the costs of hypothetical environmental asset replacement, restoration or transplan-
tation schemes (Buckley, 1989) to yield prices for the environmental costs of a
proposed project. While it has been argued that such methods provide useful in-
formation for the appraisal of projects, policies or courses of action (Turner et al.,
1992), pricing techniques re¬‚ect the costs of protecting or providing environmental


15
Figure 2.1. Methods for the monetary assessment of non-market and environmental goods. (Source: Bateman, 1999.)
Recreation: valuation methods 17

assets but not the bene¬ts of doing so. In considering only prices rather than values,
decision-makers are in danger of making incorrect choices.1 Certainly such infor-
mation is insuf¬cient for adequate CBA appraisals. We therefore reject the use of
pricing techniques and turn to consider the more theoretically rigorous valuation
methods.
The valuation methods all ultimately rely upon individual preferences. However,
within this genre two distinct categories of approach can be de¬ned: methods based
upon preferences which are revealed through purchases by individuals of market-
priced allied goods; and methods which rely upon expressed preferences elicited
through questionnaire surveys. Both of these variants provide measures of value
which are valid according to economic theory. However, the same theory shows
that these measures need not be identical even when the same change in provision
of a non-market good is considered (further discussion of this issue is provided
in any basic microeconomics text, for example Laidler and Estrin, 1989; a simple
overview is given here).
Revealed preference techniques typically cannot be applied directly to the val-
uation of environmental goods because of the lack of an observable market price.
One solution is to investigate a surrogate market and this approach is adopted by the
travel cost (TC) method. Here the costs of a visit to a recreation site are calculated
as some combination of any entry charge (typically zero for UK forests), travel
expenditure (e.g. petrol costs) and the opportunity cost of travel time (i.e. the value
of the time devoted to travelling to the site; this might be wages forgone or the lost
opportunity to enjoy some other activity during that time).2 By comparing these
travel costs with the number of visits, we observe that as costs increase (e.g. the
further an individual has to travel to a wood), fewer visits are made. This negative re-
lationship maps out a ˜demand curve™, the area under which provides an estimate of
the value of visits to the site which is known as the ˜consumer surplus™.3 While this
is a useful measure it is in fact the sum of two components: the substitution effect
(which measures the increased consumption of any good when its price falls) and
the income effect (which shows the change in consumption due to the increase in
purchasing power or ˜real™ income which occurs when the price of a good falls).
While the substitution effect is positive4 for a reduction in travel costs, the income

1 As an interesting example of how pricing methods may give little practical guidance to a decision, Medley (1992)
refers to the Department of Transport™s pricing of a motorway tunnel to avoid a cutting through the Twyford
Down Site of Special Scienti¬c Interest in Hampshire. At £70 million this was considered too expensive and
abandoned without any appraisal of the bene¬ts of such an alternative being undertaken.
2 Brief discussion of how these travel costs are estimated is provided subsequently.
3 In essence the reader can think of the consumer surplus value being estimated as the sum of what the individual
visitor would pay, if required, for each of the visits to a woodland. In an attempt to widen readership we have
avoided various technicalities in this and subsequent descriptions. References to further reading are provided
below.
4 Strictly speaking this effect is non-negative rather than absolutely positive.
18 Applied Environmental Economics

Table 2.1. Welfare change measures obtained from expressed preference measures

Change in provision
Gain Loss
WTP measure WTP to ensure that the proposed WTP to avoid the proposed
gain occurs loss occurring
WTA measure WTA compensation if proposed WTA compensation if proposed
gain does not occur loss does occur


effect of such a cost reduction may be either positive or negative depending upon
how the individual varies their consumption of the good as the purchasing power
of their income changes. Given this uncertainty, consumer surplus might provide
an imperfect estimate of the ˜income-compensated™ welfare provided by a given
recreational site.
This problem is at least in theory addressed through the application of expressed
preference approaches such as the contingent valuation (CV) method. Here respon-
dents are directly asked to state the change in income which would just offset a
proposed change in the provision of the good under investigation. Respondents
might be asked to consider either a gain or loss over the present level of provision
and in either case be asked questions concerning how much they might be willing
to pay (WTP) or willing to accept in compensation (WTA) to just offset the relevant
welfare change. Table 2.1 illustrates the four welfare measures so de¬ned.5
The income-compensated values estimated by the expressed preference methods
can therefore claim some theoretical superiority as welfare measures compared to
the consumer surplus estimates provided by revealed preference approaches. How-
ever, expressed preference methods have been the subject of considerable criticism
regarding the ability of respondents to articulate values for complex goods such as
those provided by the environment (Kahneman and Knetsch, 1992; Hausman, 1993;
Diamond and Hausman, 1994). In practice there is evidence that both TC-based
consumer surplus estimates and CV-based WTP values are reasonably similar,6 and
our research uses both methods, as they are, respectively, the most commonly ap-
plied revealed and expressed preference techniques for valuing woodland recreation
bene¬ts.7

5 For further discussion see Just et al. (1982), Johansson (1987) or any similar intermediate microeconomics text.
For an empirical comparison of all four measures, see Bateman et al. (2000a).
6 Carson et al. (1996) review 83 studies from which 616 comparisons of CV to revealed preference (RP) estimates
are drawn, yielding a whole sample mean CV:RP ratio of 0.89 (95 per cent con¬dence interval = 0.81 to 0.96).
This suggests that, while statistically different from each other (and, as we will see subsequently, on occasion
strongly dissimilar), revealed and expressed preference measures do on average produce estimates which fall
within the same broad range.
7 For applications of the hedonic pricing revealed preference method to the valuation of woodland landscape
amenity, see Garrod and Willis (1992a). In our own recent research we have examined the potential for improving
Recreation: valuation methods 19

The remainder of this chapter presents brief reviews of the CV and TC methods,
concentrating on areas of particular interest to this study. Given the focus of this
research, these reviews are far from exhaustive and are deliberately written in
a non-technical and introductory style. For further reading concerning the CV
method, see Mitchell and Carson (1989), Bjornstad and Kahn (1996), Bateman and
Willis (1999) and Bateman et al. (2002), while for the TC method, see Hufschmidt
et al. (1983), Bockstael et al. (1991), Freeman (1993) and Herriges and Kling
(1999); an introduction to all non-market valuation techniques is given in Champ
et al. (forthcoming).


The contingent valuation method
Introduction: applying the CV method
The implementation of a CV study involves a number of distinct stages. In the
¬rst, preparatory, stage a ˜hypothetical™ or ˜contingent™ market is set up in which
individuals are asked how much they are either WTP or WTA in respect of the pro-
posed change in provision of the good under investigation. These questions may be
framed using a variety of elicitation methods. In a WTP study the major alterna-
tives are: (i) open-ended (OE), in which the respondent is asked ˜how much are you
willing to pay?™, an approach which produces a bid response which is truncated at
zero but is otherwise continuous;8 (ii) dichotomous choice (DC), where respondents
are asked ˜are you willing to pay £X?™, the amount X being systematically varied
across the sample to test individuals™ responses to different bid levels. This ap-
proach produces a discrete bid response variable and may be iterated using higher or
lower bid amounts depending upon the respondents™ replies to previous amounts;9
(iii) iterative bidding (IB), in which a series of DC-type questions are followed
by a ¬nal OE question; (iv) payment card (PC), in which respondents select their
maximum WTP amount from a list of possible sums presented on a card to them.10
The respondent also requires information regarding the nature of the good under
evaluation, the proposed quantity/quality change in provision of the good, who
will pay for and who will use the good and how payment will be collected (the
˜payment vehicle™, for example higher taxes, entrance fees, donation to a charitable
trust, etc.).

hedonic pricing models of landscape and noise disamenity values through the application of GIS techniques
(see Lake et al., 1998, 2000a,b; Bateman et al., 2001a). Expressed preference methods other than CV (such
as choice experiments, contingent ranking, etc.; see Champ et al., forthcoming) have not to date been widely
applied to the study of woodland recreation values. An exception is provided by Hanley et al. (1998) who
present a choice experiment study of forest landscape values in the UK.
8 Bateman et al. (1995a) provide a comparison of OE, DC and IB formats.
9 See, for example, Hanemann et al. (1991); Langford et al. (1996); Bateman et al. (2001b).
10 See, for example, Rowe et al. (1996).
20 Applied Environmental Economics

With the questionnaire complete, the process moves to the survey stage to obtain
responses. In so doing the relevant population of either users or non-users, or a
mix of the two, must be determined. User surveys may be conducted either on or
off site while non-user surveys are restricted to the latter locations. In both cases
either face-to-face or mail/telephone surveys may be used, each of which has its
own merits and drawbacks.
Once responses have been collected, data analysis can commence. This has the
dual objective of both obtaining the required welfare measures and assessing the
validity of responses. Validation testing is complex and multifaceted (see discus-
sions in Mitchell and Carson, 1989 and Bateman et al., 2002); typically, however,
considerable emphasis is placed upon the consistency of responses with theoreti-
cal expectations, this being assessed through the estimation of bid curves linking
valuation responses to the characteristics of respondents (e.g. their income, use of
the good, etc.), and upon assessing the extent to which CV estimates converge with
those obtained by other valuation methods.
The ¬nal stage of the study is to derive aggregate welfare measures by linking
sample responses to the relevant underlying population.11 Providing that validity
tests are satisfactory, these aggregate measures may then be incorporated within
project appraisals.


Focal methodological issues
We now concentrate on topics which are central to our woodland recreation work.12
The general issue under consideration here is the extent to which design issues
affect elicited values. However, to set this in context, we begin by considering the
process by which individuals form stated responses to CV questions. This discussion
allows a consideration of the impacts which choices regarding survey design may
have upon valuation responses. Areas highlighted for subsequent research include
the effect of varying the elicitation method and changing the payment vehicle,
the impact of asking respondents to consider budget constraints and the effect of
varying the order of questions within a survey instrument. Some of these issues
are tackled through non-woodland applications, results from which are presented
in this chapter. Findings from woodland studies are presented in the following
chapter.

11 This is rarely as straightforward as it may appear. See discussions in Bateman et al. (2000b, 2002).
12 This approach precludes discussion of a number of CV issues which we address in other contexts, including
the ability of respondents to distinguish adequately between a conglomerate good (e.g. all natural areas) and
its constituent parts (just one of those areas) (see Bateman et al., 1997a) and the role of ˜reference points™ of
prior provision in in¬‚uencing the commonly observed asymmetry between WTP and WTA measures of the
same change in provision (see Bateman et al., 1997b, 2000a). This is of course far from an exhaustive list of
current CV issues, for which the interested reader should consult the literature cited previously.
Recreation: valuation methods 21




Figure 2.2. The value formation process.

The valuation process and its in¬‚uences
Consideration of the process through which respondents derive valuation responses
to CV questions can be traced back to the beliefs“attitudes“behaviour models pro-
posed by Fishbein and Ajzen (1975) and Ajzen and Fishbein (1977). Recent re-
search has suggested that this process may be highly complex, reaching beyond
the somewhat simple models of self-interested rationality underpinning much eco-
nomic theory. Figure 2.2 draws upon a number of sources to summarise recent
thinking in this area.13

13 This section draws upon a variety of sources including Fishbein and Ajzen (1975); Hoehn and Randall (1987);
Brown and Slovic (1988); Mitchell and Carson (1989); Dake (1991); Harris and Brown (1992); Bateman and
Turner (1993); Schkade and Payne (1994); Marris et al. (1996); Hanemann (1999).
22 Applied Environmental Economics

The model presented in Figure 2.2 emphasises the pre-survey base-state as a vital
determinant in the valuation process irrespective of the good under investigation.
A variety of base-state in¬‚uences are identi¬ed. These include the individual-level
factors emphasised by traditional economic models of preferences, as well as world
views (for example, whether individuals see the relationship between the economy
and the environment as ultimately benign or degrading14 ), social factors (such as
work and family in¬‚uences), cultural in¬‚uences (such as the typologies identi¬ed in
recent empirical studies: Dake, 1991; Sj¨ berg, 1995; Marris et al., 1996; Langford
o
et al., 2000) and contextual factors which may distinguish otherwise identical
changes in provision of public goods. These elements combine to yield the base-
state positive and normative beliefs which any individual brings to a valuation
experiment.15
It is tempting to see the schism between so-called individual and citizen pref-
erences (Blamey, 1995, 1996) as being re¬‚ected in differences between positive
and normative beliefs and there is some evidence to support such a de¬nition
(Peterson et al., 1996). However, the complex and uncertain nature of this argu-
ment prohibits us from taking the matter further. Rather we can see these beliefs
as the base-line points of reference from which the individual enters the CV exper-
iment. Here the respondent is presented with new information which will be used
to update the belief set. These beliefs will then form the individual™s attitudes and
norms concerning behaviour. Information, beliefs and attitudes all subsequently
feed into motivation. It is arguable that non-use (existence and bequest) values
arise from non-use motives such as altruism (Randall, 1987) drawing upon norma-
tive beliefs, whereas use values arise from positivist beliefs and attitudes. However,
while these are likely to be the main routes of in¬‚uence, we can also imagine norms
concerning instrumental goods and positivist ideas concerning non-use values.
These use and non-use motives combine and are expressed as the WTP sum within
the CV valuation process. This statement of value and the CV experience itself
then feed back either via behaviour (an actual payment) or, more usually, directly
into the individual™s positive and normative beliefs, such that values for the same
good may change if a CV study is iteratively repeated using the same respondents
(see Coursey et al., 1986). As a simple investigation of the impact of use and
non-use motives upon stated values it was decided that an initial and preliminary
objective of our empirical woodland research would be to examine variations in
WTP values between users and non-users of woodland recreation as well as to
examine the WTA compensation levels demanded by the potential providers of

14 These views relate, respectively, to O™Riordan™s (1976) technocentric man and ecocentric man.
15 Interestingly, Spash (1997) argues that individuals may also hold beliefs about the valuation process itself
and that these may result in the preferences of those opposed to the valuation process being systematically
under-represented.
Recreation: valuation methods 23

woodland recreation (farmers) who were together the target of the wider research
described in this volume.
The transition from formulated to stated value is the subject of theoretical analysis
by Hoehn and Randall (1987) and Carson et al. (1999). Here the CV respondent is
seen as undertaking a two-stage task of (i) value formulation and (ii) value statement.
In moving from formulated to stated value the respondent has the opportunity
to engage in a variety of strategic behaviours including both understatement and
overstatement of formulated WTP. Hoehn and Randall and Carson et al. see these
various strategies as being chosen according to the elicitation method being used
(OE, DC, etc.) and so we consider this issue in some detail.

Elicitation effects
Different elicitation methods may either be neutral (i.e. they have no impact upon
stated WTP) or lead to either under- or overstatement of values. Reasons for such
effects are diverse and there is considerable debate regarding the impact which the
differing strategic incentives and psychological effects of each elicitation method
have upon stated values and the consequent validity of those values.16 If methods
are neutral in their effect, the elicitation issue can be ignored. However, both theory
and empirical investigation suggest this is not generally the case and so we begin
this consideration of potential bias by ¬rst considering issues surrounding value
understatement, then overstatement, after which empirical evidence is examined
and some conclusions drawn. For convenience we shall consider WTP measures
throughout the following discussion.

Understatement of WTP
If an individual feels that a good will be provided irrespective of his response to a
WTP question, or that the payments of others will be suf¬cient to secure provision,
then, given the ability to freely vary his stated valuation (e.g. in an OE elicitation
format), the individual will ˜pretend to have less interest in a given collective activity
than he really has™ (Samuelson, 1954) and will understate his WTP for that good, i.e.
he will ˜free-ride™ (Marwell and Ames, 1981; Brubaker, 1982). A similar result will
be obtained when respondents feel that actual payments will (or should) be related
to cost shares rather than to WTP (Hoehn and Randall, 1987). Here respondents
will state the expected cost, if this is less than WTP, or zero otherwise.
Mitchell and Carson (1989) review a variety of studies for priced goods in which
hypothetical OE bids were subsequently compared to actual prices paid. These stud-
ies indicated that, where a relatively weak free-rider incentive existed (e.g. where
respondents were informed that a group threshold WTP was required in order to

16 See, for example, Carson et al. (1999) and Bateman et al. (2001b, 2002).
24 Applied Environmental Economics

secure provision), then a reasonably close correspondence between hypothetical
and actual payments was found (OE bids were between 74 per cent and 96 per cent
of actual payments in the studies cited). However, where a strong free-riding in-
centive existed (e.g. by guaranteeing provision as long as respondents stated some
non-zero sum), divergence was consistently greater (OE bids being 61 per cent to
71 per cent of actual payments).

Overstatement of WTP
Bateman et al. (1995a) identify ¬ve factors which may induce a respondent to
overstate WTP in a CV experiment, each of which we discuss below:
(i) Strategic overbidding (all elicitation formats)
(ii) ˜Good respondents™ (all elicitation formats)
(iii) Upward rounding (DC formats)
(iv) Anchoring (DC formats)
(v) Starting point effects (IB formats).

(i) Strategic overbidding. In an important empirical paper, Bohm (1972) argues
that, contrary to the prediction of free-riding, respondents may overstate their WTP

. 1
( 11)



>>