The Price of Water

And the same night there fell a shower of rain, For which their mouths gaped like the cracks of earth When dried to summer dust. Till taught by pain, Men really know not what good waters worth. If you had been in Turkey or in Spain, Or with a famished boats crew had your berth, Or in the desert heard the camels bell, Youd wish yourself where truth is-in a well. Lord Byron, Don Juan

The Price of Water

Studies in Water Resource Economics and Management
Second Edition

Stephen Merrett

Published by IWA Publishing, Alliance House, 12 Caxton Street, London SW1H 0QS, UK Telephone: 44 (0) 20 7654 5500; Fax: 44 (0) 20 7654 5555; Email: publications@iwap.co.uk Web: www.iwapublishing.com First published 2007 2007 IWA Publishing
Typeset by Aptara Inc., New Delhi, India Printed by Lightning Source Cover design by www.designforpublishing.co.uk Cover image is of the Contemporary Art Quilt Spheres of Influencecourtesy of Alicia Merrett http://www.tufpark.demon.co.uk/

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the UK Copyright, Designs and Patents Act (1998), no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior permission in writing of the publisher, or, in the case of photographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of licenses issued by the appropriate reproduction rights organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to IWA Publishing at the address printed above. The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for errors or omissions that may be made. Disclaimer The information provided and the opinions given in this publication are not necessarily those of IWA or of the authors, and should not be acted upon without independent consideration and professional advice. IWA and the authors will not accept responsibility for any loss or damage suffered by any person acting or refraining from acting upon any material contained in this publication. British Library Cataloguing in Publication Data A CIP catalogue record for this book is available from the British Library Library of Congress Cataloging- in-Publication Data A catalog record for this book is available from the Library of Congress

ISBN 1 84339 177 5 ISBN 13: 978 1 84339 177 7

Contents

Preface 1 Introduction: getting to grips with water 1.1 Instream and outstream 1.2 Reuse and recycling 1.3 Catchment and region 1.4 Four interpretations of demand 1.5 Pricing the resource 1.6 Studies of household water use and the willingness-to-pay for water 1.7 Tearing up water and floating on water-rights? 1.8 Catchment water deficits 1.9 The virtual water controversy 1.10 Final remarks The regional water balance statement: a new tool for water resources planning 2.1 Introduction 2.2 The rules of the game 2.3 The supply categories 2.4 Water storage 2.5 The use categories 2.6 The change statement 2.7 The uses of regional statements 2.8 Conclusions Integrated water resources management and the hydrosocial balance 3.1 Introduction
v

x 1 1 2 2 2 3 4 5 5 7 8 9 9 10 13 14 15 16 18 19 20 20

vi

Contents

3.2 3.3 3.4 3.5 3.6 3.7 3.8 4

A new implement for the IWRM toolbox The bridge between quantity and quality The physical geography and hydrology of Jersey Water supply in Jersey Water use in Jersey Water quality in Jersey Conclusions

20 21 24 24 27 28 34 37 37 37 41 41 42 43 43 44 45 47 51 56 58 58 59 59 60 62 63 64 65 67 68 68 69 70 71 71 72 72 73 74

Sharing the benefits of the river basins water economy 4.1 Introduction 4.2 Benefits of the water economy 4.3 Basin water productivity 4.4 Sharing the benefits 4.5 Conclusions: negotiating the benefits Farm-level drought management: an irrigation case-study from the UK 5.1 Introduction 5.2 The Anglian Region 5.3 Silver Birches plc 5.4 Drought management: the infrastructural strategy 5. 5 Drought management: the informational strategy 5.6 Conclusion The potential role for economic instruments in drought management 6.1 Introduction 6.2 The water economy 6.3 The Anglian Region 6.4 The Regions water economy 6.5 The Agencys drought plan 6.6 Anglian water services drought plan 6.7 Drought plans and the water economy 6.8 Economic instruments 6.9 Conclusions Virtual water and Occams razor 7.1 Introduction 7.2 A water deficit resolved 7.3 A critique of the virtual water thesis 7.4 Occams razor 7.5 Conclusion Virtual water and the Kyoto consensus 8.1 The use of metaphor 8.2 Crops, crop water and water deficits 8.3 The Kyoto consensus

Contents

vii

The urban market for farmers water-rights 9.1 Introduction 9.2 The urban actors demand function 9.3 The farmers supply function 9.4 The limits to theory 9.5 Conclusions The demand for water: four interpretations 10.1 Introduction 10.2 The use of water 10.3 The consumption of water 10.4 The need for water 10.5 The economic demand for water 10.6 Supply-side leakage and evaporation 10.7 Conclusions The political economy of water abstraction charges 11.1 Introduction 11.2 Abstraction charges and the theory of rent 11.3 A charge-setting taxonomy 11.4 Abstraction charges and sustainable catchment management 11.5 The impact on users 11.6 Final remarks Twelve theses on the cost and use of irrigation water 12.1 Thesis 1 12.2 Thesis 2 12.3 Thesis 3 12.4 Thesis 4 12.5 Thesis 5 12.6 Thesis 6 12.7 Thesis 7 12.8 Thesis 8 12.9 Thesis 9 12.10 Thesis 10 12.11 Thesis 11 12.12 Thesis 12 Behavioural studies of the domestic demand for water services in Africa 13.1 A methodological revolution 13.2 Market networks for water 13.3 The uses of water 13.4 Objects or subjects? 13.5 The discrete choice model 13.6 Conclusions

77 77 78 80 80 85 86 86 87 88 88 88 89 90 91 91 92 93 96 98 99 101 101 101 101 102 102 102 102 102 103 103 103 103 104 104 106 108 110 113 114

10

11

12

13

viii

Contents

14

Deconstructing households willingness-to-pay for water in low-income countries 14.1 Introduction 14.2 Survey methods 14.3 Sign and behaviour 14.4 Demand theory and survey practice 14.5 The affordability question 14.6 The treatment of substitutes 14.7 Sign and sanction 14.8 Private agendas 14.9 Anchor prices 14.10 Conclusion Industrial effluent policy: economic instruments and environmental regulation 15.1 Introduction 15.2 The generation and regulation of industrial effluent 15.3 The objectives of disposal charges 15.4 The demand for waste water services 15.5 The measurement of pollution 15.6 The design of disposal charges: the utilities 15.7 The design of disposal charges: the environmental regulator 15.8 Conclusions Nitrate pollution on the Island of Jersey: managing water quality within European community directives 16.1 Introduction 16.2 Nitrate pollution of groundwater and surface waters 16.3 Water quality management by the Department of Agriculture and Fisheries 16.4 Water quality management by the JNWWC 16.5 Water quality management of domestic abstractions: the Department of Environmental Health 16.6 Water quality management of waste water discharges: the Environment and Public Services Department 16.7 The benefits of nitrate pollution management 16.8 Conclusions Catchment water deficits in the twenty-first century 17.1 Introduction 17.2 The Dwyer catchment 17.3 Redemptive options (I) 17.4 Redemptive options (II) 17.5 From surplus to deficit 17.6 Framework, theory and empirical studies 17.7 Conclusion

116 116 117 118 119 120 121 123 124 126 128 130 130 131 132 133 135 136 137 138 140 140 141 143 144 147 148 149 155 158 158 160 161 163 164 165 166

15

16

17

Contents

ix

18

Catchment water deficits: an application to Zambias Kafue river basin 18.1 Introduction: the general hypothesis 18.2 The Kafue River Basin 18.3 Groundwater abstraction 18.4 The economic demand for water 18.5 The supply of food 18.6 Environmental needs 18.7 Conclusions The Thames catchment: a River Basin at the tipping point 19.1 Introduction 19.2 The Kafue catchment 19.3 The analysis of densities 19.4 The Thames River Basin 19.5 The Thames in water deficit? 19.6 Density analysis of the Thames River Basin 19.7 Tipping deeper into deficit 19.8 Conclusion Water resource impacts of new housebuilding in the Thames Region: 20062025 20.1 Introduction 20.2 The baseline situation 20.3 The increase in homes 20062025 20.4 Addition and subtraction 20.5 Choices Beneficial impacts for the Thames River Basin of water company leakage reduction 20062025 21.1 Introduction 21.2 Leakage in the Thames Region: some basic facts 21.3 Forecasting the reduction in total leakage 21.4 Conclusions Bibliography Index

167 167 168 170 171 172 172 173 178 178 179 179 181 185 186 188 189 191 191 192 194 194 196 198 198 199 200 203 205 214

19

20

21

Preface

On the evening of September 22nd 1994, I received a telephone call from a Danish management consultancy asking me if I would be free to work in Latvia for a month on the economics of a local water utility. My reply was positive, but I was compelled to add that I knew absolutely nothing about water or water utilities. The consultant replied that a couple of months would elapse prior to the projects start-date; from previous contact he was sure that if I applied myself to the subject-area during those eight weeks, all would be well. The deal was struck. I raised my eyes to the heavens (well, the ceiling of my flat) and said: Let it be water.The next morning (!) I began writing a book an introduction to the economics of water resources, later published under that title. During the previous thirty years, my entire working life had been spent carrying out economic research in a variety of applied fields. These included: higher education and student finance, the production of fertilizers, the British civil space programme, international trade and capital flows, housing construction and finance, and land-use planning. Taking early retirement from University College London in 1994, I decided to reinvent myself as an environmental consultant. My two daughters persuasive arts, and my lifelong love of the natural world, were the roots of this choice. But environmental research embraces an extraordinary variety of subjects; so it was that chance telephone call from Denmark that made water resources my focus. By February 1995 I was on the River Daugava in Daugavpils, Latvias second city, carrying out an affordability study of price increases for the towns water and waste water services. In my view, the oldest and still the most important production activities of humankind are securing food and water, building and maintaining housing, making clothing, and providing health care and education. These activities can be thought of as forms of reproductive production. Each is simultaneously a form of human labour as well as a direct and necessary condition for the reproduction of the species. The fact that more than half my working life had passed in the study of the economics of education and the economics of housing made the shift into the study of water resources that much easier.

1 Introduction: getting to grips with water

1.1

INSTREAM AND OUTSTREAM

This volume brings together twenty-one papers, on the economics and management of water resources, most of which were first published in various professional journals between 1994 and 2007. The subject matter of water resources research can be divided into two (interdependent) blocks. The first block concerns instream water, that is, the flows of springs, streams and rivers, the lakes these flows create, and the immense hidden stores of groundwater that lie beneath our feet as aquifers. This is the world of the hydrological cycle. The second block is composed of the flows of water abstracted from instream sources, stored, distributed to water users and then discharged back to the instream medium. This is the world of the hydrosocial cycle (Merrett 1997: 67). The principal research areas in the case of instream water are hydroelectric power, navigation, fishing, conservation, recreation and environmental tourism. In contrast, for outstream water, research focuses on the capture of instream sources, their consequent channelling, storage, treatment of raw water, use by households and farmers and urban institutions, sewerage, the treatment of waste water and its discharge back to the instream medium or to coastal waters. This volumes collection of papers mainly addresses outstream themes. The reader will soon be aware that the author has a strong interest in the philosophy of science. Research programmes, in my view, should combine theory, method, fieldwork and policy development in a continuous interplay. To take four examples, chapter thirteen, on the domestic demand for water services in Africa, is a paper in method. Chapter nine on the urban market for farmers water-rights is an exercise in theory. Chapter five on farm-level drought management in Englands Anglian Region has a major field-work content. Chapters fifteen and sixteen, on the interaction between
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

The Price of Water

economic instruments and environmental regulation with respect to water treatment and waste water discharges, address policy development.

1.2

REUSE AND RECYCLING

The fundamental distinction between instream and outstream water is complemented by the related distinction between the reuse of water and the recycling of water. Reuse and recycling are two concepts repeatedly drawn on in this volume as well as in my two books on water resources economics and management (Merrett 1997, Merrett 2002a). Reused water is waste water and irrigation drainage that, prior to its return to the instream resource, is captured and used again (perhaps repeatedly). Reuse can take place within a single institution (such as a sugar-mill) or it can occur between institutions as when urban waste water is reused for irrigation purposes. Reuse may have real advantages, such as lowering abstraction costs or by reducing waste water treatment costs. Recycling refers to water that is abstracted, used by households, industry etc., and then the fraction that is not consumed as evapotranspiration flows back to the catchments rivers, aquifers and lakes. The importance of recycling to hydroeconomists is that it augments the hydrological resource from the point at which the recycling occurs. The negative characteristic is that recycled flows may pollute the resource. The proportion of water used that is consumed, and therefore is unavailable for recycling, varies between categories of use.

1.3

CATCHMENT AND REGION

Water resources research is carried out either at the catchment scale or at the regional scale, where region refers to any area with a defined boundary. The beauty of working at the catchment (river basin) scale is seen to be the unity of its hydrological flows precipitation, evapotranspiration, run-off and groundwater recharge. However, groundwater boundaries may not fall within the catchment boundary. Moreover, hydrosocial flows such as the import of water from another catchment or the export of waste water beyond the catchment also undermine the supposed unity of flows. The attraction of regional analysis is its flexibility. It can be applied to a single dwelling, a village, a city, a province, a country, an island or what you will. However, a region will usually be part of a river basin, or overlap two basins, or contain several basins, so that the linkage of hydrological and hydrosocial flows becomes extremely complex. The hydrological unity of the single river basin is lost. In 1996 the author began developing a method concept now referred to as the hydrosocial balance, which is applicable at the catchment or at the regional scale. It is a planning tool that appears in many of this volumes papers, for example in chapter three, Table 1. My own fieldwork in applying the concept has been carried out for a single house in southern Spain, the island of Jersey, Gaza, the West Bank of Palestine, the Thames catchment in England and a farm in the Anglian Region. As is argued in chapter three, the hydrosocial balance offers excellent potential for the regional-scale planning of waters civil engineering infrastructures and their capital financing.

1.4

FOUR INTERPRETATIONS OF DEMAND

During the years that the author worked on Palestinian water resource challenges, one of the draft outputs co-written with my colleague Eng. Khalil Saleh was a review

Introduction: getting to grips with water

of a number of documents assessing the future water demand of the Palestinians. In the course of this work it became clear that the term the demand for water is used with a confusing variety of meanings. For an economist with an unslaked thirst for the philosophy of science, the subject seemed to be worth pursuing for the most practical of reasons. The outcome was a short paper entitled The demand for water: four interpretations, published as chapter ten in this volume. The core of the argument is that the demand for water may refer to: i) ii) iii) iv) the use of water; the consumption of water, that is, evapotranspiration during use; the need for water; the economic demand for water, represented by the economists trusty demand function that relates quantity purchased and unit price.

This confusion of meanings, one with another, has consequences. To take demand to mean both use and consumption neglects the truth that a given volume used can be associated with wide variations in the volume consumed. This difference is at the core of much current debate in the field of irrigation engineering as the discussion lists show. To take demand to mean both use and need obscures the shortfall of use against need for more than half the worlds population. To take demand to mean both use and the economic demand for water confuses the use of water with one of its determinants, wherever costs borne by the user vary with the volume used. In the hydrosocial balance, the categories of supply include the leakage and evaporation losses that occur between the point at which human society appropriates outstream water and the point of delivery of water to user properties. Curiously such losses are almost always treated in the current literature as a form of demand. The result is that consultants forecasts of growth in demand include (as unaccountedfor-water - UFW) supply-side leakage and evaporation. Furthermore the reduction of these losses in abstraction, storage and distribution prior to the delivery of water to the user is said to be a form of demand management. This does not seem helpful. A manufacturer of refined sugar, when considering losses from output because of pilfering, or contamination while in the warehouse or destruction in a road or rail accident en route to the supermarket, would never regard this as a demand for sugar, a bizarre act of use by a consumer whom the sugar never reaches. The manufacturer would regard all of these as storage or distribution losses in the supply chain. So should it be with the supply of water. There is a more general point here. Water resource management is now widely seen as principally a form of demand management. To me such an approach seems nonsensical. This is especially the case when so much demand-management, as with reuse and the reduction of losses, turns out to be made up of supply-side initiatives. With outstream water we should always integrate in our thinking and our practice both demand-side and supply-side strategies.

1.5

PRICING THE RESOURCE

When a household, firm or farmer wishes to access water, the costs incurred by the user can take a variety of forms. The first form, the joy and delight of the economist, is that of a price paid to the supplier (such as a water utility) per unit quantity. For policy-makers committed to demand management in England, for example, a most encouraging development in the last two decades is the increase in the proportion of

The Price of Water

households that are on a priced, metered supply from about two per cent of the total number of households to more than twenty per cent. The second form again directly relates volume received to cost incurred by the user. But in this case it is because the user is also the supplier of the water. This can be called the own-supply case. Karin Kemper, for example, describes the situation in the Curu catchment of Brazil where farmers use electric motors to pump irrigation water from the river. She describes the importance of electricity tariffs to irrigators costs (Kemper 1996). The third form in which users incur costs for access to water is where there is only an indirect link to volume used. For example, in Archangel in northern Russia where the author was working in 2004, household payments for water used are based on a fixed tariff in roubles/m3 and on the assumption that households use 225 litres per capita per day (lcd) almost a quarter of a tonne! This charge per person is then multiplied by the number of family members registered as living at a given address. In fact, no one knows what is the volume of use per person. 225 lcd may be wildly inaccurate. However the variation of the tariff paid with the number of family members clearly does have an indirect relation to volume used. The fourth form of cost-use relation is where the payment made is a fixed charge for the user, invariant with volume. In Englands districts, in cases where households are not metered, they pay a water and waste water charge based on the value of the property in which they live. In this volume a number of the papers deal with charging for water. Chapter five has a case-study of the price of water charged by an English regional water utility to a local farmer and the alternative costs of water incurred were he to begin abstraction from a local drainage channel the own-supply case. Chapter six is a market-clearing proposal for raising water prices and fixed charges in the management of regional drought. The political feasibility of introducing such a management tool turns on i) the proportion of households that have metered use and ii) the protection of low-income families. This is the first time that water pricing has been suggested for drought management alongside the familiar informational, infrastructural and regulatory instruments. Chapter eleven reviews alternative methods of designing abstraction charges. Chapter twelve sets out twelve theses on the interrelation of the cost of irrigation water to farmers and the volume of water used. Chapter fifteen discusses the design of waste water charges for sewage collection, treatment and disposal. Chapter sixteen, authored together with my great friend Nick Walton (hydrochemist extraordinaire) applies economic analysis that shows how wasteful can be the setting of water quality targets that have little relation to human health.

1.6

STUDIES OF HOUSEHOLD WATER USE AND THE WILLINGNESS-TO-PAY FOR WATER

A fascinating area of research in low-income countries in the past 25 years, indelibly associated with the names of Dale Whittington and his colleagues, is that of detailed case-studies of familieswater use behaviour and their observed ability and willingnessto-pay for water. The first strength of the water demand school is its demonstration of how socially complex the networks of access and distribution can be. The second strength is the detailed examination of the dynamic and competitive markets in water

Introduction: getting to grips with water

that exist in many of Africasvillages and towns. The third strength of the water demand school is its success in estimating the proportion of domestic income absorbed by water purchases. At the same time, there are major weaknesses of the water demand schools work. The principal one is that the authors take a naturalistic standpoint in which the research into these communities behaviour is limited to the actions of silent men, women and children, moving across a landscape like so many ants in a natural-historical study of a savannah colony. Household members are not treated as intelligent, resourceful, purposive and reflexive citizens. These arguments are made in chapter thirteen of this volume. Chapter fourteen critically reviews the deployment of the willingness-to-pay concept in eleven studies published by Whittington and his fellow workers in the years 198898. The countries visited were Haiti, Ghana, Nigeria, Pakistan, the Philippines and Uganda. The critique is drawn from the branch of philosophy known as semiotics and it ends with an alternative approach to field research into households economic demand for water. The paper includes the contrasts I drew between the official version and the true story of the price of water paid by a family in Yerevan, Armenia; this is a norm-based system like the one in Archangel already referred to.

1.7

TEARING UP WATER AND FLOATING ON WATER-RIGHTS?

Water is a collection of molecules, each of which consists of two atoms of hydrogen bonded to one atom of oxygen. A water-right is a legal claim to abstract or otherwise access water. You can float on water but not on a water-right. You can abrogate a water-right but not water. Chapter nine is the only paper in this volume that addresses water-rights markets rather than water markets. It begins with pure theory on how urban actors demand function for abstraction-rights intersects with farmers supply function and thereby produces an equilibrium price and quantity traded. An account is given on how each function is determined. A number of real-life complications, such as part-sales and transaction costs, are examined and shown to be well-handled by neoclassical theory. But the empirical material suggests the modest relevance of the model to actual sales of water-rights where the absolutely predominant form of transaction is the bilateral deal. The conclusion is that, with respect to research method, fieldwork should be orientated to asymmetric power and information in the tradition of new institutional economics.

1.8

CATCHMENT WATER DEFICITS

It is likely that the papers destined to have the greatest professional impact on theory, fieldwork and policy development are chapters 17 to 21. The theme of catchment water deficits in the 21st century is now my favourite child. Chapter 17s objective is to provide a general theory of how the water resources of a river basin shift from surplus to deficit and the means by which water resource institutions can manage or reverse this shift. The article bites the bullet of defining surplus and deficit in the following way. A catchment water surplus is a situation in which, throughout the course of a spec-

The Price of Water

ified year, total precipitation in the basin is sufficient to simultaneously satisfy four conditions: Abstraction from the aquifer is maintained at a sustainable rate. Outstream water fully meets the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors. iii) The basin populations economic demand for food is fully met from domestic rainfed and irrigated farming or from domestic fisheries or from food imports financed by the basins commodity and service exports. iv) The rivers instream flows do not fall below defined minima. When a catchment water surplus does not exist, the river basin is in deficit. The analysis in chapter seventeen proceeds by illustrating the argument with a fictitious catchment called the Dwyer, which in the base-year is in surplus and in which total population increases by one-third every 25 years, that is, about 1.15% per annum. Moreover, output per capita is also rising. This combination of growth both in population and economic productivity, associated with the rise of world capitalism, is the source of deep unease amongst the professional staff of the Dwyer Catchment Authority. The Authority estimates the catchments economic output in 2025. After assuming that the ratio of outstream water use to basin output is constant, the Authority thereby derives an estimate of total water use in 2025. The Authority anticipates that, for the first time in its history, in 2025 the catchment will move into a water deficit; one or more of the four necessary conditions for surplus will have been breached. The paper then reviews twelve redemptive options. Six of these moderate the situation but do not prevent entry into water deficit; examples are importing water from another catchment, and the extension of water reuse. Six other options do prevent entry into water deficit; examples are reducing the rate of growth of population, and increasing waters productivity in terms of value added per cubic metre of water consumed. Chapter 17 re-states the implications for a river basin of finding itself in water deficit. These are one or more of the following situations: i) ii) iii) The basin is pumping its groundwater at an unsustainable rate. There is insufficient outstream water to meet the economic demand for it. The population has to import water, or food that it is unable to pay for from its exports of goods and services. iv) The basinscitizens must accept the economic and environmental losses following from its river diminishing in volume. i) ii)

The nightmare scenario is a river basin in which groundwater is being exhausted, households, farmers and other actors cannot purchase the water they require, food imports cannot be paid for, making the basin dependent on powerful allies, and the river has been destroyed. The article ends with suggestions for catchment research projects that establish whether or not any single river basin is in deficit and, if so, what are the causes and the policy options. Chapters 1821 have all been written since 2005. They provide empirical analysis that shows that the Kafue River Basin is in surplus and that the Thames River Basin is tipping into deficit. Importantly, chapter 19 sets out the five variables that determine whether or not a river basin is in deficit or surplus and goes on to measure the value of these variables for the Thames basin. The final two chapters continue the theme by i)

Introduction: getting to grips with water

demonstrating that new housebuilding in the Thames region will drive the river basin deep into deficit, and ii) reviews the likelihood that leakage reductions by the basins water utilities will weaken the move into deficit.

1.9

THE VIRTUAL WATER CONTROVERSY

In the previous section, one of the criteria for surplus at the river basin scale is that the basin populations economic demand for food is fully met from domestic rainfed and irrigated farming, or from domestic fisheries, or from food imports financed by the basins commodity and service exports. The nightmare scenario for a deficit catchment includes a situation where food imports cannot be paid for, making the basin dependent on powerful allies; the relation between Egypt and the USA is an example. In fact, it was Tony Allans work on the hydropolitics of the Middle East that first led me to include the water needs of domestic food production in the definition of water surplus (Allan 2001). During the 1990s (and even prior to that decade) Allan had developed a concept that he now refers to as virtual water. Virtual water is defined as the water needed to produce agricultural commodities, particularly in a context where they are exported. The concept is now widely used in the discussion of the need for water in food production, particularly in the arid, low-income nations. Chapters seven and eight of this volume develop a strong critique of the virtual water concept, making the case that it be abandoned. The core of Allans argument is that when a Region A discovers that the crop water requirements of food selfsufficiency are impossible to satisfy, the consequent water deficit can be resolved by the import of virtual water from Region B. To indicate the scale of these virtual water imports, we can take the example of wheat. One tonne of exported wheat requires about one thousand tonnes of virtual water (Allan 2001: 106, 126). Less than 0.1 per cent of the virtual water is physically embedded in the food grains themselves. During cultivation in Region B, more than 99.9 per cent of the virtual water returns to the irrigation cycle as farmland drainage or is lost in evapotranspiration. As Allan writes (2001:106): At the 1000 tonnes (cubic metres) of water per tonne of grain estimate of water content the [MENA] regional imports of virtual water by the mid-1980s were equivalent to the annual flow of the Nile into the Egyptian agricultural sector. My critique is four-fold. First, the term is redundant; virtual water is nothing more or less than the water needed to produce agricultural commodities. Secondly, there is absolutely nothing virtual about virtual water. Thirdly, when one approaches agriculture from the perspective of water resources there is a danger that the experience of farming is seen largely with respect to its crop water requirements. As a result, a more rounded vision is lacking, one that understands that the water theme is only one amongst many, such as soil characteristics, land rights, labour skills, pest control, farm budgets and product markets. Consequently, if we use the term the import of food, this opens up major questions rendered invisible by the import of virtual water.Have food imports led to higher population birth rates in water deficit regions than would have occurred in their absence? Do food imports weaken the farm sector of the importing country? Do food imports open the importing country to political control from the exporting country? Will the importing country be able to

The Price of Water

maintain its foreign exchange expenditure on food imports in the long run? Finally, if food imports are subsidized when they are shipped from the European Union and North America, will these subsidies be maintained in the long term? Fourthly, the confusion of water with food that accompanies the term virtual water fosters analytic errors such as that food grain purchases are transported water and that imported food brings total water self-sufficiency for the importing region (Allan 2003: Figure 3). The greatest weakness of Allans concept is that it repeatedly confuses crop outputs with the water required to produce them. If, as Allan suggests, Hoekstra and Hung (2002) conclude that 695 km3 of virtual water is traded each year, why have none of us seen the boats within which it is carried? Why are they not causing gridlock in the worlds sea channels? Virtual water, as defined by Allan, is real water; if it is indeed traded it needs to be transported to the new owner. In fact, no such trade ever takes place in the case of food exports. Hoekstra and Hung, outstripping Allan, commit a 695 km3 error. In fact, the huge economic, political and social processes that are addressed by Allan in terms of virtual water can be reset in a world where arid and semi-arid regions (and others) do not have the capacity to feed their populations and so import food. These imports mean that less food production and therefore less water is required in these regions irrigated agriculture. Where the regions exports are insufficient to finance food imports, regional politicians may deflect attention from such dependence; the availability of imported food allows them to postpone new water supply initiatives, to delay difficult decisions about the demand management of their water resources and to neglect the issues of birth control.

1.10

FINAL REMARKS

The author is convinced that, in human if not in financial terms, the most valuable areas of research in the 21st century will include humankinds understanding and management of water resources. Economists have a part to play here, alongside those who work in the fields of agriculture, development studies, engineering, environmental science, geography, hydrology and hydrogeology, law, planning, political science and sociology. Of course, the greatest difficulty in getting to grips with water is that it slips through your fingers. Nevertheless, our work as researchers should be bold, honest, should respect the labour of disciplines other than our own, and should seek to capture the glittering diamond of method, theory, fieldwork and policy development. The author believes that the century before us will witness water deficits on an unparalleled scale. He hopes that this collection of papers helps his colleagues in the global catchment to prepare for the challenges that will confront our species and the world-wide web of other life forms.

2 The regional water balance statement: a new tool for water resources planning
The king proceeded to make use of the multitudes whom he had brought with him from the conquered countries partly to dig the numerous canals with which the whole of Egypt is intersected The kings object was to supply Nile water to the inhabitants of the towns situated in the mid-country, and not lying upon the river; for previously they had been obliged, after the subsidence of the oods, to drink a brackish water which they obtained from wells. Herodotus c. 430 B.C.

2.1 INTRODUCTION
It is widely accepted that the effective planning of water resources becomes more urgent with each passing year. Some would argue that this need is driven by the increasing size and density of human populations at the catchment and urban scale. Others point out that exponential growth in economic output and consumption produces ever higher volumes of waste water (Lundqvist et al. 1985: 1). More recently, it is also asserted that global climate change will require every society to develop strategies capable of dealing with regional shifts in the mean and variance of hydrological variables such as precipitation. In Introduction to the Economics of Water Resources: An International Perspective I have proposed that such planning should take place within a framework determined by the quest for a sustainable society, and that strategy for the water sector should be balanced by programmes on both the supply- and the demand-side (Merrett 1997: 187). Moreover, while our developing understanding of the hydrological cycle
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

10

The Price of Water

provides the vital natural science input to strategy, it must be the hydrosocial cycle illustrated in Figure 2.1 which sources the language of supply and demand programmes. At this level, the key intellectual inputs come from politics, law, civil engineering, human geography, environmental studies and political economy. The objective of this paper is ambitious. Derived from the hydrosocial cycle perspective on supply and demand, the paper seeks to develop a quantitative tool for water resource planning which within a decade could be used across the globe for strategy development in respect of abstracted fresh water for outstream uses. This tool is the regional water balance statement and its derivative, the change statement.

2.2 THE RULES OF THE GAME

The paradigm formulation of the regional water balance statement is presented in Table 2.1. Before looking at the individual entries in this template, some rules of the game are required. First, the statement relates to any dened geographic space; the term region is used advisedly because of its inherent ambiguity. A region could be a continent, a country, a province, a catchment, an irrigation district, a city, a village, the site of a manufacturing rm such as a sugar mill, or, indeed, a water services company located in a dened area. Secondly, the statement is always applicable to a dened time-period. For convenience of exposition it is assumed here that the time-period is the year 1999. Alternatively it could be for the 5 years from 1995 to 1999, for example, or for the month of August averaged over the 10-year period 199099. Regional water balance statements for past or present time are referred to as baseline statements; those for future time are scenario statements. The statement has four columns: the rst two are the categories of supply and the quantity supplied per unit period of time; the second two are the categories of use and the quantity used per unit period of time. The quantitative measure of ow will be chosen on pragmatic grounds; it is assumed in Table 2.1 that we refer to ows in megalitres per day, averaged over the year. The boundary between supply ow and use ow should also be chosen on practical grounds this paper is not intended to be a visionary text. Such a boundary might be the point where the user rst possesses or has a right to the use of the water. Here, total use includes the leakage, evaporation and wastage which occurs on user properties, as well as benecial use. The calculation of ows in the regional water balance statement requires a procedure I shall call double-entry water accounting. The approach is taken from the doubleentry book-keeping rst developed in medieval Italy, and now universally used in modern nancial accounting (Dyson 1994: 4245). Within this routine, the company accountant has a large number of separate accounts such as a bank account, a creditors account, a debtors account, a cash account, a sales account and so on and so forth. Within the full set of a companys accounts, each transaction is recorded twice. In the account deemed to provide funds for the transaction, the transaction is recorded as a credit item. In the account deemed to receive the funds, the same transaction is recorded as a debit item. For example, the transaction of paying cash into the bank will be recorded as a credit in the cash account and as a debit in the bank account; the transaction of a cash receipt for the sale of goods is recorded as a debit

Salt water sources

Fresh water sources

Recycling

Disposal

Use of solids

Abstraction Waste water treatment

External reuse engineering

Storage

Fresh water treatment Storm water Storm water collection

Foul water collection

Distribution

Use

Internal reuse engineering

The regional water balance statement: a new tool

Evapotranspiration

Leakage Instream uses and agricultural return flows

Figure 2.1 A simple model of the hydrosocial cycle. 11

12

Table 2.1 Megalitres/day 14. 15. 16. 17. 18. 19. 20. 21. 22. Categories of use* Households Agriculture Mining Manufacturing Public services Commercial sectors Instream applications Other uses Total use

Regional water balance statement for the year 1999. Megalitres/day S T U V W X Y Z STYZ

Categories of supply

The price of water

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Rainwater collection Groundwater abstraction: rst time through Groundwater abstraction: recycled sources Surface water abstraction: rst time through Surface water abstraction: recycled sources Desalination of salt or brackish waters Import of water from another region Internal re-use of waste water External re-use of waste water Less: supply leakage and evaporation Less: export of water to another region Fall or rise of volume of stored water Total net supply

A B C D E F G H J K L M ABJKLM

* Includes benecial use, re-use volumes, and leakage, evaporation and wastage on user property.

The regional water balance statement: a new tool

13

item in the cash account and as a credit item in the sales account. In the process of balancing the books, because of this double-entry approach, the aggregate of all credit items in the separate accounts must equal the aggregate of all the debit items. Should this not happen, the accounts contain one or more errors of recording. The application of nancial accounting practice to the regional water balance statement must now be described. The statement is treated as a single ledger containing all the appropriate entries either as supply items on the left-hand side or as use items on the right-hand side. For any ow of water to be recorded in the statement, it must qualify as some hydrosocial category of input to the regional system. These input ows will be referred to as red molecule ows. In Table 2.1, ows 19 are red molecule ows. Such ows are parallel to the credit items described above and are entered on the left hand side of the water accounts ledger. Once so entered, each molecule must be assigned to one of the hydrosocial output ows. These output ows are referred to as blue molecule ows. In Table 2.1, ows 10 and 11 and 1421 are blue molecule ows. Such ows are parallel to the debit items described above and are entered either as negative values under the supply column or as positive values under the use column. Using these two colours, red and blue, in writing and printing any specic statement has a heuristic value water accounting throws up some puzzling questions. Because each red molecule from the supply side is re-entered as a blue molecule (negatively under supply or positively under use) total net supply is mathematically identical to total use. In the water accounts ledger, with comprehensive and accurate records, the statement always balances. It goes almost without saying that it is only as a thought-experiment that we can identify every input molecule and trace it through to its metamorphosis as a blue molecule. One hundred per cent tracing is not the point here. The crux of the technique is that, in principle, every drop of water supplied in the region in 1999 can be allocated to either supply-side losses to the region or to regional use. A single set of molecules is being categorized in two different ways, on entrance to the set of accounts and on exit from it, and this is why the mathematical identity of total net supply and total use holds true.

2.3 THE SUPPLY CATEGORIES

Let us now consider in turn each of the categories of supply in Table 2.1. Category 1 is rainwater collection. This needs no gloss; it is found from the hamlets of the Caribbean island of Providencia to the site of the Millennium Dome in Greenwich. Rainwater collection always was and always will be with us. Categories 25 refer to the abstraction of groundwater and surface water; each of these two sources is sub-divided into rst time through supply and recycled supply. The use here of the term recycling requires clarication. The hydrosocial cycle of Figure 2.1 includes the collection of storm water and foul water and (after any treatment they receive) their routing back to rivers, lakes and aquifers. This and only this I call recycling. The term does not embrace disposal to the sea. Where regional abstraction takes place downstream of recycling points, such abstraction ows can be decomposed into two parts: the abstraction of rst-time water and the abstraction of recycled water. The quantitative ratio of these two ows

14

The Price of Water

can be taken to be equal to the ratio of fresh water to recycled water in the stream ow at the abstraction point. Table 2.1 incorporates the distinction by breaking down into two parts each of the red molecule abstraction ows. But note that the recycled ow into the fresh water source is neither a red nor a blue molecule ow. Recycled water is recorded in the statement only when it is abstracted. In some regions, it may be right to ignore this break-down because it has no policy interest, no relevance to foreseeable infrastructural investment. But where sustainable water resource planning seeks to protect the hydrological cycle by the use of properly-treated recycled water, the distinction may be vital and measurement justied. Recycling adds to effective rainfall as a source of water for abstraction, as can be seen from the planning documentation of the Thames catchment (NRA 1994). The water ows of the Thames are among the most intensively used in the world. Categories 6 and 7 add the two red molecule ows: desalination of salt or brackish water, and the import of water from another region. This is big-time supply-x territory, this is the western USA, this is California. Also added, for the sake of symmetry in water transfers, is the blue molecule ow of water exported to another region, supply category 11. In addition to recycling, Figure 2.1 also includes two further green loops: internal and external reuse. Internal reuse occurs when a household or a factory or any other organization reuses its own waste water. The water volume of internal reuse is set equal to each cubic metre of fresh water supplied to the user multiplied by the average number of times it is reused. External reuse occurs when the waste water of one organization or group of households is reused by a separate body, as in the reuse of treated waste water by agriculture. These loops are included in Table 2.1 as the red molecule ows of supply categories 8 and 9. The same volumes are entered as blue molecule ows in the categories of use. In this way, the fundamental mathematical identity is retained. Total net supply including reuse supply is equal to total use including its reuse ows. My remarks above on the policy relevance of measuring recycled water apply also to reuse. Note that the distinction between recycled and reused water is that the former is water returned after its rst use to river, lake and aquifer whilst the latter goes for reuse before disposal to fresh or salt water sinks. Supply category 10 is (like 11) a blue molecule ow; these leakages and evaporation in the supply system between the points of abstraction and the supply/use boundary are deducted in the calculation of total net supply.

2.4 WATER STORAGE

Up to this point, the balancing of the water accounts ledger has derived from the notion that each and every molecule of water recorded as an input to the regional system in a given year, as part of a red molecule ow, is then recorded as an output from the system in that year as part of a blue molecule ow. The question then arises: does the existence of water storage infrastructures destroy the accounting balance? Our understanding of this issue will be strengthened if we imagine the regions storage capacity as composed of just three reservoirs, one black, one gold and one green. The black reservoir is dedicated to the storage of water abstracted in 1999 that in the same year is distributed in its entirety to users or lost to leakage and evaporation.

The regional water balance statement: a new tool

15

Clearly, storage of these pass-through molecules in the black reservoir does not change the systemic balance for 1999. The gold reservoir contains stored water abstracted during time-periods prior to 1999. These molecules are, so to speak, a gift from the past to the present. In 1999 some of the stored water is lost to supply leakage and net evaporation, some may be delivered as water exports to another region, and some is distributed to users. The fall in the quantity of water stored in the gold reservoir during 1999 is expressed in megalitres per day and is deemed to be a red molecule ow. Once again, the identity of total net supply and total use is maintained. The green reservoir is dedicated to the receipt of water abstracted during 1999, which will be stored for distribution from 2000 onwards, a gift from the present to the future. The increase in the volume stored in the green reservoir during 1999 is expressed in megalitres per day and is deemed to be a blue molecule ow. This ow precisely matches the abstraction ow pumped to the reservoir and, for the third time, the mathematical identity holds. In practice, of course, each reservoir in a real regional system combines the functions of all three reservoirs described above. What we observe is only the net outcome of the component processes, that is, either no change in 1999 in the volume of stored water, or a fall or a rise. Thus, with respect to the value of M in Table 2.1, no change in the total volume of stored water gives a value of zero; a fall is expressed at its daily rate and is recorded as a red molecule ow; and a rise in storage is expressed at its daily rate and is recorded as a blue molecule, negatively-valued ow. The analysis here of reservoir storage applies with equal force to aquifer storage and recovery.

2.5 THE USE CATEGORIES

The use categories 1421 of Table 2.1 can be swiftly dealt with. The categorization is designed to group together similar types of use so that use forecasting is facilitated. It has already been noted that the term use includes benecial use, reuse, and the leakage, evaporation and wastage occurring on user properties. In planning for any real region categories 1421 would be redesigned in the manner most effective for strategy development in that area. Only category 20 requires further comment. It is already stated above that the regional water balance statement is a tool for strategy development in respect of abstracted fresh water for outstream uses. The statement does not, in general, address instream uses of water for nature conservation, leisure pursuits, shing and navigation. These uses are not of a kind that makes double-entry water accounting relevant. However, instream ows are sometimes supplemented by abstraction. For example, in England and Wales British Waterways pumps groundwater to maintain the necessary volumetric ows in its canals without which navigation and boating is hindered. Similarly, the Royal Society for the Protection of Birds in the UK may use water service company water supplies for application to the wetlands it manages in order to maintain their environmental quality at times when rainfall and river ows are decient. Where red molecule ows are allocated in this way, these applications must be recorded in category 20 in order to retain the necessary equivalence of total net supply and total use.

16

The Price of Water

2.6 THE CHANGE STATEMENT

Regional water balance statements, as already indicated above, may be baseline statements for the past or present or they may be scenario statements for the future. When the baseline and the scenario statement have the same structure, we can subtract the entries of the former from the entries of the latter to give a regional water balance change statement. This result is illustrated in Table 2.2, where lower-case letters are used to indicate we are dealing with differences in values. Ten rules for the change statement are worth setting out; they are all derived from a single arithmetic rule of subtraction of baseline (positive, zero and negative) values from scenario (positive, zero and negative) values.

Where the value of a cell is unchanged between the baseline and the scenario statements, the change value is zero. In the case of each of the red molecule ows from rainwater collection to external
reuse, if the value in the scenario year exceeds that of the baseline year, the change value is positive. In the case of each of the red molecule ows from rainwater collection to external reuse, if the value in the scenario year falls short of that of the baseline year, the change value is negative. In the case of each of the blue molecule ows from household use to other uses, if the value in the scenario year exceeds that of the baseline year, the change value is positive. In the case of each of the blue molecule ows from household use to other uses, if the value in the scenario year falls short of that of the baseline year, the change value is negative. In the case of each of the supply-side blue molecule ows (supply leakage/ evaporation and export of water to another region), if the absolute value in the scenario year exceeds that of the baseline year, the change value is negative. For example, if we have a shift from a base year value of 7 Ml/d exported water to a scenario year value of 11 Ml/d, the value of the letter l is 4 Ml/d. In the case of each of the supply-side blue molecule ows (supply leakage/evaporation and export of water to another region), if the absolute value in the scenario year falls short of that of the baseline year, the change value is positive. In the case of stored water, the value of m is positive when a fall in the scenario year exceeds a fall in the baseline year or when a rise in the baseline year is succeeded by a fall in the scenario year. In the case of stored water, the value of m is negative when a fall in the scenario year falls short of a fall in the baseline year or when a fall in the baseline year is succeeded by a rise in the scenario year.

From these rules we can see that any single lower-case value may be positive, zero or negative. In the change statement the total change in net supply is equal to the sum of entries a to m. Similarly, the total change in use is equal to the sum of entries s to z. A nal rule of great importance can be established. Since total net supply and total use in the baseline year are identically equal, and since total net supply and total use in the scenario year are identically equal, it follows like the night the day that the change between the 2 years are identically equal. So, in the regional water balance change statement, Merretts law states:

Total change in net supply is mathematically identical with total change in use.

Table 2.2 Changes in megalitres/day a b c d e f g h j k l m ablm 14. 15. 16. 17. 18. 19. 20. 21. 22. Households Agriculture Mining Manufacturing Public services Commercial sectors Instream applications Other uses Total Change in Use Categories of use*

A regional water balance change statement for the 10-year period between 1999 and 2009. Changes in megalitres/day s t u v w x y z st yz

Categories of supply

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Rainwater collection Groundwater abstraction: rst time through Groundwater abstraction: recycled sources Surface water abstraction: rst time through Surface water abstraction: recycled sources Desalination of salt or brackish waters Import of water from another region Internal re-use of waste water External re-use of waste water Less: supply leakage and evaporation Less: export of water to another region Fall or rise of volume of stored water Total Changes in Net Supply

The regional water balance statement: a new tool

* Includes benecial use, re-use volumes, and leakage, evaporation and wastage on user property.

17

18

The Price of Water

I have suggested that the change statement is derived as the difference between the baseline and the scenario statements. In planning practice, it may be common for a change statement to be added to a baseline statement to produce a scenario statement.

2.7 THE USES OF REGIONAL STATEMENTS

The time has arrived to outline the potential value of these statements and to propose why they may become a routine tool of water resource planning in a wide range of countries over the next decade. The baseline statement can be prepared for any geographic area and provides a comprehensive, synoptic account of the scale and composition of the supply sources of water and its uses in that region. The approach is pragmatic: one can foresee an evolution in statements for a specic region from the simplest set of water accounts to more complex ones, as policy interests drive the expansion of the categories deployed and the accuracy of their measurement. The work process to make the baseline calculations will stimulate new directions for research and will deepen planners understanding of the regions hydrosocial cycle within its hydrological context. The four-column approach gives equal weight both to supply-side and demand-side information and describes them in the language of a unied conceptual framework. The recycling and reuse entries will strengthen our understanding of the cyclical components of water resource planning, make them more visible. The fundamental identity of total net supply and total use provides a consistency check on individual calculations, akin to the trial balance procedure of company accountants (Merrett 1997: 11524). The change statement, the planning tool par excellence, is applicable to short-, medium- and long-term strategy, to comparison of typical years past and future, and it claries the quantitative choices that a regional government or catchment agency will be forced to envisage in a future year of severe drought, or as a result of climatic change in hydrological variables. The change statements can provide comprehensive, synoptic, transparent alternative scenarios within a region for infrastructural investment and for shifts in the scale and pattern of uses. It is in this way that it handles the uncertainty that, as Peter Rogers writes, is a major issue in water resources planning (1991: 18), as well as side-stepping the inadequacy of project-by-project planning (Rogers 1992: 1). The change statement does not evaluate a scenario, this is the work of environmental scientists, economists and others, but it does impose a consistency check on each case, because of the fundamental identity of total change in net supply with total change in use. The approach to a regional water balance statement presented here can be contrasted with current, orthodox methods. For example, Philip Anderson (1997) adopts the conventional approach in recent work on Scotland. In his study the demand side is based on social processes while the resource side is derived from natural processes in the form of hydrological yields, and where the resource and demand totals do not, in fact, balance. As a result of this yields approach, his overview of the Scottish situation tabulates only one of the 12 supply-side variables listed in Table 2.1, that is, leakage. A second example of the current orthodoxy is the work of Miloradov & Marjanovic (1998): Guidelines for conducting water resources assessment. It is worth noting that this

The regional water balance statement: a new tool

19

publication is one of UNESCOs studies and reports in hydrology and was written as a contribution to the International Hydrological Programme. Not surprisingly, the authors approach to ows, stocks and balances is essentially hydrological, not hydrosocial. Their Chapter 4 is entitled Basic theoretical principles for processing a water resources balance. It suggests that the core component of any water resources master plan is the water resources management balance (WRMB). However, in a graphic illustrating the water resources balance approach, the supply-side variables are predominantly hydrological and the uses of water are not represented at all. Moreover the terms of the general WRMB equation for a given area in a given timeperiod embraces a mixture of hydrological variables (such as rainfall and aquifer ows) and engineering variables (such as articial water conduit ows and discharges by water users). The general equation and the specic equations derived from it simply do not seek to report the balance between how human society in a given area gains access to its water and how it uses it. So the WRMB, however necessary, is strikingly different from the regional water balance statement presented in this paper. My third and last example of the received wisdom is the Assessment of water resources and water availability in the world authored by I.A. Shiklomanov and published by the World Meteorological Organization (WMO) in 1997. Chapters 14 are entirely hydrological in their orientation to resource availability, with a special stress on river runoff. Chapter 5 (Water resources use) and Chapter 6 (Water availability and water resources decit in the world) contain valuable synoptic material on patterns of water consumption but the supply-side concept never moves beyond either the generic term water withdrawal or the two-fold breakdown of supply into surface water and groundwater sources. Once again the approach is unaware of the richness of the concept of supply when it is located within the hydrosocial rather than the hydrological paradigm, with the huge relevance this has to catchment planning; and at no point does the WMO text set up a regional water balance statement conceived as a mathematical identity.

2.8 CONCLUSIONS
I have argued that a regional water balance statement of the type presented in Table 2.1 and the change statement of Table 2.2 make a radical break with existing approaches, in fact complement them, and that the new tool has a strong relevance to water resource planning practice. But it is a mistake to gild the lily. The limitations of the statement must also be recognized: it is a quantitative technique that embodies no hydrological or meteorological variables; for all practical purposes it does not address instream uses; and it makes no reference to water quality. Thus the regional water balance statement should be seen only as one of the many, interdependent approaches and techniques necessary for effective water resource management.

The material in this chapter originally appeared in: The regional water balance statement: a new tool for water resources planning. Water International, 24(3): 268274, 1999.

3 Integrated water resources management and the hydrosocial balance

To see a world in a grain of sand William Blake (1803)

3.1 INTRODUCTION
Integrated water resources management (IWRM) seeks to integrate in many different ways: the social sciences with the natural sciences, planning with implementation, and groundwater with surface water, to mention just three such ways. The rst objective of this paper is briey to introduce a planning method, the hydrosocial balance, and to show how it integrates: (i) outstream water quantities supplied and used, and (ii) the present with the future. The paper then shows how the hydrosocial balance can be developed to integrate water qualities with water quantities. The papers third and nal objective is to apply this management tool to a case-study from the island of Jersey in the English Channel.

3.2 A NEW IMPLEMENT FOR THE IWRM TOOLBOX

This section recapitulates the main features of a new water resources management tool, the hydrosocial balance, developed over several years. It is a tool that until now has been limited to the analysis only of water quantities (Merrett 1997: 1522; 1999: 26874; 2002: 14853).
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

20

The hydrosocial balance is a quantitative water resources planning method applicable in principle to any space with a dened boundary. This might be a house on the coast of Andaluca, Wembley football stadium, the island of Gorgona, the catchment of the Yangtse, the State of California, or the continent of Africa. Such spaces are referred to here generically as regions and are not required to be related to traditional catchment areas. The hydrosocial balances principal distinctiveness from the water balance familiar to hydrologists is that the former incorporates (almost without exception) only outstream, hydrosocial ows, and never the hydrological ows of precipitation, groundwater recharge, run-off and rivers. The water ows that it places at the centre of analysis and measurement are those directly created by human society; the idea of a hydrosocial balance is derived from the concept of the hydrosocial cycle (Merrett 1997: 67). A hydrosocial ow represents a human activity. So the hydrosocial balance, composed as it is of many hydrosocial ows, is understood primarily through the social sciences. In contrast, the hydrological balance represents natural ows and is understood primarily through the natural sciences. To summarize, hydrological ows are of a type that exists in a state of nature, prior to the recent appearance of Homo sapiens. Hydrosocial ows, in contrast, are specic to human society. In recent centuries these two types of ow have become ever more powerfully interdependent with the growth of world population and its economic activities of production and consumption. The water balance of an area and that areas hydrosocial balance should be estimated separately prior to considering their quantitative and qualitative interdependence. The generic form of the hydrosocial balance for a specied region is given in Table 3.1. A baseline balance is for a past time-period, such as the year 2001. A scenario balance is for a future time-period such as the year 2007. The shift in the quantity in millions of cubic metres (Mcm) of any one category of supply or use between the baseline year and the scenario year can be represented both as an absolute change, vide column 4 of Table 3.1, as well as an annual rate of growth or decline, vide column 5. The baseline balance provides a comprehensive, synoptic account both of the scale and composition of the supply sources of water as well as their use in the region it covers. Where measurement is accurate and comprehensive, the total net supply is always equal to total use. Scenario balances provide options for the future, based on the forecast need for outstream water in different uses and the possible allocation conicts that may be foreseen. Once again, total net supply must be planned to equal total use. The absolute difference of supply, and of use, between the base year and any specic scenario year, together with the associated annual rate of change, provide the basic input to the planning of infrastructural investment, capital nancing and demand management.

3.3 THE BRIDGE BETWEEN QUANTITY AND QUALITY

The previous section focused entirely on quantity. In this section a bridge is built that links quantity to quality. Table 3.1, for a given base year or scenario year, has more than a dozen supply-side and demand-side ows. All these ows, plus the post-use ows of waste water (including irrigation drainage), can be reclassied into the ve groups set out in Table 3.2. Note that in a region where there is no treatment of the water supply, or of waste water, one has only groups one, three and four: the supply ow, the use ow and the waste water ow.

Table 3.1 The hydrosocial balance for a specied region in a base year and a scenario year (in millions of cubic metres, mcm). Base year Scenario year Scenario year minus base year ( or ) Annual compound rate of growth from the base year to the scenario year(%)( or ) Ga Gb Gc Gd Ge Gf Gg Gh Gj Gk Gm Gs Gt Gu Gv Gw Gx Gy Gz

Categories of Supply Rainwater collection Groundwater abstraction Surface water abstraction Desalination Import of water from other regions Internal reuse of wastewater External reuse of wastewater Total Gross Supply A1 B1 C1 D1 E1 F1 G1 H1 J1 K1 L1 M1 S1 T1 U1 V1 W1 X1 Y1 Z1 S2 T2 U2 V2 W2 X2 Y2 Z2 L2 M2 (L2) (L1) M2 M1 S2 S1 T2 T1 U2 U1 V2 V1 W2 W1 X2 X1 Y2 Y1 Z2 Z1 J2 K2 (J2) (J1) (K2) (K1) A2 B2 C2 D2 E2 F2 G2 H2 A2 A1 B2 B1 C2 C1 D2 D1 E2 E1 F2 F1 G2 G1 H2 H1

Supply leakage and evaporation Export of water to other regions Fall () or rise () in volume of water abstracted and stored Total Net Supply

Categories of Use Households Agriculture Mining Manufacturing Public services Private services Other uses Total Use

Note: Gj and Gk are calculated using absolute values of leakage and exports. Gl is not calculated because of the possible change of sign. Source: Adapted from Merrett (2002a), Tables 7.1 and 7.2.

Table 3.2

Flow types in the hydrosocial cycle. Use-ows 3 At the point of use Waste water ows 4 Prior to treatment 5 After treatment

Supply-side ows 1 Prior to treatment 2 After treatment

Note. The supply-side ows include supply leakage. The waste water ows include irrigation drainage. Different levels of treatment produce different quality products at different costs with different implications for users and the environment.

For any of the ows of Table 3.2, its complex quality can be assessed provided that a water institution has the skilled professionals and the laboratories to carry out the necessary analysis. At the most general level, a ows quality can be measured by applying four criteria to samples taken from the ow: 1. 2. 3. 4. The individual inorganics present in the samples (such as arsenic, lead and zinc). The individual organics present (such as atrazine, malathion and 2,4-D). The microbiological content of the samples (in terms such as faecal coliforms, pathogenic staphylococci and salmonella). Other indicator measures (such as biochemical oxygen demand, total suspended solids and pH).

However, if one considers the hundreds of individual characteristics that can be generated by these four criteria, it would require a prodigious hydrochemical infrastructure to process comprehensively even a single sample from a single ow. So measurement must always be targeted, principally by considering the t for purpose needs of the analysis. That is to say, one reviews what the water ow under assessment is to be used for, or to which location it is to be discharged. So, at the most elementary level, if the water that is sampled has to meet drinking water standards, as it will be pumped to domestic premises, then the water quality assessment is quite different from that for water one plans to discharge to coastal waters. Moreover, the two ows will be governed by different legislation and standards. In summary, the bridge linking the quantities of the hydrosocial balance to their qualities is built in the following manner. One recognizes rst that the quantities of the hydrosocial balance fall into the ow types of Table 3.2; second that each ows quality can be assessed in terms of the four criteria listed above; and third that the specication of the assessment should be based on the resources available to carry it out, the legislative requirements to meet prescribed standards, and the t-forpurpose requirements of the hydrosocial balance ows themselves. This suggests a new term is required. When one has a cross-tabulation for a specic hydrosocial ow with: (i) rows that refer to that ows qualitative characteristics, and (ii) columns setting out the number of samples made and the measured concentration per litre or measured value of each characteristic, this will be referred to as a quality matrix. Up to this point the text is at a high order of generality. Sections 3.43.7 record an attempt to apply this analytic framework, or meta-theory, to the island of Jersey. Beginning with the geographical and hydrological background, the paper moves on to cover the supply-side of the hydrosocial balance, then the use of water in Jersey, and nally the complex issues of water quality. The case-study benets from recent

hydrogeological publications of great quality, as well as a series of interviews with key personnel.

3.4 THE PHYSICAL GEOGRAPHY AND HYDROLOGY OF JERSEY

The island of Jersey is located in the English Channel, some 60 km north of the French coastal town of St Malo. It has an area of 117 km2 and in 2001 had a resident population of about 85,000 persons, giving a density of 726 persons/km2. As Robins (2000: 23) points out, the island comprises a plateau lying at an elevation of 60120 m divided by a series of valleys running from north to south. The bedrock beneath Jersey mostly consists of ancient metamorphosed shales of Precambrian age intruded by younger igneous rocks. Unconsolidated deposits of Quaternary age partially cover the bedrock, whilst post-glacial peat, alluvium and sand are also present in valleys and low-lying coastal areas. Westerly and south-westerly winds bring moisture from the Atlantic. Long-term average rainfall for the island is 877 mm, annually varying from 600 to 1100 mm in the past decade. Mean annual potential evapotranspiration ranges from 648 to 784 mm. The 136-year monthly average rainfall shows that the bulk of precipitation occurs in the 7 months SeptemberMarch (Jersey New Waterworks Company 2002: 4). Flow along the valley streams is north to south and, in the absence of a dry winter, takes place throughout the year. Groundwater discharges both to these streams and at the coast. During prolonged dry weather, saline intrusion may occur locally. Robins writes (2000: 8): There is a notable variation in run-off and inltration from year to year. Poor winter rains had a marked effect on values for the notorious dry years 1975/76, and the periods 1989 to 1992 and 1995/96. These dry years are signicant, as it is these years of water stress that the Island needs to be able to cope with in terms of surface water storage capacity and conjunctive abstraction of groundwater. Robins also shows that the islands base ow index has an estimated value of 0.58, indicating that 58% of streamow is derived from the groundwater baseow (Robins 2000: Table 1; Blackie et al. 1996). This quantity will be of importance in the discussion of water quality in Section 3.7.

3.5 WATER SUPPLY IN JERSEY

At this point one can begin to review the principal actors in Jersey that supply water and that use it. As this is done, the jigsaw of Jerseys hydrosocial balance is assembled in Table 3.3. There can be no question that the principal supply-side provider is the Jersey New Waterworks Company (JNWWC). This is the islands only water utility. However, waste water collection, treatment and discharge are the responsibility of the States of Jerseys Public Services Department (PSD). The JNWWC is a water undertaking incorporated in 1882 and its principal activities are the supply and treatment of water for domestic use and other purposes on the island. The States of Jersey, the islands government, holds 100% of the issued A

Table 3.3 The hydrosocial balance of the states of Jersey in 2001. Row 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Categories of supply Surface water and groundwater abstraction by the JNWWC Desalination by the JNWWC Groundwater abstraction by farmers, households, etc. Surface water abstraction by farmers Rainwater collection net of evaporation Internal reuse External reuse Imports from other regions Total Gross Supply Less: exports to other regions Less: supply leakage and evaporation Change in volume of water abstracted and stored Total Net Supply Categories of Use Sourced by the JNWWC Households Agriculture Manufacturing Public services Private services Other uses Sub-Total Sourced by groundwater pumped by farmers, households, etc. Households Agriculture Industry Hotels and hospitals Leisure Other Sub-Total Irrigation water use sourced by farmers surface water abstraction Irrigation water use sourced by rainwater collection Error term Total Use Mcm 6.2 1.1 3.6 not known 0.1 0.1 0.1 0.0 10.9 0.0 0.7 0.1 10.2 Mcm 4.5 0.1 0.1 0.6 1.3 0.1 6.5

0.9 1.4 0.1 0.1 0.8 0.3 3.6 not known 0.1 0.1 10.2

ordinary shares. However, the utility acts like a private company. It has a desalination plant, six reservoirs, an extensive network of raw water mains, two treatment plants and a distribution infrastructure to the islands users. The total length of the trunk mains is 70 km and of the service mains 415 km. In 2001 the company had a turnover of 11.0 million, it employed 112 persons and the balance sheet shows that its xed assets, excluding landholdings, were valued at 47.7 million (JNWWC 2002). Of the annual volume of water abstracted by the JNWWC, the overwhelming share is from its surface water reservoirs. Total abstraction in 2001 was 6.2 Mcm to

which one should add 1.1 Mcm output from the desalination plant at La Rosire. This is shown in Table 3.3, rows 3 and 4. Note that, for management purposes, there is a good deal of switching of surface water between reservoirs, but from the perspective of the island as a whole, this can be ignored in calculating the abstraction total. The low reliance on groundwater as a primary source of supply, in contrast to surface water, may be because the prime costs of bulk surface water abstraction are lower than for groundwater, given that the main aquifers are of only poor to moderate transmissivity. Another explanation is that the company believes groundwater quality to be inferior to that of surface water in Jersey (see Section 3.7 below). If the JNWWC is the principal supply-side provider, the second most signicant supply source is a group made up of individual farmers and households. Agricultures needs for water are met largely from supply points on the farms themselves. These are either boreholes or ponds fed by small streams. The nancial costs of pumping from these boreholes and ponds are borne by the farmer. At the present time, boreholes are neither registered nor licensed nor metered, nor is any abstraction charge levied by the States. The same holds true for abstraction from farm ponds; in this case farmers usually have mobile diesel engines to pump the water from pond to eld or to farm buildings (Vint, pers. comm.). About 15% of the islands resident households are unconnected to the JNWWC network and about 17% have no link to the PSDs sewerage. There is a large overlap between these two groups of the more isolated households. Families off the water supply network use boreholes and wells for their domestic supplies. Once again these individual supply points are unregistered, unlicensed, unmetered and there is no charge for their abstraction. One can say that access to the aquifer by these individual farmers and households is outside the sphere of the States management of the islands water resources. This unmeasured and unmanaged characteristic of surface water and groundwater abstraction by most farmers and a minority of households raises obvious difculties in estimating Jerseys hydrosocial balance. However, estimates of groundwater abstraction and springow interception were made for the dry years 198991, when borehole meters were installed following the State of Emergency powers granted during the 1989 drought (Robins 2000: 15). The number of supply sources at that time were: 500 in agriculture, 4000 in the domestic sector, 50 in the leisure sector and about 80 others. Robins suggests that the average use of groundwater remained generally stable during the 1990s and is likely to be little changed to this day. This gives an annual rate of groundwater abstraction by private actors of approximately 3.6 Mcm in Table 3.3, row 5. Rainwater collection is now considered. It is known that some householders not connected to the mains supply collect water on their roofs and route it to a cistern. The scale of this is believed to be small. More important is the greenhouse sector. This includes polytunnels and glasshouses. Some farmers have very large glasshouses for the commercial production of owers, vegetables and soft fruit. They capture runoff from the roofs by channelling it along the roof-gutters and through downpipes to water tanks. The Jersey Department of Agriculture and Fisheries (DAF) suggest that the total area of glasshouse roofs is some 51,000 m2. Net of evaporative loss, rainfall on these roofs was 537 mm in 2001, giving rainwater collection in this case of about 27,000 m3; this also is recorded in Table 3.3, row 7. The evaporative loss is low because rainfall is channelled quickly and directly to the collection tanks. Internal and external reuse in Table 3.3 now needs to be considered. Reuse refers to water resource ows that, after abstraction, are rst used in the domestic, industrial

or irrigation sectors and then are used for a second time or, indeed, time and time again prior to recycling back to lake, river and aquifer (Merrett 2002a: 8). Internal reuse occurs when the repeated use is by the same institution where the water was rst used. This is familiar in the industrial sector, and increasingly in the domestic sector where grey water is reused to irrigate gardens and to ush toilets. External reuse takes place when the repeated use is by a different institution, such as the use of waste water from the cities for the irrigation of farmland; the cases of Ghana and Israel can be cited here. The importance of internal and external reuse to the hydroeconomist is that it multiplies the productivity of a given volume of water appropriated as a base ow. The water quality of the efuent reused and the degree of treatment (if any) that it undergoes are critical issues. In response to questions on this subject to a number of water resource professionals in Jersey there was a common answer: cases of reuse can be found but the aggregate scale on the island is negligibly small. The most common case is one of internal reuse: households that use their kitchen water to irrigate their gardens. The last entry under total gross supply in Table 3.3 is the bulk import of water to Jersey. This simply does not take place. One can now estimate the total gross supply; the gure is 10.9 Mcm for the year 2001. However, this total excludes the volume of one non-negligible entry for which no estimate exists: surface water abstraction by farmers from the ponds and pools on their land that are fed by small streams. In future the Water Resources (Jersey) Law that has recently been drafted will address this. To move from total gross supply to total net supply, three adjustments have to be made: for exports, losses and storage change. Exports of bulk water by Jersey to other regions, as with imports, are zero. However, supply-side leakage and evaporation of abstracted water between the point of abstraction up to the boundary of users properties is of real importance. The JNWWC has a number of staff who work to keep losses down to an economic level. The company estimates that in 2001 their losses were 6% of gross supply, an extraordinarily low gure by British and world standards. In the complete absence of other data, the same percentage is also used here for groundwater abstraction by farmers, households and others. This gives an entry in Table 3.3, row 13 of 0.7 Mcm. Lastly, row 14 in Table 3.3 refers to changes in the volume of water abstracted and stored. The logical place of this category in the hydrosocial balance is described at length elsewhere (Merrett 2002a: 1501). It is sufcient to say here that the entry refers only to changes in the volume of stored water that has already been captured, appropriated or abstracted in some way. It does not refer to hydrological changes in the volume of water in a lake or behind a barrage resulting from variation in a catchments river ows. In the case of Jersey, the volume of abstracted water in storage in this sense is quite small at any point in time, so that changes in its volume over the course of a year are negligibly small and can be ignored. The end result is that, for the time being, the best estimate of total net supply in Jersey for the year 2001 is 10.2 million cubic metres.

3.6 WATER USE IN JERSEY

This section addresses water use on Jersey in the baseline year of 2001. The use data in Table 3.3 can be seen to fall into two principal blocks: uses sourced by the JNWWC

and those sourced by private abstraction of groundwater. Examination of the tables categories of use and the volumes known to be used lead to a clear conclusion: 88% of the known total is made up by just three groups of users. In descending order these are: households, private services and agriculture. Use by households is the largest category by far, totalling 5.4 Mcm. In the year 2001, the resident population of Jersey was about 85,000. There are individual houses and villages all over the island but the largest population concentration is on the south coast, particularly in Jerseys capital, St Helier. Dividing use by resident population gives a gure of 174 litres per head per day (lhd) for domestic use. This is rather high compared with the UK average of 147 lhd, which in any case is (mistakenly) inclusive of supply-side losses (Environment Agency 2002). Of the 5.4 million total, 4.1 Mcm is used by JNWWC customers who are not metered. Their domestic water supply charges are based on the capital value of their property. So there exists no price signal to encourage demand management in these cases. This may explain, partially at least, the high domestic use average. Private services rank second among the users of water in Table 3.3. The JNWWCs water demand staff provide a useful breakdown of this category. It includes, in descending order of volume: hotels and guest houses; ofces, banks and shops; public houses; restaurants; sports facilities; stores, garages and car parks; clubs, parish halls and day centres; laundries; and hair and beauty salons. The two dominant economic contributors to Jerseys gross domestic product are the nancial services sector and tourism; this explains the salience of private services as a user of water. At the height of the summer there are around 35,000 tourists on the island; this generates a strong summer surge in demand. All commercial properties are metered. However, there is no metering of tourists and visitors individual water use within their hotels and lodging houses. Even excluding the irrigation water pumped from on-farm ponds, agriculture is the third largest user of water. Uses include net irrigation requirements, livestock watering, cleansing of farm buildings and machinery, and product washing. Traditionally the Jersey economy has been regarded as quintessentially agricultural (Frigot 2001). Even in 2001, farmland constituted 49.6% of the islands area, with 352 farmholdings. The principal outputs from the sector are Jersey Royal potatoes, tomatoes, and the dairy products of the islands 4550 Jersey cows and heifers (DAF 2002). Jersey soils are inherently low in organic matter and in most cases multi-cropping requires the addition of fertilizer. Returning to Table 3.3, one sees sub-totals of 5.4 Mcm for household use (rows 18 and 26), 2.1 Mcm for private services (22 and 30) and 1.4 Mcm for agriculture (19 and 27). Adding in the remaining uses specied, as well as an error term of 0.1 Mcm to offset the difference between estimated total net supply and total recorded use, gives total use of 10.2 Mcm.

3.7 WATER QUALITY IN JERSEY

The water quality of Jersey is a large subject. Here the discussion is limited to ve hydrosocial ows selected because of their particular interest for water resources management. These ve ows are the irrigation returns of farm use, the abstraction of groundwater by households unconnected to the JNWWC s water supply, the domestic waste water from households unconnected to the PSDs sewerage, the

JNWWC s distribution to households, private and public services, etc. and the PSDs discharges to coastal waters. Quality matrices will be deployed below for the last two of these ve ows, to illustrate the link of quality to quantity.

3.7.1 Irrigation returns

The water requirements of Jerseys crops are met both by rainfall and by irrigation, as is the case with the water needs of its parks and golf courses. Part of this joint ow is consumed as evapotranspiration. The remainder ows by gravity as eld drainage, either as run-off to local streams or by percolating down to the aquifer. At present, no water quality analysis is directly made of agricultural drainage in the island on a regular basis. However, a number of studies exist that are relevant to linkage between farmland drainage and groundwater quality. Robins & Smedley (1998) have presented the evidence in great detail. Jerseys farmers do not stint in their use of fertilizers, particularly for early potatoes. These include the traditional 201010 NPK product as well as livestock slurry, the manure of grazing cattle and, now on a much-reduced scale in comparison with the past, vraic the local term for seaweed collected from the beaches. These inputs contain inter alia nitrates, phosphorus and potassium. Robins (2000: 26), reporting on work by Lott et al. (1999), states: the continued zealous use of nitrate fertilizer and application of farm slurry remains largely above recommended levels for the UK . Chilton & Bird (1994) did carry out research specically into the pore water below farmers elds. They noted correlations with fertilizer use not only of nitrates but also of the other common fertilizer ions of ammonium, chloride and sulphate. Since Lotts report, the Department of Agriculture and Fisheries, in a personal communication, writes that there is increasing evidence that a majority of the farming industry has now adopted changed practices. Farmers also use herbicides and fungicides, and their residues and metabolites exist in farmland drainage. Until it was banned in Jersey in 1998, the pesticide chlorthal dimethyl was an ingredient in pesticides applied to brassicas and soft fruits. A principal degradation product, the metabolite chlorthal, is environmentally persistent. Matthews & Carter (1999) modelled its movements and concluded that by 2002 farmland drainage would in all cases have concentrations less than the critical level of 0.1 micrograms per litre. However, by October 2002 it was still well above the critical level in surface water and groundwater (G. Jackson, pers. comm.). Certainly there exists a Water Pollution Law (PSD 2000) to safeguard surface water and groundwater from contamination. This is targeted principally at point sources but Articles 1216 can tackle diffuse pollution issues. The DAF also has an agricultureenvironment scheme that aims at the better management of slurry, reduced fertilizer use and increased care in the use of biocides. The Water Pollution Law will also be used to introduce water management areas and these areas may themselves stimulate action to reduce the contamination that originates from farmland drainage. Such drainage is said to be the origin of some 75% of the nitrate contamination of the islands groundwater (JNWWC 2002: 5). In the case of this rst ow, farmland drainage sourced by irrigation returns and rainfall, there is a strong case for wishing to examine its current quality matrix, but no such matrix appears to exist.

3.7.2 Groundwater abstracted by households for their domestic needs

We have already seen that about 15% of the islands households have no water supply connection to the JNWWC. They rely on their own wells and boreholes and are thereby exposed to the raw state of the aquifer and thereby to the diffuse contamination of groundwater by farming practice. There are believed to be approximately 5200 domestic premises dependent on groundwater sources. Robins and Smedley write: Under the UK Environment Agencys classication, all aquifers in Jersey would be classied as highly vulnerable because of the importance of fracturing (by-pass ow), lack of impermeable cover and the shallow depth to the water table (Robins & Smedley 1998: 17). To reduce their exposure to groundwater contaminants some households have invested in private treatment systems. The public health aspects of domestic borehole supply lies with the Department of Environmental Health. A sample of 52 boreholes and wells across the island are monitored in May and November of each year for inorganics, 16 of which are further analysed for pesticides and microbiological contamination.

3.7.3

Discharges by households of their own waste water

This is the second case of households unconnected to the islands dominant infrastructures. It is said that about 17% of households have no access to the PSDs sewerage. The special position of this group lies in the quality of its domestic waste water discharges. In fact, the PSD reports (2000: 1) Sewage pollution of water from malfunctioning private drainage systems is common. Robins & Smedley (1998: 13) suggest that the average daily discharge per household is about 600 litres. An analysis of the Val de la Mare catchments nitrogen export coefcients by Lott et al. (1999) showed that 10% of the nitrogen was sourced by domestic water from septic tanks and soakaways, which was as much as the contamination by livestock. The Water Pollution (Jersey) Law 2000 requires that owners of septic tanks, tight tanks, soakaways or private sewage treatment plants have a discharge permit only where such discharges result in or are likely to result in the contamination of controlled waters. Gass et al. (1996: 45) recommend that all soakaways and septic tanks should be taken out of use and sewerage extended to the whole island population. However, they do not say why effective private treatment by households should be eliminated, nor do they estimate the economic costs of extending the PSDs sewerage network.

3.7.4 The JNWWCs supply to its customers

In earlier sections of this paper it has been shown that the JNWWC supplies two-thirds of Jerseys total gross supply, excluding surface abstraction by farmers. The qualitative assessment of its treated water supply is detailed and appears in Table 3.4. The quality matrices published by the JNWWC (2002: 2830) cover altogether 33 characteristics of the treated water supply, two of which are for microbiological standards. In the case of the other three quality categories individual inorganics, individual organics and other measures four or ve examples from each are selected, including all those cases where one or more samples exceed the maximum admissible concentration (MAC). Table 3.4 is a baseline statement of quantities in respect of 16 separate qualitative characteristics of a hydrosocial ow. Just as the hydrosocial balance can be assigned a scenario balance, so too can a quality matrix be assigned a scenario matrix. The parallel between the hydrosocial balance and the quality matrix is strong here and

Table 3.4 A quality matrix for the JNWWC S treated water supply in Jersey in 2001. Concentration Maximum admissible concentration or value (MAC)2 50 milligrams NO3/litre 0.1 milligrams NO2/litre 400 milligrams Cl/litre 50 micrograms Mn/litre 50 micrograms Pb/litre 0.1 micrograms/litre 0.1 micrograms/litre 0.1 micrograms/litre 0.1 micrograms/litre 0.1 micrograms/litre 0.01 0.01 0.01 0.01 0.01 0.013 0.012 0.12 0.02 0.02 0.01 0.01 0.02 0.01 0.01 10 10 52 50 9 523 523 6.59.5 4 N.T.U. 20 Hazen units 1500 milligrams/litre 7.2 0.08 0.5 230 7.4 0.27 4.3 389 8.3 1.5 5.0 485 229 154 155 153 28.5 0.001 54 20 1 46.7 0.037 73 20 5 69.0 0.219 90 65 53 157 155 155 157 74 31 12 0 1 1 0 0 1 0 0 1 0 0 0 0 0 Minimum Mean Maximum Number of samples taken % of samples MAC

Individual Inorganics Nitrate Nitrite Chloride Manganese Lead Individual Organics Atrazine Simazine Cyanazine Mecoprop Dalapon Microbiological Standards Total coliforms1 Faecal coliforms1 Other Measures pH Turbidity (suspended solids) Colour Dissolved solids

Notes: 1 Zone 1 East. Random consumer taps and xed points. 2 These all appear to be EU values but the source does not make this clear. Source: JNWWC (2002): 2830.

shows that, in an IWRM context, both the balance and the matrix can be treated in parallel ways, and can have a baseline table calculated and a scenario table set as a target for water resource managers. The principal difference between balance and matrix is that a region has only a single baseline hydrosocial balance but it has a baseline quality matrix for every hydrosocial ow that the water management authority wishes to measure and to plan for. In this paper only one of the rows of Table 3.4 is examined, that for nitrate. Nitrates principal point of entry into Jerseys hydrological and hydrosocial ows has already been discussed above in Section 3.7.1. The downstream outcome is that in 2001 some 31% of the treated water supply to the islands population exceeded the European Unions MAC of 50 milligrams of NO3 per litre (mg/l). During the rst 9 months of 2002, there was no case of the MAC being exceeded. Here it should be pointed out that the measure of nitrogen present in water because of nitrate contamination may take two forms: 1. 2. The measured mass of the complete nitrate ion in the water, referred to as milligrams of NO3/litre. The mass of nitrogen in the water that is locked into the measured mass of the nitrate ion, referred to as milligrams of NO3-N/litre.

For a given sample, the ratio of (1) to (2) is 4.42 to 1. Therefore, it is absolutely essential that any document or discussion of nitrate contamination makes clear which of these two alternative measures is being used. Here the data and the European Union (EU) MAC are expressed in terms of (1) above. Why in the past did the government of Jersey permit nitrate contamination of the islands drinking water supplies? Politically, it was because government was not willing to introduce measures that would be contrary to the farming interest. Farmers, in fact, made up more than their proportionate share of Jerseys political class. This is now changing fast. Some of the measures introduced or about to be introduced, which promise an amelioration of the current contamination of groundwater, have already been referred to above. With specic reference to samples in excess of the nitrate MAC (see above), the formal defence is made on the basis of an argument that cannot lightly be dismissed. In the past, World Health Organization standards for drinking water were to recommend less than 50 mg/l and to regard as acceptable 50100 mg/l. But in 1996 the WHO fell into line with the EUs criterion, rst introduced in 1980, of 50 mg/l. The States of Jersey are not a member of the EU and are under no legal or constitutional obligation to comply with its water quality directives. Moreover, to do so in respect of 6.2 million cubic metres of water distributed per year (Table 3.3) would be costly. The JNWWC already desalinates water at some expense in order to blend it with surface water abstracted so that the mean concentration of NO3 drops below 50 mg/l (see Table 3.4). The health argument advanced by the States is based on the views of medical experts in the UK. The Joint Committee on the Medical Aspects of Water Quality appointed by the Department of the Environment and the Department of Health and Social Services stated that There is no compelling evidence to suggest that signicant risks to health are encountered when water containing between 50 and 100 mg/l nitrate is supplied to the public (Gass et al. 1996: 3536; JNWWC 2002: 29; Robins 2000: 1517). In Table 3.4 the MAC is also exceeded with respect to nitrite and, by a small margin, with manganese, lead, cyanazine, and total coliforms in the east of the island. It is noteworthy that the EU, on health evidence, plans to reduce the MAC for lead from

50 down to 25 micrograms per litre in December 2003, and further down to 10 micrograms by 2013 (N. Walton, pers. comm.). Particularly in the case of nitrate, nitrite, lead and the metabolite chlorthal, it seems clear that Jerseys water resource managers need to agree on a scenario quality matrix for the treated water supply. A recent consulting contract let by the States of Jersey under the title Water Quality Objectives may lead to just such a matrix.

3.7.5 The PSDs discharges to coastal waters

As has already been mentioned, waste water collection, treatment and discharge are the responsibility of the States of Jerseys Public Services Department. The volume of such discharges (including stormwater) in 2001 was some 12 Mcm. There are two treatment plants, one at Bellozanne in the south of the island and a second, small satellite plant recently commissioned on the north coast at Bonne Nuit Bay. Bellozanne consists of two inlet 3 mm step screens, fat/grit separation at the inlet works, four primary clariers, an activated sludge treatment plant, 12 nal clariers and a tertiary ultraviolet disinfection system. Discharges from the two plants are regulated within the terms and conditions of a discharge certicate that requires regular sampling to monitor compliance levels for the efuent leaving the works. The conditions in the discharge certicate for the Bellozanne sewage treatment works (STW) are based on the requirements of the EUs Urban Waste Water Treatment Directive (European Commission 1991) but prior to 2003 were less stringent than those set out in that Directive. The STW at the end of 2002 was undergoing conversion to a Pegazur system in order to reduce nitrogen inputs into the nearby St. Aubins Bay. In addition, the ultraviolet-applied dose will be raised from a minimum of 14 mW s/cm2 in 2002 up to 24 mW s/cm2 from July 2003. The current Jersey maximum admissible concentrations of waste water contaminants are given in Table 3.5 as
Table 3.5 A quality matrix for the PSDs waste water discharges to Jerseys coastal waters in 2002. Concentration Measure MAC in 2002 Biochemical milligrams 50 oxygen per litre demand Chemical oxygen demand Suspended solids Total nitrogen milligrams 250 per litre milligrams 150 per litre milligrams none per litre 2003 25 2004 25 Actual 2002 Minimum Mean 6.3 8.2 Maximum 13.4

125

125

21.0

46.0

64.0

45 20

35

11.5

18.5

29.8

10 or 15 not not not measured measured measured

Note: Actual 2002 data are based on monthly averages for JanuaryOctober. Source: PSD (pers. comm.).

well as the minimum, mean and maximum actual values achieved for the period JanuaryOctober 2002. From 2003, the PSD will add total nitrogen to biochemical oxygen demand, chemical oxygen demand and suspended solids as the criteria of the quality matrix for waste water and stormwater discharges.

3.8 CONCLUSIONS
With respect to the outstream ows of any dened region, whether or not it is a catchment, this paper has a central thesis: integrated water resources management is the planned transformation over time in the ow quantities of the regions hydrosocial balance alongside the purposive change of the quality matrix of each policy-priority ow. It is hoped that this paper indicates the usefulness of this approach to the management of outstream ows. What remains to be done is to draw some lessons from the Jersey casestudy that will benet planners who in the future follow this methodological path.

3.8.1 The hydrosocial balance

Jersey is a region in the language of this paper, i.e. a space with a dened boundary. The fact that it is composed of many small water catchments creates no problem. The year selected as the baseline is the most recent for which data are available. The scenario year does not appear in this case-study because, at present, the islands water resource planners do not use this method and so there is no scenario year. As a general rule, the scenario year should not be so close to the baseline year that it gives too little time for plan implementation nor so distant that the projection becomes an implausibility in the misty distance. A 57-year plan period is about right. The hydrological assumptions for the scenario year should reect the regions long-term average. Since these assumptions are likely to be different from the actual conditions in the baseline year, the changes between say 2001 and 2007 (column 4 of Table 3.1) will, in part, reect the difference. The row structure of the hydrosocial balance is always the same: the components of supply making up total gross supply, the three entries for losses and storage change, and the categories of use. The column structure of base year, scenario year, absolute difference and rate of change are also a standard feature. With respect to supply, the Jersey data of Table 3.3 are structured principally by the actors responsible. This has advantages in policy formulation. The main difculty is the unavailability of 2001 data on groundwater abstraction by private persons and surface water abstraction by farmers. This leads to a policy recommendation: that the islands authorities should begin to estimate these data on a regular basis, either by sample metering or by an interview-based survey. Constructing the hydrosocial balance for a region always leads to proposals for additional data collection. The same argument holds true for rainwater harvesting and for reuse, sources often neglected that may deserve policy initiatives. If these categories are not entered into the hydrosocial balance because the values are not known, they are invisible to the planner. Turning to exports, losses and storage change, the Jersey story appears simple. There are no exports. Storage change is negligible (see Section 3.5). Loss estimation and reduction are a priority for the main supplier, the JNWWC, but no loss estimate exists for private abstraction; the water utilitys 6% datum is used in this case.

With respect to use, the data are available separately for the two dominant sources of supply, i.e. the JNWWC and private actors. From both an analytic and a policy point of view this has its advantages. Unfortunately, the classication of types of use differs between the sources, creating an adding-up problem. The researchers choice of supply and use categories should always reect the particularities of the region in which the work takes place as well as its planning focus. It was surprising to discover that private services has the second largest recorded use after households. This is explained by the vital role of tourism and nancial services in the island as well as by the fact that (outside the polytunnel and glasshouse sector) irrigation is supplementary to the islands (877 mm average) rainfall. The main weakness of the use data is that for private actors it is an estimate based on a sample last made in 198991. The lack of registration, licensing and metering of private groundwater abstraction in Jersey deserves to be reviewed by the States government. In fact the Water Resources (Jersey) Law to control abstraction and impoundment will be presented to the States in 2003. The nal comment on Table 3.3 is that an error term of 0.1 Mcm was added to total use to give equality with the total net supply gure of 10.2 Mcm. Properly dened and accurately and comprehensively measured, total net supply is mathematically identical to total use. In practice, any real-life calculation will always contain a disparity as water measurements are never exact. Unfortunately, with respect to Table 3.3, one cannot conclude that the error is only of the order of 1% (0.1/10.2). This is because the table is sure to contain mutually-cancelling errors.

3.8.2 Quality matrices

The principal methodological objective of this paper is to integrate quality with quantity, that is, to bring together quality matrices with the hydrosocial balance. In deploying matrices in the eld the biggest challenge is to specify which of the potentially innumerable ows should be measured, given the high overhead and prime costs of quality assessment. Key guidelines to such a choice are:

The importance of any specic ows quality in understanding how other downstream ows are polluted. The impact of any specic ows quality on the environmental health of the regions population. The signicance of a specic ows quality on the natural environment. The requirements of the governing legislation, standards and guidelines for that ow.
Using Tables 3.2 and 3.3, just ve ows were selected for discussion in the Jersey case-study. The rst is irrigation returns with its recognized wide externalities. Here it is immediately clear that the hydrosocial ow has to be combined with farmland drainage sourced by rainfall; the two are not separable. Much is available on the sources of contamination of farm drainage and the relative importance of such contamination island-wide. However, there appears to be no up-to-date quality matrix for this specic ow. Jerseys planners may wish to take action here, particularly in the light of the EUs nitrate-sensitive areas designations. The second selected ow is groundwater abstraction by households for their domestic needs, because they are unconnected to the JNWWC s water supply network. These families are exposed to the aquifer in its polluted condition. This ow is monitored twice per year for its quality. The third ow is households discharge of their own waste water via septic tanks and soakaways and concerns households unconnected to the PSDs sewerage network.

They may thereby pollute groundwater, for example with ammonia, detergent residues and faecal coli. The fourth and fth ows are the JNWWC s water supply and the PSDs discharge of waste water and stormwater to the sea. These are the dominant hydrosocial ows of the island and are discussed at greater length in this paper. In both cases the quality matrices are detailed and informative. Currently, the main policy issues are the nitrate content of the water utilitys output and the move towards more stringent maximum admissible concentrations for the sewage treatment works at Bellozanne. With these comments on the practical construction of the hydrosocial balance and the quality matrices of some of its ows, this paper is now complete. It is hoped that the research may be of practical use to the people and institutions of Jersey as well as suggesting, to professional colleagues, new ways of integrating water resources management in respect of the relationship between the quantity and quality of outstream, hydrosocial ows.

The material in this chapter originally appeared in: Integrated water resources management and the hydrosocial balance. Water International, 29(2): 148157, 2004.

4 Sharing the benefits of the river basins water economy

4.1

INTRODUCTION

Water is essential to all forms of life as well as to all the activities of human society. But if we limit ourselves to the part it plays within a specified economy, we can distinguish just three broad sectors. First, there are the rainfed areas of farming, forestry and pastoralism. Secondly, there is the instream sector with its navigation, fishing and hydroelectric power production, as well as its conservation, recreation and tourism services. Thirdly, we have the outstream sector where water meets the needs not only of households but also those of agriculture, mining, manufacturing, construction, as well as public and private services. In this paper the term the water economy is used to refer to a river basins economy its production of goods and services from the point of view of the dependence of output on water flows in the rainfed, instream and outstream sectors.

4.2

BENEFITS OF THE WATER ECONOMY

In the rainfed sector, the benefits of the water economy can be estimated as the total output value of rainfed farms, forestry and pastoralism (see Table 4.1). If it is useful, we can deduct from total output value the measured costs of production, giving the value of net output. Of course, it is not suggested here that water is the only input to production, just that it is a necessary condition of production like land and human labour.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

37

38

The Price of Water Table 4.1 Catchments rainfed water economy

Categories of production Rainfed farming Rainfed forestry Pastoralism Total production Total output ()

Table 4.2 A catchments instream water economy

Categories of production River navigation River fishing Hydroelectric power production River conservation, recreation and tourism services Total production Total output ()

The outputs of the instream sector can be similarly estimated in terms of the total output value of river navigation, the basins fishing industry, hydroelectric power production and the flow of services from riverine conservation, recreation and tourism (see Table 4.2). As with the rainfed sector, the value of net output can also be estimated. Turning to the outstream sector we can use the hydrosocial balance. This is a quantitative water resources planning tool applicable in principle to any space with a defined boundary. This might be a house on the coast of Andaluc a, Wembley football stadium, the island of Gorgona, the Thames Estuary, the catchment of the Yangtse, the State of California, or the continent of Africa. Such spaces are referred to here generically as areas or regions.In this specific case, where we consider the water economy of a river basin, the area or region is defined as the basin itself. The distinction of the hydrosocial balance from the water balance familiar to hydrologists is that the former incorporates only outstream, hydrosocial flows, and never the hydrological flows of precipitation, groundwater recharge and run-off. The water flows that it places at the centre of analysis and measurement are those directly created by human society; the idea of a hydrosocial balance is derived from the concept of the hydrosocial cycle (Merrett 1997: 6-7). A hydrosocial flow represents a human activity. So the hydrosocial balance, composed as it is of many hydrosocial flows, is understood primarily through the social sciences. In contrast, the hydrological balance represents natural flows and is understood primarily through the natural sciences. To summarize, hydrological flows are of a type that exist in a state of nature. Hydrosocial flows, in contrast, are specific to human society. In recent centuries these two types of flow have become ever more powerfully interdependent with the growth of world population and its economic activities of production and use. The hydrological balance of an area and that areas hydrosocial balance should be estimated separately prior to considering their quantitative and qualitative interdependence. The generic form of the hydrosocial balance for a specified catchment is given in Table 4.3. A baseline balance is for a past time-period, such as the year 2004. A scenario

Sharing the benefits of the river basins water economy Table 4.3 A catchments hydrosocial balance
Base year volumes Categories of supply (positive) Rainwater collection Groundwater abstraction Surface water abstraction Desalination of sea water Imports of water from other areas Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored Total supply Categories of supply (negative) Supply-side evapotranspiration losses Supply-side leakage Exports of water to other areas Net rise in water abstracted and stored Total negative values (supply) Total net supply Categories of use (positive) Households Agriculture (including irrigation requirements) Mining Manufacturing Construction Public services Private services Other uses Total use Categories of use (negative) Evaporation losses on users properties Leakage on users properties Total negative values (use) Total net use a b c d e f g h i j k l m n i-n o p q r s t u v w x y z w-z

39

balance is for a future time-period such as the year 2010. The shift in the quantity in millions of cubic metres of any one category of supply or use between the baseline year and the scenario year can be represented either as an absolute change or as an annual rate of growth (see Table 4.4). The baseline balance provides a comprehensive, synoptic account both of the scale and composition of the supply sources of water as well as their use in the region it covers. Where measurement is accurate and comprehensive, total net supply is always equal to total use. Scenario balances provide options for the future, based on the forecast need for outstream water in different uses and the possible allocation conflicts that may be foreseen. Once again, total net supply must be planned to equal total use. The absolute difference of supply, and of use, between the base year and any

40

The Price of Water

Table 4.4 The hydrosocial balance for a specified region in a base year and a scenario year
In millions of cubic metres

Base year
Categories of supply Rainwater collection Groundwater abstraction Surface water abstraction Desalination Import of water from other regions Internal reuse of wastewater External reuse of wastewater Total gross supply Supply leakage and evaporation Export of water to other regions Fall (+) or rise () in volume of water abstracted and stored Total net supply Categories of use Households Agriculture Mining Manufacturing Public services Private services Other uses Total use A1 B1 C1 D1 E1 F1 G1 H1 J1 K1 +/L1

Annual compound rate of growth Scenario year from the base year Scenario minus base to the scenario year year year (+ or ) (%)(+ or )
A2 B2 C2 D2 E2 F2 G2 H2 J2 K2 +/L2 A2A1 B2B1 C2C1 D2D1 E2E1 F2F1 G2G1 H2H1 (J2)(J1) (K2)(K1) (+/L2) (+/L1) M2-M1 S2S1 T2T1 U2U1 V2V1 W2W1 X2X1 Y2Y1 Z2Z1 Ga Gb Gc Gd Ge Gf Gg Gh Gj Gk

M1 S1 T1 U1 V1 W1 X1 Y1 Z1

M2 S2 T2 U2 V2 W2 X2 Y2 Z2

Gm Gs Gt Gu Gv Gw Gx Gy Gz

Note: Gj and Gk are calculated using absolute values of leakage and exports. Gl is not calculated because of the possible change of sign.

specific scenario year, together with the associated annual rate of change, provide the basic input to the planning of infrastructural investment, capital financing and demand management. We have already seen in Tables 4.1 and 4.2 that the value of total output in the rainfed sector and the instream sector can be calculated by sub-sector. Of course, we can do the same thing for the use categories of the hydrosocial balance (see Table 4.5). In this case we have not only output by type of use, such as the agriculture and mining sectors, we also have output per cubic metre of water used. There is one characteristic of the hydrosocial balance that should be made explicit in estimating the water economy of a river basin. If our interest in sharing the waters

Sharing the benefits of the river basins water economy Table 4.5 A catchments outstream water economy
Categories of use Household production Agriculture Mining Manufacturing Construction Public services Private services Other uses Total use in production Total output () ho1 ag1 mi1 ma1 co1 pu1 pr1 ot1 tup1 Base year use volume (m3) ho2 ag2 mi2 ma2 co2 pu2 pr2 ot2 tup2

41

Total output/ volume of use (/m3) ho3 qg3 mi3 ma3 co3 pu3 pr3 ot3 tup3

of the basin is the driver of the analysis, then the water flows that generate output but that are not catchment flows should be excluded from the output calculations of Table 4.4. In particular, output flows should be excluded where these are sourced by i) groundwater abstracted outside the catchment area, ii) the desalination of seawater, and iii) imports of water from other catchments.

4.3

BASIN WATER PRODUCTIVITY

The water flows that are discussed above are all dependent on the basins total precipitation. This is perfectly obvious in the case of the rainfed sector. With respect to the instream sector, this is fed by rainfall and snow-melt. Finally, outstream flows are themselves fed by capture of the instream flows. A statistic of great interest for the quantitative flows of the river and its tributaries is the ratio of total basin output to the evapotranspiration (ET) losses of the outstream sector. This is especially critical in the case where the irrigation sector is a large water user. This measure of basin-level water productivity can be expressed as outstream output value per cubic metre of the ET loss. Merrett has calculated the value for the Thames River Basin in a study of catchment water deficits. In this case, water productivity was 1730/cubic metre (see chapter 19 of this book).

4.4

SHARING THE BENEFITS

To sum up the argument so far we can say that the catchments water economy is defined as the production of goods and services within a river basin from the point of view of the dependence of output on water flows in the rainfed, instream and outstream sectors. Moreover these three types of output flow are measurable. We now have a clear understanding of the river systems economic benefits. The flow of water begets the flow of economic output. But how are these economic benefits shared, for example, between the upstream and downstream regions? In principle, this can be estimated in a quite straightforward way. We begin by measuring separately the rainfed output of the upstream and the downstream area. Then we measure separately the instream output of the river and its tributaries in the upstream and the downstream area. Finally we estimate separately the hydrosocial balance of each area and the associated output flows of each outstream

42

The Price of Water

water economy. Note that, where appropriate, the upstream-downstream categories can be defined as the freshwater river basin and the rivers estuary. This is regarded as important in the case of the River Thames, for example. At the end of this river basin survey of water flows and their dependent economic production flows, it will be possible to set out separately, for the upstream area and for the downstream area, the economic benefits of its water economy in terms of the rainfed, instream and outstream sectors output values. Where conflicts of interest exist between upstream and downstream neighbours, the political process of conflict resolution will be informed and facilitated by a clear, base-year analysis of each areaswater economy. At the same time, a potential agreement between the upstream and downstream neighbours with respect to changes in the flow regime of the river can be cross-checked for a future year on the likely consequential changes in output values in each area between the base year and the scenario year.

4.5

CONCLUSIONS: NEGOTIATING THE BENEFITS

The water economys benefits are shared in many different ways within a single catchment. Tables 4.1, 4.2 and 4.5 record what the benefits are in the base year in the rainfed sector, the instream sector and the outstream sector. It has been shown too that the outputs of these three sectors of the water economy can be broken down in terms of upstream and downstream partners. Alternatively, they can be shown for the separate riparians of the Left Bank and the Right Bank, if that is the issue. Now, a shift over time in benefit values can be estimated for any well-defined water resource management innovation. Alternative innovations will produce different shifts in total benefit and in its three-sector and upstream-downstream spatial distribution. Agreement between spatial groups on which one of a number of innovation alternatives should be selected is clearly a political process. In an ideal world negotiators would take a basin-wide perspective on the best choice, the optimum, if this can be identified. Whittington, Wu and Sadoff (2005), for example, have recently presented the results of the first economic model designed to optimize the water resources of the entire Nile Basin. Heterodox economists would suggest a satisficing approach rather than an optimizing one. If it is feasible, the alternative selected should bring advantages to both areas and be seen by negotiators as reasonably fair. However, in the case where a water resource management innovation does not bring broadly equal advantages to both areas, the area that gains most should financially compensate the area that gains least. These financial transfers can be based on the water economy estimates of Tables 4.1, 4.2 and 4.5. Of considerable interest would be to see how the optimizing model of Whittington and his colleagues could be integrated with the water economy approach of this paper.

5 Farm-level drought management: an irrigation case-study from the UK

There is special providence in the fall of a sparrow. If it be now, tis not to come; if it be not to come, it will be now; if it be not now, yet it will come. The readiness is all. William Shakespeare, Hamlet

5.1 INTRODUCTION
The farmers world is encompassed by risk and no risk is more devastating than a prolonged drought. Yet, just as risk can be managed, drought can be planned for. The purpose of this paper is to use a case-study from the Anglian Region in England to show the preparations being taken on a specic farm for the regions next severe drought, and how risk analysis permitted an evaluation of two alternative, farm-level drought management strategies. The broader context of the case-study is the introduction of a Water Bill into the UK Parliament in February 2003. The Bill, once enacted, promises a more effective water resources planning system and specically creates an obligation on water companies to develop drought plans for submission to the Environment Agency (House of Lords 2003 clauses 602). This Agency is the statutory body in England and Wales for strategic water resources management. A key pressure to which it seeks to respond is climate change. As van Hofwegen and Svendsen (2000: 910) write: The gradual warming of the earth, 1.0C in the past 50 years, is leading to glacial recession, declining snow cover, and rising sea levels. Precipitation patterns

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

43

are likely to alter, reducing water availability in some regions and increasing it in others. Increased variability in precipitation patterns will accompany this shift with a huge impact on both irrigated and non-irrigated agriculture. Precipitation patterns will include a greater proportion of extreme events, leading to higher and more frequent ooding and lower dry season ows in rivers. Within this broader context referred to above, it is important in the UK to understand how farmers themselves will manage drought risk in a water companys drought plan area. The structure of the paper is straightforward. It begins with a brief description of the Anglian Region itself, followed by a general overview of the farms activities. Thereafter an infrastructural strategy for drought years is compared with an informational strategy based on long-term rainfall data; the alternatives are evaluated in terms of relative cost. The paper ends with a brief summary and conclusions.

5.2

THE ANGLIAN REGION

The Environment Agency is divided into eight regions, one of which is Anglian. This region stretches from the Humber Estuary in the north to the River Thames in the south, from the Norfolk coast in the east to Northampton in the west. It covers 27,000 km2 and ve million people live in the area. The region has extensive, sparsely populated rural areas, particularly in the north and east, a long coastline and interspersed urban centres around which industry has developed. As communication links continue to improve (especially with London) it is likely to be one of the highest population growth areas in the country. The Anglian Region has 58% of the most productive agricultural land in England and Wales and agriculture has considerable inuence on the rural economy and communities. The sustainable use of water is crucial given the competing demands for it, particularly because this region is the driest in the UK. Low ows in late summer and increased demand especially for agriculture and garden use are likely, say the Environment Agency, to cause stress (www.environment-agency.gov.uk/) In some summers, irrigation can make up 50% of total use. The second principal actor in water resources management in the region is Anglian Water Services Limited (AWS), which serves the needs of over ve million industrial, commercial and domestic consumers (www.anglianwater.co.uk/). It is this company that will have to submit its drought plan to the Environment Agency once the Water Bill is enacted by the UK Parliament. In fact already in the spring of 2003 the company had submitted an Anglian Water Services Drought Plan March 2003 to the Environment Agency. AWS is the regional utility and its principal activities are the provision of drinking water and the treatment of waste water. The company employs 3700 full-time equivalent persons and in 2002 had a turnover of 724 million (AWS 2002). Its principal infrastructural assets are mains and sewers, impounding and pumped raw water storage reservoirs, dams, sludge pipelines and sea outfalls. The companys capital investment plan in the 5 years through 2004/5 is equal to 1.5 billion and covers further improvements to drinking water, bathing and river water quality, as well as network improvements to maintain serviceability and meet new demands.

5.3

SILVER BIRCHES plc

In this section are reviewed the activities of the agribusiness at the centre of this casestudy, an enterprise referred to here as Silver Birches plc. The farm is located in the Ely area of the Anglian Region on soils typical of the district around Cambridge (Hodge & Seale 1966). Silver Birches plc is 57 hectares in size. The product is container trees and the business is the largest such nursery in Europe. Maximum capacity in terms of the stock at any point in time is 130,000 trees ranging from 810 cm girth in 45 litre pots through to 2530 cm girth in 300 litre pots. The company sells trees of about 330 different varieties in its unique, easy-to-handle, white containers. The amenity market is its customer and this includes universities, hospitals, retail, residential and ofce developments, town centres, golf clubs and football grounds. In order to understand the farms access to and use of irrigation water, the planning concept of the hydrosocial balance is set out in Table 5.1. This requires considerable adaptation for the specic case of Silver Birches plc. Beginning with the positive cate-

Table 5.1 The hydrosocial balance for a specied area in a base year. Base year volumes (mcm) Categories of Supply (Positive) Rainwater collection Groundwater abstraction Surface water abstraction Desalination Imports of water from other areas Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored Total Gross Supply Categories of Supply (Negative) Supply leakage and evaporation Exports of water to other regions Net rise in water abstracted and stored Total Negative Values Total Net Supply Categories of Use Households Agriculture Mining Manufacturing Public services Private services Other uses Total Use a b c d e f g h j k l m n jn s t u v w x y z

Note. A net fall (rise) in the storage of abstracted water has a positive (negative) value. Source: Adapted from Merrett (2002a) Tables 7.1 and 7.2.

Table 5.2 The hydrosocial balance for Silver Birches plc in 2002. Row 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Categories of Supply (Positive) Rainwater collection Groundwater abstraction Surface water abstraction Desalination Mains water Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored Total Gross Supply Categories of Supply (Negative) Supply leakage and evaporation Exports of water to other regions Net rise in water abstracted and stored Total Negative Values Total Net Supply Categories of Use Buildings Irrigation Total Use Quantity (cubic metres) 0 0 0 0 39,315 0 0 Negligible 39,315 400 0 Negligible 400 38,915 350 38,565 38,915

gories of supply, Table 5.2 shows that in the year 2002 there was no rainwater collection, no groundwater or surface water abstraction, no desalination and no reuse of waste water. Moreover, with respect to Table 5.1s categories of net fall or rise in storage, at Silver Birches plc on-farm storage of purchased mains water is sufcient only for 48 hours use. This is necessary in case of a brief failure in the mains supply, but any change in storage over the course of a whole year is negligibly small. Two days storage is only 215 m3. The absolutely dominant supply source is mains water purchases from Anglian Water Services. AWS has a 12-inch main running close to the farm and off this a link is made to a 3-inch and a 6-inch pipe; the latter runs like a spinal cord down the centre of the farm. With respect to the negative categories of supply, on-farm losses of mains water are low: 1% is the current estimate and this is entered in row 17 of Table 5.2. Losses are monitored using the 3-inch meter at the connection to the utilitys main, as well as the meters and pressure gauges inside each of the farms four pump-houses. In most cases on-farm supply pipes are electrowelded and Silver Birches plc has two skilled industrial plumbers on its staff who are on-call on a 24/7 basis. There are no exports of water. As a result, total net supply was about 39,000 m3. In principle, total net supply is mathematically identical to total use (Merrett 2002: 150). As Table 5.2 shows, use is simply divided into buildings and irrigation. Farm management has not developed a split of use into different species of tree or different locations of tree or different container sizes. Such a division of use is not (at present) considered to have any value in the management process. Ninety-eight per cent of containers at the farm are divided roughly equally between the 45- and 85-litre sizes. The enterprise deploys a Netam pressure compensating non-leakage (CNL)

Table 5.3 Silver Birches plc: monthly Anglian Water Services supply 2002 (cubic metres). Irrigation Row 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Month billed January February March April May June July August September October November December (estimate) Total Average Central Services January December Grand Total Volume supplied 22 35 248 3426 4328 8449 6127 5379 7771 2975 132 77 38,969 Number of metered days 35 29 34 28 28 35 28 28 36 27 28 30 366 106 346 39,315 3247 Volume per metered day 1 1 7 122 155 241 219 192 216 110 5 3 Volume for adjusted month 19 37 222 3720 4699 7339 6652 5840 6562 3350 143 78

Note: The number of metered days in column 3 is the number of days that the end-of-month invoice covered. Volume for the adjusted month applies the volume per day of column 4 to a standard month of 30.4 days.

drip system to each separate tree where each container has one manifold, one CNL dripper, microtubes and four arrow-stakes. Averaged over the whole stock, each tree received 1.12 litres per day in the farming year September 2001 to August 2002, when the mid-year stock equalled 94,000 trees, equal to about 2 years sales. The variation in the rate of irrigation over the calendar year 2002 is given in Table 5.3. The volume of water delivered to the trees in the 5 months MaySeptember is 80% of the total. An interesting statistic can be derived from these data, once we realize that the average surface area of the two dominant container-types is 0.159 m2. At 1.12 litres/day, the average container consumes 2600 mm/year of irrigation water.

5.4

DROUGHT MANAGEMENT: THE INFRASTRUCTURAL STRATEGY

Silver Birches plc has grown rapidly in the past 14 years to attain its European preeminence as a supplier of container trees. But by the turn of the millennium the managing director (MD) had become increasingly concerned about the impact a severe and

Table 5.4
September 0 0 0 0 12,000 12,000 12,000 October November December January February March

Supply planning at Silver Birches plc with an 18,000 m3 safeguard against drought. April 0 May 0 June 0 July 0 August 0

Year

Stock and ow (m3)

Year 1 20034 0 0 0 0 0 40

230

3720

4700

7340

1970

6570

3350

150

80

20

4690

5840

6570 0 0 0 0 0 0 0 0

3350

150

80

20 12,000 0

40 18,000 5960

230 18,000 17,730

3720

4700

7340

6660 18,000 18,000 18,000 18,000 14,010 9310 1970 0

5840 18,000 0

Abstraction to reservoir by end of month Irrigation supply from non-drought storage Irrigation supply from Anglian Water Total irrigation supply Drought storage by end of month Non-drought storage by end of month 0 0 0 0 12,000 23,960

Total storage

35,730

32,010 27,310 19,970 18,000

18,000

Year 2 20045 0 0 150 80 20 40 230 3720 4700 7340 2240 0

7200

7200

3850

40

230

6570

3350

4420

5840

6570 18,000 0 0 7050 14,170 18,000 18,000 18,000 14,280 18,000 18,000 18,000 18,000 18,000 18,000

3350

150

80

20

40

230

3720

4700

7340

6660

5840 18,000

18,000 18,000 18,000 18,000 9580 2240 0

Abstraction to reservoir by end of month Irrigation supply from non-drought storage Irrigation supply from Anglian Water Total irrigation supply Drought storage by end of month Non-drought storage by end of month 0 18,000 18,000 25,050 32,170 36,000 36,000 36,000 32,280 27,580 20,240 18,000

Total storage

18,000

Year 3 As for 20045 20056 to 202223

extended drought might have on the business. A young tree in a 45- or 85-litre container rapidly becomes stressed without irrigation, so much so that in the absence of water for 15 days it is irreparably damaged. MDs concern is not directly linked to a regional drought because, as we have seen, at Silver Birches plc it is mains water that sources the soil moisture required by the growing trees. However, the risk exists that a severe drought would lead AWS to cut its supplies to the farm. In a conict over access to scarce water, the order of sectors in the UK in terms of relative political strength is probably households, public services, industry, private services and, nally, the agricultural sector. MDs judgement is that the National Farmers Union in the Anglian Region does not have the muscle adequately to defend the agricultural sector from mains supply reductions. In these circumstances MD realized that conjunctive supply could play a role in farm-level drought management. The Anglian Region is famously at and land drainage is a matter of great importance (Merrett 2002a: 10512, Talbot & Whiteman 1996). As a result the landscape is criss-crossed with innumerable unlined, surface drains. One such drain makes up the farms northern boundary. MD secured an abstraction licence for this drain from the Environment Agency. It gives Silver Birches plc the right to abstract during the ve winter months from 1 November to 31 March. The duration of the right is from May 2002 through March 2027. The licensed means of abstraction is a centrifugal pump of a maximum output not exceeding 29 litres/second. The maximum quantity permitted to be abstracted is 100 m3/hour, 1000 m3/day and 36,000 m3 during the ve winter months. Low ow provisions exist to protect the environment of the downstream South Level system. These are that the abstraction right is suspended when the combined ow of the Ely Ouse river and the cut-off channel at the Denver Sluices does not exceed 318,226 m3/day (1 November28 February) or 113,652 m3/day (1 March31 March). The abstraction ow will go straight into a winter storage reservoir on the farm. The contract for this has now been signed and the reservoir should be ready to receive the off-take water from January 2004. It will have a storage capacity of 36,000 m3. The new infrastructure will combine the reservoir, pump station and sand ltration. Table 5.4 models the relative role of the winter storage supply and the mains supply in meeting the businesss crop water requirements over the 20 farming years September 2003 through August 2023. This is here called the infrastructural strategy. Key assumptions in developing the model are:

Monthly crop water requirements are equal to those of the adjusted monthly volume from Table 5.3. Abstraction during the winter months is spread equally over the months when it
is permitted (in 20034 it begins only in January when the reservoir has been completed). Abstraction to the reservoir does not take place when the reservoir is full. Half the reservoirs capacity (18,000 m3) is held back as an irrigation source should there be a cut-off in the mains supply. This is drought storage. Irrigation supply from the reservoir begins only after the reservoir has reached its drought storage level. Irrigation supply from AWS is used only when there is no non-drought storage in the reservoir. The nal assumption is that, in fact, there turns out to be no drought so severe over the twenty years that the drought storage is required.

5.5

DROUGHT MANAGEMENT: THE INFORMATIONAL STRATEGY

There are two problems with the infrastructural strategy. First, it permanently dedicates half the capacity of the winter storage reservoir to drought storage when the severe and prolonged drought event that it protects against is extremely rare. Second, it reduces the irrigation supply that can be sourced from winter storage and therefore increases the irrigation supply sourced from AWS. The additional prime cost (marginal cost) of the AWS supply is 0.59/m3 whereas from winter storage it is only 0.15/m3. In the UK farmers pay only 0.15 per litre of diesel fuel. Table 5.5 models the informational strategy. Its key assumptions are:

Monthly crop water requirements are equal to those of the adjusted monthly volume from Table 5.3. Abstraction during the winter months is spread equally over the months when it
is permitted (in 20034 it begins only in January when the reservoir has been completed). Abstraction to the reservoir does not take place when the reservoir is full. There is no provision in this model for drought storage. Irrigation supply from AWS is used only when there is no storage in the reservoir. The nal assumption is that, in fact, there turns out to be no drought so severe over the 20 years that the drought storage is required.

No delivery contracts or other legal mechanisms invalidate the key assumptions referred to above. The central difference between Tables 5.4 and 5.5 is of course that in Table 5.5 the whole of winter storage is applied for irrigation, thereby reducing purchases from AWS. Table 5.6 shows these volumetric differences and their cost implications. The total excess of water purchases from AWS with the infrastructural strategy in comparison with the informational strategy is 347,000 m3 over the 20 years 2003/42022/23. In terms of expenditure this makes the infrastructural strategy more expensive by a sum of 153,000. Discounting the stream of differential expenditure at the interest rate charged on Silver Birches plcs overdraft base rate plus 2% i.e. 5.75% we have a discounted sum of 88,000. In nancial terms there is no doubt of the relative attraction of the informational strategy. However, the reader will surely have noted a drawback of the informational strategy: thus far it has no provision for drought planning! In the remainder of this section it is shown how such planning is possible and why this leads me to call the second approach the informational strategy. We have already seen that low rainfall currently provides no direct threat to Silver Birches plcs farming practice. Virtually all of its crop water requirements are met from the mains supply. The drought risk takes the form of a possible reduction in the AWS supply. The rms MD believes that this would most likely be during a dry summer that succeeded a dry winter, each of great severity. So it made sense to review rainfall records for the area, kindly provided by the Environment Agency for the Isleham Pumping Station. The calendar year data show that the driest year since 1963 was in 1996 with total rainfall equal to 391 mm compared to the 19632002 average of 548 mm. Yet in 1996 AWS imposed no restrictions of any kind on the mains supply to its customers.

Table 5.5
September October November 0 0 0 0 12,000 12,000 12,000 0 0 December January February March April May June

Supply planning at Silver Birches plc with no xed safeguard against drought.
July 0 0 August 0

Year

Stock and ow (m3)

Year 1 20034 0 0 0 0 20 40 230 3720

4700

7340

6660 5840

6570 6570 0 0 0 0 0 7200 7200 0 0 0 11,980 7200 0 0 0 11,980 0 0 0 0 0 23,940 23,940 7200 3350 150 80 20 40

3350

150

80

0 230 0 35,710 35,710 7200

0 3720 0 31,990 31,990 0

0 4700 0

0 7340 0 27,290 19,950 27,290 19,950 0 0

0 6660 5840 0 0 13,290 7450 13,290 7450 0 0

Abstraction to reservoir by end of month Irrigation supply from non-drought storage Irrigation supply from Anglian Water Total irrigation supply Drought storage by end of month Non-drought storage by end of month Total storage

Year 2 20045 6570 880 150 80

20

40

230

3720

4700

7340

6660 5840

Abstraction to reservoir by end of month Irrigation supply from non-drought storage Irrigation supply from Anglian Water 0 2470 0 0

6570 0 880 880 0 7200 7200 7200 7200 7200 0 0 0 7050 14,170 21,350 28,510 35,480 31,760 27,060 19,720 0 0 7050 14,170 21,350 28,510 35,480 31,760 27,060 19,720 0 0 0 0 0 0 0 0 0 0 0

3350

150

80

20

40

230

3720

4700

7340

6660 5840

Total irrigation supply Drought storage by end of month Non-drought storage by end of month 13,060 7220 13,060 7220 0 0

Total storage

Year 3 20056 650 150 80 20 40 230 3720 4700

7340

6660 5840

2700 3350 0 0 0 7050 14,170 21,350 7050 14,170 21,350 28,510 28,510 0 0 0 0 150 80 20 40 230 0 35,480 35,480

0 3720 0 31,760 31,760

0 4700 0

0 7340 0 27,060 19,720 27,060 19,720

0 6660 5840 0 0 13,060 7220 13,060 7220

Abstraction to 0 reservoir by end of month Irrigation supply 6570 from non-drought storage Irrigation supply 0 from Anglian Water Total irrigation 6570 supply Drought storage by 0 end of month Non-drought storage 650 by end of month 650

Total storage

Year 4 As for 20056 20067 to 202223

Table 5.6 Purchases of irrigation water from Anglian Water Services under two alternative strategies.
September October November December 6570 6570 0.44 3350 3350 0.44 150 150 0.44 80 80 0.44 20 0 0.44 0 0 0.44 0 0 0.44 0 0 0.44 0 0 0.44 0 0 0.44 January February March April May June July 4690 0 0.44 August 5840 0 0.44

Year

Irrigation water purchases

Year 1 20034

0 6570 0 0.44 3350 2470 0.44 0 0 0.44 0 0 0.44 0 0 0.44 0 0 0.44 0 0 0.44

0 0 0 0.44

0 0 0 0.44

0 0 0 0.44

2064 4420 0 0.44

2570 5840 0 0.44

Year 2 20045

2891 6570 0 0.44 3350 2700 0.44 0 0 0.44 0 0 0.44 0 0 0.44

387

0 0 0 0.44

0 0 0 0.44

0 0 0 0.44

0 0 0 0.44

0 0 0 0.44

1945 2570 4420 0 0.44 5840 0 0.44

Year 3 20056

Year 4 20067 to 202223

Infrastructure strategy (m3) Information strategy (m3) Marginal cost excess of AWS supply over winter storage supply (/m3) Cost savings of information strategy () Infrastructure strategy (m3) Information strategy (m3) Marginal cost excess of AWS supply over winter storage supply (/m3) Cost savings of information strategy () Infrastructure strategy (m3) Information strategy (m3) Marginal cost excess of AWS supply over winter storage supply (/m3) Cost savings of information strategy () As for 20056 2891 286 0 0 0 0 0 0 0 0

1945 2570

Table 5.7

Silver Birches plc: drought warning chart 20034. September October November December January February March April May June July August 319 351 351 351 351 125 319 319 319 114

Row

tbi 391 430 tbi 391 430 tbi 391 430 tbi 391 430 tbi 391 430 tbi

tbi

tbi

tbi

tbi

tbi 391 430 tbi

tbi 391 430 tbi

tbi 391 430 tbi

tbi 391 430 tbi

tbi 391 430 tbi

tbi

tbi 391 391 430 430 tbi tbi

4 5 6 7 8 9 10 11 12 13 14

Required minimum combined ow Ely Ouse and Denver Sluices (000 m3) Required minimum combined ow Ely Ouse and Denver Sluices 10% (000 m3) Actual combined ow Ely Ouse plus Denver Sluices (000 m3) 1996 rainfall (mm) 1996 rainfall 10% (mm) Actual rainfall during last 12 months inclusive of current month (mm)

Note: rainfall data should be for the whole of the Anglian Region. tbi: to be inserted.

Next a drought warning chart was prepared for Silver Birches plc and this is set out in Table 5.7. In respect of pumping to the winter storage reservoir, rows 5 and 6 are the most important. If the actual values of row 6 prior to the pumping season (NovemberMarch) fall short of 351,000 m3 (10% higher than the minimum required for pumping to take place), management action is necessary. The consultant recommended that the rate of pumping to storage should be as early and as high as permitted and practical from 1 November onwards in order to secure the 36,000 m3 quantity permitted under the licence. If on 1 November the actual combined ow is below the Environment Agencys required minimum of 319,000 m3, then the farm would have to await the time when pumping is permitted. In the meantime in any months between November and March when the Agency prohibits abstraction to the reservoir, the mains supply must carry the maximum burden of 520 m3 irrigation over those 5 months (see Table 5.5 for the year 20056). In respect of rainfall, rows 11 and 12 are the most important. Row 12 (like row 6) is intended as a trigger mechanism for the farms adaptive management. When actual rainfall falls short of 430 mm, it warns management that the Anglian Regions rainfall in the preceding 12 months has been only 10% in excess of the lowest level in the past 40 years. The management team should then decide whether to stay with the conjunctive supply balance of Table 5.5 or whether to prepare for that most unlikely event a mains cut-off. If the latter decision is made, then the team should specify at what level of actual precipitation to take action and for what maximum length of duration of the cut-off they should prepare. Let us suppose that if the row 12 datum sinks below 411 mm (1996 5%), Silver Birches plc management takes action to guard against a 2-month mains cut-off. What it must then do is immediately to purchase sufcient water from AWS add to the existing stock of water in the winter storage reservoir so that there is sufcient water in storage to cover the entire irrigation requirements of the next 2 months. In this case, provided that any cut-off is indeed less than 2 months, the irrigation supply will remain unaffected by the drought. In that year, the annual advantage in AWS water purchased of the informational strategy over the infrastructural strategy will be reduced. Between 1963 and 2002 there were only two calendar years when rainfall fell below 411 mm; these were 1972 (409 mm) and 1996 itself (391 mm).

5.6

CONCLUSION

The purpose of this paper has been to show that drought management is not only a responsibility of public-sector water policy but can also be carried through by agribusiness itself with respect to net irrigation requirements. Farm-level planning of this kind becomes more important as climate change brings with it a greater proportion of extreme events and more frequent low dry season ows. The case-study is set in the Anglian Region of the England and Wales Environment Agency, where 58% of the two countries most productive agricultural land is found. This region is also the driest in the UK. The farm that is the focus of the paper is the largest producer of container trees in Europe. The growth of its trees in their containers is currently almost entirely dependent on irrigation water sourced from the mains supply of AWS, the private regional water utility. Silver Birches plc has some 94,000 trees. An interruption of the mains supply during a drought would irreparably damage the entire stock within 15 days. Although

such a cut-off is unlikely, the cost to the farm would be so great immediate nancial ruin that drought management planning is now under way. The core element is to introduce conjunctive supply based on a farm-level winter storage reservoir. Moreover such water promises to be much cheaper in prime cost terms than the mains supply. Farm managements initial intention was to set aside 50% of the reservoirs capacity on a permanent basis so as to provide a fall-back in time of drought. This was called the infrastructural strategy. A simple model was developed to show the monthly pattern of conjunctive supplies in this case. An alternative informational strategy was next developed using local and regional time-series data on surface water ows and precipitation. Based on these two variables, a drought warning chart was drawn up that allows farm management in most years to take full advantage of its winter storage reservoir. In drought periods it stimulates early action to protect the reservoirs capacity as a defence against a mains cutoff. The use of adaptive management triggers in this chart has, for the economist, striking parallels with the crisis signals used in models of international currency trading (Kaminsky et al. 1998; Dow 2002). Comparing the two strategies, the use of a drought warning chart offers considerable savings in expenditure on mains water. Silver Birches plc is now setting up a small, on-farm weather station to underpin the informational strategy as well as to strengthen management understanding of crop water requirements for optimal irrigation scheduling. The conclusion of this paper is that farm-level drought management should draw not only on civil engineering initiatives but also on an informed understanding of the often complex linkage of regional hydrology to the farms hydrosocial supplies. Informational strategies, combined with infrastructural investment, can be cost-effective and by the intelligent use of adaptive management based on trigger routines can reduce or even eliminate the impact of drought on farm output.

6 The potential role for economic instruments in drought management

And thus the whirligig of time brings in his revenges. William Shakespeare, Twelfth Night

6.1

INTRODUCTION

It is now widely accepted that human society generates greenhouse gases on such a substantial and increasing scale that climate change and global warming in the future will be of the greatest importance in our lives as well as in the lives of succeeding generations (Watson et al. 2001). Climate change, it is believed, will have marked impacts on water stocks and ows and this in turn will require innovations in the management of coastal and estuarine defence, drought, ooding, irrigated farming, nature conservation, storm and waste water systems, and water resources (Willows & Connell 2003: Table 3.1). In particular, the mean and variance of climate variables in the future may be associated with shorter return periods of climate extremes than in the past, for example of frost days or heavy rainfall events. The focus of this paper is on drought management and it recognizes the risk that the return frequency of drought may rise during the 21st century (van Hofwegen & Svendsen 2000: 910). Drought is here dened as a sustained and regionally extensive occurrence of below average precipitation. The articles focus is sharpened by narrowing attention primarily to the economic instruments of drought management within the context of the economic impacts of drought events.

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

58

6.2

THE WATER ECONOMY

Water is essential to all forms of life, of course, as well as to all the activities of human society. But if we limit ourselves only to the part it plays within a specied economy, we can distinguish just three broad sectors. First, there are the rainfed areas of farming, forestry and pastoralism. Secondly, there is the instream sector with its navigation, shing, conservation, recreation, tourism services and hydroelectric power production. Thirdly, we have the outstream sector where water meets the needs not only of households but also those of agriculture, mining, manufacturing, construction, public and private services. In this paper the author will use the term the water economy to refer to a region or countrys economy, i.e. its production of goods and services, from the point of view of the dependence of output on water stocks and ows in the rainfed, instream and outstream sectors. The article in fact focuses only on the outstream sector. The impact of a drought on the outstream water economy can be twofold. It may be that there is an increase in the economic demand for water. This is true in the case of the irrigation sector, for example, as well as households use of hosepipes and sprinklers. However, drought may simultaneously lead to restrictions on water supplies from the local utility with sharp negative impacts on companies sales and on households welfare. So the orientation of what follows is to explore: 1. 2. Drought managements degree of understanding of the impacts of drought on the outstream water economy. How drought managers might in the future deploy the economic instrument of water charges to achieve their objectives.

The paper illustrates its principal theses from Englands Anglian Region.

6.3 THE ANGLIAN REGION

For environmental policy purposes, England and Wales has a public sector body known as the Environment Agency, with its head ofce in Bristol. The Agency operates through eight regions, one of which is Anglian. The Environment Agency Anglian Region (EAAR) stretches from the Humber Estuary in the north to the River Thames in the south, from the Norfolk coast in the east to Northampton in the west. It covers 27,000 km2 and ve million people live in the area. The region has extensive, sparsely populated rural areas, particularly in the north and east, a long coastline and interspersed urban centres around which industry has developed. As communication links continue to improve (especially with London) it is likely to be one of the highest population growth areas in the country. The hydrology of the region is of considerable interest. Annual average rainfall is 595 mm compared with a national average for England and Wales of 897 mm. Effective rainfall is only a quarter of the total at 147 mm and long, dry summers during which evaporation exceeds rainfall are a normal part of the regions climate. Plotting the cumulative difference between rainfall and potential evapotranspiration, the Agency shows that in the 100 years beginning in 1899, there were 32 years exhibiting a rainfall decit (EAAR 2003: Fig. 3.6).

Twenty per cent of the area lies below sea level, including the extensive area of the Fens with slow-owing or ponded rivers and extensive washlands. The main aquifers are the chalk underlying large parts of the east, limestones in the west and sandstones in parts of the centre and east (ibid. 202). The Anglian Region has 58% of the most productive agricultural land in England and Wales and agriculture has considerable inuence on the rural economy and communities. The sustainable use of water is crucial given the competing demands for it, particularly because this region is the driest in the UK. Low ows in late summer and increased demand especially for agriculture and garden use are likely, says the Environment Agency, to cause stress (www.environment-agency.gov.uk/). In some summers, irrigation can make up 50% of total use. Alongside the Environment Agency, the second principal actor in water resources management in the region is Anglian Water Services Limited (AWS), which serves the needs of over ve million industrial, commercial and domestic consumers (www.anglianwater.co.uk/). AWS is the main regional utility and its principal activities are the provision of drinking water and the treatment of waste water. The company employs 3700 full-time equivalent persons and in 2002 had a turnover of 724 million (AWS 2002). Its principal infrastructural assets are mains and sewers, impounding and pumped raw water storage reservoirs, dams, sludge pipelines and sea outfalls. The companys capital investment plan in the 5 years through 2004/5 is equal to 1.5 billion and covers further improvements to drinking water, bathing and river water quality, as well as network improvements to maintain serviceability and meet new demands. In addition to AWS there are four other small water utilities in the region. They are Cambridge Water Company, Essex and Suffolk Water, Tendring Hundred Water Services, and Three Valleys Water. Overall, for these ve utilities, 40% of public water supplies are provided from groundwater, mainly from the chalk and limestone aquifers, and 60% from surface water.

6.4

THE REGIONS WATER ECONOMY

The water economy of the region with respect to its outstream ows is best understood by examining its principal components.

6.4.1

Households

Anglian region has a population of some ve millions persons. At the end of the 1990s, public water supply leakage and losses were one-fth of the total public supply. With respect to use net of these losses, households took two-thirds of the total. Private domestic abstraction also exists, particularly in rural Norfolk. There is a signicant difference in estimated use in litres per head per day (lhd) between metered and unmetered customers. In 199798 the data were 145 and 163 lhd, respectively (EAAR 2001: Fig. 6.3). National data for household use show the rounded breakdown between microcomponents as personal washing (33%), toilet (25%), clothes washing (14%), miscellaneous (13%), dish washing (8%), garden watering (6%), and car washing (1%) (ibid. Fig. A3.1).

Table 6.1 Anglian region irrigated agriculture. Crop type Potatoes Field vegetables Sugar beet Cereals Fruit Grass Total Water use (%) 57 17 15 5 3 3 100 Area irrigated (%) 41 20 22 11 3 3 100

Source: EAAR 2001, Figure 3.13.

6.4.2 Agriculture
This sector ranks second in Anglian water use, accounting for about 10% of the total. Farmers use water for general purposes, livestock watering (particularly in pig rearing), and for irrigation the largest consumer. The proportion of water used that is recycled back after use to the Anglian catchments is low in comparison with all other users. The supply sources are both mains provision as well as direct (surface and groundwater) abstraction. An interesting feature of the region is the operation of the Internal Drainage Boards. In the winter, water is pumped out of the drains to maintain the elds in a workable condition. In the summer water is diverted from the rivers to the drains whence it is distributed to the elds to maintain high water tables for irrigation pumping (EAAR 2001: 26, 49; Merrett 2002a: 10512). Spray irrigation is the principal technology. Water is required mainly in the summer months when plant growth is most rapid and when transpiration cooling is critical. This is precisely the season when effective rainfall is at its lowest, as we have seen. Abstraction has to be even higher in drought years when the farmer loses his rainfed source. Table 6.1 shows the irrigated sectors crops by area and irrigation water use. Potatoes, eld vegetables and sugar beet predominate.

6.4.3

Other sectors

Table 6.2 shows the other sectors use of water. The volumes supplied in 1998 were about the same from the regions water utilities as from direct abstraction by users. Today the ratio is closer to two-to-one. Retail services, the production of food and drink, the machinery sector, chemicals, and the extractive sector are the main players. The public water supply is of drinking water standard and more costly than direct abstraction sources. Chalk groundwater is also favoured by industry when it seeks a high-quality supply. For cooling or lower quality washing requirements, river water is acceptable. The extraction industrys dewatering activities are mostly exempt from abstraction licensing; they also engage in direct abstraction for the washing of sands and gravels. In the case of these other sectors, as with households, their consumption losses are low.

Table 6.2 Other sectors use of water in 1998 in the Anglian Region. Sector Retail Food and drink Machinery Chemicals Education and health Hotels Other* Transport Extractive Construction Total Public water supply (%) 32 21 16 8 8 7 5 3 0 0 100 Direct abstraction (%) 0 34 1 19 0 0 4 0 38 4 100

* Includes paper, metals, textiles and utilities. Source: EAAR 2001: Figure 3.12.

6.5

THE AGENCYS DROUGHT PLAN

The Environment Agencys Anglian Region staff are, of course, familiar with the risk of drought in their area. Droughts may take the form of hot, dry summers or of a longer period of low rainfall and dry winters, as in 198992. Extended dry periods affect rivers quickly but ows are maintained by groundwater baseow and wetlands. As these sources are exhausted river ows drop to very low levels, their temperature increases, there is less dilution of efuent and oxygen levels fall. Fish spawning grounds may be damaged and water supplies are threatened at the same time as irrigation needs are increasing greatly. The return of rainfall does not end the low ows for the dry soil soaks up the rain and many weeks may pass before there is a sustained rise in river and groundwater levels. The Agency has pointed out the strong relation between climate change and a regional drought. Changes in climate will change groundwater and river ow regimes and therefore the availability of water for abstraction. Current estimates of climate change suggest that by the 2020s throughout the Anglian Region there will be, on average, more winter rainfall and less summer rainfall. Effectively, this means that the climate will be less predictable, with both more dry years and more wet years. This in turn means that low ows will probably occur more often. Evidence of the possibility of longer droughts is unclear; the best available view appears to be that increased variability makes drought that lasts over several years slightly less likely. However, it is important to note that the understanding of changes in extreme events is more limited than that of changes in average climate. (Environment Agency Anglian Region 2001: 47) The Environment Agency Anglian Region published its drought plan as recently as May 2003 (EAAR 2003). Drought plans are required as part of the Agencys statutory

duty to secure the proper use of water resources. The plans chapter heads are: (i) drought management teams, (ii) drought monitoring arrangements, (iii) actions and mitigation, (iv) corporate affairs, and (v) reporting. Careful reading of the text shows that the Agencys principal policy instruments in drought management are informational, infrastructural and regulatory. These include, for example, public relations media releases, regional transfer schemes such as the Ely Ouse-Essex Scheme, and abstraction limitations. There are no economic instruments for drought management of any kind. Indeed the Plan specically states (ibid. 14): Abstraction charges are set at a higher rate than would be merited by moderate to wet conditions to ensure surplus money is available for nancing the operation of [transfer schemes] during dry years. This avoids large increases in rates that would be required when a drought occurs. With respect to this papers interest in assessment of the water economy, there are many indications of the Agencys concern for specic sectors. This builds on the Regions excellent 2001 Water Resources for the Future: a Strategy for Anglian Region. There is an evident wish to provide timely warnings, particularly for spray irrigators in relation to the likely need for implementation of restrictions on the use of their licences under Section 57 of the Water Resources Act 1991. A leaet Prospects for Spray Irrigation sets out the communication channels and procedures for introducing irrigation restrictions. The importance of communications between Agency staff and the farming community is highlighted and, when drought occurs, local meetings will allow the Agency to discuss directly with abstractors how best to instigate spray irrigation restrictions to minimize impact on the farming community (EAAR 2003: 27, 41). Restrictions from surface water often take the form of each licensee being able to use his/her daily quantity on alternate days. Night-time irrigation is encouraged. The Agency also expects those irrigators with summer licences and a winter storage reservoir to stop summer abstractions and use their stored winter water. Moreover, trialling of a new cost-benet method for assessing spray irrigation restrictions has been conducted on susceptible sites across the region. Water companies such as Anglian Water Services are also permitted to seek authorization to increase their rate of abstraction prior to a likely drought event. The Agency also states that it is committed to encouraging Water Companies to adopt measures for demand management and efcient use of water (ibid. 24). The drought plan makes only the briefest reference to other groups of abstractors. However, it is pointed out that many have been previously targeted to be more efcient with their water use. Local authority Environmental Health Ofcers contact households with non-licensable domestic sources in rural areas where the possibility of dry wells exists. Overall the Agency sees itself as balancing demand for a limited supply from various quarters, while being viewed publicly as arbitrators on behalf of the environment (ibid. 45).

6.6

ANGLIAN WATER SERVICES DROUGHT PLAN

AWS published its drought plan in March 2003 (AWS 2003). AWS is required by the UK governments Department of Environment, Food and Rural Affairs (DEFRA) to

agree updated drought plans with the Environment Agency. Water company drought plans have become a statutory requirement in the Water Bill 2003. The Plans section heads are: (i) introduction, (ii) background to previous droughts, (iii) water resources planning for a drought, (iv) drought management (surface water and groundwater), (v) demand-side management, (vi) supply-side management, (vii) mitigation of environmental impact and (viii) summary. Careful reading of the text shows that the AWSs principal policy instruments in drought management, just as with the Agency, are informational, infrastructural and regulatory. These include, for example, publicity campaigns, the construction of pumped, surface water storage reservoirs with long retention periods, and hosepipe bans. There are no economic instruments for drought management of any kind. This is in spite of the fact that AWS is clearly convinced of the general effectiveness of the management of economic demand. Between 1989 and 2003 the population of the AWS area increased by 400,000 persons, and 20,000 new properties have been connected each year. However the total quantity put into supply for customers has shown no upward (or downward) trend. In part this is due to the extension of household metering; 50% of AWS households now pay on the basis of volume used. With respect to this papers interest in assessment of the water economy, it is remarkable that (unlike the Agency) the AWS drought plan gives no indication of its concern for specic sectors. This is true even though the plan briey refers to the possibility of the introduction of demand restrictions on surface water supplies in the event that reservoirs are drawn down to control curves (ibid. 7). In the preparation of its drought plan, the AWS was required to use the Drought Plan Guideline published at the national level by the Environment Agency (EA 2002). This Guideline in fact makes no reference to economic instruments of drought management nor does it suggest that utilities should review the economic impact on their customers of regulatory action.

6.7

DROUGHT PLANS AND THE WATER ECONOMY

Two principal institutions lie at the heart of this analysis, the Environment Agency Anglian Region and AWS. The Agency is the statutory body with a duty for strategic water resources planning. Its role is to protect the long-term future of the water environment while encouraging sustainable development. Its vision for water resources in the next 25 years is abstraction of water that is environmentally and economically sustainable, providing the right amount of water for people, agriculture, commerce and industry, and an improved water-related environment (EAAR 2001: 13). In contrast, AWS is a private limited company that supplies water and waste water services to its customers within the area of its operation and thereby earns prots for its shareholders. Despite this immense difference in roles between the two institutions, we can see certain characteristics common to their separate drought plans. In the rst place we have a clear commitment by each to cooperate with the other in drought management. Secondly, in managing a drought event both institutions deploy policy instruments that are informational, infrastructural and regulatory. Thirdly, neither institution deploys an economic instrument in the management process. Fourthly, both institutions discriminate between user types in reducing the volume of water used. The Agency may permit AWS to increase its abstractions prior to a likely

drought event; and spray irrigators may encounter restrictions on the use of their licences. AWS may institute hosepipe and sprinkler bans (at the request of the Agency) as well as bans on non-essential use such as vehicle washing, irrigation of ornamental gardens, etc. (EAAR 2003: 24). Finally, the drought plans of neither institution seek to demonstrate that their selection of specic actors for regulatory restriction is demonstrably the optimal solution or a satiscing arrangement that, in some sense, minimizes the aggregate welfare loss of such restrictions.

6.8

ECONOMIC INSTRUMENTS

In this section I propose that the three principal national institutions for water resources management should review and evaluate the feasibility, costs, benets, risks and uncertainties of replacing the regulatory instruments of drought management by economic instruments. These three institutions are DEFRA, the Environment Agency and the Ofce of Water Services (Ofwat). The nal sections of this paper will provide a preliminary overview of six key issues that are likely to arise in the review/evaluation for which I call.

6.8.1

Legislation

In this context economic instrument refers to any one of three ways in which the EAAR or AWS (as well as the four other regional water utilities) require outstream water users to pay for their water supplies. These forms of payment are:

A charge for abstraction, based on the metered volume withdrawn. A price per cubic metre of metered mains water supplied to users. A charge for unmetered mains water supplied to a households house or at,
where the charge is based on the capital value of their dwelling. The proposal here is that, in comparison with the non-drought situation, these three forms of payment, which are already practically universal in England and Wales, would be successively raised at what the Environment Agency calls the potential drought stage, then again at the established drought stage and nally again at the severe drought stage (EAAR 2003: Table 1). The charge increases would be removed as soon as the postdrought wind-down is declared. Authority for the Environment Agency and the water utilities to take such action would require fresh legislation in the British parliament as well as new powers for the Ofce of Water Services in Birmingham.

6.8.2

Technical issues: time

The charges of types 1 and 2 above are based on metered ows. They can be introduced only where it is technically feasible to measure the volume of water supplied to users during each of the three drought phases referred to above (potential, established, severe). This requires meters capable of providing the volumetric ow on each day of the month. A variety of such real-time meters are available on the market. For charge-type 3 above, which at present applies to 50% of Anglian Region households, the additional charge would be based on the number of days that fell within the three

drought stages during the billing period. It has to be said that the economic management of demand is always more difcult in the absence of metered supplies (Merrett 2002b). Therefore, with an unmetered supply, there exists no economic incentive for the household to moderate the quantity used during a drought, even though the daily charge is raised.

6.8.3

Technical issues: place

In England and Wales the location of each abstraction point for surface water or groundwater is precisely known. It is also true that a feature of droughts is that they vary in severity and environmental impact across any given region as well as varying in severity and impact with source type. For example, for water resource management purposes the Agency divides the Anglian Region into 183 surface water catchments and 64 groundwater units (EAAR 2003: 1517). It would be both possible and desirable, therefore, to vary the increase in charge-type 1 above with the location of the point of abstraction.

6.8.4

Economic: rationale

The basic principle is that the regulatory instrument of rationing specic user-types (such as spray irrigators or households with gardens) be abolished and replaced by the economic instrument. Environment Agency and water company decisions on who should bear the brunt of supply shortfalls during a drought would be replaced by an instrument whereby those who are most willing and able to pay for the reduced volume supplied would receive it. Regulatory at would be replaced by users decisions. However, it is essential that households do not face a price or charge increase for their lifeline supply.

6.8.5

Economic: demand response

Deployment of the economic instrument would permit the EAAR and the AWS to vary the water prices and charges it sets throughout the droughts duration and across the regions surface water catchments and groundwater units. The change in volume of water purchased following a change in price of the metered supply is described by economists as the elasticity of demand (Merrett 2002a: 235). Both the Agency and the utilities would wish to carry out new research into the economic demand for water in the region by different users so as to be able to estimate the price rises they would apply at each stage of the drought. The main actors should be able to vary their volumetric prices and charges within a high upper bound set jointly by the Environment Agency and Ofwat. Morris and his colleagues at Craneld Universitys Institute of Water and Environment have recently shown that in Englands irrigated agriculture although increased water charges can reduce demand in low value applications, very large increases in charges are needed to reduce water demand where the nancial returns to water are high, resulting in large decreases in farm incomes before consumption signicantly changes (Morris et al. 2003: 623). Meanwhile, Chapter 4 in this book shows how economically devastating regulatory action would be for a specic Anglian farm during a drought, as well as the climate adaptation options that farm management can pursue.

6.8.6

Political

Is this proposal likely to be acceptable to the people of England and Wales? I believe that the shift of power from regulator to user will be welcome. However, two issues of particular importance would have to be addressed. First, the price paid by households for their lifeline supply should be unchanged, for social reasons. Secondly, what is a reasonable but effective way to handle unmetered users? The importance of this fades the more successful is national government in increasing the proportion of households that are metered. As David King, Director of Water Management of the Environment Agency, writes in response to the rst draft of this paper: You are right in thinking that at present drought management makes little use of economic incentives to control demands. For households far more meter penetration would be necessary before this became a viable option. However, we will keep this under review and we will consider all options that can help with effective water resources management (pers. com. 3/12/2003). George Day, Head of Supply/Demand Balance at Ofwat expresses a similar opinion in a letter to me.

6.9 CONCLUSIONS
Climate change will likely be associated in the future with shorter return periods of climate extremes, including drought. So effective drought management is set to become even more necessary in the future than it is already. This in turn demands a review of the current instruments of policy that we deploy as well as a better understanding of their impact, not only on the rainfed and instream sectors, but also on the outstream water economy. The argument of this paper is developed for Englands Anglian region, particularly with respect to the roles of the Environment Agency and of AWS. Review of these two institutions drought plans shows that their instruments of policy are informational, infrastructural and regulatory. There are no economic instruments of any kind. Moreover, the Environment Agency at the national level in its drought plan guideline makes reference neither to utilities deploying economic instruments in demand management nor to a need for utilities to review the impact on their customers of regulatory action. Yet both the Environment Agency at the regional level as well as the AWS discriminate between user types in reducing the volume of water used. But neither institution seeks to demonstrate that its selection of specic actors for regulatory restriction is demonstrably the optimal solution or a satiscing arrangement that, in some sense, minimizes the aggregate welfare loss of such restrictions. With respect to outstream water, the author concludes that the three principal national institutions for water resources management, DEFRA, the Environment Agency and Ofwat, should review and evaluate the feasibility, costs, benets, risks and uncertainties of replacing the regulatory instruments of drought management by economic instruments. If that review and evaluation opts for economic instruments, they would shape both the location of abstraction as well as the relative shares between outstream water users, based on the responsiveness to charge and price changes for the abstraction and use of water.
The material in this chapter originally appeared in: The potential role for economic instruments in drought management. Irrigation and Drainage, 53: 19, 2004.

7 Virtual water and Occams razor

Until then I had thought each book spoke of the things, human or divine, that lie outside books. Now I realized that not infrequently books speak of books: it is as if they spoke among themselves. Umberto Eco, The Name of the Rose

7.1

INTRODUCTION

In 2001 Professor Tony Allan of the University of Londons School of Oriental and African Studies saw published his magnum opus entitled The Middle East Water Question: Hydropolitics and the Global Economy. The book marks the nal step in his long march from geography to the politics of water resources. It also provides the most complete statement of his views on the rle of virtual water in linking regions of water scarcity with regions rich in water. In fact, his main interest is the Middle East and North Africa (MENA). The concept of virtual water, virtually synonymous with Allans name, has been taken up increasingly widely in recent years and the time is ripe for a critical review of its relevance to our understanding of the manner in which the water needs of semi-arid countries are met. Such a review is the object of this paper. Section 7.2 provides a rsum of the main propositions of the virtual water thesis. Section 7.3 deploys the main critique. Section 7.4 introduces Occams razor. Section 7.5 presents my conclusion. Before getting under way it is worth noting that the disciplinary foundation of this paper is philosophy. So a useful philosophical convention will be used. Whenever I am discussing a concept, the word will be placed in single inverted commas. For example: Spinoza tells us that the concept dog cannot bark. Whenever I refer to the thing or activity to which a concept such as dog refers, there are no inverted commas. For example: Hagrids dog Fluffy barked without ceasing.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

68

Two other features of a philosophical paper are worth noting. First, philosophers within the English tradition are primarily concerned with the relation between language and truth, so they commonly write in a dense and compact style where the text needs to be read at a measured pace to be understood. Second, non-philosophers may take the view that philosophy is merely rhetorical quibbling, nothing but words, words, words. My view is different. Water resources management has traditionally been founded on engineering and hydrology with a history of immense achievements. But we now recognize that social and environmental scientists should ally themselves with the hydrologist and the engineer in the challenges that the future holds for us all. This requires that political scientists and economists, for example, develop an appropriate language for their work. They are now in the stage where such a language is still in the furnace. The role of philosophy is to assist in the forging of this new language, in which the objects of criticism may be water efciency in irrigation or water demand in urban planning or virtual water in global trade or the contingent value of a buttercup. In a nutshell, the successful advance of science is critically dependent on progress in its language.

7.2

A WATER DEFICIT RESOLVED

A brief, neutral statement of the virtual water thesis is set out in this section, beginning with Allans concept of a regions water decit. Region here refers to any relevant geographic area including a country, a province or a catchment. For any region it is possible to estimate whether or not it suffers from a water decit in the sense that there is not enough surface water, ground water and soil water to meet the domestic, industrial, municipal and food needs of its population (T. Allan pers. comm.). In the specic case of what will be referred to as Region A, its surface water, groundwater and soil water are insufcient to meet the needs described above. This shortfall of supply with respect to need is Region As water decit (Allan 2001: 30). In fact the needs of households and industry can and are comfortably met by regional supply. It is the crop water requirements of food self-sufciency that are impossible to satisfy. As a consequence the water decit is resolved by the import of virtual water from Region B. Virtual water refers to the soil water sourced by precipitation and irrigation that meets the crop water requirements of the food grown in Region B that is subsequently exported to Region A. Region As imports of such food are sufcient, alongside domestic production, to meet entirely the food needs of its population. To indicate the scale of these virtual water imports, we can take the example of wheat. One tonne of exported wheat requires about one thousand tonnes of virtual water (Allan 2001: 106, 126). Less than 0.1% of the virtual water is physically embedded in the food grains themselves. During cultivation in Region B, more than 99.9% of the virtual water returns to the irrigation cycle as farmland drainage or is lost in evapotranspiration. As Allan writes (2001: 106): At the 1000 tonnes (cubic metres) of water per tonne of grain estimate of water content the [MENA] regional imports of virtual water by the mid-1980s were equivalent to the annual ow of the Nile into the Egyptian agricultural sector.

7.3

A CRITIQUE OF THE VIRTUAL WATER THESIS

Allans thesis summarized in section two is, I suggest, awed in a number of ways. The critique will be set out with respect to the use of the term virtual water, the import of virtual water and, nally, the farm sectors of Regions A and B. Virtual water. In everyday English we use the word virtual (as in virtual reality) to mean something parallel to or imitative of a real-life entity or process. But Allans virtual water does not do this. This is because virtual water is real water; it is the soil water of Region B used in meeting the crop water requirements of that Regions food exports. This neglect of a well-established linguistic routine has the result that the central concept of the argument, virtual water, is not virtual in any sense. It is not good science to build theory on terms that are inherently misleading.

7.3.1 The import of virtual water

In The Middle East Water Question: Hydropolitics and the Global Economy we repeatedly read that virtual water is imported into the MENA region. So one expects to see each single ship carrying a full cargo of wheat, maize or rice imported from the wide world and heading for the sunlit harbours of the Mediterranean and the Red Sea to be followed by one thousand additional ships each carrying in its hold a full cargo of the virtual water used for the imports crop water requirements. A magnicent sight but one we are unlikely to witness in our time. It is as if, having identied a water decit in Region A, virtual water must be imported to eliminate it. It may be that Allan does not wish us to believe literally that this happens. Clearly propositions of the type water security for the MENA region is achieved by virtual water imports are false, and as misleading as the tabulation of 40 billion cubic metres of cost-free water transported into the region each year (Allan 2001: Table 2.1). But if such statements are, indeed, merely metaphorical, they cannot at the same time be read as part of a scientic argument on international trade in food and water.

7.3.2 Farming in Regions A and B

When one approaches agriculture from the perspective of water resources there is a danger that the experience of farming is seen largely with respect to its crop water and net irrigation requirements. As a result, a more rounded vision is lacking, one that understands that the water theme is only one amongst many, such as soil characteristics, land rights, labour skills, pest control, farm budgets and product markets. Consequently, if we use the term the import of food, this opens up major questions rendered invisible by the import of virtual water. Have food imports led to higher population birth rates in water decit regions than would have occurred in their absence? Do food imports weaken the farm sector of the importing country? Do food imports open the importing country to political control from the exporting country? Will the importing country be able to maintain its foreign exchange expenditure on food imports in the long run? Finally, if food imports are subsidized when they are shipped from the European Union and North America, will these subsidies be maintained in the long term?

7.4

OCCAMS RAZOR

So we come to William of Occam, a Franciscan scholar and Aristotelian philosopher who lived between circa 1295 and 1349, the year of the Black Death. He took part forcefully in the great debates of his time on the poverty of the clergy and on transubstantiation and was excommunicated by Pope John XXII in 1328. Bertrand Russell writes of him: Occam is best known for a maxim which is not to be found in his works, but has acquired the name of Occams razor. This maxim says: Entities are not to be multiplied without necessity. Although he did not say this, he said something which has much the same effect, namely: It is vain to do with more what can be done with fewer. That is to say, if everything in some science can be interpreted without assuming this or that hypothetical entity, there is no ground for assuming it. I have myself found this a most fruitful principle in logical analysis. (Russell 1996: 4623) I now wish to explore whether Occams razor can properly be applied in excising two of Allans core concepts: virtual water and the import of virtual water. Virtual water. This term can be replaced by the crop water requirements of food exports. The import of virtual water can be replaced by the import of food. More generally, the huge economic, political and social processes that are addressed by Allan in the terminology already described can be reset in a world where semi-arid (and other) regions do not have the capacity to feed their populations and so import food. These imports mean that less production and therefore less water is required in these regions irrigated agriculture. Regional politicians may deect attention from such dependence; the availability of imported food allows them both to postpone new water supply initiatives and to delay difcult decisions about the demand management of their water resources.

7.5 CONCLUSION
In this paper it is argued that:

Virtual water refers to real water there is nothing virtual about it. It denotes the crop water requirements of food exports. The import of virtual water is a metaphorical term, not a scientic one, and its use leads to statements that are plainly false. It denotes the import of food. In its policy applications, the import of virtual water leads to a neglect of the curMy conclusion is that water resource researchers and policy-makers should apply Occams razor to the virtual water thesis.

rent and future status of the agricultural sectors of the countries importing and exporting food.

The material in this chapter originally appeared in: Virtual water and Occams razor. Water International, 28(1): 103105, 2003.

8 Virtual water and the Kyoto consensus

A man may imagine things that are false, but he can only understand things that are true. Isaac Newton, Theological Manuscripts

8.1

THE USE OF METAPHOR

Three papers were published in the March 2003 edition of Water International that sought to clarify and evaluate the use of the concept virtual water in our understanding of water resources management in water-decit catchments, regions and countries (Allan 2003, Lant 2003, Merrett 2003). In the following discussion note this author seeks to capture the main points of his continuing differences with Tony Allan. Allan and I are in complete agreement that phrases such as the import of virtual water are metaphors, not propositions that can be simply said to be true or false. Moreover, we now have a clear denition to work with. Virtual water is the water needed to produce agricultural commodities (Allan 2003: 107). It follows that Allan accepts that there is absolutely nothing virtual about virtual water. Virtual water is real water. It is, rst, the water needs of livestock. Second, it is the soil water required to grow crops. Soil water is sourced by rainfall, irrigation practices and seasonal ooding such as that of the River Nile (Allan 1995: Table 1, Merrett 2002: Table 3.1). In the original paper I suggested that water resources management is increasingly becoming a eld in which social scientists play a part, alongside our colleagues from engineering and hydrology for example. The language we use to understand that most simple, yet complex of entities water is still being forged; and the successful advance of science is critically dependent on progress in its language.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

72

Here we come to the crux of the argument. I believe of course that there is a place for metaphor and other linguistic tricks in published papers and books as a means to entertain, lighten the text and communicate in a telling manner. This paper itself contains a number of metaphors. The question is, do we advance our capacity to understand water resources by introducing a new term when we already have a perfectly good one? If we can speak of the water requirements of agricultural commodities, why do we need virtual water? Moreover, can it be wise to base our analysis on a single, powerful metaphor that virtual water is imported?

8.2

CROPS, CROP WATER AND WATER DEFICITS

But there is a second arrow to my bow. Virtual water, as a central and constantly reiterated concept in the discussion of arid and semi-arid countries water requirements, can mislead us, create serious errors of analysis. I illustrated this in my March 2003 paper by pointing out that Allan in his book The Middle East Water Question: Hydropolitics and the Global Economy (2001: Table 2.1) includes in his typology of Middle East and North African waters 40 billion m3 per year of transported water in the form of virtual water, alongside the ows of pipelines, tankers and bags. His use of the virtual water concept led directly to a 40 billion cubic metre error. This is because the 40 billion cubic metres to which he refers are the crop water requirements (primarily of foodgrains) sourced by rainfall and irrigation in the foodgrainexporting areas such as Australia, Canada, the European Union and the USA. Clearly these are not Middle East and North African waters and they are not transported to the MENA region. Allan confuses crops with crop water; it is the former that are transported not the latter. This inbuilt propensity to error appears in a new form in Allans 2003 paper, as I shall show. We would all of us probably accept that there is some usefulness in the idea of a region or country suffering from a water decit. This is dened by Allan, for example, as a situation in which there is not enough surface water, ground water and soil water to meet the domestic, industrial, municipal and food needs of its population (Merrett 2003: 104). Many of us would say that one course of action, given such a present decit, is to import food. For example, the London region has done this successfully for 2000 years. For Allan, this misses the point. In his 2003 paper he suggests that access to virtual water remedies water decits and, with freshwater and desalinated water, achieves water security. 10 percent or so of water needed for drinking, domestic, industrial, and service uses must come from freshwater or from manufactured sources. The 90 percent of water need for food and other agricultural production can come from freshwater, from soil water or it can be accessed in effect via food imports. Virtual water and manufactured water are the very successful means by which water decit economies can remedy their decits. (Allan 2003: 108) Once again we see a dangerous crossover between talking about food imports as if they were water imports. In matter of fact, food imports tend to increase a regions water decit. This is of great importance in semi-arid countries with dense populations. The growth in the water decit arises from the faster rate of growth of population that

imported food permits. In addition, lower levels of national production of food follow from the lowering of food prices brought about by imports. This is particularly true when imported food is subsidized and when the importing countrys exchange rate is over-valued, making imports even cheaper. In summary, I believe we should dispense entirely with the term virtual water. This is for ve reasons: 1. 2. 3. The term is redundant; virtual water is nothing more or less than the water needed to produce agricultural commodities. There is nothing virtual about virtual water. The use of the term leads in particular to a neglect of the impact of food imports on a countrys agricultural sector, because food imports are falsely represented as water imports. The confusion of water with food that accompanies the term virtual water like a dark shadow fosters analytic errors such as that food grain purchases are transported water, that food imports remedy water decits in the importing countries and that imported food brings total water self-sufciency for the importing region (Allan 2003: Figure 3). The greatest weakness of Allans concept is that it repeatedly confuses crop outputs with the water required to produce them. If, as Allan suggests, Hoekstra & Hung (2002) conclude that 695 km3 of virtual water is traded each year, why have none of us seen the boats within which it is carried? Why are they not causing gridlock in the worlds sea channels? Virtual water is real water; if it is indeed traded it needs to be transported to the new owner. In fact, no such trade ever takes place in the case of food exports. Hoekstra and Hung, outstripping Allan, commit a 695 km3 error. Social science, if it is constructed on a metaphor, is built on sand.

4.

5.

8.3

THE KYOTO CONSENSUS

The nal issue raised here is whether or not virtual water has now entered the emerging consensus shaping current approaches to water resources management signalled in the World Water Forum in Kyoto, the Kyoto consensus as I shall call it. Key features are that big dams are bad whereas small dams are good; that water resources management should be carried out on the catchment scale; that irrigation projects should be turned over to water user associations; that public water utilities are ripe for privatization; that water resources management should focus on demand management; and that water use should everywhere be priced. A thoroughly modern schema set, unfortunately, in a post-modernist world. The Kyoto consensus is best understood as the waterish extension of the Washington consensus. What follows bears comparison with Christopher Lants insightful paper, both his suggestion that virtual water refers to a water resources management strategy and his comments on the economic linkage between the MENA region and the dominant grain-exporting regions (Lant 2003: 11315). The reasoning is this. The Washington consensus is that human society throughout the globe should, in its economic activities, be organized around free markets in which the dominant players are private companies with governments playing only a limited, regulatory role. Those free markets should include the export and import of goods and services between countries. The hegemonic political and economic power in all of this

is to be wielded by the dominant classes of the USA and the European Union and their institutions, in particular the International Monetary Fund and the World Bank. Grain exports from Australia, Canada, Europe and the USA should be encouraged in this context, particularly to semi-arid countries with dense populations. Two major problems arise from this specic agenda. First, as indicated earlier in this rejoinder, food imports by the South weaken the domestic sector that produces food by driving down the market price of grain output. The decline of agriculture drives down rural incomes and stimulates rural-urban migration. Second, where the South has the capacity to export services and manufactured goods on a substantial scale, food imports can be nanced out of export earnings. But if the South cannot pay for its imported food in this way, it becomes dependent on North America and Europe for subsidy of these imported foodgrains. This means that, along with other forms of nancial and military dependence, the future of these countries is shaped by the North. In this case such countries, far from looking to a future of total water selfsufciency as the virtual water theorists argue, face an indenite journey of economic dependence. Here we nd conrmation that the most important characteristic of economic systems is the complex forms that exist of the interdependence of supply and demand.

The material in this chapter originally appeared in: Virtual water and the Kyoto consensus. Water International 28(4): 540542, 2003.

9 The urban market for farmers water-rights

9.1 INTRODUCTION
In the eld of water resources management a widely held belief exists that allocation stress is to be found in many parts of the world and is set to become more intense in the future because of global population growth and climate change (Meinzen-Dick & Rosegrant 2001a). Allocation stress refers to access conicts between the agricultural, domestic, industrial, urban service and environmental uses of outstream and instream water ows. If the belief that allocation stress will intensify is well founded then, because of the dominant role of irrigation water use at the global level, it is imperative to explore the possibilities of reducing farmers use of water or, at the very least, of slowing its growth (Merrett 2002a). One process by which the scale of irrigation is reduced occurs when farmers choose to sell their abstraction rights (or other rights to access water) in perpetuity to actors that apply these released ows in towns and cities for household, manufacturing and urban service uses. It is this market that is the subject of the present paper and the term market is used here to refer to those institutions that provide the context for the purchase and sale of commodities. A benecial allocation multiplier exists here; if water supply for irrigation and for urban purposes divides in the ratio 70:30, for example, then a 15% transfer of agricultures total creates a 35% increase in the urban total. Note that the paper does not address environmental and hydropower needs or the transfer of abstraction rights between farmers. Neither does the paper review the broad-ranging arguments for and against re-allocating water from farming to the urban sphere (vide Merrett 2002a: 14882, Rosegrant & Ringler 1998).

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

77

The objectives of the paper can now be stated. These are: rst, to set out an orthodox microeconomic account of a water-rights market in equilibrium; second, to ascertain whether that approach can deal with a number of empirical complexities in real-life trading. The nal introductory comment concerns the conceptual distinction between water and water-rights. Water is a collection of molecules, each of which consists of two atoms of hydrogen bonded to one atom of oxygen. A water-right is a legal claim to abstract or otherwise access water. You can oat on water but not on a water-right. You can abrogate a water-right but not water. This paper addresses water-rights markets, not water markets.

9.2 THE URBAN ACTORS DEMAND FUNCTION

We begin with a dened region composed of a metropolitan area located within an agricultural landscape in which farmers abstract surface water and pump groundwater for irrigation purposes. Abstraction takes place within a framework of rights that are formalized in national laws, or are recognized by customary practice, or exist in situations of legal pluralism (Bruns & Meinzen-Dick 2001: 1). The city already accesses water from a variety of sources, as the hydrosocial balance shows. These sources include: 1. 2. 3. 4. Rainwater harvesting in cisterns by households. Groundwater and surface water abstraction by the metropolitan water utility, urban developers and by manufacturing industry. The utilitys import of water from a neighbouring region. The manufacturing sectors reuse of its own waste water.

The population, gross domestic product, income and size of the city are all growing as is the demand for water. Consequently the metropolitan water utility as well as the enterprises that capture their own water needs (together the urban actors) are seeking to expand their separate water supply capacities. In particular, the water utility knows that supply expansion requires capital nance and infrastructural construction. This will push up the rms overhead costs and prime costs and thereby total costs (Merrett 1997: 325). But the utility is sure that if the average total cost per cubic metre of water delivered remains close to its existing level, the price charged to customers will be one that they are willing and able to pay. The utility and those companies self-supplying their own water requirements begin to take an interest in the abstraction rights enjoyed by farmers in the surrounding countryside. They know that these rights can legally be transferred to urban actors. For any single project the urban actor would have to pay a lump sum to the farmer for the transfer of the abstraction right, as well as meeting accompanying transaction costs. In addition, the urban actors would have to make capital investments in headwork and network infrastructures in order to capture the water they are now permitted to abstract and to deliver it to the city. The quality of water from any specic abstraction point would likely impact on headwork treatment costs, particularly with respect to the waters potability. So what is the maximum payment an urban actor would be willing to pay to the farmer for the transfer of his abstraction right, that is to say, what is the urban actors

X Demand function (urban actors)

Abstraction rights price ($/m3)

Z E

Supply function (farmers) Abstraction rights volume (m3)

Figure 9.1 Demand and supply functions for farmers water-rights. maximum bid price? One answer is that it would be equal to a cash sum that when added to transaction costs and to the accompanying capital expenditure on infrastructure would generate an average total cost per cubic metre of the same order as the existing level of average total cost. A second answer, a stronger response, is that the maximum bid price plus the other capital spending would not produce an average total cost per cubic metre higher than the average total cost of the best alternative supply scheme (Kemper & Olson 2000: 352). For example, such alternatives might be an increased capacity of the utility to import supplies from another region or the expansion of a manufacturing companys waste water reuse facility. In the calculation of average total cost, the volume used as the divisor should be that delivered to the customer. In this way storage and distribution losses are properly accounted for. This point is made strongly by Schifer (1997: 368). There may also exist urban institutions requiring abstraction rights for whom no feasible alternative supply source exists. In this case the maximum bid price would be that which, after the capital payments associated with the rights transfer, still leaves the investment project with an acceptable nancial rate of return. As an alternative to supply expansion, the utility could also consider a tariff rise. In any given year there are a number of urban actors each with one or more potential abstraction rights purchases in preparation but in which the nal purchase price has not yet been legally agreed. For urban actors as a group, a function of the type represented in Figure 9.1 can be conceived to exist. The vertical axis measures, for urban actors, their maximum bid price per cubic metre in 2003. The horizontal axis measures the abstraction rights volume purchased. We have a standard demand curve. For a modest addition to the urban actors abstraction capacity, high prices are bid. This is because such additions are necessary to meet clearly identied needs over, for example, the years 20042008 (Rosegrant & Binswanger 1994: 1617). Much lower

prices are bid for a large volume transfer of rights because they are not required in the short-medium term.

9.3 THE FARMERS SUPPLY FUNCTION

The farmer who sells his abstraction right is here assumed to relinquish his entire legal claim on water use for irrigation purposes. Hereafter farmer can be taken to refer to a farming family, an agribusiness, and an irrigation district or farmers association that possesses water-rights. As Hearne and Easter point out (1995: 41): These rights can be stipulated by volume or by percentage of river ow. But in areas where water supplies are highly variable, it is necessary to designate how water is allocated during times of scarcity. Rosegrant & Gazmuri Schleyer (1996: 2701) make the same point. The farmer who sells his abstraction rights reverts to cultivation activities that are feasible on a rainfed basis. This may make it necessary that he incurs some additional capital expenditure and it will certainly entail a loss of income from produce sales, because of the fall in yields under the new conditions in which crop water requirements cannot be met by irrigation. This reduction in farm income (net of any reduction in prime costs) plus any new capital expenditure required (including transaction costs) must be compensated through the price that the farmer receives from his sale of the abstraction right. He sets his minimum release price equal to new capital expenditure plus a multiple (here referred to as m) of the annual loss of net farm income. (There is a parallel here with the payback period required by some investors in order to go ahead with a new investment project.) From the point of view of theory, the fact that the farmer may set m equal to (say) 10 years or less demonstrates his discounting of future returns. The value of m can be expected to vary with crop type; and estimated annual income loss may include a risk premium for cultivation dependent entirely on rainfall. Figure 9.1 represents the supply function. The abstraction rights volume offered by farmers at the lowest price is small but increases at higher prices. There is an equilibrium price E at which the market is cleared. At this price the abstraction rights volume sought by the metropolitan water utility and other urban enterprises is precisely equal to the volume farmers are willing to release. Note that at the price E urban actors total expenditure on the acquisition of abstraction rights is lower by the area XEZ than the demand price function (built up from individual urban actors maximum bid prices) shows they would have been prepared to pay. At the same time farmers total income from the sale of their abstraction rights is greater by the area YEZ than the supply function (built up from individual farmers minimum release prices) shows they would have been prepared to accept. Everybodys happy. Economists will recognize that the supply and demand analysis used above is of a type that most basic orthodox economics texts deploy.

9.4 THE LIMITS TO THEORY

The previous sections have presented a simple theory of a ruralurban market in abstraction rights. But simple and elegant societies to which such a theory might apply do not exist. The point has come in the argument where it is necessary to

muddy these crystal waters with real-life impurities if our market theory is to have application to actual transactions.

9.4.1

Time-scale

The analytic exposition is based on the trade of water-rights in perpetuity. However, Rosegrant & Meinzen-Dick point out: While tradable water rights should be permanent, or very long term, to ensure security of the right, transfer of water rights need not be permanent: water rights can be leased for a season, a year, or many years. (Rosegrant & MeinzenDick 1996: 47) The present paper does limit itself to long-term or permanent trading. In the case of the transfer of water-rights, the shorter the lease, the closer one approximates a market in water itself. Rosegrant & Gazmuri Schleyer (1996: 2767) give examples from Mexico both of a one crop season transfer in Monterrey and a 50-year trade in Guanajuato. The Mexican Comisin Nacional del Agua permits concessions and grants for periods from 5 to 50 years, with terms exceeding 30 years the norm, to ensure security of the water right (ibid. 269). It is not wise to present the theoretical approach above as a basis for all trades and transfers, whatever their length. From the point of view of both the farmer and the urban actor a one-year lease, for example, is an asset utterly different from a permanent sale. Only in the latter case does the farmer give up for all time the use of this ow of water for irrigation; and only in this case does the urban actor secure for the indenite future its water source. However, it must be admitted that the short term may be the prelude to the long term. Haddad (2000: 95116) has described a fallowing agreement in respect of about 118 million cubic metres of water per year for the 2 years ending in July 1994 between the Palo Verde Irrigation District and the Metropolitan Water District of Southern California. In 2002 the two parties were negotiating a contract to fallow up to 29% of the irrigated farmland for a 35-year period (A. Quist, pers. comm.).

9.4.2

Concessions

The general theory is framed in terms of a market in abstraction rights. But the use of water by farmers is also made possible when they enjoy a legal right to access water ows from an organization that itself is the abstractor. The obvious case is where a public irrigation authority assigns farmers the right to draw off water from a canal. The model of Sections 9.2 and 9.3 can incorporate such concessions (as they are sometimes called) without amendment and permits me to refer to it as a theory of the market in tradeable water-rights rather than one limited to trades in abstractionrights. Trading in the surface water concessions of a public authority may also often fall within a leasing model (see Section 9.4.1 above).

9.4.3 Part-sales
Here we consider the situation where the farmer sells only a fraction of his waterrights, retaining the rest for irrigation on a smaller scale. The theory incorporates this

with no difculty. The partial sale of the farmers water-rights is still represented by the supply function; the specied volume is smaller but the release price per cubic metre may be little changed. A cultivator may choose to use the payment he receives to invest in water-saving technology for the areas he continues to irrigate (Gazmuri Schleyer & Rosegrant 1996: 41). However, from the point of view of the urban actor, to attain a targeted volume from a group of partial sales rather than a single, highvolume trade implies increased storage, distribution and transaction costs.

9.4.4 Sale of land

Here we have a situation where the farmer sells not only his water-rights but also his farm and the land on which it is set. With respect to the cultivators release price, he now has to face a future in which he derives no income stream from his property. The minimum release price is now set at his transaction costs plus a multiple n of the total annual surplus, that is, the farms turnover less the sum of prime and overhead costs. The minimum release price for the joint sale of land and water-rights exceeds that where the sale is only of the water-rights, because in the former case alone is all farming income lost. A farmer receiving a capital sum for the sale of his land and waterrights may well use the money to clear any debts he may have. With respect to the maximum bid price of the urban actor, he now purchases two assets the farm property as well as the water-rights. There may be signicant advantages here. The purchaser may believe that the risk of legal challenge to his waterrights is considerably diminished by virtue of his land ownership status. Moreover, access to his new headwork and network infrastructures is now entirely within his control. Finally, parts of the farm unrelated to the abstraction and distribution process can be sold off, immediately or in the future. As a result the urban actors bid price for the twin assets enjoys a premium over the bid price for the water-rights alone. These remarks on urban actors position can be put more forcefully. In many countries water-rights and rights to land are indissolubly linked; in order to access the former it is necessary to buy the latter. In such cases a metropolitan area pursues its search for water by means of land purchases within its region. The most famous example is that of Los Angeles and Owens Valley (Reisner 1993: 5996). These remarks on concessions, part-sales and the sale of land address some of the complexities of the real world. They also signicantly widen the scope of the theory yet leave the theoretical approach more or less intact.

9.4.5 The legal context

For the theory to have an application to inter-sectoral markets, a strong condition is the existence of secure water-rights within the ruling legal framework. These must be rights for the farmer to access water (by abstraction or concession) from an aquifer, river, lake or canal at dened times and in dened quantities and they must dene how access will be reduced in times of drought. Rights may take the form of a share in stream ow. The legal system must also recognize that the farmer is permitted to transfer this right to an urban actor. In the absence of such a legal basis a market may still exist, but it will be arcane. The current market in Mexicos Baja California Sur is an example (E. Troyo: pers. comm.). In such circumstances urban actors bid prices will be lower because of the perceived risk that they may in future lose the access to their water source following a private law

suit or through a government ban. In fact, Bruns & Meinzen-Dick (2001: 3) suggest that industries rely on their economic and political power to obtain water from farmers, legitimately and illegitimately. Institutional frameworks for carrying out reallocation through voluntary agreements among users are usually non-existent. They suggest a dual approach to providing a legal basis for trade. Where transactions are likely to be sparse, the framework would simply enable transfers to be negotiated and allow challenges by those who feel they may be harmed. In other cases laws and agencies would establish a cadastre, introduce requirements for public notice, provide expert assessment of third-party impacts (see below) and assist the less educated and powerful in the trading process (Bruns & Meinzen-Dick 2000). In this context it is suggested that: the establishment of markets in tradable property rights does not imply free markets in water. Rather, the system would be one of managed trade, with institutions in place to protect against third-party effects and possible negative environmental effects (Meinzen-Dick & Rosegrant 1997: 317)

9.4.6 Third party effects

When a farmer irrigates, some proportion of the water used is recycled downstream or back to the aquifer. Shiklomanov (2000, Table 5) estimates at the global scale that 70% of agricultural withdrawals are consumed and the remaining 30% are recycled. Following the sale of an abstraction right or a concession to an urban actor, the water is physically transferred to the water utility, to a residential developer or to an industrial enterprise. As a result the recycling (at an increased but perhaps polluted volume) takes place downstream of the new users households, manufacturing or urban service companies. The upshot is that the third party interests of farmers and other actors downstream of the original point of irrigation are affected. They lose the recycled ow they had enjoyed prior to the inter-sectoral transfer. But note that some farmers may experience greater recycled ows if the city in question is upstream of the farm selling its water-rights. A third party effect also exists when an urban actor purchases a water right not utilized fully by the farmer and then exploits that right 100%. The downstream ow is correspondingly reduced. As Kemper (2001) writes: [the downstream neighbours] may go to court to prevent the sale, thereby causing high transaction costs, or courts may not be available and they will have to accept the loss. A mechanism is therefore needed to either negate thirdparty interests or to mitigate the impacts of water trading on the different stakeholders, including an effective conict resolution system. So the approach to the return ows issue can stand at two extremes. In one case, there are no private rights to return ows; government retains them. Here the sale of a water-right by a farmer to an urban actor that results in a diminution of return ows cannot be challenged by the downstream party. This has long been the case, for example in Chile, in Mexico and in the Northern Colorado Water Conservancy District of the USA. At the other extreme farmers and other instream water beneciaries have a right to the return ows they receive and therefore can seek to block a development that reduces these ows or, at least, to be compensated for their loss. The transaction costs that must be met in agreeing a water-rights transfer are consequently higher (Rosegrant & Binswanger 1994: 1619, Gazmuri Schleyer & Rosegrant 1996: 37).

There is another source of third-party effects. In section one it is stated that this paper does not review the broad-ranging arguments for and against re-allocating water from farming to the urban sphere. However, it is clear that a large-volume reduction of irrigation ows can have powerful, negative impacts on a rural economy in which farming and agriculture-based services, manufacturing and transport are major employers. Water-rights trades can dry up the economy itself. In New Mexico, for example, some acequas have gone to court to block sales because they threaten the community and local livelihoods (Meinzen-Dick, pers. comm.).

9.4.7 The market

The next real life complexity reviewed here concerns the market within which intersectoral trades take place. A physical market-place (such as a particular caf) may exist, or transactions simultaneously involving a number of farmers and urban actors may occur through an open, printed or digital form. Here potential buyers and sellers can meet and gather information on the rights that are in trade. In the absence of such a market, deal-making is likely to be on a bilateral basis. A specic urban actor seeks out a farmer who may be willing to sell his water-rights and they conclude an agreement in the partial or complete absence of information about other potential purchasers and vendors. Here there is a strong likelihood of asymmetric information with the urban actor better informed about the farmers negotiation position than vice versa. The haggling process tips in favour of the water utility or the urban enterprise. Karin Kemper writes (2001): Socio-economic asymmetries need to be considered because of the different types of water users, who will vary in their educational background, culture, and economic power. When water rights were allocated and made tradable in Chile, electricity companies bought up a large number of them to be held for future use, to the detriment of smaller users who at the outset did not understand the implications of their selling the rights. In regions where bilateral deals are dominant there is neither a market nor an equilibrium price. Hearne & Easters valuable study on Chilean water rights markets clearly indicate that the number of intersectoral transfers was low. This was true of the upper Maipo valley, the Azapa valley and the Elqui valley. In the Limar valley transactions are fairly frequent, but the text suggests these are predominantly transfers within agriculture (1995: 326, 503). In contrast a study of the water register of Santiago County in Chile for 19934 indicated there were some 500 permanent, agriculture-urban trades. They equalled only about 0.1% of the total water rights held in the area (Gazmuri Schleyer & Rosegrant 1996: 423). In fact the best quantitative expression for the importance of water-rights trading, past and present, to the water requirements of a city is the volume of water comprised in such trades over time divided by the total volume of water the city receives in the current year.

9.4.8

Transaction costs

These have already been referred to several times and are fully incorporated into the general theory. The legal context strongly inuences the transaction cost associated with any rights transfer. The greater the cost, the less likely it is that a deal will be

struck. Transaction costs include the time and money expended on searching for a buyer or seller of water-rights, negotiating and legalizing a contract, validating the legal ownership of the water-rights, payment of any government tax on transfers, and enforcing the contract. In his Rivers of Gold Haddad (2000) makes clear in a set of fascinating and detailed case studies from California that the management of risk lies at the heart of transaction costs. Note that these costs are those incurred in the transfer of water-rights, not those incurred in the physical capture and distribution of water. The latter is a commonplace supply cost (Merrett 1997: 540).

9.5 CONCLUSIONS
Allocation stress and the policy responses to the dilemmas it poses will be a prime source of debate in the eld of water resources management during the next decades. In the broad sweep of world history, the transfer of water-rights from farmers to urban actors has been one form in which allocation stress has been managed. This has taken place through the sale of such rights or their seizure by force majeure; in both cases such sales have frequently been accompanied by the sale or seizure of riparian land. This paper set out to accomplish two objectives. The rst was to set out an orthodox microeconomic account of a water-rights market in equilibrium. This was done by representing urban actors requirements for farmers water-rights as a demand function based on each urban actors individual maximum bid price. The rural response was represented as a supply function based on each farmers individual minimum release price. The second objective was to ascertain whether this neo-classical approach could deal with a number of empirical complexities in real-life trading. The conclusions are that the theory capably handles not only abstraction rights but also concessions, part-sales, accompanying sales of land, and transaction costs. However, the theorys applicability requires a strong enabling condition: that the legal system fully incorporates such transfers (including a clear stance on third party rights) and that an open, well-informed market exists. But the empirical material cited shows that the absolutely predominant form of transaction is the bilateral deal. Bilateral transactions predominate because the total number of annual sales in a dened market area is usually small, eliminating the possibility of a thriving market process involving multiple sellers and multiple buyers. Rather than a competitive and open market we typically see only one or a small numbers of sellers, that is, a rural monopoly or oligopoly. Similarly we nd only one or a small number of purchasers, a monopsony or oligopsony. The conclusion is that a theory of the neoclassical type does not well represent actual social processes. To understand these processes, research is required into one-off, scattered, fragmentary, arcane deals, often of dubious legality and marked by information asymmetries. However, it seems that the concepts of a farmers minimum release price and an urban actors maximum bid price are still appropriate, as is the account of how these prices are shaped. Similarly, a rural monopolyurban monopsony approach can also deal not only with abstraction rights but also concessions, as well as part-sales, land sales and the transaction costs structured by formal rights, customary practice or legal pluralism.
The material in this chapter originally appeared in: The urban market for farmers water rights. Irrigation and Drainage 52(4): 319326, 2003.

10 The demand for water: four interpretations

In Geometry (which is the only science that it hath pleased God hitherto to bestow on mankind) men begin at settling the signications of their words; which they call Denitions. Thomas Hobbes, Leviathan, 1651

10.1 INTRODUCTION
An extraordinary characteristic of the current management of global water resources is the breadth of disciplines on which it draws. To name but a few one can cite agronomy, civil and hydraulic engineering, economics, environmental science, geology, human and physical geography, hydrology, land-use planning, meteorology, political science and sociology. Most readers of this paper will have received a professional training in at least one of these subjects. Each of these disciplines has its own language and not even a Darwin, a Keynes or a Newton could dream of mastering all of them. Yet every historian of science recognizes how critical is the language of a discipline to its success and how vital the development of the language of science is to sciences progress (Kuhn 1970). As Dow writes (2002: 5) It is normal in scientic thought for terms to change meaning over time. A feature of water resource debates at the present time is the wide application of the term demand. It may be that civil engineers rst gave it real weight in water resource planning. Civil engineers saw themselves above all as responsible for water supply and therefore (this is an hypothesis) assigned the term demand to refer to those outstream ows that began at the point where engineers put into supply the water they had abstracted from lake, river and aquifer.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

86

What is not in doubt is that demand is a concept applied by professionals and policy-makers from many disciplines. It is the objective of this technical note rst to show that the word has a number of quite distinct meanings that it is unhelpful to confuse, and second to show that distinguishing these meanings from one another will strengthen planning practice.

10.2 THE USE OF WATER

The thesis presented here is that there are four demand-side concepts with respect to outstream water: use, consumption, need, and economic demand. Water use is the Table 10.1 The hydrosocial balance for a dened region in 2003. Row 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Title Categories of Supply Rainwater collection Groundwater abstraction Surface water abstraction Desalination Imports of water from other areas Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored Total Gross Supply Categories of Supply (Negative) Supply leakage and evaporation Exports of water to other regions Net rise in water abstracted and stored1 Total Negative Values Total Net Supply Categories of Use Households Agriculture Mining Manufacturing Public services Private services Other uses Total Gross Use Categories of Use (Negative) Leakage and evaporation during use Net rise in water received and stored1 Total Negative Values Total Net Use Base year volumes (Mcm) a b c d e f g h j k l m n jn o p q r s t u v w x y vy

Note: 1 A net fall (rise) in storage of water has a positive (negative) value. Source: Adapted from Merrett (2002a) Tables 8.1 and 8.2.

rst and easiest concept to grasp. A generic way of tabulating the regional users of outstream water is in terms of households, agriculture, mining, manufacturing, public services, private services, and other uses. In each case we are referring to the quantitative ow per unit time of water arriving at a users property (a dwelling, a farm, a mine, a factory, a hospital, an ofce block, etc.). That ow may be used benecially or it may be wasted in various ways. Directly or indirectly, water use is the basis of human life and it makes up one of the two building blocks of the hydrosocial balance in a base year (such as 2003) or in a future scenario year. The generic form of the balance is set out in Table 10.1.

10.3 THE CONSUMPTION OF WATER

During the course of using water, some part of it returns to the hydrological cycle through evapotranspiration; this is the second concept of interest to us water consumed. What is left after the use process is complete is waste water and irrigation returns. These ows are reused internally or externally by user groups, or recycle as run-off to surface water, or recharge groundwater. The consumption of water is of particular interest because the ratio of consumption to waste water plus returns varies signicantly between types of use (Shiklomanov 2000). To take an example, a Colombian sugar-mill in the Cauca Valley that internally reuses its own waste water to avoid abstraction charges seeks a low level of consumption. In contrast, an English farm in the Anglian region irrigating containerized trees has 100% consumption all its water is evaporated or transpired. The ratio of consumption to returns has generated a major debate on irrigation efciency (Perry 1996).

10.4 THE NEED FOR WATER

The third of our concepts is water need. For households water need is a social, cultural and health-related concept; it refers to desirable or recommended levels of use for drinking, cooking, personal washing, disposal of urine and faeces, cleaning and washing in the house and watering of a kitchen garden. In the farm sector water need refers to net irrigation requirements. For mining, manufacturing and the services sectors, their need for water is technically and institutionally determined. So the central difference between water use and water need is that the former represents an actual ow to users whereas the latter represents a desired or recommended ow. Returning to the hydrosocial balance, estimated water need in a baseline year can be compared with water use, possibly leading to recognition of unmet need. For a scenario year, water need can provide the starting-point for planning an expansion in supply.

10.5 THE ECONOMIC DEMAND FOR WATER

The fourth concept is the economic demand for water. For economists the concept of demand refers to the relationship, at a given time and within a dened market,

between price per unit of any product or service and the quantity in each time period that users are willing to purchase at each price. The demand function is conventionally represented graphically with price on the vertical axis and quantity on the horizontal axis, showing the difference in quantity purchased at each price (e.g. Merrett 1997: Fig. 4.1). The cultural and economic context of any demand function is referred to as the conditions of demand. The rst of these conditions is the tastes and habits of users. The second is the price, quality and availability of commodities that users consider to be substitutes for the product. These two together should account for the users willingness to purchase. The third condition is the incomes, assets and access to credit of users, which account for the ability to purchase. Where the conditions of demand are stable, the graphical function can be used to represent not only price and quantity differences within a given time-period, as above, but also price and quantity changes in successive time-periods. This is legitimate only if expectations about the future are stable. Almost invariably, under these conditions, higher prices are associated with lower quantities and so the demand function slopes downward from left to right. A matter of great interest in analyzing the economic demand for water is users responsiveness to price differences. With any given difference (or change) in price, is the response large or small in terms of quantity purchased? One measure of this concept of responsiveness is called the price elasticity of demand and is equal to the proportionate difference in quantity purchased divided by the proportionate difference in price paid. In applying the concept of the economic demand for water, it is necessary for users to face a volumetric price or cost in accessing their supply. For example, on the Palestinian West Bank water utilities such as the Jerusalem Water Undertaking meter their supplies to households, to industry and to urban services. Moreover, rural households and farmers, although not on a networked supply, still face costs varying with volume when they pump groundwater.

10.6 SUPPLY-SIDE LEAKAGE AND EVAPORATION

In Table 10.1 the categories of supply include in row 16 leakage and evaporation losses that occur between the point at which human society appropriates outstream water and the point of delivery of water to user properties. Curiously such losses are almost always treated in the current literature as a form of demand. The result is that consultants forecasts of growth in demand include (as unaccounted-for-water) supply-side leakage and evaporation. Furthermore the reduction of these losses in abstraction, storage and distribution prior to the delivery of water to the user is said to be a form of demand management. This does not seem helpful. A manufacturer of rened sugar, when considering losses from output because of pilfering, or contamination while in the warehouse or destruction in a road or rail accident en route to the supermarket, would never regard this as a demand for sugar, a bizarre act of use by a consumer whom the sugar never reaches. The manufacturer would regard all of these as storage or distribution losses in the supply chain. So should it be with the supply of water. There is a more general point here. Water resource management is now widely seen as principally a form of demand management. To me such an approach seems

nonsensical. This is especially the case when so much demand-management, as with reuse and the reduction of losses, turns out to be made up of supply-side initiatives. With outstream water we should always integrate in our thinking and our practice both demand-side and supply-side strategies.

10.7 CONCLUSIONS
This technical note concludes by illustrating some of the confusions that can arise when the concept of demand is deployed in the catch-all manner here criticized.

To take demand to mean both use and consumption neglects the truth that a given

volume used can be associated with wide variations in the volume consumed. This difference is at the core of much current debate in the eld of irrigation engineering as the discussion lists show. To take demand to mean both use and need obscures the shortfall of use against need for more than half the worlds population. To take demand to mean both use and the economic demand for water confuses the use of water with one of its determinants whenever costs borne by the user vary with the volume used.

If we professionals and policy-makers in water resources management cannot distinguish the use, consumption, need and economic demand for water from one another who will? And if not now, when?

The material in this chapter originally appeared in: The demand for water: four interpretations. Water International, 29(1): 2729, 2004.

11 The political economy of water abstraction charges

11.1 INTRODUCTION
Fresh water is a fundamental necessity for human society for the individual citizen, for households and for all those sectors which contribute to the economys output of goods and services. The principal sources of this water are threefold: rst, directly from precipitation itself, most obviously in forestry and agriculture; secondly, the abstraction of surface water, that is, fresh water from rivers and lakes; thirdly, the abstraction of ground water, that is, from subterranean aquifers. Note that the desalination of sea water, at the global level, contributes only a small proportion of total supply (McDonald & Kay 1988: 834). The focus of this article is the charges levied by government on individuals and institutions for the right to abstract surface and ground water. One objective is to provide a general economic analysis of the subject capable of international application to specic case studies. The second objective is to recommend the most appropriate basis for government abstraction tariffs. The scope of abstraction charging includes water consumers abstracting directly for their own use, such as farmers, mining companies and manufacturers. One also includes water service companies that abstract not for their own use, but in order to provide the public water supply. In the Thames catchment in the UK, for example, the largest abstraction permissions provided by the Environment Agency are for six statutory water companies, plus a number of industrial users, such as Didcot power station, that are direct abstractors. Excluding precipitation itself, direct abstraction is also the main source of water for agriculture in the Thames Valley (Merrett 1997: 1522).

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

91

The instream users of water are not addressed here, such as shing, navigation and leisure, where no abstraction of water is required for such use. Nor does the paper deal with tradeable abstraction rights, such as those in Australia and Chile, because these constitute market transactions between private parties, with government merely setting a regulatory framework (Pigram et al. 1992, Hearne & Easter 1995).

11.2 ABSTRACTION CHARGES AND THE THEORY OF RENT

A useful analytic starting point is to establish that abstraction charges are a form of economic rent. The modern theory of rent was rst developed in the early nineteenth century (Robinson & Eatwell 1974: 1117). Its application was to British agriculture and the common situation where capitalist farmers used land owned by the aristocracy and the gentry, and paid rent for that right. At that time, agriculture contributed about one-third of Britains total output. The classical school argued that the appropriation of land has the consequent effect of the creation of rent. Thus, Ricardo writes: Rent is that portion of the produce of the earth, which is paid to the landlord for the use of the original and indestructible powers of the soil (Ricardo 1821: 67). It is true that we may now doubt that any power of the soil is indestructible. Nevertheless, since Ricardos time, economists have used rent theory whenever they are dealing with a natural resource of economic value, which can be privately appropriated, and which is in restricted supply. Clearly this makes rent theory applicable to natural supplies of fresh water. These ows are of economic value. Rights in their abstraction can be privately appropriated. Lastly, surface and ground water is in restricted supply in any given year. Winpenny (1994: 5467) provides examples of rental markets for water in India and the USA. However, the supposed inelasticity of raw water supply deserves closer scrutiny. Within a catchment, short-term elasticity is perversely high, where substantial stocks are held in a catchments reservoirs or where aquifer stocks are high and ground water pumping capacity is not fully utilized. But it is the elasticity of the planning supply function of raw water over the medium- to long-term which is at issue here. It is certainly true that total rainfall in the catchment less evaporation sets a hydrological constraint on long-term catchment abstraction, as Dubourg suggests (1993: 3). But the recycling and reuse of water weakens this natural barrier whilst the desalination of sea water, and abstracted water imports from another catchment can smash it. The problem is that the wider one casts the net for additional supply, the greater are its unit costs. It is the exponential character of this planning function which maintains the truth of the statement that in the majority of the worlds catchments with substantial populations, abstracted water is indeed in restricted supply (Turner & Dubourg 1993: 5). Let us now return to the classical theory of differential rent. This assumed that competition existed between farmers for access to land, and between landowners in supplying land. In such a situation, it is theorized that land rent on the least fertile tract of land worth cultivating would be zero, and here the farmer would earn the going rate of prot on capital for the the private sectors of the economy as a whole. Land which was more fertile (or better located) would bring the owner a higher ow of rental receipts such that the farmer, after paying this higher rent, would still receive only the

going rate of prot. Rent could therefore be seen as the transformation of the surplus prots of capital, driven by the competition for land of differential fertility. Ricardo had only a limited interest in non-competitive markets in the supply of land for rent, albeit he did describe a case where monopoly rent, rather than differential rent, is paid (Merrett & Sharp 1991: 303). As a consequence, his theory is essentially demand-driven. This neglect of the supply-side determinants of land rent restricts the scope of differential rent theory in its application to abstraction charges levied by government. Such charges exist typically where an agency of the state licences abstraction to specied individuals and institutions, and where no property right exists to abstract ground and surface water without such a licence. Here, property rights on the supply side are assigned by a state monopoly. Thus with government abstraction charges, analysis certainly requires an understanding of the effective demand for water by water service companies and direct abstractors. Differential rent theory is unquestionably valuable here. But analysis also needs a means to interpret the legislative practice of charge-setting by state institutions in different catchments, regions or countries.

11.3 A CHARGE-SETTING TAXONOMY

Having argued that abstraction charges are a form of economic rent, but that differential rent theory does not provide a basis to explain charge-setting by the state, the best way to proceed may be by developing a taxonomy of charge-setting principles. Such principles can then provide the basis for interpretive empirical work in specic catchments, as well as a starting point for policy review of existing legislative practice. Eight charge-setting principles are set out below and each is glossed in turn. No charge. This is the lower limiting case. In the majority of the worlds countries, government abstraction charges simply do not exist. In some cases the argument is that, since surface and ground water are a gift of Nature or of God, the state has no business in taxing it. This may explain the situation in Scotland, for example (Fowler 1995: 178). In other cases, what is lacking is the administrative capacity to levy the tax. A revenue-maximizing charge. This is the upper limiting case. The long-term effective demand curve for water is probably cubic, reecting the absolute necessity for the good, so in principle government could raise all its revenue requirements from this single tax (Merrett 1997: 538). In practice, of course, such a principle is never applied. A market-clearing charge. In countries and regions which are arid, or where levels of water consumption are high in relation to effective rainfall, users may wish to gain access to more water than is available, at least in the absence of inter-catchment transfers. In this case, government may put in place a demand-management policy in which a general abstraction charge is applied which, although it is not revenuemaximizing, does just match the demand by abstractors to the annual ows available. No pure examples exist of this, the closest being the public auction of tickets for xed time and ow in the centuries-old water market of Alicante in Spain (Winpenny 1994: 557). The adoption of demand-management policies, including water resource pricing of various types, is gaining ground on a global scale. For example, Schifer and his colleagues at the German Development Institute have shown for the case of Jordan that supply-x policies are under considerable pressure and that a demand-

management philosophy is taking shape (Schifer et al. 1994: 136). Politically, it is impossible at the present time in Jordan to tax water abstracted for use by farmers, but an abstraction tax on industry has been introduced there. Winpenny argues that one of the worlds longest established and most successful water markets is in the Colorado-Big Thompson scheme, where the Northern Colorado Water Conservation District since the mid-fties has used water pricing to transfer abstraction supplies from agricultural to urban uses, thereby raising allocative efciency (1994: 5961). An environmental regulation charge. In this case one is considering a country which has a well-developed national policy for water resource management. The necessity for a regulator of the fresh water environment is accepted and abstraction charges are levied and hypothecated to nance the costs of regulation. (Hypothecation is a term widely used in the UK to refer to cases where government income from a dened tax is reserved for a specic expenditure category.) A clear example here is the Environment Agency for England and Wales. This institutions riverine management costs on current account are specically recovered by charges made on licensed abstractors under a scheme of abstraction charges approved by the Secretary of State for the Environment (National Rivers Authority, 1993: 78, 1995). The Agency may revoke or revise licences but funds for compensation have to be found from its regional water resources budget (Rees & Williams 1993: 256). An average total cost charge. In addition to regulation costs, the state may bear other costs in making fresh water resources available in a catchment. The abstraction of fresh water can require infrastructures provided by the state, sometimes on a considerable scale, as was the case in Early Mesopotamia. In such cases, a charge may be levied per cubic metre of water abstracted in order to defray in whole or in part the average total cost of these infrastructures. An example here is the abstraction charges levied on farmers in the Goulburn-Murray Irrigation District of Victoria in Australia (Pigram et al. 1992: 91120). Another cost is the compensation it may be necessary to pay to existing holders of abstraction rights, when a catchment authority wishes to modify or rescind such rights in order to achieve the objectives of its catchment management plan. An average total cost charge is equal to total nancial costs of regulation, infrastructural investment, compensation payments and any other miscellaneous items, divided by the total volume of water abstracted during the year. Its introduction would have many opponents amongst water users and water suppliers, despite the clear feasibility of such a charge being scally neutral. Moreover, in the opening up of new areas for irrigation as part of a regional development strategy, a charge on users reecting the full cost of the water supplied to them would deal the coup de grce to any such ambition. An interesting account of successful full recovery of deep tube-well installation and maintenance costs from farmers through abstraction charging is the Barind Multipurpose Development Authoritys experience in Bangladesh since 1985 (Zaman 1996). A marginal cost charge. Some writers working within the orthodox paradigm, such as Dubourg (1995), have considered the appropriateness of an abstraction charge which reects marginal cost. There are two serious problems with such proposals. First, it has to be said that catchment authorities frequently have little or no interest in marginal additions to abstraction capacity. Marginal change is not the point, it is non-marginal changes in capacity that we need to address, as Dubourg himself points out (1995: 8). This moves the discussion, for example, to average cost per cubic metre of additional water abstracted and to the traditional utility price analysis dilemma in which large surpluses are generated when average additional cost exceeds the

average cost of existing capacity and large decits in the reverse case. Secondly, a catchment will have many infrastructural options in terms of scale and of type, capacity increasing, nature conservation, water quality improvement etc. So, many different average incremental costs exist, some of which bring no increase in the volume abstracted or even bring reductions in that volume. The average incremental cost per cubic metre of water in the no-change situation is innite and with a cut in output is negative! A Pigovian charge. Here, the principle is that government should set a charge which is differentiated according to the external costs imposed on society by each class of abstractors (Pigou 1932). K. William Kapp later formulated the parallel concept of social cost, covering all direct and indirect losses sustained by third persons or the general public as a result of unrestrained economic activities. (Steppacher 1994: 440). An example of the kinds of damage done by over-abstraction is that of the Hueco Bolsn aquifer on the MexicanUS border: over a period of 70 years the water table fell by 25 metres, resulting in increased pumping costs, subsidence, and contamination by increased ows of saline and polluted waters into the fresh water source (McDonald & Kay 1988: 513). The second source of external costs is the pollution of abstracted water by use and the subsequent discharge of that water back into surface and ground water sources, at outfalls and even through injection wells. But this second type of externality does not, strictly speaking, arise from abstraction but from consumption. As a general rule, whilst abstraction charges may include a component in recognition of the polluter-pays principle with respect to over-abstraction and, sometimes, pollution by use, the differentiation of the charge on the basis of the monetary evaluation of the external costs never takes place, because of the extreme difculty of measuring them. Indeed, political economy denies even that such measurement has any meaning. As Sderbaum (1994: 197) says: the idea of a correct level of pollution control connected with some level of economic activity where marginal net private benets intersect with marginal external cost are questioned, as is a correct level of (Pigovian) taxes which leads us to the optimal level of externality. While there is a monetary aspect in damage to ecosystems and human health from pollution, the institutionalist questions the reduction of damages and negative impacts to money and also the idea that theory can single out correct prices and correct charges as part of societal valuation. Kraemer (1995: 2312) refers to the overwhelming problems of applying a Pigovian tax in his introduction to the development of German abstraction charging between 1988 and the present. An incentives charge. Incentive charging in a catchment or region can be dened as the use of a water tariff a table of xed charges to give price signals to abstractors which reinforce water resource management based on environmental standards and regulatory controls. The best developed system of incentive charging is probably that of Germany. In contrast, sources within the Environment Agency suggest that abstraction charges in England and Wales are so low, they are incapable of signicantly inuencing abstractor behaviour. An incentive tariff can include the following components:

A licence fee to have a new water abstraction installation approved by the regula-

tory authority. For example, Foster et al. (1993) in their comprehensive introduction to the hydrogeological, legal and administrative aspects of ground water

licensing in Latin America indicate that installation fees are common and are hypothecated. A charge per unit volume. This may be invariable with total volume consumed. Alternatively, it may rise with the volume drawn off. The total charge payable may be based either on the licensed volume or the actual volume abstracted. As Schifer et al. (1994: 15) point out, a signicant drawback of a rising block tariff is that it hits hardest those abstractors requiring large volumes simply because of the size of the farm or factory. It should also be observed that abstraction charging is never imposed where the installations capacity falls below a minimum level set by government. The effect of this is that direct abstraction by households to meet their needs for drinking, washing, cooking and sanitation are outside the regulatory system. A charge reduction for the quantity of water directly returned to surface waters after use. To take the example of Didcot power station again, this has a licence to abstract 142 million litres of water per day. The licence requires 5066% of the water abstracted to be returned to the river, depending on ow conditions. Unit price is lower because of this non-consumptive use. In contrast, spray irrigation provides virtually no return ow to river or aquifer and so is undiscounted. A charge which is greater for higher quality water. This is found in Germany, for water drawn from deep aquifers, where the average total cost per cubic metre abstracted is inevitably higher. A charge which varies with the seasons. The volumetric rate is higher in those months when effective demand is greater and higher too when precipitation is less. A charge which is greater for certain locations. These include upstream sources, because the length of the river exposed to abstraction impacts is greater; rivers, lakes and aquifers most threatened by past or present overdraft; and regions with lower effective rainfall.

11.4 ABSTRACTION CHARGES AND SUSTAINABLE CATCHMENT MANAGEMENT

In this section, the subject of abstraction charging is related to the debate on sustainability, beginning with the approach of Richard Dubourg (1993, 1995), whose contributions to the study of hydroeconomics are within the neoclassical framework and are widely inuential within the UK. Dubourg suggests that aggregate sustainability is a situation where natural capital as well as non-natural capital are non-declining and specically where critical capital such as water, within the natural capital category, is non-declining. From these denitions he deduces impeccably that aggregate sustainability is consistent with catchment management policies in which abstraction of surface and ground water is equal to effective rainfall, that is, the entire hydrological ow known as runoff, dened by hydrologists as total precipitation in a catchment less evaporation (McDonald & Kay 1988: 819). It has to be said that the adoption of such a denition of sustainable abstraction is extraordinarily dangerous. Abstraction at a rate equal to runoff is certainly capable of leaving surface and ground water stocks unchanged over the course of a year the critical capital stock is, indeed, maintained. But a rate of abstraction at this maximum rate would have the effect of capturing the entire river ow at some point or points in the catchment. In effect, Dubourgs

denition ignores completely what in the US is termed the environmental need for water in river systems all on the basis of three mathematical constraints! In Introduction to the economics of water resources (Merrett 1997: 1478) I have suggested an alternative, multi-dimensional approach. After dening the concept of a sustainable society, it is suggested that water resource planning in such a society has six principal elds of action:

the protection of waters hydrocyclical capacity to renew its ground and surface water ows and stocks; the conservation of societys species and natural habitats in all their fresh and salt water environments; the husbandry of water in its supply and use; the supply of fresh water sufcient to meet the biological, cultural and economic needs of societys human populations; the purication of water from domestic, agricultural and industrial efuents; the drainage of storm water, and the protection of rural and urban communities
against oods. Abstraction charges can contribute positively to the rst three of these elds and below it is shown how this can be done, using what I shall call full incentive charging. The preparation of a tariff scheme should take as its starting point the regulatory controls put in place by the catchment authority in carrying out its responsibilities. In the development of these controls, there is an important role for social cost effectiveness analysis and social cost benet analysis. These techniques have a place, alongside environmental impact assessment, in comparing alternative regulatory options. The design of environmental regulation is, then, in part an economic process as well as an ecological one. The tariff rgime introduced to any specic catchment (or region) will be contingent on its physical geography, its habitats and species, its human settlement patterns, its economy and the power relationships which hold between various social and economic groups. So what we require are criteria for tariff design which can be applied in the appropriate way to any single area with all its unique characteristics. I propose three such criteria for full incentive charging. First, the annual income from abstraction charges should be hypothecated to the environmental regulator such that, when added to the income received from discharge fees, shing licences, navigation permits etc., abstraction fee income is sufcient to cover all the states capital and current account expenditures on environmental regulation, water infrastructures, research and database development, compensation payments, etc. The justication for this criterion is that water users, who in the nal analysis are certain to bear the charge, should pay the full nancial cost of water provision. In providing the regulator with hypothecated income, it is likely to increase their relative power as an agency of government. As Kraemer (1995: 238) writes of Germany since 1988: On the whole, the water resource taxes contributed to capacity building within the water management administration in the German Lnder and thus partly overcame the implementation decit in water resource management. It will also provide a budget to nance the legal costs of modifying or terminating abstraction licences which threaten the hydrocycle or undermine nature conservation. Higher abstraction charges will also give a price signal to water companies and direct abstractors to reduce their storage and distribution leakage between the points of abstraction and the users gate.

Secondly, where the full cost abstraction tariff still leaves an excess of demand for abstracted water over its licensed supply, the charge should be raised so that market clearing takes place. Thirdly, specication of the components of the abstraction charge should provide incentives for abstraction behaviour that is economically efcient and that avoids environmental degradation. Price per unit volume should be invariant with total volume abstracted, unless there are countervailing economic or environmental arguments. Here, one accepts the argument in section three above by Schifer et al. against a rising block tariff. Price should be discounted where abstractors recycle their off-take to surface or ground water sources. (Discharge fees should be used to handle the water quality aspects of recycled water.) Charges should be higher for abstracted water of higher quality. Seasonal variations in effective rainfall and economic demand should be dealt with through the licensed volume provisions laid down by the regulator and by the market clearing criterion. Charges should be higher for upstream sources and for abstraction in locations where species and habitats are more threatened by abstraction. Charges should be higher in parts of the catchment (or region) where waterrelated infrastructural costs borne by the state are higher and should be higher to reect the greater infrastructural costs generated to meet peak demands.

11.5 THE IMPACT ON USERS

What can one say about the impact of full incentive charging on the purchase of abstracted water? Here the analysis is best conducted in terms of the big battalions water companies, industry and agriculture. Water companies with a statutory responsibility to provide the public water supply would readily acknowledge that the current abstraction charges they face are passed on to nal consumers. In any case, such charges make up only a small part of total annual costs. The annual expenditures of a water company are dominated by the cost of capital, staff wages and salaries, electric power and materials such as chlorine (Merrett 1997: 12834). So even a two- or ve-fold increase in the abstraction charge per cubic metre of water would have little effect on the price to consumers of the public water supply. The principal response to a change in the charge is likely to be a review by the company of the cost effectiveness of reducing the leakage rate in its production and distribution systems. The response of consumers and of the water company would, nevertheless, reduce the quantity the company seeks to abstract. Unfortunately, no empirically based estimation of the elasticity of demand for water abstracted by statutory companies exists, as far as I am aware. Of course, in countries such as the UK where most households pay a xed charge for water consumed, not a price per unit volume, there would be no demand response whatsoever. Now let us consider the case of direct abstraction by industry, such as in power generation, mining, construction, manufacturing and the services sectors. As a general rule one can say that where industrial companies use only small quantities of water, their reduction of quantity purchased in response to full incentive pricing is likely to be negligible. Rees et al. (1993: 17, 40) have suggested that the costs of water to UK manufacturers are usually a maximum of 0.2% of total costs and that industry exhibits a very low level of knowledge of waters volumetric price even when purchased from water companies. However, where industrial processes require large

volumes of water, there may be a strong interest in water reuse within the company, with a consequentially large fall in the quantity which needs to be abstracted. Reuse also has the advantage of cutting rms fees for the discharge of industrial efuent. Throughout the world, water is also directly abstracted by individual farmers or by organizations acting for them. Water for irrigated agriculture constitutes more than 90% of total consumption in a large group of low-income countries (Kinnersley 1994: 181). In these cases, full incentive charging may have a powerful impact on farmers budgets. Where farmers incomes are low and full incentive charging would make water unaffordable for them, and where government or a delegated agency wishes to continue to subsidize the agricultural sector, such subsidies should be transparent. Where full incentive charging is applied, it is likely to stimulate husbandry in the application of water as well as the production of crops with high water productivities, as Schifer et al. (1994) and Gleick et al. (1995) have shown for Jordan and California respectively.

11.6

FINAL REMARKS

I wish to end this paper by suggesting the institutional framework which would be most appropriate for full incentive charging. These ideas have been strongly inuenced by Karin Kempers The cost of free water (1996). The basic approach is a negotiation model in which there exists a public sector catchment agency that, through a negotiating forum, develops its policies with the advice of water companies, direct abstractors, environmental organizations and water user associations representing the domestic sector, irrigated agriculture, mining and manufacturing, etc. The original prototype for such negotiation models is the French water parliaments (Tuddenham 1995). Central government retains the statutory right to determine which public and private institutions may enjoy the right to abstract water, on what scale, in what locations, at what time of year, at what price but it delegates such rights to the catchment agency. The agency, with the assistance of its partners in the negotiating forum and with a full understanding of existing customary rights, then assigns formal abstraction rights on a time-limited basis to specic abstractors. The time limit would be 10 years, let us say, rolled over each year except where the agency allows the licence to expire. Where these 10-year rights need to be modied or rescinded for hydrological, environmental or economic reasons, compensation would be payable to the abstractor so affected. Rules would also exist to address Third Party impacts. Abstraction rights assigned to abstractors could be freely traded between abstractors provided the agency had approved such transfers in the light of their environmental impacts. Full incentive charging would be the basis of the price paid by abstractors for their water and has the objectives of underpinning environmental regulation with a hypothecated income source, of requiring abstractors (and therefore nal users) to cover the full costs of regulation and state infrastructural provision, of bringing the quantity of water demanded by abstractors into line with regulatory limits, and of giving price signals which promote both allocative efciency as well as abstraction practices which avoid damage to riverine eco-systems. No charge would be made for abstractions below a minimum scale; the transaction costs of such charges would be high relative to the volumes abstracted.

Abstraction infrastructures would be constantly reviewed. Where they require rehabilitation or new investment, this may be carried out either by the agency or the abstractors themselves. In the former case, the nancial costs would be reected in the agencys charges, as already indicated. Arrangements would be made to monitor abstraction with respect to its location, time and quantity, as well as to invoice abstractors, to collect the charges owed and to enforce all agreements. Such transaction costs would be included in the full costing tariff of abstraction charges. The creation of this institutional framework imposes social, economic and political costs on the parties concerned, that is, structural costs of change both real and nancial in their nature. Therefore the negotiating forum may agree that it is sensible for full incentive charging to be phased in gradually. One nal issue is briey considered here. It is argued above that regulatory control is the context within which abstraction charging is deployed. This may suggest that the environmental constraints are designed and specied and only thereafter is the incentive tariff formulated. In practice, work on both should proceed simultaneously and dialectically, because regulation and environmental pricing are interdependent activities within a single planning process.

The material in this chapter originally appeared in: The political economy of water abstraction charges. Review of Political Economy, 11(4), 1999.

12 Twelve theses on the cost and use of irrigation water

12.1 THESIS 1
In the eld of water resource management a widely held belief exists that allocation stress is found in many parts of the world and is set to become more intense in the future both because of climate change as well as global growth in population and income. Allocation stress refers to access conicts between the agricultural, domestic, industrial, urban service and environmental uses of outstream and instream water ows. If the belief that allocation stress will intensify is well founded then, because of the dominant role of irrigation water use at the global level, it is worthwhile exploring the possibility of slowing the growth of farmers use of such water.

12.2 THESIS 2
To some degree the deceleration of irrigation water use can be achieved by raising the cost of water to farmers. The higher cost may lead them to reduce the cultivated area under irrigation, or to favour crops that use less irrigation water, or to apply water more productively, or some combination of these three actions.

12.3 THESIS 3
Raising the cost of water to farmers is unfavourable to them as a class. It is likely to reduce their gross margins, or require investment in water-saving technology, or reduce the market value of their holdings, or in some cases drive them out of business. Farmers can be expected to resist the new cost policy more-or-less forcefully.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

101

12.4 THESIS 4
In all cases, today and since the dawn of irrigation in Mesopotamia, farmers costs of irrigation water have taken one of four forms: own-supply cost, volumetric price, indirect fee or revenue-only fee. Hybrid types also can be found. Own-supply cost refers to any case where the farmer provides (at a cost) his own supply of irrigation water, such as in the case of building and maintaining his own well. Volumetric price refers to those cases where a farmer purchases irrigation water at a price per unit volume. Indirect fee indicates the many and varied cases where a farmer pays for his water under rules in which the cost he bears has an indirect relation with the volume used. Revenue-only fee indicates cases where a farmer pays for his water but where no relation exists between the payment and the volume of irrigation water used.

12.5 THESIS 5
Own-supply costs, certainly with respect to their prime cost component, vary with the volume of water used by the farmer, that is, the farming familys or agribusinesss eld supply. This binding of cost and volume may be loosened by implicit subsidies, particularly cheap energy. Here the cost of irrigation water to farmers can be raised by reducing energy subsidies. It can also be increased by the introduction of abstraction charges.

12.6 THESIS 6
Volumetric price also binds cost and volume. But the price may fall short of the fullcost of the eld supply. Indeed it may fall short of operation, maintenance and management costs. Here the cost of irrigation water to farmers can be raised by increasing tariffs. However, the introduction of volumetric pricing faces grave difculties: ow meters require delivery through a pipe resulting in head loss that may reduce water availability at the farm level; and the cost of installing meters, monitoring them and collecting fees from smallholders may be prohibitive.

12.7 THESIS 7
Indirect fees are charges for irrigation water that have an indirect binding of cost and volume. They include charges that vary with the farmers irrigable area, or for a water turn of given time duration, or with crop composition where wet crops incur heavier charges than dry crops. Indirect fees are often accompanied by a quota regime. Here again the cost of irrigation water to farmers can be increased by raising tariffs.

12.8 THESIS 8
Revenue-only fees are those that do not meet the binding principle but merely act to raise income for an irrigation authority. They can be used to reduce irrigation water use only by driving farmers out of business.

12.9 THESIS 9
Where the binding principle operates (see Thesis 5) the possibility exists that an increase in irrigation water costs may lead to a deceleration in the growth of use (or a fall in use) in comparison with the no-cost-increase alternative.

12.10 THESIS 10
A policy to decelerate signicantly irrigation water use by raising its cost to farmers requires that a minimum absolute value of the elasticity of demand exists, for example a value of 0.5. Such a minimum value requires some combination of the following conditions:

The binding principle operates. The farmer exercises control over the eld supply volume. Irrigation water costs are a visible component of prime costs to the farmer. The farmer understands that a proportion of the water applied with the current

technology (and at specic stages of the crop year) enjoys a low return in terms of crop yield. The farmer understands that new techniques are available that will improve the productivity of water applied, such as its gross margin efciency. The farmer understands that reducing the volume of irrigation water used does not materially increase the risks s/he faces. The farmer understands that the losses of non-crop returns from irrigation water reduction are not large. Crops are not feed crops, that is, they are not cultivated for the farming families subsistence, or for feeding to their livestock, nor are crops grown by agribusiness as an input to its own manufacturing processes. Under conditions of general (even if modest) ination, the cost of water to the farmer must outstrip the wholesale price of the crop.

12.11 THESIS 11
At the basin level the good sense of water management policies tackling allocation stress by reducing abstraction for irrigation holds only if the water thereby released actually becomes available for domestic, industrial, urban service of environmental uses.

12.12 THESIS 12
The reduction of the base supply to the irrigation sector reduces the return ows of water to farmland, river, lake and aquifer. In some cases these lost returns consequent on diminished abstraction can wipe out the advantages of reallocation.
The material in this chapter originally appeared in: Twelve theses on the cost and use of irrigation water. Irrigation and Drainage, 51(3): 265268, 2002.

13 Behavioural studies of the domestic demand for water services in Africa

there is no such entity as a distinctive type of structural explanation in the social sciences; all explanations will involve at least implicit reference both to the purposive, reasoning behaviour of agents and to its intersection with constraining and enabling features of the social and material contexts of that behaviour. Anthony Giddens, The Constitution of Society, 1984

13.1

A METHODOLOGICAL REVOLUTION

The purpose of this paper is to contribute to the modern history of hydroeconomics. From the mid-1980s a very small group of university academics, and professionals from the international banks, launched a methodological revolution in the study of domestic water and waste water services in the low-income countries. This revolution resulted in a permanent transformation of the way the water resource professions approach the household sector there. Here I wish critically to review this literature in order to strengthen future research. The paper does not cover techniques for forecasting future price and volume outcomes based on willingness-to-pay eld work; it is limited to the study of existing behaviour. Willingness-to-pay surveys are the subject of a future article. The three principal authors of this change were Dale Whittington from both the University of North Carolina at Chapel Hill and the World Bank, Xinming Mu of the Asian Development Bank, and Donald T. Lauria, also from the University of North Carolina. Other authors were Briscoe, Roche and Wright of the World Bank, Choe, Hughes, Okun and Swarna all at Chapel Hill, Okorafor and Okore at the University
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

104

Table 13.1 Seven behavioural studies of African domestic water services. Authors Whittington, Lauria, Okun, Mu Whittington, Mu, Roche Mu, Whittington, Briscoe Whittington, Lauria, Mu Whittington, Okorafor, Okore, McPhail Whittington, Smith, Okorafor, Okore, Liu, McPhail Whittington, Lauria, Choe, Hughes, Swarna, Wright
Source: see references.

Case study location Ukunda, Kenya Ukunda, Kenya Ukunda, Kenya Onitsha, Nigeria Nsukka, Nigeria Nsukka, Nigeria Kumasi, Ghana

Year of eld work 1986 1986 1986 1987 1989 1989 1989

Year of publication 1989 1990 1990 1991 1990 1992 1993

of Nigeria, Liu and Smith at North Carolina State University, and McPhail at Johns Hopkins University. The key publications are listed in Table 13.1. As can be seen, the eld work was carried out in the period 19869 in Africa, and the papers were published in the period 198993. Four areas were researched: Ukunda in Kenya, Nsukka district in Nigeria, Onitsha in Nigeria and Kumasi in Ghana. The focus in these rst three areas is water services; in the fourth it is waste water services. Whittington and his colleagues show a strong interest in policy, suggesting that the low-income countries in general have suffered from repeated failures in programmes devoted to domestic water and sanitation needs. There are simply too many leaking taps, abandoned water systems, and defunct village water committees for anyone to be sanguine about the current rate of progress (Mu et al. 1990: 521). It is forcefully argued that these weaknesses ow from traditional approaches that are out of touch with present demographic and nancial realities (Whittington et al. 1993: 733) in which designs for new systems are generally made and projects constructed with little understanding of household water demand behaviour (Whittington et al. 1991: 179). Master plans of sewers and treatment plants as well as engineerdominated supply-side philosophies are criticized for their neglect of the demands of beneciaries. The view, ascribed to van Damme & White (1984) that design specications should not require households to pay more than 35% of their income on water and waste water services repeatedly comes in for attack. In fact the setting of water charges within the 35% bracket goes back at least as far as 1975 (World Bank 1975). As a consequence, the authors argue that the way forward is to undertake detailed case studies of families actual water use behaviour and their observed ability and willingness to pay for water services. This information can provide a well-grounded basis for estimating the uptake of new programmes (Whittington et al. 1989, 1990a). In their rst Nigerian study, Whittington et al. conclude (1991: 194): Most people in Onitsha are already paying high prices for water from the vending system for service which is inferior to that which could be provided by a well-run piped distribution system; and the local water and sewerage authority should view itself as a regulated utility, not as an agency providing a social service. In what follows, I consider rst the principal insights of these valuable contributions by Whittington et al. and their colleagues and then develop a critique of some aspects

of their work. Two tables that may be useful for future studies are proposed: an Attribute Matrix and a check-list of some information requirements for household surveys.

13.2 MARKET NETWORKS FOR WATER

The water services in any region can be represented as a three-layer cake. The bottom layer is the forms of access to fresh water; the middle layer is the modes of distribution of the volumes initially generated; and the top layer is the pattern of uses of water. Taking all categories of use together, the total access volume less evaporation and leakage provides a net supply mathematically identical to the total volume distributed which, after again deducting losses, is identical to the total volume used. These identities provide the type of internal accounting checks I have employed in developing the concept of the hydrosocial balance (Merrett 1997, 1999, 2002). The rst strength of the water demand school that has been introduced above is its demonstration of how socially complex these networks of access and distribution can be. The 1986 study of the village of Ukunda with its population of 5000 demonstrates this well. Some households queue to collect their water directly from local wells and hand-pumps, performing all three roles of abstractor, distributor and user. In other instances a very small number of households have piped water to their homes sourced by the local water utility. In the third case kiosk owners receive a piped supply which is then sold directly to households who carry it back to their dwellings in 20-litre jerrycans. In the fourth case kiosk owners sell on the water to retailing vendors who themselves then distribute it to households front doors, using carts or bicycles to carry the jerrycans (Whittington et al. 1989). The water demand schools work on the cities of Kumasi and Onitsha, with their populations of 600,000 and 700,000, respectively, showed each had its own distinctive access-distribution-use system. In these two urban cases, water tankers played a major role. In demonstrating these social complexities, Whittington and his colleagues also contribute greatly to our understanding of the economics of the supply processes. The second strength of these research studies is their detailed examination of the dynamic and competitive markets in water that exist in many of Africas villages and towns. In the Ukunda and Onitsha case studies it was possible to draw a chart of the access and distribution network to households, to collect data on the unit price for each branch of the network, to estimate the volumetric ows of abstraction and distribution and therefore to calculate the turnover along that branch. Table 13.2 illustrates the case of Onitsha. With respect to households there are 11 relevant access/distribution branches. The variation of volumetric ow between branches is high. For example, the lowest dry season branch ow (shallow well collection by doorstep vendors) is a mere 10,000 imperial gallons per day (2200 litres/day), only 0.3% of the highest ow (private borehole sales to tanker trucks) at 3 million g/d (660,000 l/d). Unit prices are also marked by their high variance. Shallow well collection incurs no monetary charge, although the collection time (the time spent queuing, drawing water and the journey to and from the well) can be lengthy and exhausting. Even limiting our scope to priced supplies, there is still a marked divergence. For example, the unit price of the public piped supply at 0.003 naira/gallon (0.0007 n/l) is only 2.4% of the price charged by doorstep vendors to households at 0.127 n/g (0.028 n/l).

Table 13.2

Water and money ows in Onitsha 1987 dry season. Water volume (million gallons per day)1 0.30 3.00 1.00 1.70 1.55 0.05 0.05 0.01 0.11 0.30 1.50 9.57 Unit price (naira/gallon)2 0.027 0.003 0.020 0.020 0.050 0.040 0.020 0.000 0.127 0.000 0.003 0.018 Turnover (naira) 8000 10000 20000 34000 78000 2000 1000 0 14000 0 5000 172000

Access/distribution branch

Household collection from private boreholes Private borehole sales to tanker trucks Truck sales to households Truck sales to shops Shop sales to households Shop sales to doorstep vendors Private borehole sales to doorstep vendors Shallow well collection by doorstep vendors Doorstep vendor sales to households Shallow well collection by households Public piped supply to households Total or weighted average3

Source: Whittington et al. 1991, Figure 1. Notes: 1 One imperial gallon 4.546 litres. 2 In 1987 the rate of exchange of the US$ to the naira was 1:4.3. 3 Tables of this branch type inherently embody double-counting. The total supply to households is 4.76 million g/d and total turnover from sales to households is 125,000 naira.

The most extraordinary aspect of the monetary data, for those unfamiliar with African water markets, is that the proportion of water supplied to households by the public sector Anambra State Water Corporation in the 1987 dry season was 32% of the total but its share of total turnover was only 4%. In Onitsha, Whittington et al. also note the powerful effects on the market of the changing seasons, dry and wet. The main difference is that in the rainy season households manage to capture 2.3 million g/d (10.5 million l/d) of rainwater with the effect that their purchases from the water vending system fall to about half the dry season total. Rainy season prices are lower, of course, with the downward shift in the demand schedule. Seasonal variation in the structure of the market is also observed in Ukunda, where rainy season prices are half those of the dry season, as well as Nsukka district. The third strength of the water demand school is their success in estimating the proportion of domestic income absorbed by water purchases. In Ukunda annual per capita income is put at US$350 in 1986, annual domestic expenditure on water per head is about US$30, giving a per capita outlay/income ratio of some nine per cent. Those who use vendors enjoy a high level of service; good quality water is delivered to their doorstep on demand. Although vending does not provide a level of service comparable to house connections from a well-run piped distribution system, it is far superior to that available in most rural communities in Kenya. This high level of service is, however, expensive (Whittington et al. 1989: 164). In Onitsha in 1987, average annual household income was about US$1600, although about 25% of households received less than US$600. Average household size was six to seven people. Whittington et al. do not present a single value for the outlay/income ratio over the whole year and for all income classes, but their published information

allows this to be estimated at about 12% in the dry season, 5% in the rainy season with an annual average of about 8% (Whittington et al. 1991). What is also clear is that the poor pay a far higher proportion of their income on water: 18% in the dry season and 8% in the rainy season. In this case the income elasticity of demand for water services is positive but less than one. In the past, it has been commonly assumed that households could only afford to pay 35% of their income for improved water services (ibid. 189). But, the data from Ukunda and Onitsha, as well as studies by Linn (1983) in Ethiopia, by Fass (1988) in Haiti, and by Cairncross and Kinnear in Sudan (1991, 1992) refute this assumption completely. For Nsukka district and Kumasi the treatment of the outlay/income ratio is much slimmer. In the former case, annual water purchases from vendors are probably between 6% and 10% of household income, but the estimate is judged to be highly speculative (Whittington et al. 1990, 1903). In the latter case, average household expenditure on water and sanitation combined is about 3% of average household income but about 10% of households spend more than 8%. Abraham Maslows lexicographic approach to a ve-level hierarchy in human need satisfaction receives conrmation here: life-or-death need for the service, high service cost and low household income combine to give a staggering proportion of income devoted to water and waste water services (Maslow 1954). This conveniently brings us to the fourth strength of these authors research: an innovative application of the access/distribution chart (see above) specically to sanitation (Whittington et al. 1993). Kumasis 600,000 persons are estimated to fall into ve groups: those using public latrines (38%), those using bucket latrines (25%), those with access to a WC emptying into a septic tank (25%), those using traditional pit latrines (7%) and those who urinate and defecate on open ground (5%). The monthly total of this feculent mass is some 25,000 m3, which goes to open street drains, streams, neighbourhood dumps and landll. Unit costs are estimated for the use of alternative sanitary facilities as well as the ow of funds to sanitation actors, including rent payments to the local Committees for the Defence of the Revolution. The authors conclude (ibid. 745): Our survey of sanitation conditions in Kumasi revealed an appalling and, from a public health perspective, dangerous situation. Households are currently generating about 25,000 m3 of human waste per month (including ush water for WCs), but only about 10% of it is removed from the city. The rest, 90%, is left in the urban environment until it decomposes, is carried away by small streams or drainage ditches, or dries and becomes airborne.

13.3 THE USES OF WATER

Above I have reviewed the principal strengths of the water demand school in these studies published in 198993. What follows is a critique of some aspects of their methodology, beginning with the treatment of water consumption. In terms of economic theory the demand for a product (or service) can be dened as the relationship, at a given time and within a dened market, between price per unit of the product and the quantity in each time period that consumers are estimated

to be willing to purchase at each price. Demand is conventionally represented graphically, with price on the vertical axis and quantity on the horizontal axis, showing the difference in quantity purchased at each price. The cultural and economic context of any demand function is referred to as the conditions of demand. The rst of these conditions is the tastes and habits of consumers, that is, the nature of need for the commodity. The second is the price, quality and availability of commodities that consumers consider to be substitutes for the product. These two together should account for the consumers willingness to purchase. The third condition is the incomes, assets and access to credit of consumers, which account for the ability to purchase. With respect to the rst of these conditions of demand, the nature of the need for water is complex because, even when we limit our interest to the domestic sector, the uses of water are extraordinarily diverse. This accounts for the huge variance in household water consumption in litres per head per day (lhd) within and between societies. At the most basic level, there is a biological drive, a physiological necessity, because water is the basic solvent necessary for all the bodys functions, both at the biochemical and the cellular level. This suggests that a lexicographic classication of the domestic uses of water is appropriate (Earl 1995: 412). This could be:

drinking; cooking; washing ourselves; disposal of urine and faeces; cleaning and washing in the house; garden watering; cleaning and washing of a car and lling of a swimming pool.
In general in the low-income countries, the rst step in understanding use is to distinguish instream use (such as bathing or washing clothes in a pond) from outstream use. The greater the availability and convenience of instream supply, the less will be the demand for outstream water in the home. The second step is to exclude from household use the water that is applied for the irrigation of crops and the watering and cleaning of livestock. At the centre of an urban area this is unlikely often to be an issue, but on the urban periphery or in villages it can give misleading totals for the use types listed above. The third step is to take a view on the total volume of water used by a household per time period (allowance made for seasonal change) and its proximate distribution between the lexicographic uses. The fourth step is to examine what relation exists between the multiple uses of water and its multiple sourcing; for example high-priced, high-quality water may be reserved for drinking and cooking whereas turbid or malodorous water may be limited to waste disposal or garden watering. The fth (and not the least important) step is to determine which members of the family use the water. Commonly it is the women who not only fetch water but who also wash the children, cook, clean the house and wash the clothes. It is they who best understand the complex choices to be made about water supply and water use and it is they who provide the best source of information on household behaviour. The sixth (and here the nal) step is to examine household water use efciency, the scale of reuse and how water is stored in order to balance the temporal uctuations of supply and demand. Use efciency is likely to correlate directly with unit cost, particularly when user and collector are the same person. The ellipsis of the water demand school is that no account whatsoever is given in the published papers referred to in Table 13.1 of how African households in general,

or their survey respondents in particular, use water. Simply put, to understand the demand for water requires, as a necessary condition, that the researcher knows how it is used. Whittington and his colleagues fail to meet this condition.

13.4 OBJECTS OR SUBJECTS?

The water demand school is open to a second methodological attack, related to but separate from their neglect of water consumption. I shall call this their naturalistic bias. Given their policy interests, Whittington et al. need to understand why it is that households make specic decisions on the sources from which they choose to gain access to water, the volumes they purchase, and (as above) how water is used. A useful technique in such explanation would be an attribute matrix of the type set out in Table 13.3, but such a matrix is never employed. Their approach is to collect a large dataset that covers, for each household, social variables (such as the number of adults and children in the household), economic variables (such as the households total monthly cash income) and water practice variables (such as the volume of water accessed from each alternative source). These classicatory and quantitative datasets are then used as inputs to econometric modelling. Table 13.4 provides a possible template for such information requirements, including coverage of the uses of water. An alternative approach, that would itself have provided the basis for designing an attribute matrix and designing an econometric model, would have been set within the framework of academics such as Veblen (1919) and Simon (1959, 1992). Peter Earl (1995) provides an excellent review of this paradigm. This approach would regard water access and use decisions by households as a process of problem-solving during which their views of the world and their wants may undergo considerable evolution, yet where at the date of any cross-section survey domestic behaviour takes the form of routinised, habitual action (Hodgson 1994). The door to the households mind-set and actions in this regard would have been through semi-structured interviews with its members in which Africans speak for themselves, they voice the constraints they may face, they articulate their own accounts of what their routines are and why they satisfy a familys needs (and change with the seasons) rather than alternative habitual actions. Household members, in this case, are regarded as active, intelligent, resourceful, purposive, reexive subjects, not merely as objects exhibiting quantiable behavioural patterns. A quotation from Earls critique of orthodox economics applies with full force to the water demand school (1995: 61): Neoclassical economists working on consumer behaviour have preferred to steer well clear of this approach Instead, they have tended to conne themselves to studies of choice in relation to characteristics that are easy to measure in supposedly objective terms. Anthony Giddens in The Constitution of Society (1984) captures the method of authors like Whittington to submit it to a higher level of ontological critique. He suggests that social scientists who use frameworks that treat human subjects as objects are committing a naturalistic error. They turn their back on the active, reexive character of human conduct. Giddens also rejects orthodox social theorys vision of human behaviour as the result of forces that actors neither control nor comprehend. He denes the naturalistic standpoint as that which looks to the natural sciences, including biology, as a model for social science.

Table 13.3 An attribute matrix for household water services. Attribute Price per litre Payment arrangements1 Collection and queuing time per trip Odour Turbidity Other Pressure qualities Service times Service reliability

Source

Household collects rainwater Household collects from house well Household collects from other well Household collects from stand pipe or pump Purchase from doorstep vendor Purchase from doorstep tanker Purchase from shop or kiosk Piped supply to house

Such as a xed monthly charge or volumetric price collected at the time of sale.

Table 13.4 Water and waste water services for households: information requirements. The Survey Area Map including residential locations and water sourcing points (wells, kiosks, etc.) Number of resident persons Number of resident households Brief review of the survey areas economic base, its social character and its transport links to external areas of employment Annual rainfall Seasonal rainfall Note: in cases of strong seasonal differences, data will be required separately for each season Household and Water Use Information Date/time/place of interview Position in family of person(s) interviewed Household size Gender and age composition of household Gender and age of persons collecting water Education level of the household head Occupation of all adults in household Total household income per week (alternatively: total weekly expenditure) Total weekly expenditure on water services Address/location within the survey area Ownership of water-using xtures: bath, shower, WC, sink, dish-washer, washing machine, garden-watering items Volume of abstracted water used in litres per head per day (lhd) Composition of use (%): cooking and drinking; bathing, showers and personal washing; washing of clothes, house, vehicles, etc.; sanitation; garden; other Types of use that are specic to types of source (if any) Information on water reuse: waste water source to type of second use; no. of cycles, volume of reuse as multiple of initial use Water storage: type of storage by volume Attribute Matrix for Water Services See Table 13.3 Waste Water Type of service: water closet with piped connection to sewer; own latrine; communal latrine; septic tank; open areas Type of treatment at point of service Means of removal of urine and faeces: ushing to sewer or septic tank; collection from latrine; direct to lake or river; other Charges for: service, collection, treatment and disposal Total weekly expenditure on waste water services Other Information Owners/managers of water sources and waste water services Action taken by owners/managers in event of non-payment Water and waste water tariffs and billing arrangements established by source owners/managers Type of dwelling, including storey height in case of ats Number of rooms Total weekly expenditure on power (gas, electricity, other fuels).

13.5 THE DISCRETE CHOICE MODEL

We have already seen that in Ukunda and Onitsha a considerable volume of quantitative data had been collected from households and market institutions. Modelling household behaviour could have been carried out by designing the equations in the light of social theoretical, semi-structured household interviews a subject-oriented approach. Whittington and his collaborators neglect of the reexive character of household members did not permit a design of the model in this way. In practice they rejected multiple regression equations and opted for a conditional multinomial logit model where the dependent variable is the probability that any specic water source of several available is chosen by the household. This probability is expressed as a function of the price of water from each source, its collection time, its taste, household income, education of the household head and the number of women in the household. The model was used in the Ukunda study where waters price and collection time were both highly signicant (Mu et al. 1990). On rst sight, the approach seems useful. But on careful consideration it appears to rely on the assumption that households choose only one of the alternatives available. This does not square with the fact of use of multiple sources by each household, amply illustrated in the Nsukka district. Nor does it address the volume of water taken from one or other source, the traditional dependent variable in modelling household water demand. Nor (as we saw above) is it based on an understanding of water use, the dog that does not bark in Whittingtons world. The rationale for the orientation to a discrete choice model is clearly policy-based: in an African village the rst element in a water demand model must be a description of the likelihood, under different conditions, that a consumer would choose to use an improved source rather than continue to use the existing traditional source(s) (ibid. 521). The decision to use the discrete choice model, containing as it does no quantityprice relation, also has the regrettable outcome that Mu et al. cannot pursue mainstream interests such as the nature of the demand function for water (linear, exponential, constant elasticity, or cubic?) and the values of the demand elasticities. The principal challenge offered by multi-sourced water, where cash price per litre can be traded off against collection time per litre, is how to incorporate subjective perspectives into a modelling framework that seeks to explain price and volume data and recognises explicitly:

the multi-source environment; the simultaneous use by households of more than one source; water quality differences; the proportion of household members who are adult women; the female labour force participation rate in the survey area; the average hourly income for women from wage-labour or self-employment in
the local market economy. There is a further difculty with the water demand schools approach and this can be illustrated by a quotation from their rst publication (Whittington et al. 1989: 165): information on the water vending system can serve as a useful indicator of a communitys ability and willingness to pay for a piped distribution system.

The fact that yard taps do not already exist throughout Ukunda indicates an inability on the part of the community or water authority to mobilize resources, not an inability or unwillingness of the population to pay for the cost of the improved service. But such a conclusion is hazardous. Whittington and his collaborators here fail to recognise the force of two key economic characteristics of the system they describe:

The tariff for water in every case is volumetric, whether the cost is in collection time or in Kenyan shillings. The penalty for unwillingness or inability to meet these costs is an immediate cutoff in supply. Only in the most unlikely case that a piped system is metered and where billing sanctions are draconian could a smooth transition from the existing multi-source system to a piped network be guaranteed. In the absence of these necessary conditions, one would expect a simultaneous fall in turnover, and a surge in volume consumed if service levels are capable of meeting the demand. It is only in their 1990 publication on Nsukka district that the critical importance of volumetric pricing is recognized, following trenchant criticism of the performance of the Anambra State Water Corporations cost recovery practice.

13.6 CONCLUSIONS
This paper provides a critical review of the early publications of what I call the water demand school with respect to their behavioural studies of the domestic demand for water and waste water services in Ukunda (Kenya), Onitsha and Nsukka district (Nigeria) and Kumasi (Ghana). The schools work has resulted in a permanent transformation of the way the water resource professions approach the domestic sector in the developing countries and has had a major impact on policy formation in institutions such as the World Bank. The strengths of the work of Whittington et al. and their collaborators are fourfold. They demonstrate how socially complex are the access/distribution networks in the villages and towns of Africa. They give a rich account of volumetric ows, unit prices and turnover along each branch of the network and indicate the dynamism and competitiveness of local water markets. They provide data on the ratio of water service outlays to household income which knock out expectations that these ratios should not exceed an affordability quotient of 5%. They extend their source/distribution network analysis with great success to waste water disposal with respect to volume, unit price and turnover. The weaknesses of the water demand school are also four-fold. First, to understand the domestic demand for water we need to know how it is used in the household; but these seven studies give little account of how their survey respondents consume the water they access at such great cost relative to their incomes. Secondly, the authors take a naturalistic standpoint in which the research into these communities behaviour is limited to the actions of silent men, women and children, moving across a landscape like so many ants in a natural-historical study of a savannah colony. Household members are not treated as intelligent, resourceful, purposive and reexive citizens. Thirdly, the conditional multinomial logit cabbala chosen for the

econometric work is hardly transparent, side-steps traditional and justiable interest in demand functions and their elasticities, and abstracts completely from the uses of water. Fourthly, in their policy work in Ukunda, Onitsha and Kumasi, the water demand school seems to forget that amongst the fundamental attributes of the systems they describe are: costs to the user which are volumetrically determined, and immediate closure of access to water in the case of a household which is unwilling or cannot meet those supply costs. The principal conclusions of this paper are that behavioural studies into the domestic demand for water and waste water services in the low-income countries should be set within the complex access/distribution networks of a quantitative, whittingtonian type and the price information that these embody; this research should be based on semi-structured interviews, usually with adult females, seeking from these reexive subjects their account of the origins and character of their serviceaccess routines; the investigation should incorporate the scale and composition of use and reuse and their relation to the quality of water; and the econometric work should act as a servant to this methodological approach.

The material in this chapter originally appeared in: Behavioural studies of the domestic demand for water services in Africa. Water Policy, 4(1): 6981, 2002.

14 Deconstructing households willingness-to-pay for water in low-income countries

14.1 INTRODUCTION
During the early 1980s growing disquiet began to be expressed over the outcomes of publicly-funded water and sanitation programmes for households living in lowincome countries. Many systems had fallen into disrepair and had been abandoned; others were used by only a small part of the original target population; the proportion of costs recovered was low; and expansion to meet the needs of growing populations was too slow (MacRae & Whittington 1988: 247). As time passed, a new paradigm was broached and then developed by what has been called the water demand school. It suggested that embodying the expressed preferences of individual households is critical to successful project design. It rejected the view that households can afford to pay no more than 35% of their income on fresh and waste water services. The new approach highlighted the widespread existence of complex and vigorous water markets. A disciplinary shift took place away from sanitary engineering with its supply-side emphasis to the economic analysis of the demand for water. Finally, the new paradigm suggested that domestic supply programmes could be sustained and replicated only if they responded to the potential market for water services. The mood of the time is well-illustrated in the following quotation (Whittington et al. 1990a: 294): If rural water projects are to be both sustainable and replicable, an improved planning methodology is required that includes a procedure for eliciting information
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

116

on the value placed on different levels of service, and tariffs must be designed so that at least operation and maintenance costs (and preferably capital costs) can be recovered. A key concept in such an improved planning methodology is that of willingness to pay. If people are willing to pay for the full costs of a particular service, then it is a clear indication that the service is valued (and therefore will most likely be used and maintained) and that it will be possible to generate the funds required to sustain and even replicate the project. The purpose of the present paper is to critically review the willingness-to-pay concepts use in the 11 studies listed in Table 14.1 and thereafter to set out the research and policy implications of this critique. As the table shows, the case studies were carried out in the Caribbean, Africa and Asia over the 9-year period from 1986 through 1994 and published in various academic journals in 1988 through 1998. What they have in common, in addition to addressing households willingness-to-pay for water services in the low income countries, is that Dale Whittington of the University of North Carolina is one of the joint authors in every case. Without a shadow of doubt Whittington is the doyen of this branch of scholarship and so these eleven studies provide a comprehensive account of the new paradigm, making them an appropriate subject for critical review.

14.2 SURVEY METHODS

Before initiating the main argument, it is useful here to sketch the lineaments of the willingness-to-pay surveys carried out in Laurent, Onitsha, the Punjab, Nsukka, Kumasi, Davao, Calamba and Lugazi (see Table 14.1). The pre-requisites of a survey are a budget, the professional staff to plan and carry out the work, identication of the study area, the recruitment and training of local staff as enumerators, and securing whatever institutional support is necessary for the research to go forward. With the team in place, the interview questionnaire is designed and pre-tested, usually Table 14.1 Eleven studies of the willingness-to-pay for domestic water. Authors MacRae, Whittington Whittington, Briscoe, Mu, Barron Whittington, Lauria, Mu Altaf, Whittington, Jamal, Smith Whittington, Okorafor, Okore, McPhail Whittington, Smith, Okorafor, Okore, Liu, McPhail Whittington, Lauria, Wright, Choe, Hughes, Swarna Choe, Whittington, Lauria Whittington, Choe, Lauria Davis, Whittington Whittington, Davis, McClelland
Source: see bibliography.

Case study location Laurent, Haiti Laurent, Haiti Onitsha, Nigeria Punjab, Pakistan Nsukka, Nigeria Nsukka, Nigeria Kumasi, Ghana Davao, Philippines Calamba, Philippines Lugazi, Uganda Lugazi, Uganda

Year of Year of eld work publication 1986 1986 1987 1988 1989 1989 1989 1992 1992 1994 1994 1988 1990 1991 1993 1990 1992 1993 1996 1997 1998 1998

drawing on discussions with local families or community leaders. The sampling technique is explored and selected; this is always ad hoc because the statistical universe of households is never known. Inevitably, the sample does not conform to the rigorous demands of statistical theory. However, the view is rightly taken that it is better to have surveyed imperfectly than never to have surveyed at all. The individual interview may be a once-for-all event; or an initial interview can be followed a day or two later by the collection of answers to specic questions where the household is given time to think more carefully about their response; or the interview is followed up later to give respondents the chance to revise their rst reply to key questions. Close supervision of the enumerators is advisable. In Lugazi the survey was followed by a community meeting to discuss the issues raised in the questionnaire (Davis & Whittington 1998). Fieldwork duration varied widely: from as little as two weeks in Onitsha and Lugazi to ve months in Kumasi. The number of households interviewed ranged from 170 in Laurent to 1200 in Kumasi. The population size of the survey areas was as little as 1500 persons in Laurent, up to 700,000 in Onitsha where the rapid reconnaissance survey period was 10 days (Whittington et al. 1991: 1802). In a World Bank review of techniques for systematic client consultation, based on a background paper by Whittington and Davis, 24 weeks eldwork is recommended (Owen 1994: 16). The survey questionnaires differed between case studies, of course, but their content can be considered to fall into six categories. First is the date of the interview and the address of the household or its location in the study area. Secondly, we have the age, gender and household status of the respondent. Thirdly, household information would be collected such as the number of persons and their age/gender, their health, education and occupation, and household income, expenditure and assets. Fourthly, data are collected on the quality of the dwelling, whether it is owned or rented, and its costs. Fifthly comes the sources of the water used in the dwelling, its quality (but not its quantity), conicts over access, seasonality of supply and demand, the price or charge payable for each separate fresh or waste water source used, household expenditure on water per time period, water storage facilities in the dwelling, current sanitation facilities, the uses to which fresh water is applied, and the degree of satisfaction with the existing service. Sixthly, a bidding game (or a referendum technique) would be used to establish the households willingness-to-pay for alternative water or sanitation projects and this would require some kind of description of the new scenario.

14.3 SIGN AND BEHAVIOUR

The deconstruction of the willingness-to-pay can now get under way. The principal propositions to establish immediately concern critical ontological features of the work reviewed, that is, the 11 studies of Table 14.1. The (im)precise meaning of the willingness-to-pay will be addressed in Section 14.4. But whatever sense is agreed upon, it should be clear that the willingness-to-pay captured by Whittington and his colleagues in their eldwork is a sign, or rather, a collection of signs. These are: the oral reply or replies given by the respondent during the bidding game; the written record of these replies noted down by the enumerator on his answer sheet; the digital information stored on the professionals laptop which becomes manifest on his screen, after transcription from the answer sheet; and the printed information produced in the teams reports and any subsequent publications.

At a quite separate ontological level we have the behaviour of the household in accessing water or in using sanitation facilities at some later date in those cases where the project scenario set out in the original interview is implemented. Our specic interest is in the actual payments made by households for those realised projects fresh and waste water services. So we have sign and behaviour. I believe that the most interesting and important relationship in this policy eld is the degree of consonance which exists between the willingness-to-pay and market behaviour in those cases where scenario projects are in fact implemented. The remainder of this paper explores the conditions which must be met for such sign-behaviour consonance to exist.

14.4 DEMAND THEORY AND SURVEY PRACTICE

The rst condition for consonance between sign and behaviour, word matched by deed, is that the willingness-to-pay for a projected fresh or waste water service is both a meaningful concept from a social science perspective and that the survey respondent understands that meaning. In economic analysis, whether of the institutional school (within which this paper is located) or the neo-classical school, the demand for a water service can be dened as the relationship, at a given time and within a dened market, between price per unit of the service and the quantity in each time period that consumers would purchase at each price. Demand is, of course, conventionally represented graphically with price on the vertical axis and quantity on the horizontal axis, showing the difference in quantity purchased at each price. Almost invariably higher prices are associated with lower quantities and so the demand function slopes downwards from left to right; it can be represented as a linear, log-linear, exponential or cubic function. The cultural and economic context of demand can be analysed under three rubrics. The rst of these is the tastes and habits of consumers, that is, the nature of need for the service or product. The second is the price, quality and availability of services and commodities that consumers consider to be substitutes. These two together should account for the consumers willingness-to-purchase. The third rubric is the incomes, assets and access to credit of consumers, which account for the ability-to-purchase (Robinson & Eatwell 1974: 149, Merrett 1997: 538). There are a number of serious problems in the application of demand theory to the survey practice of Whittington and his colleagues. These are: the use of a maximum willingness-to-pay price (wtpmax), the lack of clarity in the way affordability is dealt with, the treatment of substitutes, and the relevance of sanctions in the case of non-payment of water service bills. I shall deal with each of these in turn. In all of the studies listed in Table 14.1, it quickly becomes clear that the object of the bidding game section of the questionnaire is not to elicit an answer to a question of the sort: In the case of project X going ahead, would you be willing to pay price Y per drum for the water supplied? In fact, what the bidding game elicits is an answer to a question such as: In the case of project X going ahead, what is the maximum price you would be willing to pay per drum for the water supplied? So the entire approach seeks data on wtpmax.

There is a puzzle here which, to solve, requires us to distinguish between two fundamental ways in which water service payments can be made. The rst is by means of a unit price, that is, a payment based on the quantity of water received, or (say) for each visit to a public latrine. The second is by means of a xed charge per unit time period for access to a service, such as a yard tap connection to a public supply network. In the rst case a maximum price simply has no meaning; the household adjusts downwards the quantity purchased the higher the price. In the second case, however, there certainly will be a maximum charge above which the user declines the use of the service. This crucial distinction is not discussed by Whittington and his colleagues with reference to wtpmax; moreover, Whittington neglects the distinction in his behavioural studies of household demand in African settlements (Merrett 2001). It would also be theoretically meaningful to ask what would be the maximum unit price a household would pay per month for a specic volume of water received. But the water demand school never does this. As a consequence, wtpmax is not a meaningful concept from a social science perspective if it is applied to water demand analysis of projects where payment is based on unit prices. Even in the case of projects based on a xed charge, where wtpmax is meaningful, the questionnaire design must elicit answers in terms of a xed charge, not a unit price. Moreover, if the baseline payment mode is of a unit price type, and the scenario payment is a xed charge, the capacity of the respondent to respond to the wtpmax questions in the bidding game will be diminished because of the unfamiliarity of that payment mode for water services (see also Owen 1994: 16). It is unsurprising then that the World Bank Water Demand Research team states (1993: 49): it was hard to convey the notion of what was meant by the maximum an individual would be willing to pay. A respondent in Haiti asked an enumerator, What do you mean the maximum I would be willing to pay? You mean when someone has a gun to my head?. Elsewhere Whittington himself has made the same point (1998: 22). It is worth noting with respect to both the baseline and the scenario situations, that the willingness-to-pay literature shows that a key advantage of unit price payments is that they alone give the household a degree of control over its water service expenditure during the course of the year, by means of varying their daily consumption. Such control can be valuable to families whose incomes are low and variable (Whittington et al . 1990b: 1907). It is a curious feature of the water demand schools work that, whilst it has a fascination for the price of water, it says virtually nothing about the quantity purchased by individual households nor how it is used (Whittington & Swarna 1994: 32, 60, Merrett 2001).

14.5 THE AFFORDABILITY QUESTION

The previous section pointed to the conventional distinction between the willingness-topay for a commodity and the ability-to-pay for it. Of course, this is familiar to the water demand school (Whittington 1998: Table 1). However, the distinction is only once made explicit to the households interviewed in the surveys listed in Table 14.1. In practice, the concept of the willingness-to-pay is used by the water demand researchers to mean both willingness- and ability-to-pay; the separate concepts are rolled up in one. Unfortunately this opens the real possibility that some respondents will understand willingness-to-pay as inclusive of affordability, and others to take it as exclusive.

So their responses become ambiguous, we do not know quite what they mean. In a 1994 report to the Asian Development Bank, Whittington and Swarna admitted as much in their discussion of eld methods that use a question of the type: What is the maximum amount of money you would be willing to pay (for a specied good or service)? They write (Whittington & Swarna 1994: 50): If respondents could always provide accurate, reliable answers to such a question, this clearly would be the preferred question format. Unfortunately, for a variety of reasons, this often seems not to be the case. It is often difcult to convey the notion of the most (or the maximum) that one would be (freely) willing to pay, that is, able to pay if willing to do so. Some respondents misinterpret direct, open-ended questions to mean What is the most you would like to pay? or What is the most you think you should pay. Both of these nuances are clearly not what is meant to be conveyed. With clarication like this, who needs confusion? In the eld, the research method seems to have been as follows:

Laurent: an explicit reference to affordability in the bidding games last question. Onitsha, Punjab, Nsukka: no reference to affordability. Kumasi, Davao, Calamba, Lugazi: no explicit reference to affordability, but
strongly implicit.

14.6 THE TREATMENT OF SUBSTITUTES

In Section 14.4 it was suggested that the rst condition for consonance between sign and behaviour requires that the respondent understands the meaning of the wtpmax question. It was also pointed out that in demand theory the willingness-to-pay for a service or product is shaped by the price, quality and availability of services and commodities that consumers consider to be substitutes. In fact, in the low income countries there is often a rich variety of means of access to water services for households. This is illustrated in Table 14.2 for the case of Table 14.2 Water and money ows in Onitsha 1987 dry season. Water volume (million gallons per day) (1) 0.30 1.00 1.55 0.11 0.30 1.50 4.76 Unit price (naira/gallon) (2) 0.027 0.020 0.050 0.127 0.000 0.003 0.026 Turnover (naira) 8000 20000 78000 14000 0 5000 125000

Access branch

Household collection from private boreholes Truck sales to households Shop sales to households Doorstep vendor sales to households Shallow well collection by households Public piped supply to households Total or weighted average

Source: Whittington et al. 1991, Figure 1. Notes: 1. One imperial gallon 4.546 litres. 2. In 1987 the rate of exchange of the US$ to the naira was 1:4.3.

Table 14.3 An attribute matrix for household water services. Price per litre Odour Turbidity Other qualities Payment1 Collection arrangements and queuing time per trip Pressure Service Service times reliability

Source

Household collects rainwater Household collects from house well Household collects from other well Household collects from stand pipe or pump Purchase from doorstep vendor Purchase from doorstep tanker Purchase from shop or kiosk Piped supply to house

Such as a xed monthly charge or volumetric price collected at the time of sale.

Onitsha. There are six means of access for the citys households. Volume used varies from as little as 0.11 millions of gallons per day (0.5 millions of litres per day) in doorstep vendor sales to 1.55 million (7.05 million litres) in shop sales. The prices households are willing and able to pay vary from as little as zero naira/gallon (n/g) for shallow well collection by households, through 0.003 n/g (0.0007 n/l) for the public piped supply up to 0.127 n/g (0.028 n/l) for doorstep vendor sales, a variation of 1:42 in terms of the positive price range. This creates some difculty for the wtpmax question. The project to which the question refers can certainly be clearly described using attributes of the type set out in Table 14.3. But the research team is never certain that the project will be delivered with the attributes the team describes to the household; and the team is never able to provide an attributes matrix for the substitutes for that projects service. As a result the respondent household cannot know at the time of the interview the nature of the real choice that will become available to it. It cannot therefore make the kind of commitment necessary for consonance between word and deed. So the true preferences that the water demand school explicitly seeks in response to the wtpmax question do not exist (Davis & Whittington 1998: 7, Whittington & Choe 1992: 57, Whittington et al. 1992: 212, 1993: 1544).

14.7 SIGN AND SANCTION

There is a special case concerning the meaning of the commitment by a household to pay the wtpmax. A household may wish to enjoy a service, be able to afford it and be willing to pay for it, yet when push comes to shove, the household does not pay for it. This occurs in many, many countries and in my own eld experience is widespread in, for example, Armenia, Bangladesh and Peru. The phenomenon is found where there is no sanction for non-payment of the water bill. It invariably takes place only where the service is provided by a public sector institution; private companies that cannot collect on their invoices go out of business. The optimum conditions for this species of non-consonance to breed and multiply exist: where economic life is hard so that households need to take the greatest care over their domestic expenditure; where there is a widely-held view that certain public services should be free; where persons or parties in political life give their support to non-payment; where the quality of the public service is poor; where the government is so manifestly corrupt that payments for public services are known to line the pockets of the power elite; and where neither the government nor the public water utility is willing to exercise sanctions against non-payment because of the likely political and/or public health consequences. Two examples are worth citing from the work of the water demand school. In the Nsukka district of Nigeria, the Anambra State Water Corporation had nanced borehole drilling but was unable to recover even the operations and maintenance costs of borehole pumping (Whittington et al. 1990b: 1905). In the Punjab nonpayment of the at monthly charge was estimated at 40% but Households that do not pay are seldom disconnected from the system (Altaf et al. 1993: 1905). In Yerevan, the capital of Armenia with a population greater than one million persons, survey research that I reviewed in 1999 indicates that about 85% of households do not pay their monthly bill to the public sector Yerevan Water and Sewerage Enterprise. Box 14.1 gives a real-life account of the situation of one family I interviewed.

Box 14.1

The price of water in Yerevan.

The ofcial version Family Alpha lives in Omega district in Yerevan. It is composed of paternal grandmother, father, mother, son and daughter (the respondent). They are owner-occupiers of a small house which is 90 m2 in size. It has one cold water connection to the Yerevan Water and Sewerage Enterprise supply and the interior points of water use are one ush toilet, one bath, one shower, one kitchen sink, one washing machine, and one wash-hand basin. There is no dish washer and no garden tap. The dwelling receives no hot water supply from any exterior source and is kept warm with an oil heater. Family Alpha also purchases one 0.5-litre bottle of mineral water each day. Their water supply is not metered. They consume 250 litres per head per day, a total for the family of 1250 litres. They pay per month per person 420 drams for their water, 2100 drams per month. The payment is made in cash at the municipal area ofce. The unit price of water is 55 drams per cubic metre. The true story Much of the ofcial version is true but there are some twists. There are only three residents because grandmother has moved to the country and the son is living in Moscow. The family has not reported their departure to the Central Police Register because of the legal difculties this could create for the Alphas. They have estimated for our study that, as best they can guess, the familys real daily consumption is 350 litres, not 1250. Mother is very careful by nature in husbanding the familys resources and sees that everyone else toes the line. So the cost to the family of the fresh and waste water service is, in fact, 196 drams per cubic metre, well over three times the ofcial version. The family is seriously considering going over to a metered supply, especially with rumours that the ofcial price will rise under the new International Water Operator which will begin operations in the year 2000. After moving to the Echmiadzin area, grandmother has decided to stop paying her local bills there: her neighbours say no-one else pays their water bill and there are no sanctions from the water utility. The service is lousy and they are all charged for water they do not get.

14.8 PRIVATE AGENDAS

Sections 14.414.7 above dealt with the rst condition for consonance between sign and act, that the willingness-to-pay for a projected fresh or waste water service is a meaningful concept from a social science perspective and that the survey respondent understands that meaning. The second condition relates to the effects the respondent seeks in giving his or her answer. In a sibling paper to the present one, critically reviewing behavioural studies of the domestic demand for water services in Africa, I have argued that household members interviewed by water resource research teams should be regarded as active, intelligent, resourceful, purposive, reexive subjects, not merely as objects exhibiting quantiable behavioural patterns (Merrett 2001). If this approach is accepted, it follows that such subjects will certainly have private agendas of their own when replying to wtpmax questions. Indeed, early criticism of what is widely called the contingent valuation approach of willingness-to-pay studies, that it is inferior to market data or virtually useless or that responses are self-seeking, reect a deep-seated concern with private agendas (Whittington et al. 1990a: 294).I suggest that research in this eld indicates four categories of respondent in respect of private agendas: the cynic, the strategist, the diplomat and the idealist. The cynic is a person who has little faith in human sincerity and goodness. He (or she) believes the research is a waste of time and the research team a waste of space.

He is likely to refuse the interview, or give an unreasonably low bid price. Cynicism may be specically directed at government institutions and their performance rather than a sweeping condemnation of all humanity. The presence of the cynic has been identied by Whittington and his colleagues. For example, the 1986 study of Laurent had a 14% non-response to interview questions on the wtpmax for standposts and 25% non-response for private connections (Whittington et al. 1990a: 302). Many of these were probably cynics. In the 1989 Nsukka district study one half of the sample was re-interviewed one day after their rst interview in order to see whether they wished to revise their bid. Almost one-fth did do so, mostly downwards, and these downward revisions lay in the range of a 5380% cut in the wtpmax. The research team took the view that this was due to a lack of trust by respondents in the government water utility to deliver and maintain the project. Indeed, there was an abandoned elevated storage tank and a capped borehole in one of the villages where interviews took place, bearing witness to public sector incompetence (Whittington et al. 1990b: 19067, 1992: 21920). In the 1992 work in Calamba, the researchers suggested that the perceived ineffectiveness of government programmes in the Philippines led respondents to give a higher wtpmax for projects with a lower scope (Whittington et al. 1997: 232). The strategist takes the view that her or his wtpmax response carries a commitment to the project and is likely to inuence its outcome. She determines that her response should be one most favouring the interests of her family and (perhaps) the local community. Most commonly there is a bias to low prices in these cases, where the aim is to hold down the tariff the water utility will eventually set. However, respondents may also veer to high prices if they believe this is the best way to ensure the project goes ahead. My reading of the Laurent case study is that downward strategic bias was present: the survey offered a standposts project with no risks for individual households, and they were happy to respond positively to this but offered a low wtpmax to keep down the tariff. Similarly in Onitsha, where the scenario was based on volumetric pricing, the wtpmax was about 0.2 n/g for the rainy season, less than one-sixth of the price charged by distributing vendors (Whittington et al. 1991: 196). In Nsukka, the timeto-think opportunities brought lower bids; there, strategic behaviour was recognised by the research team as one possible explanation, the others being the search for a just price or the sobering effects of time to consider (Whittington et al. 1992: 220). The diplomat is aware that behind the enumerator lies a group of highly-paid professionals who have own in perhaps from North America or Europe and who have the support of local bigwigs and the state government. These are people making offers it is unwise to refuse. So the diplomat gives answers which he (or she) believes to be those the enumerator wants and where any hint of what might be a favoured response is rapidly accepted by the interviewee. The clearest examples of diplomat bias come from surveys in Indonesia (Whittington 1998: 8). Referendum techniques for determining wtpmax are especially likely to give evidence of diplomacy. When a referendum elicitation procedure is used, respondents themselves do not need to be asked a question about the maximum they would be willing to pay for a proposed good or service. Instead, split-sample techniques with variation between the samples in the referendum price quoted to the interviewee can be utilised to randomly selected respondents (ibid. 22). The statistical distribution of yes answers over the price range can give fat tails, that is, an implausibly high number of persons who say that they are willing to pay the higher prices.

The idealist is a woman (or a man) who regards the enumerators and the research team behind them as good, serious people; she believes the survey could contribute positively to her family and the local community; accordingly she gives it her best shot, answering all the questions put to her as accurately as they permit. The Laurent survey came across one person of almost Franciscan charity whose wtpmax was so high that it exceeded 10% of local annual per capita income, on the grounds not that he would use the new CARE fountains but because the community needs it (MacRae & Whittington 1988: 255). Contingent valuation studies often have a moralistic tone when referring to the requirement that respondents, having true preferences, should reply truthfully and honestly to the questions designed for them by the research group. The working assumption is that those surveyed are idealists but that survey design and socioeconometrics are worthwhile in testing for strategic bias, for example. A more openminded approach would be to consider respondents as reexive actors composed of cynics, strategists, diplomats, idealists and, indeed, of persons of mixed afliation, as was clearly the case in Nsukka. Survey design and analysis, in this case, would include an attempt to identify the proportion of actors who fall into these agenda categories.

14.9 ANCHOR PRICES

In Section 14.3 above it is noted that the willingness-to-pay captured in eld surveys is a sign, in contrast to the behaviour constituted by paying for water where a scenario project described in the survey is actually delivered. There it is suggested that consideration of the determinants of consonance between sign and behaviour is the most valuable contribution of methodology, and hermeneutics in particular, in this eld of hydroeconomics. Here it is worth stressing that a research teams modelling of its own survey results is not a determinant of consonance, for the relationship between word and deed cannot be shaped by such research. For that reason the socioeconometrics of willingness-to-pay studies (the production of signs interpreting signs) is given little attention in this paper; the modelling of wtpmax responses deserves a critical review of its own. The deconstruction in Sections 14.414.8 of the willingness-to-pay literature begins by arguing that the willingness-to-pay a maximum price has no locus in the theory of the demand function for water unless it refers to a specic volume received per time period. Pari passu, it lacks any clear meaning for a respondent who faces a volumetric pricing tariff for the scenario project. wtpmax can have a practical signicance only when it applies, for example, to a monthly xed charge for access to a water supply. Even here the results should be represented as that xed charge and not as a unit price, for the only sense attributable to the latter in this case is the statistical ratio of the xed charge to the unknown volume of water used, a ratio the user never computes. The next attack (in Section 14.5) is that rolling up ability-to-pay and willingnessto-pay together into the wtpmax concept, but then rarely clarifying this explicitly, creates for the respondent household an ambiguity at the heart of the survey process. Whittington and his colleagues never say so, but perhaps their unwillingness to make the affordability issue explicit is founded on the fear of thereby encouraging strategic bias.

The third broadside (in Section 14.6) is that the expression of the households true preferences in a wtpmax is not empirically possible since neither the research team nor the household know what substitutes enjoying what attributes may compete with the scenario project over its working life. The next critique (in Section 14.7) is that, where the appropriate conditions exist, a family may commit itself to a wtpmax, desire, be able to afford and enjoy the service but not in fact pay for it. The necessary condition for such behaviour is that nonpayment attracts no penalty. The fth and nal blow (in Section 14.8) aimed at the literature under review is that, if we recognize survey households as intelligent, resourceful, reexive actors, then the response to the questions supposedly eliciting wtpmax will be pregnant with the private agendas of the persons interviewed. Consider a question of the type: We have explained our work and described the proposed project; can you now tell us what is the maximum price you would be willing to pay for this service? If the preceding, ve-fold critique of willingness-to-pay studies is well-founded, one would expect that answers to such a question to be the outcome of confusion, uncertainty, disbelief and very private motivations. As a set, the responses would be incoherent, providing no useful wtpmax results. In fact, at least in the published studies, respondents answers are coherent enough to be statistically modelled and tested, for example for the presence of private agendas as well as the variation of wtpmax with household income or assets. However, I suggest that the tractability of the data to numerical testing, providing at least some plausible results, springs precisely from the fact that the type of question posed in the preceding paragraph is never asked. Instead we have a bidding game where the enumerator presents a plausible wtpmax and then gives the household the chance to move above it or below it. Whittington and his colleagues in their publications almost always record the structure of the bidding game (e.g. Whittington et al. 1990a: 302, 1991: 196, 1992: 2223, 1993: 15578). In the papers listed in Table 14.1 the water demand school repeatedly suggests that a bidding game format is more familiar to respondents because it is like the bargaining process taking place in local markets. This is an absurd claim. Market bargaining concerns a transaction which, once successfully concluded, is immediately implemented; it touches on a commodity lying within the customers reach; and the potential purchaser never, ever, reveals the maximum price she is willing to pay. The breach of these routines and conventions, with near-fatal results, is the subject of the famous haggling scene published by Chapman et al. (1979). The use of anchor prices of this type in contingent valuation research, particularly with respect to environmental projects, takes place precisely because, without such a method, the valuation question generates absurd and intractable responses with high variance in the value of the independent variable, wtpmax. Elsewhere I have published another critique of the methods application to water resource projects, in that case those seeking to protect the Norfolk Broads from ooding (Merrett 1997: 1704). There the attack focused primarily on the counterfactual nature of contingent values. It is only a slight exaggeration to say that contingent valuation provides model answers to its own questions. More precisely, in the context of the present paper, such

research generates questions eliciting a single, quantitative answer, when the subject demands a complex textual response. In order to successfully elicit a coherent set of signs from the sampled population, the researcher provides the interviewees with anchor prices. It may be that it was recognition of the leading question character of anchor prices that led some willingness-to-pay studies to try the referendum approach referred to in Section 14.8. The severe drawbacks of the referendum method in low-income countries are exposure to diplomat bias, loss of condence by respondents and enumerators in the procedure when posing extreme values in the interview, and confusion and even anger in tight-knit communities when different families are assigned widely-varying prices (Whittington 1998: 46). Whittington & Swarna (1994: 8) have written: the project analyst does not actually have to know why an individual values a project, or even which of the many consequences or effects of the water project that an individual cares about most. It is enough to have a measure of the strength of the individuals preferences for the project; it is assumed that the individual knows his (or her) own interests and is the best judge of what the project is worth to him (or her). If this is true, why is it necessary to provide him (or her) with an anchor price?

14.10 CONCLUSION
This review of eleven studies in the willingness-to-pay literature is now complete. Next I shall present my own positive views on the demand-side applications of economics to the water resource challenges facing developing country NGOs and governments. In spite of the critical nature of this paper, the policy prescriptions set out below, for the design of fresh and waste water projects aimed at the household sector, do not mark a clean break with the new paradigm described in Section 14.1. Rather, they are a contribution to the evolutionary development of that paradigm. In the rst place, the design and development of water and sanitation projects in a local area need to be based upon a good understanding of the existing local market for these services. The water demand school, particularly Dale Whittington, have made a great contribution here to our comprehension of these hydrosocial processes from the point of abstraction to the moment of water purchase by the household. This now needs to be complemented by an attempt to grasp how much water is used by the family, by whom and to what purpose (Merrett 2001). Without this baseline behavioural groundwork, projects in the scenario year are exposed to a greater risk of failure. Secondly, our understanding of future household behaviour requires semistructured interviews primarily with families female adults. It is women who play the primary role in the collection and purchase of water, as well as its internal use in cooking, washing and cleaning. For this reason alone, the professional team should include a female sociologist or socio-economist. In large surveys the bulk of the data collection will be done by trained enumerators (ideally women). But the professionals in the research team should play the central part in at least a sub-sample of the interviews. This also calls for national professionals to form part of the research team. Surveys where enumeration is the sole responsibility of local secondary-school gradu-

ates, but where data analysis and report writing is the preserve only of international staff, should be strictly avoided. Furthermore, respondents should be recognised not as objects sourcing numerical signs but as intelligent, purposive, reexive subjects, rich in their textual accounts of their neighbourhood and their region. One should also appreciate that respondents are made up of persons with a variety of private agendas: cynics, strategists, diplomats and idealists. Thirdly, the scenario project option(s) can be developed prior to the survey or, with the sampled population, by the research team itself. In either case the expressed preferences of the targetted groups are vital, as the new paradigm insists so eloquently. The project option should be costed in terms of capital investment as well as ongoing outlays. The use of wtpmax in the survey questionnaire and the employment of anchor prices would cease, to be replaced by a question of the type: We have explained our work and described the proposed project, including the means that households would pay for it. To cover its full costs the project would require a price of x naira/gallon. Can you now tell us whether your family would be willing to pay and able to afford x naira/gallon? Where appropriate, a monthly xed charge of y naira per family would replace a price of x naira/gallon. If the project would be nanced and approved only on the basis of a full-cost tariff, the survey respondent would be informed of this. This approach would face the same difculties as the current paradigm in characterizing scenario substitutes (vide Section 14.6). The nal report would stress that the project is likely to succeed nancially only if there are penalties for the non-payment of water bills. Points 13 suggest the professional team for this twilight zone would be composed of a statistician/data analyst, a female socio-economist and a public health engineer.

The material in this chapter originally appeared in: Deconstructing households willingness-to pay for water in low-income countries. Water Policy, 4(2): 157172, 2002.

15 Industrial efuent policy: economic instruments and environmental regulation

Falstaff: Page: Sirrah, you giant, what says the doctor to my water? He said, sir, the water itself was a good healthy water; but, for the party that owed it, he might have more diseases than he knew for. William Shakespeare, Henry IV, Part Two, Ii

15.1 INTRODUCTION
The uses of outstream water in human society can be classied into three grand orders: by households for domestic purposes; by farmers for the irrigation of land; and by industry in the widest sense of the term. The industrial uses of water include heating and cooling processes, steam production, washing and cleaning, factory-scale cooking, incorporation in the product itself as in beer manufacture, the drinking water of animals in factory farming, the life-medium of sh as in trout farming, hot water disinfection, dyeing, transporting inputs, products and waste within industrial premises, and re-ghting. These industrial uses in manufacturing, livestock factories, hospitals, commerce, ofces, mining, petroleum rening, power generation, railway companies etc. inevitably generate industrial efuent, that is, aqueous waste. In the 250 years since the Industrial Revolution, such waste has been disposed of predominantly by dumping it untreated into rivers, lakes, the sea and even underground. The environmental impact of the use in this way of the natural world as a sink has been to destroy animal and plant species, degrade habitats and spread illness and death within human populations.

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

130

In the quest for a sustainable society, it has been proposed that water resource management comprises six elds of action, one of which is the purication of water prior to and after its use (Merrett 1997: 14852). An important economic instrument of relevance to such action is disposal charges, also known as discharge fees, for industrial efuent. Bhatia et al. (1994: 3) suggest that the effective use of such charges is rare in developing nations. But they are certainly widely employed in high-income countries. This paper briey considers the network of institutional, economic and regulatory relationships encompassing efuent generation and disposal, then turns to the objectives and design of discharge fees using examples drawn from the European Union, and nally compares and contrasts the power, precision, and interrelation of discharge fees and environmental licensing.

15.2 THE GENERATION AND REGULATION OF INDUSTRIAL EFFLUENT

The generation and disposal of aqueous waste is illustrated in Figure 15.1. Abstracted water is supplied to industry for the multifarious uses already described and this leads to the production of raw efuent. In some cases this is collected and discharged directly to rivers, lakes, estuaries and coastal waters, or to underground locations. In other cases the waste water is discharged to sewer for treatment in a sewage treatment works. The third possibility is that the efuent receives pre-treatment on the premises and then is disposed of to water course or sewer or is cycled back for internal reuse within the institution whence it came. With respect to efuent discharged to sewers and collected for transfer to a sewage treatment works, this works has two outputs: sewage sludge and treated waste water. The sludge may be used in agriculture or incinerated or sent to landll sites or dumped

Landfill

Water supply

Industrial use Internal re-use

Raw effluent

Incineration Aquifer, river, lake, estuary, coastal waters

Pre-treatment

Collection Agriculture External re-use Waste water

Sludge

Disposal

Treatment

Figure 15.1 The generation and disposal of aqueous waste.

into water courses. Waste water is disposed of to surface waters and the sea. Alternatively it is reused externally in other industrial organizations or as irrigation water. In the European Union, North America and other regions of the globe, environmental regulators control the disposal of raw efuent, sewage sludge and treated waste water. As already suggested, the polluting content of these solid and liquid wastes has the potential to damage the natural environment and bring illness and disease to urban and rural populations. For example, this is true of complex efuents containing phenols, ammonia, chlorine, heavy metals and organo-phosphates. The governments regulatory agency imposes a ban on some toxins. In other cases it limits their maximum level by means of consent conditions. Specically with respect to the European Union, the 1991 Urban Waste Water Treatment Directive lays down uniform emission standards for sewage treatment for all sewage treatment works serving populations of two thousand or more. The Directive also covers industrys direct disposal of efuent. In the UK, the most dangerous substances which can be found in industrial efuent and domestic sewage are on the Environment Agencys Red List. Environmental regulation raises the cost of pre-treatment by industry and the cost of waste water treatment by water service companies. Industrial companies may therefore oppose the introduction of pollution legislation or may attempt to evade it when it is in place. As a result, the regulator requires systematic monitoring of the chemical composition of discharges to water courses and sewers. The prohibitions and consent conditions imposed on water company discharges in turn may lead the company to set quality standards on industrial discharges to its sewers. This is because specic industrial efuents can raise sewage treatment works costs substantially, within a dened regulatory framework, or disable some treatment processes. The water service company needs to shield its assets from damage, protect its operatives, keep its production costs as low as possible and ensure its treated waste water discharge and sludge output meets the regulators standards. As John Hills points out in his valuable book, Cutting Water and Efuent Costs, a water service company ban on complex efuents in turn places pressure on industry for strengthened on-site pre-treatment facilities (Hills 1995). The interdependence of industry, water utility and environmental regulator is readily illustrated. The bulk of sewage sludge in the UK is used on farmland, but heavy metals in the sludge can make it unacceptable for agricultural purposes. Conventional treatment of efuent does not remove these pollutants. As a result, the Environment Agency and the water utilities impose strict limits on the heavy metal content of discharges to water courses and sewers, for example by the electroplating and photographic lm processing industries (Hills 1995: 267). So we can see that the introduction of vigorous environmental regulation of industrial efuent disposal to fresh and salt water sinks places new technological demands both on water utility treatment of aqueous waste as well as on the pre-treatment of its efuent by industry itself.

15.3 THE OBJECTIVES OF DISPOSAL CHARGES

A discharge fee on industrial efuent can be dened as a payment levied on an industrial organization by a government regulator or by a water utility in respect of the volume

of efuent and/or the pollutant mass it discharges to sewer, aquifer, river, lake, estuary or coastal waters. To understand the objectives of the institutions which set and collect disposal charges, a vital analytic distinction is necessary based on the ow chart of Figure 15.1. Two types of organization impose discharge fees: 1. 2. Water service companies levy charges on industrial organizations (as well as domestic consumers) which discharge their efuent to the companies sewers; and the environmental regulator levies charges on industrial organizations, including water service companies, which discharge their efuent to water courses.

In the rst case above, the utility provides a waste water service which embraces the collection of efuent through the sewerage network, its treatment, and the disposal of waste water and sewage sludge. The objective of the utility in setting the disposal charge is to cover the prime and overhead costs of providing this sanitation service (Merrett 1997: Figure 3.2). Disposal fees of this type are universally referred to as sewage charges. In 1999 a typical charge in the UK is of the order of US $0.80 per cubic metre of waste water. In the second case, that of the environmental regulator, the objectives are more complex. The control of industrial efuent discharges to water courses imposes a variety of nancial costs on the regulatory authority, such as for research, policy-making, the preparation of discharge licences, and monitoring adherence to the standards set. Htte et al. (1995: 220) indicate that the original objective of the levy on discharges to surface waters in the Netherlands was fund-raising for water quality management by government authorities. We can call this the hypothecation objective. Such a goal is probably the origin of discharge fees in every regional or catchment authority in the world, where regulatory charges are collected. The parallel here with fee payments for abstraction licences is evident (Merrett 1999a). But another goal for discharge fees seems to have developed in an evolutionary manner from the hypothecation charge. At the most general level, the objective of regulatory standards on industrial efuents has been to reduce industrial pollution of the natural environment. This has benets not only of an environmental nature but also can increase gross domestic product: by reducing the costs of instream water users such as in shing, leisure pursuits and other ecosystem activities; and by cutting the costs of downstream water abstractors. In applying hypothecation fees, it became evident to regulatory authorities that industry was cutting back on the efuent mass it discharged in order to reduce its exposure to these now internalized costs. Once again, the case of the Netherlands is instructive here. So the charge raised to nance the command and control system was now acting independently as an economic instrument in achieving the prime goal of regulation itself. As a consequence, with the passage of time, the second objective of discharge fees, pollution reduction, became established.

15.4 THE DEMAND FOR WASTE WATER SERVICES

Waste water services are a public good, that is, where the benet derived from the good by one consumer does not diminish the benet derived by consumers in general. The standard textbook example is a countrys armed forces (Sandmo 1987). In a laissez-faire

economy, it is impossible to price or to sell services for the treatment and subsequent disposal of industrial efuent, for no purchaser of such services privately appropriates their benets. Where the state intervenes to defend the public realm against industrial pollution, a space is opened up by environmental legislation within which sewerage utilities (private or municipal) can operate. As we have seen, sewage charges payable to the utility and discharge fees levied by the regulator for the use of water courses as a sink, create a response from industry. Organizations confronted by disposal charges on the volume of their efuent and/or on the pollutant mass will investigate ways of cutting the charge payable, provided that the fees are not negligibly low. This requires operational research combining managerial, engineering and economic skills. The relation between charge and industrial action can be measured by the elasticity of response of the pollutant mass/ discharge fee function (see below). One method is to reduce the total volume of abstracted water consumed or the total mass of a specic chemical used for the production process. For example, a site survey can be used to calculate the organizations own regional water balance statement, to revise water need calculations and to check the functioning of the water distribution system so as to cut internal water supply volumes and reduce on-site leakage (Merrett 1999b). Another example, from the leather industry, is advances in the technology of the tanning process which permitted a fall in chromium use because of an increased percentage uptake in the hides. A second method for industry to reduce its disposal charges is pre-treatment by the institution of its own efuent so as to diminish the pollutant mass discharged to sewer or to water course. Pre-treatment includes air otation for fats, special separators for oil, and bacterial processes for organic efuents. Pre-treatment is likely to be introduced only if the nancial rate of return is attractive when its costs are assessed against the reduced bills for disposal. A potentially benecial spin-off from pre-treatment is by-product manufacture for sale to niche markets, such as the production of animal feed in efuent treatment by whisky distillers. The third method for cutting discharge fees is for industry to reuse its water and/or its process chemicals. Often this demands pre-treatment as a rst step. The impact on water volumes of internal reuse is both to cut its water intake requirements, with an associated fall in water supply costs, as well as to reduce its efuent volume, thereby diminishing its disposal charges. These opportunities frequently occur in process washing where, as Hills points out (1995: 85): the cleanest water washes the nearly clean product and the reused water washes the dirty incoming product. With respect to a process chemicals example, equipment exists to handle the rinse waters from chrome and nickel electroplating drag-out tanks, enabling the plating chemicals to be returned to the plating bath. The effect of disposal charges on industrial rms behaviour is illustrated in Figure 15.2. The vertical axis measures the pollutant mass discharged per unit time period, that is, the variable M of equation (1) below. The horizontal axis measures the independent variable P, the price charged as a disposal fee per unit of pollutant mass. The shape of the function is that of a quadratic equation: M aP2 bP c (1)

where a, b and c are the equations parameters and a is negative. Figure 15.2 embodies the hypothesis that at low prices the mass discharged is virtually unaffected by price differences. Here managers take no interest in the invoices

1200

1000 Pollutant mass in kilograms

800

600

400

200

0 0 2 4 6 8 Disposal price per kilogram 10 12

Figure 15.2 The effective demand for pollutant disposal services. they receive from the environmental regulator or the water utility. In a higher price range, the discharge fee really bites and the slope of the curve is steeper. At very high prices, pollutant discharge ceases as the industrial institution is forced for nancial reasons either to change its technology or discontinue production. The price at which M falls to zero has the same outcome as that where a regulatory agency places a complete ban on the pollutants discharge.

15.5 THE MEASUREMENT OF POLLUTION

Before an understanding of the design of discharge fees is possible, it is necessary to examine how pollutant ows are calculated. One can only price what one can measure. Clearly, the starting point is the ow volume of industrial efuent, for example in cubic metres discharged per day. A second measure is the concentration of pollutants in the efuent. This can be measured as the mass of a specic pollutant per unit volume, for example measured in grams per cubic metre, equal to milligrams per litre. This gives: MVC (2)

where M is here measured in grams/day, V is efuent volume and C is pollution concentration. However, pollutant-specic fees entail high transaction costs because of the technical difculties of measuring the pollutant concentration and monitoring how it changes over time (Green 1990: 9). Efuent is a complex waste. In the steel city of Jamshedpur in India, for example, Bhatia et al. (1994: 1011) showed that untreated

industrial discharges into the Subernarekha River contained phenols, oil, grease, ammoniacal nitrogen, cyanide, chromic and other acids, and hexavalent chromium. The discharge volume is so high that in the low ow month of April the ratio of the efuent volume to the river ow is 1:3.75. These inquinating cocktails need grouping categories to facilitate the measurement and pricing routines, such as biological oxygen demand (BOD), chemical oxygen demand (COD) and suspended solids, all quantiable in grams/cubic metre. BOD and COD are directly related to dissolved oxygen, the principal measure of water quality in terms of its ability to sustain plants, sh and other biota. In France, Tuddenham (1995: 2012) reports that pollution charges are based on the quantity of pollution generated on an ordinary day in the month of maximum discharge. There are eight grouping categories: suspended solids, oxidizable substances, soluble salts, nitrogen, phosphorus, organohalogenated compounds, metals and metalloids, and toxins. The rst seven are measured in grams and the eighth in toxic units. In the Netherlands, the principal measures are COD and the heavy metal pollutants cadmium, mercury, arsenic, copper, nickel, zinc and lead (Htte et al. 1995: 2212).

15.6 THE DESIGN OF DISPOSAL CHARGES: THE UTILITIES

Discharge fees should be designed to achieve the objectives pursued in setting them. For that reason, one must deal separately with the water service company and the environmental regulator. The water utility seeks to cover the costs of providing its sewerage service to industry, such costs dened in a broad sense to include the cost of capital. Its prime costs per unit output comprise wages, salaries, power, materials, spare parts and other consumables. The gross margin per unit output meets overhead costs and net prot. Overheads are made up by the rent of land and buildings, leasing costs, payments to directors, interest on loans and amortization. Net prot is split between taxes, dividends and retained prots. Pricing strategy is likely to be based on average cost, as Lee (1994) proposes. I have suggested elsewhere that the most appropriate average cost model for the waste water industry is mark-up pricing. In this case average prime cost per cubic metre of industrial efuent at normal capacity is calculated and a mark-up added (Merrett 1997: 5961). The mark-up may be in cents per cubic metre or it may be a percentage addition to prime cost. Where prices are controlled by a regulator, these too will be average cost based. There is no space here for prices set at short-run marginal cost nor should there be. Similarly, there is no role for the average total cost of new investment in network and headwork infrastructures, otherwise known as long-run marginal cost. Short-run marginal cost is negligibly low and pricing on this basis would ruin the utility. The average total cost of additional capacity as a price determinant would generate huge losses (or surpluses) wherever average total cost was substantially lower (or higher) than the average total cost of existing capacity. Marginal cost pricing, if combined with a standing charge, discriminates against small-scale users of the sewerage system. This approach to cost recovery pricing in the collection and treatment of waste water, embracing both prime and overhead costs, is embodied in the European Unions Water Framework Directive, which also requires the design of the most cost

effective measures to achieve local reductions of substances prioritized on the basis of risk (Bjerregaard 1998: 7880). The starting-point for industrial disposal charge setting will be the average cost per cubic metre of collection, treatment and disposal of all sewage. Where the concentration of pollutants from specic organizations is calculated to be higher than the average and where this difference gives rise to higher treatment or disposal costs, a higher price should be set, such that the resulting price difference matches the higher cost, on the basis of cost engineers calculations. An example of such customary practice is the modern Mogden formula (Mogden is the largest sewage treatment works in London): K R Q B(Ot/Os) S(St/Ss) (3)

where K is the charge per cubic metre; R are the costs for conveyance through sewers; Q is the volumetric treatment cost covering screening, primary settlement, tertiary treatment, and outfalls of treated sewage; B is the cost of biological treatment plus a proportion of secondary sludge treatment and sludge disposal; S is the cost of treatment and disposal of primary sludges; Ot/Os is the COD in milligrams per litre after 1 hours settlement at pH 7 of the companys industrial efuent divided by the regional average of the same measure; and St/Ss is the suspended solids in milligrams per litre after 1 hours of settlement at pH 7, again of the companys industrial efuent divided by the regional average of the same measure. Here, then, the multipliers are based on the concentration of chemical oxygen demand and suspended solids. In the UK, Stewart has successfully developed a logarithmic model of the functional expenditure of large sewage treatment works that shows clear economies of scale and uses separate terms for maximum BOD and ammonia consents (Ofwat 1994).

15.7 THE DESIGN OF DISPOSAL CHARGES: THE ENVIRONMENTAL REGULATOR

The regulators two objectives in charge-setting, discussed in Section 15.3 above, are hypothecatory and pollution-reducing. These will be reviewed in turn. But rst it is necessary to make a simple point, often neglected, with great bearing on the regulators charges. At the simplest level, industrial efuent is composed of two parts: water and pollutants. The ratio of the rst to the second is large. For example, Tuddenham produces data that show that in France the mass of pollutants produced per inhabitant per day is 166 grams (1995: 203). The volume of water used per inhabitant per day is of the order of 150 litres. Therefore the ratio of the total mass of pollutant to the mass of water is about 1:900. The discharge of water to river, lake and estuary imposes no costs on the environment. Indeed, it is a welcome addition within the hydrosocial cycle to the catchments effective rainfall. In contrast, pollutants do impose such costs. Thus the regulators golden rule should be to levy disposal fees only on the pollutant mass, never on the water volume. The hypothecatory objective requires that total receipts from the discharge fees levied on industry should equal the total costs of the regulation of industrial efuent. In practice, these two variables are interdependent. The rst determines the second which determines the rst. To understand the contingencies of their relation requires an economic history of the catchment and an account of its environmental politics.

The French basin agencies provide a clear case of the hypothecatory principle (Tuddenham 1995). The pollution-reducing objective implies that the fee per unit measure of pollution (see Section 15.5) should vary with the environmental cost it is assessed to impose. In some cases it may be possible to measure the nancial cost suffered by in-stream uses such as navigation, recreation and shing, or by organizations abstracting water for subsequent use. However, I know of no case where such measured nancial cost has in fact shaped the discharge tariff. In any case, the principal harm done by pollution is perceived to be environmental, rather than GDP-reducing. In my view such impacts cannot usefully be measured in dollar terms (Merrett 1997: 16870). In such cases the design of the components of the tariff is best carried out on the basis of environmental impact assessments. Such impacts vary with pollutant type, the ecology of the water course at the point of disposal and downstream thereof, the quality and volume of the river ow at the point of disposal, the season, and the concentration of pollutants within the efuent discharged. Finally, in this section, it is appropriate to refer briey to a pricing mechanism for the release of pollutants to water course sinks which environmental regulators may wish to consider. This is the introduction of tradeable disposal rights, a parallel concept to tradeable abstraction rights. Such an arrangement would see the regulator auction permits to pollute to the highest bidder. Such rights would be recognized, for example, for 5 or 10 years and would state the maximum mass of pollutants permitted to be discharged. They would be tradeable between organizations. Unfortunately, the polluting institutions of industry each has a pollutant mass they wish to discharge annually which differs in total mass, pollutant composition, timing of discharge and location of disposal. Thus it would be impossible for a regulator to design a permit with a known environmental impact which would be of value as a commodity, for it could not be of commercial interest to more than one institution. As Bhatia et al. write: efuent permit trading schemes have been implemented only rarely for water pollution (1994: 7).

15.8 CONCLUSIONS
On the basis of the preceding discussion of the disposal of industrial efuent to water courses, there seem to be eleven conclusions an environmental agency might draw on the inter-relation of environmental regulation and economic instruments in the quest for a sustainable society. 1. In terms only of efuent discharged, disposal fees in the course of time can have the same broad outcome as regulatory limits or bans. 2. Both regulation and discharge fees stimulate pre-treatment and reuse within industry, the regulation of industrial discharges to sewer by the water service companies, and specic forms of sewage treatment by those same utilities. 3. Both regulation and disposal charges engender substantial transaction costs for the environmental agency in monitoring industry and waste water utility behaviour. This is because waste water treatment and the disposal of its products are costly public goods. Grouping categories for the measurement of industrial pollution, such as biological oxygen demand, help reduce these transaction costs for both approaches.

4. For both command and control as well as economic instruments, the golden rule must be to target pollutant mass, or its derivatives such as chemical oxygen demand, never the volume of industrial efuent, in order to reduce pollution without diminishing the recycling of water. 5. The design of regulatory proscriptions and consents, as well as of discharge fees, should be based on environmental impact assessment, not on vain attempts to give a meaning to the dollar valuation of environmental costs. 6. Disposal charges generate a hypothecated income for the environmental agency. Regulation does not. 7. Regulation has precision, speed, directness and forcefulness unmatched by discharge fees. 8. The prime and overhead costs of the collection, treatment and disposal of industrial efuents can be prodigious and should be carefully reviewed before regulatory bans and consents are set (Briscoe 1995). 9. Disposal charges permit a exibility of response by industry unmatched by regulation. This exibility can be manipulated by the agency so that the industrial costs of conforming to required standards are minimized. 10. The objective of regulation has always been to reduce the industrial pollution of the environment, whereas the objective of discharge fees has been hypothecatory. However, in the course of time, discharge fees have evolved so as also to serve a pollution-reducing objective. 11. Regulation works best where draconian and absolute action is required. Economic instruments work best where partial restrictions shared amongst a group of stakeholders are appropriate. and where limits on pollution behaviour are more costly to society if undertaken by some organizations rather than by others. Under these assumptions, economic instruments can allocate the permission partially to pollute in the most efcient manner. In conclusion, my judgment is that both approaches are valuable and should complement each other, provided that they are introduced in a society where the design of rules and of prices are appropriate to their objectives, where it is possible to implement regulation effectively, and where disposal charges can be correctly assessed and are actually paid.

The material in this chapter originally appeared in: Industrial efuent policy: economic instruments and environmental regulation. Water Policy, 2, 201211, 2000.

16 Nitrate pollution on the Island of Jersey: managing water quality within European community directives
We must have some comradeship with imperfection.

George Eliot (1876)

This paper was co-authored with my friend Nick Walton

16.1

INTRODUCTION

The objectives of this paper are two-fold. First, the authors wish to review the anthropogenic processes by which the Island of Jerseys groundwater and surface waters have been polluted by nitrates over a broad span of time. Secondly, the costs and benefits are assessed of the principal management innovations by the government sector to reduce nitrate pollution. Jersey is an island in the English Channel and is situated off the north-west coast of France, only 23 kilometres from the Normandy coast. Its political relationship to England, France and the European Community (EC) is complex for historical reasons. Prior to the Norman Conquest in 1066, Jersey and the other Channel Islands were part of the territory of the Duchy of Normandy. But when continental Normandy was freed from English rule in 1204, Jersey retained its allegiance to the King of England.

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

140

Nitrate pollution on the Island of Jersey

141

Thereafter successive English monarchs have ruled the Island through their claim to the title of Duke of Normandy. As a result, the Islandsjudicial systems are independent of the English courts and at no time has Jersey been subjected to the administrative systems of the United Kingdom government (Chief Advisers Office 1997). Jersey has its own legislative assembly and comprehensive legal, fiscal and administrative systems. Most of the laws by which it is governed emanate from the States of Jersey (the representative assembly); committees of that assembly administer most of the public services. With respect to the European Community, the Island enjoys free movement of manufactured and agricultural goods, but the provisions of the Treaty of Rome and subsequent treaties are not formally applicable to Jersey. Jersey has an area of 117 km2 and in 2005 had a resident population of about 87,000 persons, giving a density of 744 persons/km2 . As Robins (2000: 23) points out, the island comprises a gently sloping plateau lying at an elevation of 60-120 metres divided by a series of valleys running from north to south. Westerly and south-westerly winds bring moisture from the Atlantic. Long-term average rainfall for the island is 877 mm, annually varying from 600 mm to 1100 mm in the last decade. Annual potential evapotranspiration ranges from 648 mm to 784 mm. The 136-year monthly average rainfall shows that the bulk of precipitation occurs in the seven months September-March (Jersey New Waterworks Company 2002: 4). Flow along the valley streams is predominantly north to south and, in the absence of a dry winter, takes place throughout the year. Groundwater discharges both to these streams and at the coast. During prolonged dry weather, saline intrusion may occur locally. Robins has shown that the islands base flow index has an estimated value of 0.58, indicating that 58 per cent of streamflow is derived from the groundwater baseflow (Robins 2000: Table 16.1, Blackie et al. 1996). This quantity will be of importance in the discussion of water quality.

16.2

NITRATE POLLUTION OF GROUNDWATER AND SURFACE WATERS

Nitrate contamination of both surface and groundwater has been a topical issue throughout the western world since the 1970s when it was recognised in Europe and the USA that fertiliser-intensive agricultural activities, which were tripling food yields, were also producing an undesirable excess of nitrate run-off into rivers and aquifers (Holmes 1979). In Britain, agricultural, surface water and groundwater field studies were producing a large number of reports from various agricultural and water research institutions throughout the 1970s (e.g. Anon 1975, Young and Gray 1978, Foster et al. 1976 and 1977). These led to the UK Government instigating first a Steering Committee, and then a Nitrate Co-ordination Group that in 1986 issued a major holistic report on the subject as DOE Pollution Paper No. 26 (DoE 1986). The original World Health Organisation (WHO) international drinking water quality standards (1970) decreed <50 mg/l as a recommended level and 50100 mg/l as acceptable for nitrate (NO3 ), which equates to 11.322.6 mg/l when expressed as NO3 -N. However, the WHO (European) standards, gave the lower level of 50 mg/l as a recommended maximum, and subsequently this level was incorporated into the EC drinking water standards (80/778/EC) and thereafter as a maximum admissible

142

The Price of Water

concentration (MAC) in all member countries legislation. The epidemiological and scientific reasoning behind this change from an acceptable level of 100 mg/l to a MAC of 50 mg/l has long been contested. In any case, the EC directive, along with the 91/676/EC Nitrates Directive, has necessitated EC states investing substantially in costly nitrate reduction measures and nitrate removal plants. The nitrate problem is complex, and many studies have shown that simple leaching of applied nitrate fertiliser is only a part of the problem. The leaching process is location-dependent in terms of climate, soil type, bedrock geology, drainage characteristics and vegetation. On top of this are the anthropogenic factors such as farming practices, cropping type and intensity, fertiliser application rates and timings etc., as well as inputs from the septic soakaway systems of non-sewered dwellings. The two principal driving forces behind nitrate leaching are the extreme solubility of nitrates, and the clear relationship between applications of nitrate fertiliser and increased crop yields that, in a modern capitalist market economy, means the difference between economic survival and extinction for the farming community. There are well-researched Government guidelines for determining the optimal level of Nfertiliser to apply for each crop, and these can be refined with local field soil-nitrogen knowledge. However, the complex, climate-driven, biogeochemical reactions that both store and release nitrogen from the soil zone, and the unpredictable timing and effects of rainfall, all combine to make prediction of N-leaching difficult. In fact, the generally thin, clay-poor soils over hard-rock basement throughout Jersey mean that their leaching potential and aquifer vulnerability are both high. One of the earlier papers addressing the Jersey nitrate problem identified the excessive use of artificial fertilisers, especially on the many smaller farms, as being a significant contributor to the elevated nitrate levels being experienced within most surface and ground waters in Jersey (Foster et al. 1989). Farming studies have shown that both potato cropping and grassland dairying require some of the highest rates of fertiliser application for optimum/maximum yields (Holmes 1979). Moreover, optimum cropping methods tend to promote high leaching rates, especially in thin, clay-poor soils. This combination of soil, geology, climate, cropping and farming practice, as applied to the Jersey situation where potatoes and dairying historically dominated the agricultural sector, produced conditions for maximum nitrate generation and leaching. It should therefore be no surprise that Jersey has suffered from perennially high nitrate levels in all its water bodies, especially as over-fertilisation of crops by up to 60% of farmers has been historically practised (Foster et al. 1989). Lott et al. (1999) state that if Jerseys fertiliser application rates were to be reduced to the levels jointly recommended by the UKs Ministry of Agriculture, Food and Fisheries (MAFF) and the Agricultural Development Advisory Service (ADAS) (and now by DEFRA the Department of the Environment, Food and Rural Affairs), then nitrate levels in Jerseys waters could fall by up to 34% (DEFRA 2000). Additionally, establishing water catchment controls, Nitrate Vulnerable Zones and incorporating non-spreading buffer zones along watercourses to prevent direct leaching and support natural denitrifying activity, could all help to resolve Jerseys nitrate problem. Although agricultural drainage is the predominant route for nitrate leaching into Jerseys surface and groundwater bodies, there is also a small input from the 5,000 or so rural unsewered dwellings, the septic tanks of which discharge organic nitrogen and ammonium into the shallow soil zone, which then oxidises to nitrate and adds to

Nitrate pollution on the Island of Jersey

143

the nitrate concentrations in groundwater. Although ground waters supply less than 3% of the public water supply from the Jersey New Waterworks Company (JNWWC), base flow from Jerseys shallow aquifers provides some 58% of annual average flows in Jerseys surface waters that are then combined into reservoir storage for JNWWC water supplies. However, Lott et al. (1999), using a simple model based on nitrogen export coefficients, found less than 10% of the total nitrogen loading to be derived from domestic waste water and livestock. This was supported in the study of Green et al. in 1998 who used nitrogen isotopes to characterise the different sources of nitrogen inputs, and found only a small input from these non-agricultural sources. Finally, there are always point source leakages of nitrate from broken sewers, silage clamps, farmyard drainage and storage areas etc. These can be minimised by good planning, inspection and maintenance programmes, but never completely eliminated. Under the EC-driven UK legislation on Nitrate Vulnerable Zones, improvement grants are available to farmers to re-locate, upgrade and improve all farm areas where nitrate leaching could be a problem; this includes farmyard drainage, silage clamps, silos etc.

16.3

WATER QUALITY MANAGEMENT BY THE DEPARTMENT OF AGRICULTURE AND FISHERIES

The preceding section shows that the principal processes of nitrate pollution in Jersey take place in the agricultural sector. In the last five years the Water Pollution (Jersey) Law 2000 has provided the legal basis for new forms of water quality management. The Law contains pollution prevention provisions and Article 17 makes it an offence to cause or knowingly permit pollution of any controlled waters, unless it is done under the conditions of a discharge permit. Controlled waters refers to the territorial sea adjacent to the Island, coastal waters, inland waters and ground waters. With respect to agriculture, farmers and employees and contractors can be prosecuted for causing pollution, with a maximum penalty of an unlimited fine and/or two years imprisonment. In the case of a prosecution, a farmer can defend himself by showing due diligence in his operations. A due diligence defence can be based on the 2004 Code of Good Agricultural Practice for the Protection of Water. This Water Code sets out recommended practices that, if followed, will reduce nitrate and other forms of water pollution. Table 1 shows the Water Codes sections and for each of these an example of its content. As the Water Code states (p.7) You are therefore strongly recommended to comply with this Code of Practice especially in view of the Article 18(5) of the Law in relation to the defence of due diligence. In brief, the 2004 Code of Good Agricultural Practice for the Protection of Water has teeth. If it is not complied with, a farmer who pollutes controlled waters can be fined and/or sent to prison. The Water Code is fifty pages long and contains a wide variety of recommended practices that have direct relevance to the nitrate pollution of controlled waters. With respect to nitrates the most important recommendations are to be found in Section 2 (Farm manure and waste management planning), Section 3 (Slurry), Section 5 (Solid manure), Section 10 (Liquid fertiliser) and Section 11 (Nitrate and phosphorus). Appendix 1 of the Code includes nitrogen compounds as amongst the high-risk substances of the second schedule of Water Pollution (Jersey) Law 2000.

144

The Price of Water

It is worth mentioning that the Department of Agriculture and Fisheries has ceased to exist since January 1st 2004. The technical and scientific part of its work was transferred to the Environment and Public Services Department on that date. In this way, the agricultural activities of Jerseysgovernment institutions have been drawn closer to their environmental responsibilities. By 2005 it had become clear that Jerseys long battle to redeem the quality of its surface and groundwater flows was beginning to record clear successes. Changing agricultural practices through the 1990s with a greater number of better-managed, larger farms, reductions in potato and dairy farming, and the gradual introduction of optimised ploughing, plant-rotation and cover-cropping practices have together had a positive effect on the nitrate levels in Jerseys waters in recent years. Implementing the 2000 Water Pollution Law was always going to be easier for point source infractions than for diffuse pollution, because of the greater visibility of the former. Two farmers have been taken to court over silage spills under the Law but none for non-point source breaches of the Law. However, the combination of the regulatory threat and the governments successes with friendly persuasion on nitrate applications began to take effect. Nitrate use has been falling for 56 years and there has been a 25% fall in the potato area between 1999 and 2004. All the available evidence in 2005 suggests a decline in the nitrate content of surface and groundwater. Borehole tests provide the evidence on groundwater. Moreover, a quinquennial review of Jerseys flowing water based on macro-invertebrate populations showed striking advances in quality between 199798 and 200204. For example, bad quality incidence fell from 37% to 8% (Langley & Kett 2005).

16.4

WATER QUALITY MANAGEMENT BY THE JNWWC

a) The water utility

The Jersey New Waterworks Company Ltd. (JNWWC) is the oldest registered company in Jersey (founded 1882), although the States of Jersey took a major shareholding in 1981 to guarantee a public voice in the essential supply of piped water to over 85% of the island population. Water from all JNWWC stream, spring, borehole and desalination sources is collected and stored in six reservoirs prior to pumping for treatment at one of the two centrally-situated water treatment works at Handois and Augres. Treatment at both plants consists of the basic standard water works practice of clarification, filtration and chlorination, prior to distribution to customers via some 400 km of service mains (JNWWC, 1998). By these practices, the varying qualities of the different water sources are blendedout and the water quality output from both treatment plants is generally very similar, giving a broadly uniform water quality throughout most of the public supply network. The quality of water supplied by JNWWC aims to comply with the EC criteria for public supply drinking waters, although it has always had problems meeting the 50 mg/l NO3 MAC, with up to 30% of samples exceeding this limit in 2001 (JNWWC, 2002). Since Jersey is not a member of the EC there has, until recently, been no legal requirement to keep to this MAC. However, a recent (2004) amendment to the Jersey Water (1972) Law has incorporated the UK Water Quality (2000) regulations, which implements the EC potable water quality standards, including the 50 mg/l MAC for

Nitrate pollution on the Island of Jersey

145

nitrate. However, the States, recognising the extent of the nitrate problem in Jersey, has allowed dispensations until 2008 for up to 33% of samples annually to exceed the nitrate MAC, as long as they are less than 70mg/l NO3 .This pragmatic measure gives the JNWWC time to plan for ways in which it can blend-in low nitrate water sources to reduce the nitrate concentration in all its piped supplies to below the 50 mg/l MAC by the end of 2008. Other JNWWC water quality problems are few and minor, involving occasional minor excedences for nitrite, lead, manganese, cyanazine, coliforms and chlorthal (JNWWC Annual Reports).

b) Seawater desalination
Jersey is not particularly short of water in any absolute sense, since total abstractions are only around 25% of the average annual effective rainfall of the island. However, its typical island limitations and the annual summer influx of tourists, which can swell the population by up to 50%, mean that its six storage reservoirs (holding up to 30% of annual public demand) can become seriously stressed during successive dry years or drought conditions like 1975/76, 1989/92 and 1995/96 (Robins 2000). Prior to the JNWWC commissioning the UKs first seawater desalination plant in 1970, drought conditions, like 1959, had necessitated severe water rationing for the islands population. There were also optimistic forecasts of growth in the Islands population and its economy. The 6,000 m3 /d multi-stage flash (MSF) distillation plant, commissioned in 1970 at La Rosi` ere in the south-west of the island, was designed in the context of this optimism as well as to alleviate occasional drought situations and, although it operated only on a demand-only basis for some 1000 days over its 28 year existence, it became a life-saver at those particularly dry times. MSF distillation technology of the 1960s was energy intensive and inefficient by modern standards and also environmentally polluting due to the burning of some 45,000 litres/day of heavy grade fuel oil and the consequent emissions of 2.2 tonnes/day of sulphur dioxide (Howard 1999). In 1997 a decision was taken by the JNWWC to demolish this old MSF plant and replace it with a much more energy efficient, less polluting and cheaper reverse osmosis (RO) desalination plant, which produces a similar volume of around 6000 m3 /d of freshwater for approximately 50% of the running costs of the old MSF plant. The new RO plant desalinates filtered and chemically pre-conditioned seawater by a pressure-driven membrane process which uses electrical pumping energy, at some 65 bars pressure, to force pure water molecules through a semi-permeable membrane leaving the concentrated brine salts to be rejected and disposed of back into the sea. The system uses modern, high efficiency, thin film composite (TFC) membranes in a spiral wound configuration that operate at a 45% recovery rate. Additional energy efficiency is gained by recovering pressure from the reject brine stream through an energy recovery turbine (Howard 1999). The new RO plant uses the same seawater intake, pipelines and holding reservoir as the old MSF plant, but requires additional pre-filtration and chemical conditioning stages prior to feeding the seawater into the sensitive RO membranes to prevent membrane fouling. The plant contains four independent RO trains, so has great flexibility in its operation, cleaning and stand-by modes. Overall, the advantages of the new RO plant over the old MSF process are lower energy requirements, no fuel-handling operations, short start-up and shut-down times,

146

The Price of Water

operational flexibility, lower maintenance requirements, lower operational cost and no atmospheric pollution. The flexibility of the RO plant now enabled its output to be used for blending purposes by pumping variable quantities of desalinated fresh water into the Val de la Mare storage reservoir, significantly decreasing the final mixed waters nitrate content at a time when nitrate levels were beginning to rise above the ECs potable water quality standard of 50 mg/l. Since the desalination plant can produce up to one quarter of the summer peak daily demand for potable water from the JNWWC, running this plant with its near-zero nitrate concentration could, theoretically, reduce all of the JNWWC water supplies to below the ECs MAC of 50 mg/l NO3 . (In practice this would occur primarily only in the western part, due to the pipe-networks delivery of all the desalinated water into the Val de la Mare storage reservoir in the west where the Handois water treatment plant is located. Desalinated water can be routed to the Augres water treatment plant in the east of the Island, but there the nitrate levels are, in any case, lower.) Running the desalination plant at full capacity would mean an expensive, uneconomic production of a surplus of potable water. However it has now become possible to arrive at a compromise position whereby the desalination plant is run for certain periods of time, like seasonably high nitrate peaks, in a nitrate-reduction blending mode, rather than just an emergency water production mode in drought times.

c) Costs of RO Desalinated Water

Continued advances in membrane technology and energy recovery systems have brought a substantial decrease in the total costs of RO desalinated seawater internationally from around $2/m3 in the late 1980s to around $0.7/m3 currently. Costing is complex because it depends upon fluctuating energy costs, the degree of energy recovery possible, manpower costs, the economies of scale achieved by continuous running of the plant, and the capital amortisation rate. In Jersey, the island situation requires the import of all energy and consumables that, together with high manpower costs and the intermittent running of a relatively small plant, all push towards higher costs than those achieved internationally. The capital cost of the 6000 m3 /day RO plant was 5.1 million. Amortised over 10 years, this implies 1400/day or 23 pence per cubic metre (p/m3 ) of product water at full capacity output. Running costs are approximately 3600/day or 60 p/m3 of product water. Average total cost of desalinated water is thus approximately 83 p/m3 when the plant runs at full capacity. This compares unfavourably with the approximate average total cost of 15 p/m3 for the main surface water supplies from the JNWWC. These, of course, suffer from some degree of nitrate pollution. Thus, although desalinated water appears relatively expensive, this price becomes acceptable in the event of drought conditions. Since the capital and labour costs are effectively fixed, the marginal cost in running the plant at other times to produce zeronitrate water for blending purposes, is only the 33 p/m3 due to consumables (mainly electrical energy) costs. Recent improvements in energy recovery technologies if retrofitted to the reverse osmosis plant could reduce this energy cost significantly. The most recent developments have been very positive. In the year 2004 it was not necessary to run the desalination plant at all, either for drought functions or for the blending function. This is ascribed partly to the decline in nitrate levels in Jerseys

Nitrate pollution on the Island of Jersey

147

water as the work of gentle persuasion and regulatory action has brought changes in farm practice in nitrate application. Farm output itself is now in gentle decline. Jersey now projects itself successfully as a financial centre and as a location of environmental and seaside tourism.

d) RO De-Nitrification Plant
Following the successful inauguration of the new RO plant at La Rosi` ere in 1999, and pilot scale trials of a de-nitrification plant at Handois water treatment works, JNWWC decided to invest in a reverse osmosis freshwater de-nitrification plant for the Le Mourier stream in the north-west of the island, which has some of the highest nitrate concentrations in stream water on the island typically exceeding 100 mg/l nitrate for much of the year. This plant, commissioned in 2001, contains two independent, 1000 m3 /d RO trains with low-pressure spiral-wound TFC membranes. It has an Ultra-Filtration membrane pre-treatment plant to remove particulates from the stream water inlet, and post-treatment consists of soda ash dosing for pH correction and chlorine disinfection. The stated objectives of the plant are mixed: partly one-off trial plant, partly to capture a perennial, high-nitrate surface water source. The treated water is channelled to JNWWCs supply network, whereas the nitrate-rich effluent is discharged back to the Le Mourier stream. What is denitrified is the public water supply, not the flowing stream. The environmental pollution is thereby made worse (because of the reduced streamflow but unchanged nitrate volume) but now a second source of nitrate-free water now exists in addition to the desalination plant. Capital costs were 1.2 M to build from scratch in this small valley. Amortised over ten years, this implies about 330/day or 16 p/m3 of product water. Running costs are small, due to shared manpower from the La Rosi` ere plant and the low operating pressure (10 bar), which only requires 0.46 kWhr/m3 of product water. Electricity costs are 10p/m3 . This gives an overall cost of 26 p/m3 for de-nitrification of this freshwater when the plant runs at full capacity. However, when the plant operates only at a low capacity, the capital costs per cubic metre are very much higher. One hours operation per day, rather than 24 hours per day, increases the capital cost per cubic metre by a factor of 24 from 16 pence to 384 pence. With respect to marginal costs, those of the desalination plant are 33 p/m3 and those of the denitrification plant are 10 p/m3 .

16.5

WATER QUALITY MANAGEMENT OF DOMESTIC ABSTRACTIONS: THE DEPARTMENT OF ENVIRONMENTAL HEALTH

About 15% of households are not connected to the JNWWC mains supply (see paper 3). These rural families and farms use their own wells, springs or borehole supplies for historical or geographical reasons. These groundwater sources, estimated to be around 5200 in number, are unlicensed and unmetered. In the great majority of cases, such households are also unconnected to the mains sewerage network and so return their wastewater back into the ground via the usual septic tank and soakaway systems (see below).

148

The Price of Water

Households that are not connected to the mains supply install and operate their own water treatment systems. Prior to 2004, the public health aspects were dealt with by the Department of Environmental Health (DEH) through its monitoring programme of domestic wells and boreholes. Every year, in May and November, the DEH took a sample of 52 boreholes and wells across the island. These were monitored for inorganic pollution, including nitrates. A further 16 from the sample were also analysed for pesticides and microbiological contamination. This routine programme was terminated at the beginning of 2004 when departmental priorities were re-assessed. Currently, DEH assess the water quality of such abstractions whenever a household requests them to do so. A charge is made for the service. Individual households can also monitor their own water quality by making use of private analytical laboratories, and some have subsequently invested in private household treatment systems. Such treatment systems are usually filters, of varying types, used mainly to remove particulate matter, although those incorporating charcoal, silver, or membranes can sometimes remove additional contaminants if kept clean and well maintained. However, none of these domestic filters can remove nitrates, and will not effect any significant changes to the overall inorganic quality of the water. It follows that it is only the EPSDsactivities through the Water Code(see section 3 above) that provides any protection from nitrate pollution for the Islands households unconnected to the mains supply.

16.6

WATER QUALITY MANAGEMENT OF WASTE WATER DISCHARGES: THE ENVIRONMENT AND PUBLIC SERVICES DEPARTMENT

a) Domestic discharges
Section two above showed that domestic waste water discharges are a small but significant source of the nitrate pollution of water on the island. In 1996 Gass, Robins and MacDonald recommended that all soakaways and septic tanks should be taken out of use and sewerage extended to the whole island population. However, they did not say why effective private treatment by households should be eliminated, they did not estimate the economic costs of extending the sewerage network to every rural dwelling, and they did not clarify the relative importance of such discharges as a source of nitrates. This was a clear case of a policy recommendation that lacked any sense of the costs and benefits of the proposal. The Water Pollution (Jersey) Law 2000 requires that owners of septic tanks, tight tanks, soakaways, or private sewage treatment plants have a discharge permit only where such discharges result in or are likely to result in the contamination of controlled waters. As Parkinson et al. (2001: 14) have pointed out, the efficiency of soakaways and septic tanks has been found to vary with maintenance input and with loading. By the end of 2004, the Water Resources Section of the EPSD had issued about 700 such permits (out of a possible total in excess of 4,000). These are written in a generic form, as resources do not allow for regular testing of effluent quality etc. If pollution occurs from a private system and there is a risk to human health, the Water Resources Section would work closely with its colleagues in Health Protection.

Nitrate pollution on the Island of Jersey

149

b) Waste water treatment plant discharges

Earlier in this paper reference has been made to Article 17 of the Water Pollution (Jersey) Law 2000 and its concern (amongst other issues) with discharges to coastal waters. Internationally, the flow of nitrates into coastal waters has been associated with eutrophication. One potential source of such pollution is the discharge of sewage at the coast. Jerseys collection, treatment and discharge of waste water are predominantly carried out by the EPSD itself. The volume of such discharges (including stormwater) in 2004 was some 12 million cubic metres. There are two sewage treatment plants (STPs), one major plant at Bellozanne in the south of the island and a second, small satellite plant recently commissioned on the north coast at Bonne Nuit Bay. The main plant at Bellozanne consists of two inlet 3 mm step screens, fat/grit separation at the inlet works, four primary clarifiers, an activated sludge treatment plant, 12 final clarifiers and a tertiary ultraviolet disinfection system. Discharges from the two plants are regulated within the terms and conditions of a discharge certificate that requires regular sampling to monitor compliance levels for the effluent leaving the works. The conditions in the discharge certificate for the Bellozanne STP are based on the requirements of the ECs Urban Waste Water Treatment Directive (91/271/EC) but prior to 2003 these conditions were less stringent than those set out in the Directive. The STP by the end of 2004 had undergone conversion to a Pegazur system, at a cost of 7.2 million, in order to reduce nitrogen inputs into the nearby St. Aubins Bay in addition to the other benefits of waste water treatment. The 2004 Jersey maximum admissible concentrations of waste water contaminants are given in Table 16.2. From 2003, the EPSD had added total nitrogen to biochemical oxygen demand, chemical oxygen demand and suspended solids as the criteria of the quality matrix for waste water and stormwater discharges. The total nitrogen target at Bellozanne, just north of St. Aubins Bay a favourite bathing beach was introduced in order to reduce eutrophication. There had been limited evidence both in surface and coastal waters of eutrophication in the past with the appearance, for example, of excessive green algae and sea lettuce in the late summer. The JNWWC and the EPSD are by far the largest government spenders on nitrate pollution management. The water utilitys costs of producing blending water were reported in section four above. With respect to the EPSD, its total budget in 2004 was about 500,000, of which nearly 400,000 was on staff costs. However, the proportion of this first figure that was aimed directly at nitrate management was less than 5%.

16.7

THE BENEFITS OF NITRATE POLLUTION MANAGEMENT

Up to this point the papers content has embraced the processes by which nitrates pollute groundwater and surface waters, the means by which such pollution is managed by various Jersey institutions, and the costs of nitrate pollution management. This overview is pulled together in Table 16.3. The time has now come to assess the benefits of water quality management. The outcomes of environmental management consist in the first place of what one can call intermediate objectives. Many specific examples spring to mind from the Jersey

150

The Price of Water

Table 16.1 The Jersey Water Code

Section title About this code Example of sections content

The Water Pollution (Jersey) Law 2000 contains pollution prevention provisions and allows people to be prosecuted if they pollute. Farm manure and The nitrogen value of manures will generally be considerably waste management reduced if applied in autumn or early winter due to losses planning of nitrogen by leaching (particularly on sandy or shallow soils) or denitrification (mainly on poorly drained soils). Slurry A facility for storing slurry should be designed to collect and hold slurry to cope with your farm manure, waste and dirty water. A guide to designing and building slurry storage tanks is given in British Standard 5502: Part 50: 1989. Dirty water It is extremely important to minimise the amount of dirty water to be handled. Clean water from roofs, nearby fields or clean concrete, running onto dirty concrete will increase the amount of dirty water which you need to store and dispose of carefully. Solid manure Stores specially built for solid manure will reduce the risk of pollution through run-off and will make it easier to handle and load the stored material. Silage effluent No part of the silo, [its] tank or channels should be within 10 metres of a watercourse or field drain, which silage effluent could get into if it escaped. Fuel oil Provide bunding to contain any spillage from above-ground agricultural fuel oil tanks or areas where fuel drums are stored. The bund should be able to store 185% of the store volume. Pesticides You MUST get agreement from the Water Resources Section of the Planning and Environment Department and the Jersey New Waterworks Company Ltd, before you use herbicides to control aquatic weeds in or near water. Disposing of animal All dead animals, ie. Horses, cows, sheep etc requiring carcases disposal must be notified to the Public Services Departments Knackers yard on Tel: 619281, after having consulted their own vet as to the cause of death. Liquid fertilizers All hatches, lids and valves should be securely closed before [road] tankers or bowsers are moved, and valves should be locked when unattended. Nitrate and Do not apply extra fertilizer to be on the safe side. The phosphorus amount of nitrogen fertilizer applied should not exceed the crop requirement as this increases the amount of nitrate lost by leaching and is a waste of money. Specialised Recent research has shown that the nitrate concentration in horticulture feeds for tomatoes can be partially replaced by chloride, resulting in reduced nitrate in the run-off.

Table 16.2 A quality matrix for the EPSDs Waste water discharges to Jerseys Coastal Waters in 2004
Concentration Measure milligrams per litre milligrams per litre milligrams per litre milligrams per litre 150 none 45 20 35 15 250 125 125 not available not available not available 50 25 25 not available MAC in 2002 MAC in 2003 MAC in 2004 Actual 2004 Minimum Actual 2004 Mean not available not available not available not available Actual 2004 Maximum not available not available not available not available

Biochemical oxygen demand Chemical oxygen demand Suspended solids Total nitrogen

Source: PSD pers.comm.

Table 16.3 Nitrate Pollution Management on Jersey in 2005

Pollution source Policy instruments (1) Cost type Agricultural fertilizers Agricultural fertilizers Agricultural fertilizers Unsewered domestic properties Agricultural fertilizers Water Pollution (Jersey) Law 2000. Discharge permits where deemed appropriate. Monitoring of domestic wells and boreholes on request. In addition, households install water treatment systems. Pegazur sewage treatment plant reducing total nitrogen in waste water discharges. One desalination plant and one de-nitrification plant producing blending water. Water Pollution (Jersey) Law 2000, Articles 17&18. 2004 Water Code. Benefit type Human health

Institutional manager

Jersey New Waterworks Company Environment and Public Services Department Environment and Public Services Department

Environment and Public Services Department Department of Environmental Health

Infrastructure costs plus maintenance and operation. Advisory services, monitoring, legal action. Infrastructure costs plus maintenance, operation and monitoring. Monitoring.

Environmental quality and human health Environmental quality. Environmental quality and human health Human health

Monitoring plus household expenditure on treatment systems.

1. Note that these instruments fall into three categories only: regulatory, infrastructural, and programmes of friendly persuasion. No economic instruments are used that might reduce nitrate fertilizer use, such as a tax on imported fertilizers. The traditional political strength of the farmer lobby would have made this difficult to receive approval by the Islands legislative bodies. Finally, water pricing in the household sector would reduce families water use, for example on gardening, and therefore reduce the need for blending water production.

Nitrate pollution on the Island of Jersey

153

case-study. The Department of Agriculture and Fisheries efforts leads to the construction of more facilities for storing slurry that are based on British Standard 5502 (see Table 16.1). The Jersey New Waterworks Company reduces the nitrate content of the water supply by blending in the desalinated water output of La Rosi` ere. A Department of Environmental Healthstest identifies nitrate pollution of a family well and the household ceases to use it for drinking water and for cooking. The Environment and Public Services Departments monitoring programme leads to the repair of a defective septic tank and the same Departmentsconversion of the Bellozanne sewage treatment works to the Pegazur system reduces total nitrogen inputs into St. Aubins Bay. But there is no sense in shouldering the economic costs of water quality management for its own sake. Intermediate objectives, from the micro-scale of a repaired septic tank through to the macro-scale of reducing total nitrogen inputs to the sea, are nothing more than means to an end. That end we can call final benefits. The final benefits of water quality management are three-fold: enhanced environmental quality, improved human health, and any additional economic benefits not covered by the health and environment categories. We shall consider these three benefit groups in turn with specific reference to the management and reduction of nitrogen pollution of the controlled waters of Jersey, i.e. the territorial sea adjacent to the Island as well as its coastal waters, inland waters and ground waters. In the case of neither the environmental nor the human health benefits have we assigned them monetary values. This is because such estimates require the use of contingent value surveys, which we consider invalid for reasons expressed at length in other publications (see Merrett 1997: 168-175 and paper 14).

Environmental benefits
First, the lower nitrate content of groundwater and of streamflow is considered by ecologists to have benefits in terms of providing increased bio-diversity for all forms of aquatic life. The delivery of these benefits, specifically with respect to macroinvertebrates, has been strongly supported by the work of Langley and Kett (2005). Secondly, the production of desalinated seawater reduces the necessary volumes abstracted from the islands ground and surface waters, thereby increasing streamflow, especially in the summers low-flow conditions. However, Merretts work suggests that over a full year desalination output is considerably less than 10 per cent of the total water supply (See paper 3, Table 3.3). We have already seen that in 2004 there was no desalination output. Thirdly, the reduced flow of nitrates to inland and coastal water has reduced eutrophication, both in the Islands streams as well as in the bathing waters of St. Aubins Bay. Specific future research to assess these benefits in their own terms, rather than via money values, will help the people of Jersey understand better the advantages of the measures taken to reduce nitrate pollution. <anchor role=section id=c01c01 label=/>

Human health benefits

Our focus here is the likely human health benefits of Jerseys attempts to improve drinking water quality by cutting the maximum permitted nitrate level in such water from 100 mg/l to 50 mg/l.

154

The Price of Water

The general position is that a high concentration of nitrate in the diet is linked to methaemoglobinemia in babies under six months and to stomach cancer in adults (Lott et al. 1999: 79). No such link exists where the concentration is low. It is never argued that nitrate should be completely excluded from our diets. For example, the consumption of vegetables is a source of nitrate in the human diet, as is drinking water. It is necessary, therefore, that public health legislation secures a countrys population from the ingestion of NO3 only at concentrations that endanger health. Such legislation requires in its turn a definition of the maximum admissible concentration in the drinking water supply. We have already seen in section 2 above that the MAC of nitrate in drinking water was originally set by the WHO in 1970 at an acceptable level of 100 mg/l but that EC legislation in 1980 reduced that level to 50 mg/l. In section 4 it is noted that Jersey has been under no legal requirement to comply with this standard and has struggled to do so. Only by 2008 at the earliest is it likely that Jersey will be able to fully comply with the 50 mg/l standard. Excluding waste water discharges, all the costs of nitrate pollution management in the last 25 years have been politically driven by the ECs 1980 shift to a lower acceptable level of nitrates in drinking water. Once again, we need to ask: what have been the consequential health benefits of the shift from 100 mg/l to 50 mg/l? In fact medical experts in the UK were asked to address this specific issue. A Joint Committee on the Medical Aspects of Water Quality appointed by the Department of the Environment and the Department of Health and Social Services stated that There is no compelling evidence to suggest that significant risks to health are encountered when water containing between 50 and 100 mg/l nitrate is supplied to the public (Gass et al. 1996: 35-36; JNWWC 2002: 29; Robins 2000: 15-17). If the Joint Committees views are still a true reflection of the epidemiological evidence, then there have been no health benefits from the shift from a 100 mg/l standard to a 50 mg/l standard. What is striking about the standard shift in 1980 from 100 units to 50 units is that the level was simply halved. Our working hypothesis is that the persons responsible for changing the standard had no clear evidence for change and that it was the precautionary principle that drove the decision rather than scientific research. ORiordan (1995: 9) suggests that at its simplest the concept of precaution has four meanings. The relevant meaning here is: Thoughtful action in advance of scientific proof of cause and effect based on the principles of wise management and cost effectiveness, namely better to pay a little now than possibly an awful lot later. In this sense, precaution is a receipt for action over inaction where there is a reasonable threat of irreversibility or of serious damage to life-support systems. But in the case of the 1980 standard shift, thoughtful action based on scientific evidence had already been taken prior to 1980 and the shift itself entailed paying an awful lot earlier as against not paying it at all. The only benefit of the shift that is absolutely clear is the growth in the production and sales of the water treatment and bottled water industries. In fact, epidemiological evidence on the standard shift is possible to collect and analyse in Jersey. We have seen that about 15% of the population are off the mains supply and therefore access tap water from sources that are completely untreated. Even if treated, these supplies have none of their nitrate content removed (see section 5). Here is potential evidence for any possible links between ill-health and nitrate standards in excess of the 50 mg/l prescription. But at the end of 2003, the threat was

Nitrate pollution on the Island of Jersey

155

considered to be so slight that the Department of Environmental Health terminated its routine sampling of wells and boreholes nitrate levels.

Economic benefits
An active (and well-publicised) role of the States of Jersey in environmental protection might have some additional economic benefits to the Island by securing greater tourist numbers than would otherwise be the case. In fact, Jersey because of the nitrate in drinking water issue has had a bad press in the last two decades in this respect. Noone knows or ever will know how much greater (if any) tourist numbers would have been in the absence of this critical coverage in the media. In any case, a reduction in tourism on Jersey would have brought an increase in tourism elsewhere, on Guernsey for example, with no net benefit to tourists or the European tourism industry.

16.8

CONCLUSIONS

The papers objectives were first, to review the anthropogenic processes by which the Island of Jerseys waters have been polluted by nitrates over a broad span of time and secondly, to assess the costs and benefits of the institutional management of nitrate pollution within a framework ultimately shaped by the European Community. On an island where potatoes and dairy farming have been historically dominant, Jersey has in the past suffered from perennially high nitrate levels in all its water bodies, a syndrome exacerbated by soil, geology, climate, cropping and farming practices, especially excessive fertiliser use. A minor addition to the nitrate load originates from rural, unsewered dwellings. Nitrate pollution of water bodies, above certain levels, undoubtedly creates potential threats to human health as well as environmental eutrophication. The management of nitrate pollution on the Island in 2005 is principally the responsibility of two institutions. These are the Jersey New Waterworks Company Ltd. and the Environment and Public Services Department. With respect to the role of the JNWWC, the water utility in which the States of Jersey have a majority shareholding, the States in 2005 permits up to 33% of its annual samples of distributed water to exceed the 50 mg/l MAC provided they are less than 70 mg/l. The utility can reduce the nitrate content of its abstracted water by blending in desalinated water, which has a near-zero nitrate content. The Islands first desalination plant was installed in 1970 to deal with drought years, especially in the summer months when the tourist population is high. In 1997 the old plant was demolished and replaced with a Reverse Osmosis plant with a capacity of 6000 m3 /day, using electrical pumping energy. Since the desalination plant is required for drought conditions, the supply of water for blending purposes can be costed simply on its use of consumables (mainly electric power at a cost per cubic metre of 33p), a little more than double the average total cost of the main surface water supplies and substantially more than the marginal cost of surface water. By 2004 the redefinition of the required nitrate standard and the successes of nitrate pollution management in the farm sector resulted in no requirement for desalinated waters use in blending, which now may become a thing of the past.

156

The Price of Water

The JNWWC also operates a 2,000 m3 /day denitrification plant on the north-west of the Island where some of the highest nitrate concentrations in stream water are to be found. It is unclear to the authors of this paper why this plant was built. The basis was either to add a new surface water source to the utilitys abstraction choices or it was to provide a second source of nitrate-free water for blending purposes. The plants overhead costs are about 330/day but the marginal costs are only 10 pence per cubic metre. For the year 2004 its average daily output was not available to us but was probably low and therefore its average total cost per cubic metre would have been high. With respect to the agricultural role of the EPSD, it is the Water Pollution (Jersey) Law 2000 that makes it an offence to cause or knowingly permit pollution of controlled waters. In case of prosecution, a farmer can defend himself by showing due diligence under the 2004 Code of Good Agricultural Practice for the Protection of Water. The Water Pollution Law and the Water Code together provide the States of Jersey with teeth in reducing agricultural pollution. By 2005 there was a growing body of evidence that the EPSD (and the Department of Agriculture and Fisheries prior to 2004) were winning the war against nitrate pollution with the stick of the Water Pollution Law and the carrot of environmental education and friendly persuasion. At the same time, changes in the management of farms and in the size of the potato crop were having positive effects in reducing diffuse nitrate pollution. The Environment and Public Services Department also monitors septic tanks, tight tanks, soakaways and private sewage treatment plants. These provide only a small proportion of the Islands nitrate pollution. In addition, the Department is responsible for the waste water treatment plants at Bellozanne and at Bonne Nuit Bay. Since 2003, the Bellozanne plant has been converted to the Pegazur system in order to reduce the total nitrogen content of discharges into St.Aubins Bay to a maximum admissible concentration of 15 mg/l. The impact on coastal eutrophication has been positive. Having reviewed the costs of nitrogen pollution management, the paper moves on to the benefits of such management. A distinction is drawn between the intermediate objectives of water quality management and its final benefits. The latter comprise enhanced environmental quality, improved human health, and any additional economic benefits not covered by the health and environment categories. Taking these in reverse order, there is no evidence that nitrogen pollution of Jerseys controlled waters has reduced its tourism income, although such a reduction may have existed in the past to the economic advantage of other tourism centres. With respect to human health benefits we should note that the Islands judicial systems are independent of the English courts and at no time has Jersey been subjected to the administrative systems of the United Kingdom government or the European Community. However, the Islands government has taken a series of actions since 1980 to bring Jersey into line with the ECs reduction of the maximum admissible concentration of nitrate in drinking water from 100 mg/l to 50 mg/l, that is, 50 parts per million. However, a UK Joint Committee on the Medical Aspects of Water Quality has reported that there is no compelling evidence of significant risks to human health when water containing between 50 and 100 mg/l nitrate is supplied to the public. If the Joint Committeesviews are still a true reflection of the epidemiological evidence, then there have been no human health benefits from the shift from a 100 mg/l standard to a 50 mg/l standard in Jersey or anywhere else. In that case, the only financial beneficiaries of these unwarranted costs have been the water engineering and bottled water industries.

Nitrate pollution on the Island of Jersey

157

With respect to enhanced environmental quality, the rate of reproduction of macroinvertebrate populations has already been referred to; indeed it is used as a measure of water quality. Secondly, the production of desalinated seawater when it takes place modestly reduces the necessary volumes abstracted from the islands ground and surface waters, increasing streamflow. Thirdly, the reduced flow of nitrates to inland and coastal waters has reduced eutrophication and its green algae and sea lettuce manifestations. No estimates in money terms have been made of these certain benefits, nor do the authors consider such estimates useful. However, a systematic record of the environmental advantages of nitrate pollution management may be considered valuable in showing the citizens of Jersey what benefits arise from the management of nitrate pollution.

* The authors are grateful to Gerry Jackson (EPSD), Iain Norris (EPSD), Howard Snowden (JNWWC) and Dr Duncan Nicholson (Department of Environmental Health) for detailed discussion with us in the course of preparing this paper.

17 Catchment water decits in the twenty-rst century

Ye can discern the face of the sky; but can ye not discern the signs of the times? St Matthew ch. 16, v. 3

17.1 INTRODUCTION
The objective of this paper is to provide a general theory of the manner in which the water resources of a catchment shift from surplus to decit, and the means by which water resource institutions can manage or reverse this shift. The text requires the clarication of some key concepts and this is done as the argument proceeds. Note that catchment and river basin are used interchangeably. The term watershed is avoided because it has rival and inconsistent denitions. The broad hydrological framework here is perfectly conventional. Total precipitation in a catchment, where the principal river ows to the ocean, is consumed in part as evapotranspiration; the residual is referred to as effective precipitation. This latter ow recharges the basins aquifers, the boundaries of which fall within the catchment. Effective precipitation also sources run-off, most strikingly in the case of the main river itself. The aquifers supply base-ow to the river system as well as discharging as springs. Finally, both surface water and groundwater discharge at the coast. The broad hydrosocial framework is set out in Table 17.1. It represents only outstream ows. Outstream water is that which is abstracted by human society from lakes, rivers and aquifers, to be stored and channelled for human use. As Table 17.1 shows, both supply and use may have negative as well as positive entries. Total net supply is mathematically identical to total use; this is empirically conrmed only if all the supply and use entries are comprehensively assessed and accurately measured. The evapotranspiration of outstream (or instream) water is hereafter referred to as water consumption.
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

158

Table 17.1 A catchments hydrosocial balance. Row 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Title Categories of Supply (Positive) Rainwater collection Groundwater abstraction Surface water abstraction Desalination of sea water Imports of water from other areas Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored Total Supply Categories of Supply (Negative) Supply-side evapotranspiration losses Supply-side leakage Exports of water to other regions Net rise in water abstracted and stored Total Negative Values (Supply) Total Net Supply Categories of Use (Positive) Households Agriculture, including irrigation requirements Mining Manufacturing Construction Public services Private services Other uses Total Use Categories of Use (Negative) Evaporation losses on users properties Leakage on users properties Total Negative Values (Use) Total Net Use Base year volumes a b c d e f g h i j k l m n in o p q r s t u v w x y z wz

Source: Adapted from Merrett (2002a) Table 7.1.

It has already been stated that the paper addresses river basin surpluses and decits. Therefore, these two concepts require denition. A catchment water surplus is a situation in which, throughout the course of a specied year, total precipitation in the basin is sufcient to simultaneously satisfy four conditions: 1. 2. Abstraction from the aquifer is maintained at a sustainable rate. Outstream water fully meets the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors.

3.

4.

The basin populations economic demand for food is fully met from the domestic rainfed and irrigation sectors and/or from food imports nanced by the basins commodity and service exports and/or from domestic sheries. The rivers instream ows do not fall below dened minima.

When a catchment water surplus does not exist, the river basin is in decit. Whether or not conditions 14 are met is a matter of judgement. That is the human condition. For example, sustainable aquifer management is a multi-year activity because storage naturally uctuates over time with annual effective precipitation. The presentation of the general theory below uses a thought experiment applied to a ctional catchment under threat of a water decit and continues by reviewing a wide range of redemptive options. In the light of what follows, the reader should recall that the Oxford English Dictionary cites J. Dwyer in 1847 as the rst person to use the word catchment (OUP 1979: 355).

17.2 THE DWYER CATCHMENT

Consider a river valley, with its many tributaries, that ows to the ocean. One can call it the Dwyer. Total precipitation, averaged over the years, has remained constant throughout living memory, despite climate change. The Dwyer catchment has been settled for centuries and contains a network of villages and towns as well as a single metropolis. For many decades the catchment population has grown at a steady rate such that total population increases by one-third every 25 years. Curiously, this rate of growth is equal to projected global population increase in the rst 25 years of the twenty-rst century (Merrett 2002a: 1). The Dwyer River Basin is a source of considerable economic output through its hydroelectric power schemes, its navigation activities, its shing catches and its location for conservation, recreation and environmental tourism. Moreover, diversions and pumping from the Dwyer and its tributaries, as well as spring capture and the pumping of groundwater have been sufcient throughout the basins history to meet outstream use. The catchment is recognized currently to be in surplus. The water law of 1847 protects the catchments instream ows, its environmental needs as some call them, in terms of both quantity and quality. The quantitative rule for the Dwyers annual ow to the ocean is that it should never fall short of one-third of the past 10 years average of net precipitation within the basin. This is referred to as the one-third rule. Net precipitation is the estimate of what effective precipitation would be, were there no dams, no hydrosocial supply and no hydrosocial use of water in the basin. This revision of actual effective precipitation is made in order to avoid prioritizing, for example, evaporation from reservoirs or irrigation water consumption over the Dwyers instream ows. As with many other basins, such as the Ganges/Brahmaputra and the Gambia, the measurement of the rivers discharge to the ocean is not easy. It is carried out at the rst feasible point upstream from the coast. Of course, the one-third rule requires an effective system of abstraction licensing and drought management. Instream ows are valued by the people and their government not merely because of their economic output, already referred to above, but also because of the intrinsic value of the riverine and coastal ecology.

Table 17.2 Population, productivity and output change over time. Rate of growth of population in per cent per year Rate of growth of output per head in per cent per year 0 1 2 3 4 5 0 1.000 1.010 1.020 1.030 1.040 1.050 1 1.010 1.020 1.030 1.040 1.050 1.061 2 1.020 1.030 1.040 1.051 1.061 1.071 3 1.030 1.040 1.051 1.061 1.071 1.082 4 1.040 1.050 1.061 1.071 1.082 1.092 5 1.050 1.061 1.071 1.082 1.092 1.103

Note: The statistics above, lying between 1.000 and 1.103, can be used as multipliers to show how output changes between any two successive years. The decimals give the rate of change of output. For example, 1.051 implies a rise of 5.1 per cent in output between any two years. At this rate of growth, output would double in 14 years.

It has already been seen that the catchments population is increasing. Output per capita is also rising. This combination of growth both in population and economic productivity, associated with the rise of world capitalism, is the source of deep unease amongst the professional staff of the Dwyer Catchment Authority. The Authority has used Table 17.2, and data from the Ministry of the Economy and Population, to estimate the catchments economic output in 2025. After assuming that the ratio of outstream water use to basin output is constant, the Authority also has a derived estimate of total water use in 2025 (row 33 of Table 17.1 for the scenario year 2025). The Authority anticipates that, for the rst time in its history, the catchment will move into a water decit; one or more of the four necessary conditions for surplus will have been breached.

17.3 REDEMPTIVE OPTIONS (I)

The Dwyer Catchment Authority responds by considering and starting to implement redemptive options that moderate the situation but that do not prevent entry into water decit. There are six such adaptive measures. The rst option is to amend the 1847 Act so as to lower the measure of environmental needs below the 33% oor. The decline in environmental ow (and the ecological and economic effects that ensue) is accepted by the catchment authority (if not by the environmental movement) as a price worth paying to maintain the growth of population, economic productivity and total output. The second option is to promote the reuse of abstracted water as a means to cut the rate of abstraction. Reuse can take place within a single institution (such as a sugarmill) or it can occur between institutions as when urban waste water is reused for irrigation purposes. Reuse may have real advantages, such as lowering abstraction costs or by reducing waste water treatment costs. But there is no reason to believe that reuse reduces the consumption of water supplied to and used by households, rms, etc. One can take a small sugar mill as an example. In the absence of reuse it

abstracts 1000 m3 per week, 150 m3 of which is consumed in evaporative losses leaving 850 m3 returned to the river. In the case of reuse only 500 m3 are abstracted on Monday and by Wednesday 75 m3 are consumed and replaced from storage. By Saturday the reuse stage is complete and a further 75 m3 consumed, once again to be replaced from storage. The reuse option has abstracted less than in the absence of reuse (650 m3 including storage recharge), but there is no difference in total consumption. Since reuse has not brought with it a fall in consumption it has not served to protect the instream ow. The third option is rainwater harvesting. These projects can be successful at the local level; they are specic forms of water capture closely related to surface water abstraction. As with reuse, there is no fall in consumption and no protection of the instream ow. The fourth option is to import water from another catchment. This permits the sustainable management of the Basins aquifers, sufcient outstream supply to match the economic demand for water, satisfaction of the economic demand for food and continued protection of the Dwyers environmental ows. The fth option is the desalination of sea water and other saline waters. The power requirements are considerable but recent falls in the factory-gate price per cubic metre make this option an interesting one for the Dwyer Catchment Authority. The use of sea water itself, requiring dual supply infrastructures, is also of interest to the Authority in some sectors. Hong Kong sets an example in this respect. The last of the six options that the Authority may take to win hearts and minds as the entry to water decit begins is to import food. This permits administrative or legal restrictions on the scale of irrigated agriculture such that there is now sufcient water for all other economic demands. A benecial allocation multiplier exists here. If water use for irrigation and for urban purposes divides in the ratio 70:30 (Shiklomanov 2000: Table 5), then a 15% transfer of agricultures total creates a 35% increase in the urban total. The consumption multiplier is even greater. If, water consumption by the irrigation sector and the urban sectors divides in the ratio 93:7 (ibid.), then a mere 5% transfer of agricultures total permits a 66% increase in the urban total. Each of these six options above needs to be reviewed. The rst redemptive option reduces environmental ows below what had been considered the minimum acceptable, breaching condition 4 for water surplus (see Section 17.1). The second option is welcomed on economic grounds but makes no contribution whatsoever to defending the basins water surplus. Nor does the third option. The fourth measure places reliance on the precipitation of other catchments rather than precipitation in the Dwyer River Basin. Parallels spring to mind with Israeli pumping of Lake Tiberias, the plans to divert water from the Ebro River to south-east Spain, and the southnorth transfer in China (Pearce 2003). The fth measure is noteworthy but, like option four, only compensates for a shortfall in the Dwyers precipitation. The sixth option is in breach of condition 2 for water surplus. Moreover, if the basins exports are insufcient to nance food imports, condition 3 is breached. The Dwyer places reliance on other catchments food production and, indirectly, their precipitation. It is worth noting that the argument above relies (implicitly) on a distinction between the terms reuse and recycling. Reused water is waste water and irrigation drainage that, prior to its return to the hydrological resource, is captured and used again (perhaps repeatedly) for the uses listed in Table 17.1. Recycling refers to water that is abstracted, on the supply side of the hydrosocial balance, used by households, industry etc., and then the fraction that is not consumed

ows directly back to the catchments rivers, aquifers and lakes. The importance of recycling to hydroeconomists is that it augments the hydrological resource from the point at which the recycling occurs. The negative characteristic is that recycled ows may pollute the resource. The proportion of water used that is consumed, and therefore is unavailable for recycling, varies between categories of use. Therefore the relative shift of use between Table 17.1s categories of use will change water consumption for a given volume of use. The example above of the benecial multiplier illustrates this.

17.4 REDEMPTIVE OPTIONS (II)

The Catchment Authority continues by considering redemptive options that do prevent entry into water decit. Once again, there are six such adaptive measures. They are numbered 712 to maintain continuity with redemptive options (I). The seventh option is to reduce the rate of growth of population. Other things being equal, this will reduce the rate of growth of output and thereby the rate of growth in outstream water use. A fall in demographic increase within the Dwyer River Basin would be achievable if its women become more powerful in the determination of the number of children conceived, if access to contraception is facilitated, if social provision for elderly persons is strengthened, and if limits are placed on net inward migration. Chinas population policy since 1949 has contributed enormously to the impact its society has on the available water resource, in comparison with the population totals it would now have in the absence of that policy. The eighth option is to reduce the rate of growth in output per person and thereby the rate of growth in output and outstream water use. In the Dwyer River Basin this would be difcult to defend because there is extensive poverty in the catchment. In basins with higher per capita incomes it would require a fundamental change in cultural values and economic organization, a farewell to capitalisms relentless search for output growth founded on market expansion. In the case of options seven and eight one sees that these require action that clearly falls outside the eld of water resources management. But this is not true of our nal four options. The ninth adaptive measure is to increase outstream waters productivity in terms of value-added (in US$) per cubic metre of water consumed. It is consumption, not use, which reduces the volumetric ow of the Dwyer. Water productivity augmentation is a subject of discussion, research and implementation throughout the catchment with respect to all the outstream users. The greater is waters productivity, the lower the volume of water consumed for any given level of future output. It has already been seen above that at the global scale 93% of all consumption of water withdrawals takes place within the irrigated sector. In the Dwyer Catchment it is irrigation yields (in terms of output per hectare, output per cubic metre of water used and output per cubic metre of water consumed) that are considered as the key variables in respect of option nine. The positive link between food prices and crop yields is of especial interest (Berkoff 2003). The tenth option relates to the economic demand for water use under condition 2 for water surplus. The economic demand for water is analysed by economists as a functional relationship between the price of water (money charged per unit volume) and the total volume users wish to purchase at that price. A higher price is usually associated with a lower volume purchased. If it is possible to extend water pricing in

the Dwyer River Basin as well as to raise the current level of prices, then the volume of water purchased by households, farmers, companies and other institutions will diminish. Differential pricing for users that recycle a high proportion of their supply helps reduce water consumption. The introduction of pricing, particularly in the irrigation sector, can be immensely difcult (Merrett 2002b). A subsidiary arrangement under option ten is to facilitate an urban market in farmers water-rights, taking advantage of the allocation multiplier. A positive spin-off from pricing is the stimulus it gives to higher water productivity. A negative spin-off would be a contraction of irrigation output leading to food imports that the Dwyer population is unable to nance out of non-farm exports. Water pricing can also be deployed to moderate abstraction and use on a seasonal basis; this can be of value when precipitation, evaporation and river ows vary strongly during the course of the year, as is the case with the Dwyer catchment. The eleventh option is to reduce the evaporation from dams and reservoirs. A deep dam has an advantage here over several shallow dams with the same total storage capacity. Articial aquifer recharge, substituting for surface storage, is of great relevance in those areas of the Dwyer River Basin where the rate of evaporation is high because of heat, wind and low humidity. The twelfth and last option, for society as a whole, is to establish water resource institutions (public, private and cooperative) that have an effective leadership, are properly resourced, of considerable expertise, strongly committed to the agenda of a sustainable society and, in the case of the Dwyer Catchment Authority, with the necessary regulatory powers (Abernethy 2003; Merrett 1997: 14362).

17.5 FROM SURPLUS TO DEFICIT

At this point it is necessary to consider whether or not it is important that one catchment is in decit and whether or not it is of any signicance that a second catchment is moving from water surplus towards water decit. The issue can be approached by re-stating the implications for a river basin of nding itself in water decit. These are one or all of the following: 1. 2. 3. 4. The basin is pumping its groundwater at an unsustainable rate. There is insufcient outstream water to meet economic demand. The population has to import water, or food that it is unable to pay for from its exports of goods and services. The basins citizens must accept the economic and environmental losses following from its river diminishing in volume.

The nightmare scenario is a basin in which groundwater is being exhausted, households, farmers and other actors cannot purchase the water they require, food imports cannot be paid for, making the basin dependent on powerful allies, and the river has been destroyed. Consider too the global catchment the set of all the worlds river basins. As the worlds population increases from six billions in 1999 to some eight billions in about 2025, and as world economic growth maintains an even higher rate of expansion, is it not the case that the global catchment itself, whilst still in surplus, will embrace ever

more individual river basins that are in decit? As the global catchments surplus diminishes, will not the situation of specic catchments in decit become ever more precarious? Moreover, since water ows uphill to wealth and power, will not the social classes that have neither wealth nor power feel these decits most intensely? It is this prospect that is the basis of the importance of water decits and the signicance of the shift of river basins from water surplus to decit (Lundqvist 1998).

17.6 FRAMEWORK, THEORY AND EMPIRICAL STUDIES

It may be that this approach to catchment decits and surpluses will be of interest to the international community of researchers working on water resources and their management. If that is so, the analytic framework requires testable hypotheses. The general hypothesis is: Catchment X is in water decit, where Catchment X is the specic river basin to be researched. To test the general hypothesis, the most basic information on the catchment that is required for a catchment in a specic year is:

Total precipitation. Effective precipitation. Net precipitation. The sustainable rate of abstraction of groundwater. The actual rate of abstraction of groundwater. The performance of water suppliers in meeting the economic demand for water

from households, agriculture, mining, manufacturing, construction and the services sectors. The hydrosocial balance of the catchment will prove useful here. The ability of consumers to exercise their economic demand for food. Where such food is imported, are these imports purchased at the international market price and is this nanced from the catchments foreign exchange earnings? The quantitative ow of the river at the point of measurement of its discharge to the ocean. A statement of the dened minima for the rivers instream ows and the actual flows that correspond to these environmental standards. If no environmental standards are in place, the Dwyers one-third rule may be used as a substitute.

In the light of this information, if the four conditions for water surplus of section one are met, and if this is possible without reliance on desalination or water imports, then the general hypothesis is falsied and the catchment is in surplus. If not all four conditions are met, but only as a consequence of water exports, the catchment again is in surplus, although it may suffer the symptoms of decit. The Jordan River Basin may be an example. Where the general hypothesis is veried, an appraisal of redemptive options 712 is overdue. Where a catchment is in decit, one can hypothesize that, in comparison with basins in surplus, net precipitation per head is low, or output per head is high or water productivity is low the causal hypothesis. Finally, if a number of Ph.D. students, professional researchers and institutions adopt this approach, test the hypotheses and appraise the redemptive options, a splendid panorama opens for interchange in respect of framework, theory, and empirical studies.

17.7 CONCLUSION
The paper introduces the concepts of catchment water surplus and catchment water decit. It describes how a shift from surplus to decit can take place as population and output per head in the catchment increase. Twelve redemptive water policy options to manage or halt this shift are set out and the signicance of such shifts at the global scale are considered. The paper ends by formulating two testable hypotheses derived from the general theory and invites professional colleagues to pursue, within this paradigm, investigations of specic river basins.

The material in this chapter originally appeared in the journal Water Policy, Volume 7, Number 2, 2005: pp. 141 149.

18 Catchment water deficits: an application to Zambias Kafue river basin

18.1

INTRODUCTION: THE GENERAL HYPOTHESIS

In paper 17 above I proposed that the quantitative sufficiency of water in any given river basin can be expressed as a catchment water surplus. This is a situation in which, throughout the course of a specified year, total precipitation in the basin simultaneously satisfies four conditions: 1. 2. Abstraction from the catchments aquifers is maintained at a sustainable rate. Outstream water fully meets the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors. 3. The basin populations economic demand for food is fully met from the domestic rainfed and irrigation sectors, domestic fisheries and food imports financed by the basins exports. 4. The rivers instream flows do not fall below defined minima. In brief, the catchments precipitation is sufficient to meet the rivers environmental needs, for groundwater to be pumped at a sustainable rate, for the economic demand for water to be satisfied and for the populations food requirements to be met. Where one or more of these conditions is not met, we have a catchment water deficit. Paper 17 then continues by setting out 12 redemptive options that either moderate the situation of deficit basins or that forestall the entry into water deficit. The paper concludes by setting out two hypotheses, calling for networked international research in testing them. The general hypothesis for any basin is that catchment X is
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

167

168

The Price of Water

in water deficit and the causal hypothesis is that deficit catchments are characterized by low precipitation per head, high output per head and low water productivity. With respect to the general hypothesis, it is shown to be verified the catchment is in water deficit if any one or more of the four conditions listed above is breached. Some time ago, the author was given access by Scott Wilson Pi esold, a British engineering firm, to its exhaustive study of the Kafue River Basin (Scott Wilson Pi esold 2003). This Report had not, of course, been prepared in order to test the general and causal hypotheses referred to in above. For example, the Report investigates not the Zambezi River but one of its immense tributaries, the Kafue. Moreover, Scott Wilson Pi esolds work is primarily directed at environmental issues. However, their research ranges so extensively that it seemed likely that it could, indeed, be used in a preliminary investigation of both catchment water deficit hypotheses. Moreover, an initial deskbased study, with the insights and puzzles it was likely to deliver, would also help structure future field research into other catchments. The appendix to this paper sets out the questions the author now believes are most useful as the starting point in any study of catchment water deficits and their drivers.

18.2

THE KAFUE RIVER BASIN

The Kafue River Basin is located entirely in Zambia in Africa (see Figure 18.1). It has an area of 157,000 km2 and its length is 1,300 km. At its junction with the Zambezi its discharge is approximately 300 m3 /second. The Basin is conventionally divided into four regions: the Upper Kafue, the Middle Kafue, Kafue Flats, and the Lower Kafue. The Basin has a tropical climate with two distinct seasons: a wet season between November and March and a dry season between April and October. At the end of the wet season humidity is 75% and at the end of the dry season it is 45%. Mean daily temperatures vary from 13 and 20 Centigrade in July and between 21 and 30 in November. Sunshine hours during the dry season are 13 hours per day and are 7 hours per day in the wet season. The long-term average precipitation is 1,057 mm. Pan evaporation varies between 1900 mm and 2200 mm during the year. Effective precipitation is 53 mm (Scott Wilson Pi esold 2003, Table 3.7). In respect of groundwater, the best aquifers are associated with i) limestone formations around Lusaka and at locations in the Upper Kafue, including Mpongwe, and ii) the alluvial sands and gravels of the Kafue Flats, Lukanga Swamps, and the Nanzhila, Lufupa and Luswishi rivers. Groundwater potential is reported as a depth per unit area as follows: Upper Kafue 85 mm/y, Middle Kafue 75 mm/y, Kafue Flats 70 mm/y, Lower Kafue 55 mm/y. Groundwater is deemed private. Water abstraction is predominantly of surface water, although there is some groundwater abstraction for irrigation and rural domestic water supplies. Pumping water from mines for dewatering purposes takes place on a large scale, in order to facilitate copper production. One of the Kafue Basins copper mines is said to be the wettest mine in the world. In respect of surface water, the immense discharge at the Zambezi of 300 m3 /second is very much less than total rainfall. This is because of evaporation and transpiration from areas of impeded drainage and groundwater seepage zones, and because of high solar radiation in Zambia. Amongst the flood plains, swamps and marshy areas, Kafue Flats and Lukanga Swamps are the most significant. The Kafue Flats extend from

Catchment water deficits

25 E 26 E 27 E 28 E 29 E

169

Map 1: The Kafue River Basin

12 S

12 S

Upper Kafue
13 S 13 S

Middle Kafue Mpongwe

14 S
Kafu

14 S
e River

Lukanga Swamp

15 S Kafue Flats

15 S

Lusaka
Itezhi-Tezhi Kafue Gorge

Lower Kafue Flats

Ri v

16 S

Mazabuka

Kafue

er

16 S

Lake Kariba

25 E

26 E

27 E Scale 1 : 3 000 000 0 40 90 120 km

28 E

International boundary Kafue Basin Regional boundary River

Swamp/flats Kafue Gorge Lake

Figure 18.1 The Kafue River Basin.

Za m b e

29 E

zi

170

The Price of Water

Itezhi-Tezhi dam to the head of the Kafue Gorge, a distance of 260 km with a drop of only 6 metres. The Kafue Flats have an area of 6,500 km2 and the Lukanga Swamps area is 2,600 km2 at peak water level. There are also numerous dambos (local shallow depressions). Hydroelectric power is an important regional (and national) resource. Itezhi-Tezhi reservoir and dam provides regulatory storage to enhance dry season flows and ensure firm energy availability for the downstream Kafue Gorge Upper (KGU) hydroelectric power station. In 2002 there was no power generation at Itezhi-Tezhi, which provides a live storage capacity of 6 thousand million cubic metres with a surface area of 390 km2 . Downstream, the KGU reservoir, when fully impounded, has a surface area of some 800 km2 at 976.6 metres, and 257 km2 at 975.4 metres! The raised water level extends as far west as Nyimba, 115 km upstream. The power station is of 750 MW capacity and has gross and net storage of 900 million and 770 million cubic metres respectively. KGUs energy output is estimated to be 3673 GWh/yr. With respect to demography, in 1990 the Basinspopulation was 2.9 million persons, 40 per cent of the Zambian total. The catchment is the most urbanised in Zambia. By the year 2000, Zambias population was estimated at 10.4 million; if the catchments proportion remained at 40 per cent, this would have given a basin population of 4.2 million in that year. Using the basin population for 1990 and 2000 above, there is a high estimated annual rate of growth of 3.8 percent. The rate of growth has been strongly impacted by immigration into a region offering many employment opportunities. However, the relative importance of natural growth and in-migration is not known.

18.3

GROUNDWATER ABSTRACTION

It is now appropriate to begin to test the hypothesis that the Kafue River Basin was in deficit in the year 2002. Considered in turn are groundwater abstraction, the economic demand for water, the populations food requirements, and environmental needs. Unfortunately there is no comprehensive account of groundwater abstraction for the Kafue River Basin. In 2002, data relating to pumping tests and borehole logging was not being collected. Scott Wilson Pi esold comment: Extensive aquifers are found within the basin particularly in the areas of the Copperbelt and Lusaka. Limited exploitation of these aquifers has taken place. Groundwater has been used on a local scale for urban water supply (Mazabuka and in the Copperbelt) and for rural water supply but surface water sources have formed the predominant source of supply(Scott Wilson Pi esold 2003: 3.18). In rural areas the average yield of deep wells was limited by the capacity of hand pumps to six m3 /day, which appears to be considerably below the minimum specific yield of most aquifers in the basin. The Scott Wilson Pi esold study assumed the yield of shallow wells to be 2 m3 /day. For the Basin as a whole, groundwater potential was estimated at 368 m3 /second. This compares with a total demand for surface water and groundwater in 1995 of 32 m3 /second. We can now return to testing the theory of catchment water deficits set out in section one above. Condition 1 for the truth of the deficit theory is that abstraction from the catchments aquifers was not maintained at a sustainable rate. In fact the evidence shows that groundwater abstraction was sustainable. So condition 1 does not confirm the general hypothesis that the Kafue catchment was in water deficit in 2002.

Catchment water deficits

171

Table 18.1 The Kafue River Basin in 20022003: total surface water abstraction rights by category in cubic metres/second
Sub-basin Upper Kafue Middle Kafue Kafue flats Lower Kafue TOTAL Agriculture 7.10 0.01 13.20 0.90 21.2 Mining and Industrial 2.70 0.00 0.00 0.30 3.0 Domestic and Services 3.20 0.08 0.20 3.80 7.3 Total 13.00 0.09 13.40 5.00 31.5

Source: Scott Wilson Piesold (2003) Integrated Kafue River Basin Environmental Impact Study, Ashford: SWP.

18.4

THE ECONOMIC DEMAND FOR WATER

The economic demand for water by households, agriculture, mining, manufacturing, construction and the services sector is now considered. The economic demand for water refers to the volume of water that users are willing and able to purchase at the suppliers prices or charges. Table 18.1 shows the relative distribution between sectors of surface water abstraction rights. Agriculture is the dominant rights holder. With respect to the industrial sectors, they either purchase water from the municipal water utilities or access their required supply by surface water abstraction. The utilities in 2002 made charges for water but use was not metered. Industry in general did not lack access to water, nor were water charges unaffordable. With respect to the agricultural sector, access to water was secured by farmers pumping their own surface water and groundwater requirements. These own-supply costs were affordable. With respect to the domestic sector, it is clear that in the Kafue River Basin in 2002 there was extensive poverty in the household sector (see section five below). However, there is no evidence that the purchasers of water were not able to secure what they were willing and able to purchase. What is striking about many of the African countries is the variety of forms of supply to users, particularly in urban areas, as well as the high proportion of income that the poor allocate to water purchases (Whittington et al. 1991, Merrett 2007). It is not suggested here that in some defined sense the people of the Kafue in 2002 had their needs for water met in full, in terms of quality or quantity. However, the second condition for deficit relates to economic demand, not need. Survey work in Lusaka, in the Copperbelt towns and amongst rural households would certainly find widespread shortfalls of purchases in relation to the need for quality and quantity in water. But such shortfalls are manifestations of low income, they are not driven by insufficient quantities of surface water and groundwater flows. Condition 2 for the truth of the deficit theory is that outstream water did not fully meet the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors. In fact the evidence shows that the economic demand for water was fully met. So condition 2 does not confirm the general hypothesis that the Kafue catchment was in water deficit in 2002.

172

The Price of Water

18.5

THE SUPPLY OF FOOD

In 1995, agriculture made up 17 per cent of the Kafue River Basin area, primarily on the eastern side of the basin. In the same year there were 85,000 farming households, mostly small and engaged in rainfed crops, crops and poultry, or crops, poultry and livestock. The large-scale operators are typically involved in double cropping, with rainfed cropping in the wet season (often with supplementary irrigation) and full irrigation in the dry season. There were only 7 large or medium-scale irrigation projects in the basin in 1995. The total irrigated area was only 21,000 hectares, about 1 hectare in a thousand for the basin as a whole. There are important sectors for maize and the irrigation of sugarcane (both for domestic consumption and for export). Major irrigation takes place at Mpongwe in the Upper Kafue. There is extensive fishing and livestock grazing (nonirrigated) and this is an important food source. Coffee is also produced. By 2002, there was evidence of an expansion and intensification of agriculture. In 2002 the use of irrigation water was about one litre per second per hectare. This is an annual total volume of some 660 million cubic metres (mcm). The annual river flow at the Zambezi is about 9,450 mcm. So irrigation use is equal to about seven per cent of that downstream flow. The main irrigated crops are banana, coffee, cotton, rice, soybean, sugarcane, tea and wheat. There is no evidence that the economic demand for food is not met, either by production within the Basin or by imports. Moreover, the Kafue River Basin exports copper and electricity on such a scale, in addition to its agricultural exports, that the Basin generates more than sufficient export income to finance its residual food needs from imported food at market prices. However, there is widespread hunger and poverty throughout the catchment. Food insecurity is widespread in the Kafue Basin, particularly among farm households. We can certainly say that sufficient food is not produced for the populations needs. But this is due neither to lack of water nor lack of land. Low farm productivity amongst small farmers and extensive urban poverty are the major issues. The argument here parallels that of section 4. The insufficiencies of food are not sourced by a lack of precipitation and instream water flows, they are rooted in low incomes. Condition 3 for the truth of the deficit theory is that the basin populationseconomic demand for food is not fully met from the domestic rainfed and irrigation sectors, domestic fisheries and food imports financed by the basinsexports. In fact the evidence shows that the economic demand for food was fully met. So condition 3 does not confirm the general hypothesis that the Kafue catchment was in water deficit in 2002.

18.6

ENVIRONMENTAL NEEDS

In 2002 no environmental requirements for the Kafue River Basin had been set by the Zambian water resource institutions. In that sense the rivers instream flows did not fall below defined minima and condition four for a catchment water deficit does not confirm the general hypothesis. In any case, Table 18.2 shows just how small were surface water abstraction rights in 2002-03 compared with the rivers flow and the basins groundwater potential. Were environmental requirements to be introduced, they would almost certainly be defined in terms of water levels of the basins magnificent floodplains, swamps and

Catchment water deficits

173

Table 18.2 The Kafue River Basin in 20022003: estimated total surface water abstraction rights and the available resource in cubic metres/second
Sub-basin Upper Kafue Middle Kafue Kafue flats Lower Kafue Abstraction Rights 13 <1 13 5 Average Annual flow 138 278 285 316 Groundwater Potential 62 200 91 15

Source: Scott Wilson Piesold (2003) Integrated Kafue River Basin Environmental Impact Study, Ashford: SWP. marshes. The Kafue Flats and the Lukanga Swamps are the most significant of these vast ecosystems. However, whereas there is no deficit-driven breach of environmental law, it is appropriate to refer to a different impact of the rivers development. Attention has already been drawn to the Kafue Gorge Upper and the considerable hydroelectric power generated there. To secure maximum power output, a secure flow of water into the Kafue Gorge is required. For this reason the hydroelectric sector has a powerful, long-term interest in maintaining the flow of the Kafue River. A catchment water deficit would be a disaster for the countrys electricity generation and its export of electric power. This secure flow to the KGU is dependent, as has been shown in section 2 above, both on the reservoirs at the Gorge itself and those upstream at Itezhi-Tezhi. Unfortunately the Itezhi-Tezhi dam has reduced the depth, areal extent, duration and frequency of flooding in the whole of Kafue Flats. Moreover, recommendations to preserve the ecological balance of the Kafue Flats by release of a freshet from Itezhi-Tezhi dam of 300 m3 /sec throughout the month of March each year have never been implemented. In part this is due to a smaller firm yield from the reservoir than in design documents Itezhi-Tezhi dam is too low. Note that in this specific instance, the failure to meet an environmental recommendation is not because of deficit in the rivers flow but because of the manipulation of that flow for the requirements of hydroelectric power generation. Condition 4 for the truth of the deficit theory is that the rivers instream flows do not fall below defined minima. Such minima do not exist. If they were to be introduced they would probably apply to the flooding parameters of the Kafue River Basins floodplains, swamps and marshes. In any case, condition 4 does not confirm the general hypothesis that the Kafue catchment was in water deficit in 2002.

18.7

CONCLUSIONS

This paper begins by setting out two hypotheses applicable to any of the worlds catchments and then tests the first, general, hypothesis on the Kafue River Basin. The general hypothesis in this case is that in 2002 the Kafue catchment was in water deficit. The hypothesis is verified if, in 2002, any one or more of the following propositions was true:

174
1. 2. 3.

The Price of Water

Abstraction from the catchments aquifers was not maintained at a sustainable rate. Outstream water did not fully meet the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors. The basin populations economic demand for food was not fully met from the domestic rainfed and irrigation sectors, domestic fisheries and food imports financed by the basins exports. The rivers instream flows fell below defined minima.

4.

In fact, it was shown that none of these four propositions was true. Therefore, in 2002 the Kafue River Basin was in surplus. The application of the general theory to the specific circumstances of the Kafue River Basin has been instructive because it has shown both that the theory is testable (to destruction in this case!) and also that that some of the characteristics of a catchment water deficit exist even when no deficit exists. The author has particularly in mind the unmet need for water amongst poor households, widespread hunger, and changes in the flooding characteristics of the Kafue Flats. However, none of these realities can be ascribed to the catchments precipitation being insufficient to meet the rivers environmental needs, for groundwater to be pumped at a sustainable rate, for the economic demand for water to be satisfied and for the populations food requirements to be met. The case-study has also suggested ways in which to strengthen the causal hypothesis that defines the drivers of a catchment water deficit. It has become clear that in considering the water surplus or water deficit of catchments the focus should be carried out through what can be called the analysis of densities. Four such density measures are offered here as causal factors, drivers of surplus and deficit. The first density measure we should deploy is the catchments population density in terms of persons per square kilometre. The more people and households in the resident population, the greater will be household use and consumption (evapotranspiration) of water. High density here pushes the catchment towards deficit. The second density measure is the catchments production per square kilometre. The greater the level of production, the more water is used and consumed in the production of goods and services. High density here again pushes the catchment towards deficit. The third density measure addresses irrigation head-on. Irrigations importance is that it uses about 70 per cent and consumes about 93 per cent of the worlds outstream supplies. Its density can be measured in various ways of which the best, perhaps, is irrigation water supplied to farmers, expressed as millimetres across the entire basin. Once again a high density pushes the catchment towards deficit. The fourth density measure is precipitation in the basin expressed in millimetres. Total precipitation is proposed here, although a technically better variable would be what is called net precipitation in the original paper on catchment water deficits (Merrett 2005a: 143). Here, high density pushes the catchment towards surplus. The fifth factor that is a driver towards surplus, not strictly a measure of density, is water productivity measured by net output in each economic sector divided by water consumed in the production process. Table 18.3 provides some of the required data for the Kafue River Basin. This information suggests that in 2002 the Kafue River Basin was in surplus because of its low population density, its low irrigation density and its high total precipitation. As

Catchment water deficits Table 18.3 Density measures of the Kafue River Basin in 2002
Population persons/square kilometre 71 Irrigation millimetres/year 4 Total precipitation millimetres/year 1057

175

Source: derived from Scott Wilson Piesold 2003. other studies of catchment water deficits and surpluses are completed, comparison of measurements of the five drivers discussed above will become possible.

Acknowledgements
I acknowledge permission to use the results of the Main Report from the Zambian Office for Promoting Private Power Investment, the Zambian Ministry of Energy and Water Development, and ZESCO. I am also indebted to Alan Bates of Scott Wilson Pi esold for his generous assistance in my work.

Appendix
The Generic Hypothesis: The Dwyer River Basin was in deficit in the year 2004.

River hydrology 1.1 Where is the Dwyer River Basin? 1.2 What is the area of the Dwyer River Basin? 1.3 What is the length of the Dwyer River? 1.4 What is the elevation of the river at source? 1.5 What is the elevation of the river at the rivermouth? 1.6 Was water imported to the River Basin in 2004? 1.7 Was water exported to other river basins in 2004? 1.8 Did the River Basin have any desalination plants in 2004? Precipitation 2.1 Briefly describe the Basins weather. 2.2 What has been the average, long-term, annual total precipitation in the Basin? 2.3 What has been the average, long-term, annual evaporation from the Basin? 2.4 What has been the average, long-term, annual effective precipitation in the Basin? 2.5 What was the annual effective precipitation in the Basin in 2004? 2.6 What was the annual net precipitation in the Basin in 2004? 2.7 What were the effective and the net precipitation per head of population in 2004?

176 2.8 3

The Price of Water Were there specific rainwater harvesting projects in place in the Basin in 2004?

Groundwater 3.1 What is the location of aquifers with respect to the Dwyer River Basin? 3.2 What was the annual rate of abstraction of groundwater within the Basin in 2004? 3.3 What was the sustainable yield of the aquifers below the Dwyer Basin? 3.4 What were the legal rights to abstract groundwater in the Basin in 2004? Environmental requirements and instream activities 4.1 What was the volume of water discharged at the rivermouth in 2004? 4.2 What was this mean annual flow at the rivermouth as a proportion of mean annual rainfall? 4.3 Does the river basin contain natural lakes? 4.4 Were there hydroelectric infrastructures on the river in 2004? 4.5 What was the area of each of the principal reservoirs in 2004? 4.6 What was the rate of evaporation from each reservoir in 2004? 4.7 What was the capacity output of each hydroelectric power plant in 2004? 4.8 Was there an active navigation sector in 2004? 4.9 Was there an active fishing sector in 2004? 4.10 Did the river basin generate wildlife conservation, recreation and environmental tourism services in 2004? 4.11 Did the catchment authority set a quantitative environmental requirement for the flow of the river in 2004? Economic demand for outstream water 5.1 What was the River Basins hydrosocial balance in 2004? 5.2 What were the main forms of hydrosocial supply? 5.3 Was outstream water priced and, if so, what were these prices? 5.4 In the case of the use of water that was not priced in 2004, were there costs incurred by users in accessing water and what were these? 5.5 Could users purchase as much water as they wanted and could afford at 2004 prices? 5.6 Were abstractions from the river used and then re-used in 2004? 5.7 Did tradeable water rights exist in the Basin in 2004? 5.8 Was water distributed unequally between different social classes? 5.9 What was the ratio of output per cubic metre of outstream water used in 2004 for each of the principal economic sectors? Food self-sufficiency 6.1 Briefly describe the agricultural sector of the River Basin. 6.2 Did the Basins farm sector produce sufficient food from the rainfed and irrigated sectors to meet the populations needs in 2004? 6.3 If not, were the exports of the catchment sufficient to purchase residual food needs in 2004? 6.4 Did the institutions of the Basin import food in 2004?

Catchment water deficits 6.5 6.6 6.7 6.8 Were the prices of food imports equal to the international market price in 2004? What proportion of outstream water was used for irrigation purposes in 2004? What proportion of outstream water was consumed for irrigation purposes in 2004? Were there programmes in place in 2004 to raise the ratio of value of output per unit of water used or consumed in the irrigation sector?

177

Population 7.1 What was the size of the Basin population in 2004? 7.2 What was the rate of change of the population within the catchment in 2004 and what was the relative importance of natural growth and migration? 7.3 Was there a programme in 2004 to encourage family planning? Production 8.1 What was the value of Basin output in 2004? 8.2 What was the value of output per head in 2004? 8.3 What was the rate of growth in 2004 of annual output per head of the population? Water resource institutions 9.1 How were the River Basins water resources managed in 2004?

19 The Thames catchment: a River Basin at the tipping point

There is nothing so practical as a good theory

Bertrand Russell

19.1

INTRODUCTION

In the preceding papers an account is given of what is termed a catchment water deficit. The concept is advanced in order to provide a methodological framework for research into catchments and river basins (the terms are used here interchangeably) in which the volume of river and groundwater flows are insufficient for the basin populations requirements. A catchment water deficit is a situation in which, during the course of a specified year, total precipitation in a basin is insufficient to satisfy simultaneously the following four criteria: 1. 2. Groundwater abstraction is maintained at a sustainable rate. Outstream water fully meets the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors. There is no assumption here that the set of charges and prices for water are fixed. Indeed, it may well be that in future river basin authorities will vary their charges and prices over time to deal, for example, with droughts. The capacity to adapt should be a central goal of river basin management. The basin populations economic demand for food is fully met from the domestic rainfed, irrigation and fishing sectors and/or from food imports financed by the basins commodity and service exports. The balance between food imports and food grown (and then consumed) in the basin is sure to vary from year to year. Again, the capacity to adapt is vital in water resource management. The rivers instream flows do not fall below defined minima.

3.

4.

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

178

The Thames catchment: a River Basin at the tipping point

179

Criteria 1 and 4 are hydrogeological, hydrological and environmental. Criteria 2 and 3 are social, economic, commercial and agricultural. This comprehensiveness in the specification of deficits is unusual and may be considered to give the approach strength and relevance in the 21st century. These four criteria are, perhaps, deceptively simple. With criteria 1 and 4, who is it that defines the sustainable flows? It may be that these are set by only a single institution, such as a catchment environment agency. However, when two or more organizations define the sustainable flows, the researcher may have to admit that accepting organization 1sdefinition, the catchment is in surplus, whereas it is in deficit if one accepts organization 2s definition. This author believes that, in fact, there is no acceptable, generic definition applicable to every river basin on the globe. The original paper referred to above suggested that global population growth and increases in output per head would drive ever more catchments throughout the world from surplus to deficit. By way of illustration, already in 2005 the Yellow River in Asia, the Nile in Africa, the Guadalquivir in Europe and the R o Bravo/Rio Grande in the Americas are all catchments in water deficit.

19.2

THE KAFUE CATCHMENT

General theory has its place but it needs to be tested in the crucible of empirical falsification or verification. The original paper required a specific river basin on which to test the general hypothesis that Catchment X is in water deficit.As good luck would have it, a British engineering company, Scott Wilson, had provided the author with extensive material on the Kafue River Basin and the decision was taken to use their study to match general theory with ground-level truths (Scott Wilson Pi esold 2003). The results are set out in paper 18 of this volume.

19.3

THE ANALYSIS OF DENSITIES

The significant advance achieved through the Kafue study was to strengthen understanding of the drivers to deficit and to surplus. These were weakly specified in the original paper. The author now believes that these drivers causal variables are best understood as a set of five densities, calculable for any and every river basin in the world once the necessary fieldwork is carried out. The densities are set out in Table 19.1. Density 1, population density is equal to persons resident in the catchment divided by its area in square kilometres. It is a driver for deficit and creates this potential for deficit in two quite distinct ways. First, a bigger population is associated with a larger volume of domestic water used and thereby (partially) lost to evapotranspiration. Secondly, bigger populations require more food. These food needs may lead to greater requirements for irrigation output in the basin (see Density 3). Simultaneously, high population density may lead to a breach of criterion 3 for surplus as described in section one above. The catchment population may suffer from food shortages or need to turn to the international trade in agricultural produce in search of subsidized food. Density 2, production density, is the value of output in the basin divided by its area. It is also a driver for deficit. Greater output is accompanied by greater supply and use

180

The Price of Water

Table 19.1 Density drivers for catchment water deficits and catchment water surpluses
Density drivers for deficit 1. Persons per square kilometre (a) 2. Value of output per square kilometre (b) 3. Value of irrigation output per square kilometre
a.

Density drivers for surplus 4. Effective precipitation in millimetres (c) 5. Water productivity in $/cubic metre (d)

Notes Number of persons resident in the catchment divided by catchment area. b. Value can be measured in various ways. Catchment gross value added is one such measure. Output from the basins irrigation sector should be excluded from driver 2 as it is covered by driver 3. c. Total precipitation minus catchment evapotranspiration, but excluding evapotranspiration during the course of household use and output production. It is an attempt to estimate a natural process driven by total precipitation, radiant energy, wind speed and humidity, rather than a socially-determined variable. d. The ratio of total value of output to all evapotranspiration losses in the production of that output.

of water in production, with consequential losses in evapotranspiration. Output here should exclude that from the irrigation sector as this sector appears as density 3. Density 3, irrigation density, is the value of irrigation output in the basin divided by the catchmentsarea. This is a special case of density 2 and made separate because of the intense evapotranspiration of irrigated agriculture. If we exclude evapotranspiration from reservoirs, Shiklomanovs data suggest that at the global scale 93 per cent of hydrosocial evapotranspiration can be ascribed to irrigation and only 7 per cent to the urban sectors domestic and production uses (Shiklomanov 2000: Table 5). Density 4, effective precipitation, is a driver for surplus. The greater its value, the more likely is it that the river basin will be in surplus. Effective precipitation is equal to total precipitation less the catchments evapotranspiration, other than the evapotranspiration that occurs in household use and the production of output (see Table 19.1, footnote c). Density 5, water productivity, is also a driver for surplus. It is defined as the catchments output value of goods and services divided by the evapotranspiration that takes place in producing that value. The catchment water deficit approach has a central concern with environmental flows; therefore it is primarily oriented neither to the supply of water nor to watersuse, but focuses on the losses to the river basinsflows of water as a result of evapotranspiration. The appropriate measure of these losses in production should be the evapotranspiration that takes place between the point at which water is first abstracted from river and aquifer through to the recycling point of waste water and irrigation returns back to aquifer and river. Frederiksens and Perrys critiques of water crisis solutions and irrigation efficiency studies have contributed powerfully to this orientation to evapotranspiration rather than water supply and use (Frederiksen 1996, Perry 1996). Catchment water deficits are powered by evapotranspiration.

The Thames catchment: a River Basin at the tipping point

181

19.4

THE THAMES RIVER BASIN

The original paper on catchment water deficits placed great emphasis on output growth per person and population increase driving river basins from surplus to deficit. Following completion of the work on the Kafue (see section 2 above), it made sense therefore to test deficit theory in a catchment known for its high densities of population and production. The Thames River Basin meets these criteria, yet exhibits no obvious symptoms of deficit. Karl Popper took the view that it is the falsification of theory that best drives scientific advance, so work on the Thames seemed to offer real benefits in testing the general theory (Popper 1959, Dow 2002: 8590).

Hydrology of the Thames

The Thames River Basin is located in southern England (see Figure 19.1). The source lies in Gloucestershire, below the Cotswold Hills close to Cirencester, and the riverhead is 107 metres above sea level. The river finally flows to the North Sea to the east of London. The Thamess length is 333 km from Thames Head Bridge to Shoeburyness in Essex. The Tidal Thames stretches from Teddington in the west of London to the North Sea. The mouth of the river is therefore at sea level. The river has many tributaries including the Windrush, the Kennet, the Cherwell, the Thame, the Wey, the Colne, the Mole and the Lee. There are 5,330 km of main river in the Basin and the catchments area is 13,000 km2 (www.environment-agency.gov.uk/regions/ thames/). Long-term, annual total precipitation in the Basin is 690 mm. Long-term, annual effective precipitation is 280 mm. The latter figure is based on the 1961-90 average for the Berkshire Downs, derived from the Environment Agencys use of CatchMod, their model of catchment hydrology. So the best estimate for long-term evapotranspiration is 410 mm. With respect to groundwater, the basin has extensive, water-rich, chalk and limestone aquifers. Surface water and groundwater cannot be abstracted without a permit from the Environment Agency. The right to abstract surface water is also conditional on the existence of minimal surface water flows at specified locations on the river and its tributaries.

Environment
The Basin is the location of a wide range of environmental services and these generate substantial income for the public institutions and private companies that provide them. In particular the Environment AgencysThames Region welcomes each year millions of visitors. These include anglers, boating enthusiasts, canoeists, sailors, holidaymakers in hired craft, bird watchers and other naturalists, oarsmen and oarswomen, swimmers, sub-aqua divers, walkers and cyclists. The non-tidal Thames passes through 44 picturesque locks. There is a Thames Path National Trail. It meanders for 340 km from the rivers source, through both peaceful countryside as well as the Thames Valleys villages and towns, and then on to the City of London, before ending at the Thames Barrier at Greenwich. There are also 462 Sites of Special Scientific Interest within the Thames River Basin.

Figure 19.1 Map of the Thames River Basin.

The Thames catchment: a River Basin at the tipping point Table 19.2 The hydrosocial balance of the Thames River Basin in 1984

183

VOLUME (million cubic metres) CATEGORIES OF SUPPLY (POSITIVE) Rainwater collection Groundwater abstraction Surface water abstraction Desalination of sea water Imports of water from other areas Internal reuse of wastewater External reuse of wastewater Net fall in water abstracted and stored TOTAL POSITIVE SUPPLY CATEGORIES OF SUPPLY (NEGATIVE) Supply-side evapotranspiration losses Supply-side leakage Exports of water to other regions Net rise in water abstracted and stored TOTAL NEGATIVE SUPPLY TOTAL NET SUPPLY CATEGORIES OF USE (POSITIVE) Households Agriculture, including irrigation requirements Mining Manufacturing Construction Public services Private services Other uses plus not known categories TOTAL USE Source. Merrett (1997) Table 2.1. 0 570 1260 0 50 not known not known negligible 1880 not known 310 30 negligible 340 1540 760 150 not known 190 not known not known not known 440 1540

Population and hydrosocial balance

The catchment population in 2004 was 12.5 million. The Basins principal towns are Swindon, Oxford, Reading, Maidenhead and Greater London. The population of Greater London alone is 8.1 million persons. No recent estimate has been made of the river basins hydrosocial balance but one exists for the year 1994 and appears here as Table 19.2 (Merrett 1997: 17). (The hydrosocial balance for a specific year and area tabulates the volume of outstream water supplied from different sources and the volume of outstream water used by different user-groups.) In 1994 the main characteristics of the catchment in respect of the supply and use of outstream water were:

r r r

Total volume of abstraction was about 1,880 million cubic metres. Water imports from and exports to other river basins were each about three per cent of total net supply. Surface water abstraction was the main source of supply.

184

The Price of Water

Table 19.3 Greater London Authority Gross value added by industry group in 2002 in percentages
INDUSTRY GROUP Real estate, renting and business activities Financial intermediation Wholesale and retail trade Transport, storage and communication Manufacturing Other services Health and social work Education Construction Hotels and restaurants Public administration and defence Electricity, gas and water supply Mining and quarrying Agriculture, hunting, forestry and fishing Total Source: Office for National Statistics December 2004 GROSS VALUE ADDED (%) 32 13 10 9 8 7 5 5 4 3 3 1 <1 <1 100

r r

There was substantial leakage on the supply side. The domestic sector was by far the biggest user of water.

There are a number of users of water that access their supply themselves, by abstracting from the river flow and from local aquifers. All abstractions above a minimal limit are charged for by the Agency and these charges form a part of its annual income. The major abstractors are the water utilities, which face similar controls and charges as do non-utility abstractors. In 2005 the Basins first desalination plant has been proposed for the London Borough of Newham on the Thames Estuary. At the time of writing Londons Mayor has vetoed the project on environmental grounds. If completed it would have a capacity of 150,000 m3 /day, that is, 55 million cubic metres per year.

Production
No estimates have existed heretofore for the Basins economic output. However, the principal location of economic production in the Basin is the area of Greater London and data for the Greater London Authority does exist. The Office for National Statistics (2004) reports that Londons gross value added (GVA) for 2003 was 155 billion, with a GVA/head/year of 21,000. GVA is identical with Gross Domestic Product at basic prices. The percentage breakdown of the total by industry group is given in Table 19.3, which shows that services of various types dominate with 87 per cent of the total, whereas manufacturing, construction, mining and agriculture combined make up only 13 per cent of the total. Estimates of regional GVA are on a residence basis, where the income of commuters is allocated to where they live rather than their place of work. Since commuting into London exceeds commuting out of London, the GVA statistics probably understate Greater Londons production. If we apply the 21,000 per capita data to the entire

The Thames catchment: a River Basin at the tipping point

185

12.5 million population of the Thames River Basin we have total output of about 260 billion per year at 2002 prices.

Water pricing
All outstream water supplied to the Thames River Basinscustomers is charged for. The utilities supplying the catchment are Thames Water, Three Valleys Water, South East Water, Sutton and East Surrey Water, and Essex and Suffolk Water. The charges set by Thames Water and these other private utilities are controlled by the Office of Water Services. For both households and business, provided that their supply is metered, there is both a fixed charge and a volumetric charge. In the domestic sector, if the supply is not metered then, in addition to the fixed charge, there is a charge for the supply of water based on the value of the property. In addition, there are wastewater charges. Thames Water suggests that the hot weather associated with climate change is increasing the economic demand for water and its use. Domestic use has risen from 150 litres per head per day (lhd) in the 1980s to 163 lhd in 2005. Single occupancy households are increasing and this also raises use per person. (www.thames-water.com/). In most cases in the domestic sector, water is charged for on the basis of the value of the familys house. Therefore the household water bill does not vary with use; there is no price of water. So, in these cases, no price-based, demand-side management of the use of domestic water exists.

19.5

THE THAMES IN WATER DEFICIT?

It is now appropriate to assess whether the Thames catchment is in surplus or in deficit. To begin with, this can be done by turning the four criteria of section one into four questions. First, is groundwater abstraction maintained at a sustainable rate? The Environment Agency, as we have seen, has considerable regulatory powers to set limits on groundwater abstraction so that the rate of withdrawal does not exceed the sustainable yield within the basin. However, some parts of the aquifer have reached their sustainability limit; and there are some abstractions which are thought to be unsustainable and these are being investigated (Environment Agency, pers.comm., 17th August 2005). There is a more complex issue here. Hydrogeologists agree that the estimation of an aquifers sustainable yield rarely produces a single, unambiguous statistic. As Roger Calow of the British Geological Survey says, there are widely-placed error bars on most sustainable yield estimates (personal communication 22.10.05). In the case of the Thames, different sustainable yield constraints and, correspondingly, different rates of groundwater abstraction, lead to different environmental habitats. Which of these mutually-exclusive choices should be the base-line is therefore a matter of debate, particularly in relation to recorded historical changes in habitat over long periods of time. To sum up, there are cases of unsustainable groundwater abstraction in the Thames River Basin and there are cases where the sustainable yield is thought by environmentalists to be too high in historical retrospect. Secondly, does outstream water fully meet the economic demand for water from households, agriculture, mining, manufacturing, construction and the services sectors?

186

The Price of Water

The answer is uncertain. At the time of writing, in 2006, It has been more than ten years since restrictions were imposed on use, by limitations on spray irrigation or by limiting the use of hosepipes and garden sprinklers. However, the author is informed by the Environment Agencys Head of Water Resources that You will be aware, I hope, from the Environment Agencys published reports, that Thames Water currently has a deficit of some 200 Ml/d; we are very concerned about security of supply(personal communication 12.9.05). 200 Ml/d (megalitres per day) is equal to 73 million cubic metres per year. Thirdly, is the basin populations economic demand for food fully met from the domestic rainfed, irrigation and fishing sectors and/or from food imports financed by the basins commodity and service exports? Food output in the catchment is small in comparison with food purchases by the basins population of 12.5 million. However, the regions exports of services are vast, both to the rest of the U.K. and to the world. In fact, London and New York may well be the worlds two largest urban producers of services for export. Finance, banking, tourism, the arts, research and education flourish as exporting sectors. The basins exports are more than sufficient to meet its net food requirements. Therefore the population of the Thames River Basin has no difficulty whatsoever in accessing at the market price all the food it wants and can afford to purchase. Fourthly, do the rivers instream flows meet defined minima? The author has not received any evidence from the Environment Agency that river flows regularly fall below these minima as a result of excessive surface water abstraction. However, in October 2004 an Environment Agency officer suggested that the abstraction levels in the Thames River Basin were ten per cent higher than ideal from an environmental perspective (London Assembly Environment Committee 2004: 5). But it is also necessary to consider the export and import of water in the Basin. In 2004 imports were negligible whereas exports, primarily to the Anglian Region, were about 40 million cubic metres. (Environment Agency, personal communication 17.10.2005). To sum up, in 2005 the Thames River Basin is a catchment in water deficit, but only at the margin.

19.6

DENSITY ANALYSIS OF THE THAMES RIVER BASIN

The Thames catchment is modest in size, has a large population, and has vast total output. Despite this, the basin in 2004 and 2005 appears to be only modestly in deficit. In comparative terms, the river is certainly in a far, far healthier state than, for example, the Murray-Darling, the Yellow River, the Nile, the Guadalquivir and the Rio Grande. This conundrum should be possible to explain using density analysis. Table 19.4 provides the estimates for the Thames of the five density drivers set out and defined in Table 19.1. Population density is about 960 persons per square kilometre. In comparison with the rest of the worlds river basins this is extremely high. Figure 19.1 shows how large is Greater London in comparison with the scale of the river basin. Table 19.5 shows that, amongst countries with a population of five millions or more, England is the fifth most-densely populated in the world. The population density of the Thames River Basin is 3.4 times greater than that of the rest of England. Put another way, if the

The Thames catchment: a River Basin at the tipping point Table 19.4 Density drivers for the Thames River Basin
Density Drivers for deficit 1. Persons per square kilometre (persons/km2) 2. Value of output per square kilometre (/km2) 3. Value of irrigation output per square kilometre (/km2)
a

187

Driver value 960

Density drivers for surplus 4. Effective precipitation (millimetres) 5. Water productivity (/cubic metre)(a)

Driver value 280

20,000,000

1,730

negligible

Derivation of Driver 5. Output is set at 260,000,000,000. Total use is set at 740 mcm, using the the 1994 data of Table 2 and excluding household use. Evapotranspiration losses to the catchment in the course of production are assumed to be 20 per cent of use, giving the value for production ET of about 150 mcm. Dr. Stephen Merrett, London

Table 19.5 Top ten world population densities in the year 2005: countries with a population of at least 5 million
States and territories Bangladesh Taiwan Korea (South) Netherlands England Japan El Salvador Belgium Rwanda India Population (million persons) 144 23 48 16 46 127 7 10 8 1009 Area (thousand square kilometres) 144 34 98 42 130 378 21 31 26 3288 Population density (persons/km2) 1000 676 490 381 354 336 333 323 308 307

Source: http://www.factmonster.com/ipka/A0004379.html Thames River Basin were a country it would be the second most densely populated in the world. Production density is 20 million per square kilometre, extraordinarily high. When we have a broad range of global data of this type, it may prove that the Thames has the highest production density of any catchment in the world. The first two drivers to deficit are powerful but the third, irrigation density, is extremely weak. The scale of irrigation in the catchment is small. Unfortunately there is no published output value, but we do have data on the scale of irrigation water use. The Environment Agency Thames Region (EATR) states that in 2003 the licensed volume for agricultural spray irrigation was 11,321,000 m3 (EATR personal communication 21.7.05). (The licensed quantities for general agriculture, fisheries and watercress were

188

The Price of Water

8,730,000 m3 .) Even if all of this irrigation use is lost to evapotranspiration, the flow is still slightly less than 1 mm when expressed across the river basin. Turning to the drivers for surplus, we see that effective precipitation is rather high in global comparative terms. The catchment has a humid climate with moderate average temperature and wind speeds. Finally, an attempt is made to estimate water productivity at the catchment scale. This variable, in terms of the definition given in section three, has probably never been estimated before anywhere. The data are fairly crude, particularly the volume of water used in production and the suggested 20 per cent evapotranspiration losses in use. But the value of this density driver is so large that even a substantial adjustment, following the necessary fieldwork, would still leave a value almost incredible in its size. So it is useful to restate what that value means: The ratio of the gross value added in the production of goods and services in the Thames River Basin to the number of cubic metres lost to the catchment in the evapotranspiration of water used for production probably exceeds 1700/m3 . In 2004-2005, the volumetric charge for water supplied to non-domestic customers by RWE Thames Water was 0.65 per cubic metre, or 3.25 for each five cubic metres. Therefore the ratio of gross value added to the cost of water supplied was 523 to 1! This allows for the fact that every five cubic metres supplied represents only one cubic metre consumed. Put another way, non-domestic users cost of water was some 0.2 per cent of gross value-added. There is another way of expressing water productivity at the basin scale, although it does not permit inter-basin comparisons. Across the Thames River Basin, for every millimetre of water lost to evapotranspiration during the course of output production, the value of that production is 20,000,000,000!

19.7

TIPPING DEEPER INTO DEFICIT

It is argued above that the Thames River Basin is in deficit but that the situation is not yet desperate, despite the regions high population density and production density. This is because the catchment is fairly humid and because production recycles back to the river basin about 80 per cent of the water used in the production process, with a consequentially high value of water productivity. The high rate of recycling exists in part because of the modest scale of irrigation. The minimal scale of plant production (either rainfed or irrigated) is possible because the basin population of 12.5 million exports sufficient goods and services particularly the latter so that it faces no financial or economic difficulty in importing food produced in other river basins. Those imports themselves require rainfall and/or irrigation for their growth, of course, but the evaporative losses take place in other catchments, not in the Thames River Basin. However, the author believes that we are now in the early stages of a tipping process during which the severity of the catchments water deficit will accelerate. This shift in severity is sourced by three dynamic forces: climate change, population growth and production growth. Climate change will bring with it hotter, drier summers and warmer wetter winters. Summer heat and low humidity will reduce effective precipitation, impacting severely on river flows and groundwater recharge (Collingwood 2005). Population growth in the South-East, a region that overlaps with the Thames River

The Thames catchment: a River Basin at the tipping point

189

Basin, will drive up total household use. The South-East England Regional Assembly announced in 2005 plans to build 578,000 new homes in the region (Guthrie 2005). Pressure for growth at the eastern limit of the Thames River Basin will also arise from the siting of the 2012 Olympic Games on the River Lee, the Thamess most easterly tributary. Production growth will drive and be driven by population growth. In particular this will occur in the site area of the Olympic Games as well as in Thurrock, Essex, where the biggest container port in Britain is to be constructed. Thurrock is on the Thames and, technically, a dozen miles east of the basins limit (Blitz 2005). In the light of these dynamic forces within to the Basin, redemptive options need to be examined. There are just six such options, widely varying in their political feasibility and desirability (see Paper 17): 1. 2. 3. 4. The deceleration and reversal of the rate of growth of population. The deceleration and reversal of the rate of growth of production. Further increases in waters productivity. The introduction of universal metering, to reverse the volume of household use per person. 5. The reduction of evaporative losses both from reservoirs and from water supply leakage; groundwater re-charge has advantages here. 6. An enhanced capacity of the Environment Agency in its forward planning and its policy instruments, including the routine use of the hydrosocial balance in forecasting. Certain palliative options also exist. One is to reduce the environmental standards relating to groundwater and surface water abstraction. These will, quite properly, be opposed by the environmental movement. A second is to increase water re-use; but this can only be beneficial if it reduces evapotranspiration losses and this may not occur. This is a policy area deserving new research. A third is rainwater harvesting, but this is just surface water or groundwater abstraction under another name. A fourth is importing water from other catchments. But this may drive other catchments into deficit and it is also costly in construction terms as well as demanding in terms of the electricity required for pumping. Electricity generation is not carbon-neutral. A fifth is the desalination of sea-water, as in the case of the proposed plant in the London Borough of Newham. Again this requires expense and substantial electricity inputs.

19.8

CONCLUSION

It is the opinion of the author that the two roads to redemption from the threat of much greater Thames River Basin deficits in the future are to severely limit the housebuilding programme in the catchment and to introduce the universal metering of household water use. As Jay OKeeffe suggests, water resource management should be like the construction of a cathedral; one approaches it within the perspective of the very long run.

Acknowledgements
The author is grateful to the following institutions and persons for their valuable assistance in the writing of this paper: W.S. Atkins, Dr. Sally Watson, Ian Barker (Head of

190

The Price of Water

Water Resources, the Environment Agency), Michael Coffey (formerly RWE Thames Water plc Director of Strategy), Alison McCartney (Department for Environment, Food and Rural Affairs), Melissa Bance, Aisha Burtally, Samantha Dinnage, Clare Dinnis, Mark Funnell, Doug Hill, Crystal Hinks, Kate Kerton, Karen Lingard, Alys Logan, Jenny Sampson and Alastair Wilson all from the Environment Agency, Rob Coward and Adarsh Varma of the Greater London Authority, Peter Johnson and Michael Snell of ICID.UK, Robert Palmer of the Office for National Statistics and from RWE Thames Water plc: John Haakerson, Maria Ioannous and Steve Tuck.

20 Water resource impacts of new housebuilding in the Thames Region 20062025

20.1

INTRODUCTION

If the Thames Region were a country it would be the second most densely populated in the world, amongst countries with a population of 5 millions and more. The Region is also probably the most productive river basin in the world in terms of output per head. Families as well as businesses require water supplies in large quantities. So it should come as no surprise that the Thames Regions high density of population and production together place the river basins water resources under stress. It is the Environment Agencys staff who are best placed to describe this stress on the water resource. For example, in October 2004 an Environment Agency officer suggested that the abstraction levels in the Thames River Basin were ten per cent higher than ideal from an environmental perspective (London Assembly Environment Committee 2005, p.5). In 2005 the Agency commented on excessive withdrawals of groundwater by saying there are some abstractions which are thought to be unsustainable and these are being investigated (see Paper 19). Again in 2005, the Environment Agencys Head of Water Resources wrote You will be aware, I hope, from the Environment Agencyspublished reports, that Thames Water currently has a deficit of some 200 million litres per day; we are very concerned about security of supply (personal communication to the author).

C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

191

192

The Price of Water

Table 20.1 Useful information on the Thames region

Area of the Thames River Basin (sq. kilometres) Length of the River Thames (kilometres) Length of River Thames and its tributaries (kilometres) Average annual rainfall (millimetres) Average annual evapotranspiration (millimetres) Average annual effective rainfall (millimetres) Average annual effective rainfall (thousand million litres) Population in 2001 (persons) Population density in 2001 (persons per square kilometre) Average annual effective rainfall per person (litres) Population (households) Number of dwellings (houses, flats etc.) Average number of persons per household Average number of persons per dwelling Domestic use of water in 2001 (litres per head per day) Total domestic use of water in 2001 (million litres per day) Percentage of households with a metered supply in 2001 Production ( billion gross value added) Production density ( per square kilometre) 12,867 333 5,330 690 455 235 3,022 11,790,000 916 256,000 not known 5,060,000 not known 2.33 159 1,875 16 260 20,000,000

Main sources: Environment Agency (2001) Water Resources for the Future: a Strategy for Thames Region, Reading: Environment Agency; and Paper 17. The stress on the water resource will increase in the future because of climate change. The Thames Regions natural supply of water is its rainfall and this averages 690 millimetres per year (about 27 inches). But 455 millimetres is lost annually in evapotranspiration. So net rainfall averages only about 235 mm/y. Climate change will bring with it hotter, drier summers and warmer, wetter winters. The overall, average impact is forecast to reduce net rainfall (Collingwood Environmental Planning and Land Use Consultants 2005). So climate change clearly threatens the vitality of the River Thames and its tributaries such as the Kennet, Thame, Loddon, Wey, Colne, Mole and Lee. But there is a second powerful threat and that is the plans for new housebuilding in the region. House construction itself uses water in large quantities but the permanent change it brings to the river basin is an increase in population and therefore in the economic demand for water. The objective of this paper is to examine the likely water resource impacts of new housebuilding in the Thames region over the twenty years 2006 to 2025.

20.2

THE BASELINE SITUATION

To anticipate the future, we need a good understanding of the present. The key information is contained in Table 20.1. The area of the Thames Region is almost 13,000 square kilometres and at the time of the 2001 Census the population was a little less than 12 million persons. So population density was more than 900 persons per square kilometre, approaching that of Bangladesh (see Table 20.2). This high density results in extremely low net annual rainfall per head: 0.26 millimetres per person per square kilometre. This is the slender flow that has to meet the combined requirements of:

Water resource impacts of new housebuilding

193

Table 20.2 Top ten world population densities in the year 2005: countries with a population of at least 5 million
States and territories Population (million persons) Area (thousand square kilometres) Population density (persons/km2)

Bangladesh 144 144 1000 Taiwan 23 36 639 South Korea 48 98 490 England 50 130 385 Netherlands 16 42 381 Japan 127 378 336 El Salvador 7 21 333 India 1080 3288 328 Belgium 10 31 323 Rwanda 8 26 308 Sources: http://www.factmonster.com/ipka/A0004379.html; and National Statistics Online. Thames Region 12 13 916 U.K. excluding 48 232 207 Thames region Sources: Office of National Statistics, Environment Agency, and Dr Stephen Merrett.

1. 2. 3.

The regions households. The river basins ecosystems. The regions annual economic production, valued at 260 billion.

The focus of this paper is on domestic water use and how the economic demand for water will swell with new housebuilding in the Thames river basin. For the baseline years, the Environment Agency Thames Region has provided a comprehensive overview of families use of water. The public water supply in 2006 is sourced by five companies. They are: Essex and Suffolk Water, South East Water, Sutton and East Surrey Water, Thames Water and Three Valleys Water. Seventy per cent of their total supply (net of leakage) was taken up by households in 1999. Table 20.1 shows that in 2001 domestic use equalled 1,875 million litres per day. In terms of litres used per head per day (lhd) in the household sector, there is great variation. The lhd bandwidth lies almost entirely within the range 130 to 196. In 2006 the overall average is about 160 lhd. There are many determinants of this wide variation of use around the average, such as whether the family has a garden or not, the number of persons in the family (smaller households use more litres per head), the number of white goods owned, their water efficiency as well as the efficiency of other water-using fixtures, and the environmental awareness of the family. Table 20.3 shows the micro-components of household use in 2001 (Environment Agency 2001). In that year, the heavyweight players capturing 80 per cent of the total were personal washing, use of a WC and clothes washing. An important question in domestic use is whether or not the house or flat has a metered water supply. If the answer is yes, then the user knows that the bill for water will increase with the amount used. If the answer is no, the bill does not rise with

194

The Price of Water

Table 20.3 Micro-components of householdwater use in the thames region in 2001

Microcomponent Personal washing WC Clothes washing Dishwashing Garden watering Miscellaneous Other (internal) Other (external) Total Percentage share of use 39 26 15 7 6 3 3 1 100

Source: Environment Agency (2001) Water Resources for the Future: a Strategy for Thames Region, Reading: Environment Agency. additional use. So metered properties record lower use in terms of litres per head per day. In 2001, only 16 per cent of households had a metered supply.

20.3

THE INCREASE IN HOMES 20062025

The Thames Region is one of a number of regions defined by the Environment Agency for the purposes of environmental policy and its implementation. However, there also exist a number of planning regions, with quite different boundaries from those of the Agency, and these are the basis of Englandsregional government. Each such region has its own elected regional assembly. The Environment Agency Thames Region overlaps with six of the regional assembly areas: South East England, the East of England, the East Midlands, the West Midlands, South West England and Greater London. Every regional assembly prepares a strategic plan that, amongst much else, outlines the annual additions to the housing stock that the assembly wishes to see delivered. Table 20.4 is my attempt to estimate the total number of houses and flats that will be added (net of demolitions) to the Thames Regions stock of dwellings in the 20-year period 20062025, if the plans are fully implemented. The total is about 950,000 dwellings, of which sixty per cent fall in Greater London, twenty-four per cent in South East England and thirteen per cent in the East of England. Even South West England sees an additional 25,000 houses and flats lying within the boundary of the Thames region. Houses, post-construction, do not use water. People do. The likely maximum additional population by 2025 is likely to be about 2.2 million persons, using the 2001 ratio of 2.33 persons per dwelling. And if each of these persons uses about 160 litres of water per day, we have an additional use of 352 million litres per day.

20.4

ADDITION AND SUBTRACTION

These estimates of 950 thousand homes and 352 million litres per day of water should not be regarded as scientifically precise. They are only orders of magnitude based on

Water resource impacts of new housebuilding Table 20.4 Estimated total new housebuilding in the Thames River Basin by Region 20062025 (1)
Region East England (2) East Midlands (3) Greater London (4) South East England (5) South West England (6) West Midlands (7) Total

195

Estimated total number of houses to be constructed 20062025 125,600 600 570,000 231,300 25,000 400 952,900

Notes: 1. The general procedure in estimating the total number of dwellings that will be built in the Thames River Basin during the 20-year period 20062025 is to begin with each Regional Assemblys target total for all its Districts. The target is then divided between Districts within the region on the basis of the 2001 population numbers the more persons in a district in 2001, the greater the housing estimate. Districts entirely outside the Basin are then set aside. Districts entirely within or overlapping the Basin are assigned a housing total equal to its District total multiplied by the percentage overlap of that Districts area with the Thames River Basin. 2. The East of England Draft Plan for the Region has an annual average of 23,900 net increase in dwellings in Policy H1. District-wise data not available. 3. The East Midlands Regional Housing Strategy cites a total figure for 20012021 of 97,000 dwellings on p.29. I used the same annual output for 20062025. Only Daventry and South Northants overlap the Thames River Basin. 4. In 2006 the intention is to review the London Plan at an unspecified future date. The current plan states that some 30,000 homes per year need to be provided to meet demand. 28,500 is used here in the Table, as parts of the London Boroughs of Havering, Bexley and Bromley fall outside the Thames River Basin. The Greater London total is by far the largest as its annual rate of building is relatively high AND it falls almost entirely within the Basin. 5. The South East England Regional Assemblys South East Plan sets out a vision for the region through 2026. The Draft Plan refers to an annual average provision of 28,900 dwellings. 6. The South West Regional Housing Strategy 20052016 refers to overall net housing commitments in RGP10 of an additional 20,200 units/year. 7. Regional Planning Guidance for the West Midlands in Table 1 has a median value of 15,665 as the annual rate of housing provision for 2007 to 2021. Only Stratford-on Avon District and Wychavon District overlap with the Thames River Basin. the assumptions that have been set out above in this paper. Moreover, the two key statistics (the number of new houses and the additional water used domestically) do not incorporate certain changes over time that will tend either to add to stress or to subtract from it. These changes must now be briefly considered.

Additions to stress
i) Climate change in the Thames Region will bring with it warmer, wetter winters and dryer, hotter summers. The combined effect will be to cause a decline in net rainfall

196

The Price of Water

and it is the low volume of net rainfall per person that is the key driver of the Regions water stress. Climate change will exacerbate stress. ii) Dryer, hotter summers will increase the use of baths, showers and garden sprinklers. The last of these also suffers from high rates of evapotranspiration, so that the water flow is lost to the river basin (see Paper 17). Droughts will be more frequent.

Subtractions from stress

iii) We can be sure that the number of households whose water use is metered will increase in the future. This will moderate use. Of course, all new houses and flats will be metered because the regulation of new housebuilding standards requires this. But there will be more houses and flats in the existing stock that switch to a metered supply. The impact on use per person, compared with the unmetered situation, has many determinants, including the tariff itself and, most important, the average price of water. Metering and tariffs help little when prices are low. iv) Continuing design improvements in white goods and other water-using fixtures will reduce water use. v) Additional water requirements, whatever their exact volume, may be met by interbasin transfers or desalination plants. This will serve to reduce the impact of population growth on the River Thames ecosystems, but at a financial and environmental cost in the infrastructures required, as well as the carbon emissions required for the electricity powering that infrastructure. vi) Leakage reduction will make more productive use of the water abstracted from the river or its aquifers, particularly in the Greater London area where the leakage runs uselessly into the Thames Estuary (see Paper 21).

Household size
vii) New housebuilding will add to the Thames Regions population and exacerbate water stress, as has already been said. But the anticipated fall in average household size over time will weaken the net increase in the number of persons resident in the Region. Yet again, smaller households use more water per person. The overall impact of the increase in the number of persons and higher use per person requires estimation. A final complicating factor is that Thames Waters estimates show 2.46 persons per dwelling, not the 2.33 used in the calculations up to this point. This higher figure implies that future use will be 5 per cent higher than I have calculated above.

20.5

CHOICES

I believe that the resolution of these concrete threats to the Thames River Basin environment can be achieved by draconian reductions in new housebuilding targets, particularly in the East of England, South East England and Greater London. But what will be the impact of these proposals on the housing opportunities of families living in these three regions? First, the housing market will be tighter; market rents will be higher and so will be housing prices. However, if central government and local authorities focus their

Water resource impacts of new housebuilding

197

housebuilding efforts specifically on rental housing associations, the associations access to land will be improved. Second, the tighter market in the Thames River Basin will raise out-migration of households from the area to other parts of the country. It will also reduce in-migration of households from the rest of the country, from the European Union and from the Rest of the World. This will partially reduce the tightening of the housing market already described. The author asks the individual and institutional stakeholders in these matters to consider the arguments advanced here and, if they find them persuasive, to take appropriate action. Sustainable water resource use is a regional collective action challenge that requires both environmental policy and land-use planning.

* The author also acknowledges his debt to staff of the Environment Agency and RWE Thames Water, as well as to that miraculous daily, the Financial Times. Of course, the views expressed here are my own.

21 Beneficial impacts for the Thames River Basin of water company leakage reduction 20062025

21.1

INTRODUCTION

In Paper 17 I set out a general theory to explain why, in the 21st century, we can expect to see an immense expansion in the number of the worlds rivers that suffer a catchment water deficit. This general theory was first applied to the Kafue River Basin in Zambia, which was shown to enjoy a comfortable surplus at the turn of the new millenium (Paper 18). The Zambian exercise itself led to the specification of five drivers that explain the existence of surplus or deficit in a river basin. These are set out in Table 21.1. The first two drivers to deficit, population density and output density, are both known to be to be high in the Thames River Basin. But as far as I was aware the Basin was not in deficit. So a case-study of the Thames seemed to be a worthwhile exercise in making or breaking the general theory of catchment water deficits. This led to Paper 19, which measured all the Basins five density drivers. It concluded that the Thames River Basin is in deficit in the sense set out above, but that the situation is not yet desperate, despite the regions high population density and production density. This is because the catchment is fairly humid and because production recycles back to the river basin about 80 per cent of the water used in the production process, with a consequentially high value of water productivity. The high rate of recycling exists in part because of the modest scale of irrigation. The minimal scale of plant production (either rainfed or irrigated) in the Thames Region is possible because the basin population of 12.5 million exports sufficient goods and
C 2007 IWA Publishing. The Price of Water: Studies in Water Resource Economics and Management 2nd Edition by Stephen Merrett. ISBN: 9781843391777. Published by IWA Publishing, London, UK.

198

Beneficial impacts for the Thames River Basin Table 21.1 Density drivers for catchment water deficits and catchment water surpluses
Density drivers for deficit 1. Persons per square kilometre (a) 2. Value of output per square kilometre (b) 3. Value of irrigation output per square kilometre
a

199

Density drivers for surplus 4. Effective precipitation in millimetres (c) 5. Water productivity in $/cubic metre (d)

Notes Number of persons resident in the catchment divided by catchment area. b Value can be measured in various ways. Catchment gross value added is one such measure. Output from the basins irrigation sector should be excluded from driver 2 as it is covered by driver 3. c Total precipitation minus catchment evapotranspiration, but excluding evapotranspiration during the course of household use and output production. It is an attempt to estimate a natural process driven by total precipitation, radiant energy, wind speed and humidity, rather than a socially-determined variable. d The ratio of total value of output to all evapotranspiration losses in the production of that output. Dr. Stephen Merrett, London. services particularly the latter so that it faces no financial or economic difficulty in importing food produced in other river basins. Those imports themselves require rainfall and/or irrigation for their growth, of course, but the evaporative losses take place in other catchments, not in the Thames River Basin. The next link in this chain of research papers was to assess the likely impact of new housebuilding and population growth on the regions water deficit (Paper 20). The forecast number of new houses and flats in the years 20062025 is 950,000 (net of demolitions) and the forecast number of additional residents living in the catchment area is 2.2 million. As a result, household use of water in the Thames Region would increase by more than 350 million litres per day (Ml/d). Water stress would become more severe than it is already. However, a number of changes may help offset the calamitous impact of such a seismic shift on the Thames ecosystems. These changes include a reduction in the water utilities leakage of water, the spread of domestic metering of water use, design improvements in water-using white goods and fixtures, inter-basin transfers, and desalination plants. In this paper, I examine the first of these changes, estimating the beneficial impacts for the Thames River Basin of reduction in water company leakage during 20062025.

21.2

LEAKAGE IN THE THAMES REGION: SOME BASIC FACTS

There are five water utilities in the Thames Region: Essex & Suffolk Water, South East Water, Sutton & East Surrey Water, Thames Water and Three Valleys Water.

200

The Price of Water

Of these only Thames Water provides both a water and a sewerage service. Not all the companies lie entirely within the Environment Agencys Thames Region. In fact, whereas the area of the Region is 13,000 km2 , the combined areas of the five water utilities is as much as 24,000 km2 . The best source of leakage data is the Office of Water Services. Table 21.2 shows Ofwats estimate of the scale of total leakage by company for the 5-year period beginning in 200001. For readers unfamiliar with this arcane subject it is worth pointing out that between the point at which a utility puts water into supply and the point that it enters the internal network of the user, there are two types of leakage. The first is leakage from the companys distribution system and the second is the leakage from customersunderground supply pipes. The English and Welsh companies estimate that in 200405 their total leakage was made up of 72 per cent from distribution leakage and 28 per cent from supply pipe leakage. Table 21.2 shows that, in the four-year period ending in 200405, annual total leakage by the five Thames Region companies averaged 1.2 thousand million litres per day (Ml/d). Thames Water alone was the source of about 75 per cent of the total. In terms of leakage expressed as litres per property per day or cubic metres per kilometre per day, Thames Water also performed significantly less adequately than the other four companies. Ofwat and the companies use the terms leakage and losses almost interchangeably. From their perspective this makes sense. If the Thames Regions utilities pump water from river or aquifer, store it, convey it to a treatment works, treat the water so received to drinking water standards but then see 1,200 million litres leaking every day of the year prior to reaching the customer, this is certainly a loss. But from an environmental and a hydrological point of view, leakage is not a necessarily a loss. Admittedly, it may evaporate or flow to the saline waters of the Thames Estuary. However, leakage may recycle to river and aquifer, and it may source plant growth such as urban trees, This Papersperspective on leakage is to assess the scale of the positive hydrological impact of leakage reduction in the context of a possible net increase of 950,000 houses and flats in the Thames Region over the period 20062025. The fundamental proposition here is that if the leakage flows that are reduced would otherwise have passed to the atmosphere through evapotranspiration or would have flowed to the Estuary, the beneficial effects of leakage reduction in respect of the water resource for abstraction are clear and positive. But if the leakage flows that are reduced would otherwise have recycled to the river, its tributaries and its aquifers, there are no such beneficial effects of leakage reduction. The implication of this argument is that in understanding the beneficial impacts of leakage reduction we need to know what proportion of the leakage flows reduced would otherwise have been lost to evapotranspiration or to the Estuary.

21.3

FORECASTING THE REDUCTION IN TOTAL LEAKAGE

The struggle to reduce total leakage in England and Wales has been extensively documented in the Environment Agencys Demand Management Bulletin since its first issue in July 1993. However, the best source for forecasting future reductions is Ofwats annual publication Security of Supply, Leakage and the Efficient Use of Water (Ofwat 2005).

Beneficial impacts for the Thames River Basin Table 21.2 Company estimates of total leakage by year
ESSEX & Suffolk Water (Northumbrian South) Million litres per day 20002001 20012002 20022003 20032004 20042005 72 73 67 70 67 South East Water Million litres per day 85 75 72 69 69 Million litres per day 24 24 24 24 24 Million litres per day 688 865 943 946 915 Million litres per day 140 157 152 152 149 Million litres per day 1009 1194 1258 1261 1224 Litres per property per day 147 128 122 116 116 Sutton & East Surrey Water Litres per property per day 91 91 90 91 90 Thames Water Litres per property per day 200 250 272 271 261 Three Valleys Water Litres per property per day 116 129 125 124 120 All Five Companies Litres per property per day

201

Cubic metres per kilometre per day

20002001 20012002 20022003 20032004 20042005

Cubic metres per kilometre per day 9 8 7 7 7 Cubic metres per kilometre per day 7 7 7 7 7 Cubic metres per kilometre per day 22 28 30 30 29 Cubic metres per kilometre per day 10 11 11 11 10

20002001 20012002 20022003 20032004 20042005

20002001 20012002 20022003 20032004 20042005

20002001 20012002 20022003 20032004 20042005

20002001 20012002 20022003 20032004 20042005

Source: Office of Water Services (2005).

202

The Price of Water

Ofwats interest is in the economic level of leakage, as the following quotation shows (Ofwat 2005, p.34). The water companies in England and Wales manage water distribution networks with a total length of approximately 335,000 km. In addition, there are almost 24 million connections to properties and associated customer supply pipes, which all have the capacity to leak. Eliminating leakage would be virtually impossible and enormously expensive. Therefore, target levels for leakage have to balance the needs of customers and the environment. We believe that companies should compare the cost of reducing leakage and the value of the water saved, including any associated environmental and social costs and benefits. The level of leakage at which it would cost more to make further reductions than to produce the water from another source is known as the economic level of leakage (ELL). Operating at ELL means that the total cost of supplying water is minimised and companies are operating efficiently. In 2005 the regulator published a target total leakage volume for each utility and for each of the six years 200405 to 200920. These targets are recorded in Table 21.3. The Table also shows the corresponding year-on-year reduction in leakage. In the future, targets will increase (or decrease) only if a change in the cost of reducing leakage is out of step with a change in the value of water saved, raising or lowering the economic level of leakage, as the case may be. In the absence of such change, the targets will remain steady and the year-on-year reduction after 200910 will equal zero. The era of leakage reduction will terminate. The implication of all this is that from 200405 the total reduction in leakage that is sought by the regulator for the future is the sum of the year-on-year data in Table 21.3 (row 17), i.e. 194 Ml/d. However, this figure overstates the future contribution of leakage reduction in moderating the impact of new housebuilding on the Thames River Basin for two reasons. First, the leakage reduction targets for Table 21.3 take place throughout the companies areas, whereas the Thames Region makes up only 54 per cent of those areas. If in the unlikely event that leakage reduction takes place in a uniform pattern throughout

Table 21.3 Ofwat leakage targets 200405 to 200910 for the five thames region utilities (Ml/d)
Utility Essex & Suffolk Water South East Water Sutton & East Surrey Water Thames Water Three Valleys Water Total Year-on-Year reduction Total reduction 200405 to 20092010 = 194 Ml/d 200405 200506 200607 200708 200809 200910 70 69 25 905 150 1219 n/a 69 69 25 860 150 1173 46 68 69 25 805 145 1112 61 68 69 25 770 145 1077 35 67 69 25 745 145 1051 26 66 69 25 725 140 1025 26

Source: OfwatsSecurity of Supply, Leakage and the Efficient Use of Water 20042005 Report, Table 8.

Beneficial impacts for the Thames River Basin

203

A. TOTAL LEAKAGE

B. ECONOMIC LEVEL OF LEAKAGE

C. LEAKAGE REDUCTION TARGET

D. RECYCLED LEAKAGE

E. EVAPOTRANSPIRATION AND FLOWS TO THE ESTUARY

Figure 21.1
the companies areas, the Regions target would be only 105 Ml/d. However, Thames Water lies entirely in the Thames Region and it makes up 80 per cent of the total leakage target. So a Region target for leakage reduction could be taken as about 160 Ml/d. Secondly, the water resources available for abstraction benefit from leakage reduction only in the case where such leakage would have flowed to the saline waters of the Thames Estuary or would have evapotranspirated. Figure 21.1 is useful here. If the leakage flows that are reduced would otherwise have recycled to the river, its tributaries and its aquifers, there are no such beneficial effects of leakage reduction from a resource perspective. The future contribution of leakage reduction in sourcing the needs of the 2.2 million residents who would live in the houses built under current housebuilding strategies is important but should not be over-estimated. For the present, the author suggests that in quantitative terms it is probably considerably less than 160 Ml/d. This can be compared with the estimate that the new housebuilding programmes for the Thames Region will increase household use of water by some 350 Ml/d in a river basin that is already in deficit.

21.4

CONCLUSIONS

In this Paper the likelihood is examined that Ofwats target reductions in water utility leakage provide an additional water resource to offset the growth in population in the Thames river basin. But unqualified use of these targets overestimates the beneficial impacts of leakage reduction in the Thames Region. First, the Region makes up only

204

The Price of Water

54 per cent of the area of the five companies, Secondly, a proportion of leakage in any case recycles to the river, its tributaries and its aquifers and therefore such flows are not lost to the Region from a hydrological and environmental point of view. For the present, the author suggests that in quantitative terms target leakage reductions are important but would probably contribute considerably less than 160 Ml/d in sourcing the needs generated by population growth.

Bibliography

Abernethy, C.L. (2003). Financing river basin management. In Svendsen, M., Merrey, D.J., & Shah, T. (2003). Irrigation and River Basin Management: Options for Governance and Institutions. Wallingford: CABI. Allan, T. 1995. The political economy of water: reasons for optimism but long term caution, in J.A. Allan and J.H.O. Court (eds) Water in the Jordan Catchment Countries: a Critical Evaluation of the Role of Water and Environment in Evolving Relations in the Region, London United Kingdom: School of Oriental and African Studies. Allan, T. (2001) The Middle East Water Question: Hydropolitics and the Global Economy, London: Tauris. Allan, T. 2003. Virtual water the water, food, and trade nexus useful concept or misleading metaphor?, Water International 28 (1): 106113. Allen RG., Willardson LS., Fredricksen H.D. 1997. Water use definitions and their use for assessing the impacts of water conservation, in de Jager J.M., Vermes L., Ragab R. (eds) Sustainable Irrigation in Areas of Water Scarcity and Drought. Oxford: ICID. Altaf, M.A., Whittington, D., Jamal, H. & Smith, V.K. (1993). Rethinking rural water supply policy in the Punjab, Pakistan. Water Resources Research, 29(7), 19431954. Anderson, P. 1997. Demands and resources for public water supplies in Scotland from 1991 to 2016, in Water and Environmental Management, Vol. 11, No. 3, pp. 164169. Anglian Water Services Limited (2002). Annual Report and Accounts 2002. Huntingdon, AWSL. Anglian Water Services Limited (2003). Anglian Water Services Drought Plan March 2003. Huntingdon, AWSL. Anon. (1975) Medical, health, agricultural and public water supply aspects of nitrates in water: symposium on nitrates in water, Water Treatment and Examination 24: 142223. Berkoff DJW. 1990. Irrigation Management on the Indo-Gangetic Plain. Washington DC: World Bank. Berkoff, J. (2003). Prospects for Irrigated Agriculture: Has the International Consensus Got It Right? http://www.brad.ac.uk/acad/bcid/GTP/altwater.html.

205

206

Bibliography

Bhatia, R., Rogers, P., Briscoe, J., Sinha, B. & Cestti, R. (1994). Water Conservation and Pollution Control in Indian Industries: How to Use Water Tariffs, Pollution Charges, and Fiscal Incentives, Washington: World Bank. Bjerregaard, R. (1998). Getting Europes waters cleaner: getting the citizens involved, Water Policy, 1(1), 7380. Blackie, J.R., Houghton-Carr, H.A., McCartney, M.P., Moores, J.P., Davies, H.N. (1996) The Trinity Catchment Study: Final Report, Wallingford: Institute of Hydrology. Blitz, R. (2005). 1.5 bn boost for sea freight. Financial Times 21 July 2005. Blomqvist A. 1996. Food and Fashion: Water Management and Collective Action among Irrigation Farmers and Textile Industrialists in South India. Link oping: University of Link oping. Briscoe, J. (1995). The German Water and Sewerage Sector: How Well it Works and What This Means for Developing Countries, Washington: World Bank. Bruns BR. & Meinzen-Dick RS. (2000) Negotiating Water Rights. London: IT Publications. Bruns BR. & Meinzen-Dick RS. (2001) Water rights and legal pluralism: four contexts for negotiation. Natural Resources Forum 25: 110. Cairncross, S. & Kinnear, J. (1991). Water vending in urban Sudan. Water Resources Development, 7, 267273. Cairncross, S. & Kinnear, J. (1992). Elasticity of demand for water in Khartoum, Sudan. Social Science and Medicine, 34, 183189. Chapman, G., Cleese, J., Gilliam, T., Idle, E., Jones, T. & Palin, M. (1979). Monty Pythons The Life of Brian (of Nazareth), London: Eyre Methuen. Chief Advisers Office (1997) An Introduction to Jersey: Jersey and the Channel Islands in relation to England and France, St. Helier: Chief Advisers Office. Chilton, P.J. & Bird, M.J. (1994) Nitrate in Jerseys Groundwater, Keyworth: British Geological Survey. Choe, K., Whittington, D. & Lauria, D. (1996). The economic benefits of surface water quality improvements in developing countries: a case study of Davao, Philippines. Land Economics, 72(4), 519537. Collingwood Environmental Planning and Land Use Consultants (2005). SEERA Climate Change Adaptation and Mitigation Implementation Plan: Consultation Draft. South-East England Regional Assembly, London. van Damme, H.M.G. & White, A. (1984). Technology choices for the decade. Bourne, P.G. (ed.) Water and Sanitation: Economic and Sociological Perspectives. San Diego, Academic. Davis, J. & Whittington, D. (1998). Participatoryresearch for development projects: a comparison of the community meeting and household survey techniques. Economic Development and Cultural Change, x(y), 7394. Department of Agriculture and Fisheries (2002) Statistics 2001, Trinity: DAF. Department of Agriculture and Fisheries (DAF) (2004) Code of Good Agricultural Practice for the Protection of Water (The Water Code) Jersey, Trinity: DAF. Department of the Environment (1986) Pollution Paper No. 26 by the Nitrate Coordination Group, London: HMSO. Department of the Environment, Food and the Regions (2000) Fertiliser Recommendations for Agricultural and Horticultural Crops RB 209, London: HMSO. Dinar, A. (ed.) (2000) The Political Economy of Water Pricing Reforms, New York: Oxford University Press. Dow, S.C. (2002). Economic Methodology: an Inquiry. Oxford: Oxford University Press.

Bibliography

207

Dubourg, W.R. (1993) Sustainable management of the water cycle in the United Kingdom (Norwich, Centre for Social and Economic Research on the Global Environment). Dubourg, W.R. (1995) Pricing for sustainable water abstraction in England and Wales: a comparison of theory and practice (Norwich, Centre for Social and Economic Research on the Global Environment). Dyson, J.R. 1994. Accounting for non-accounting students, London, Pitman, 3rd edn. Earl, P. E. (1995). Microeconomics for Business and Marketing. Cheltenham: Edward Elgar. Environment Agency (19932006) Demand Management Bulletin, www.environment-agency.gov.uk/ savewater Environment Agency (2002) Demand Management Bulletin, 54 (August): 3. Environment Agency. 2002. Drought Plan Guideline. Bristol: Environment Agency. Environment Agency Anglian Region. 2001. Water Resources for the Future: a Strategy for Anglian Region. Peterborough: Environment Agency. Environment Agency Anglian Region. 2003. 2003 Regional Drought Management Plan Final Plan. Peterborough: Environment Agency. European Commission (1991) Urban Waste Water Treatment Directive 91/271/EEC, Brussels: EEC. Fass, S. (1988). Political Economy in Haiti: the Drama of Survival. (New Brunswick NJ: Transaction Books). Foster, S., B. Adams, M. Morales, S. Tenjo (1993) Groundwater protection strategies: a guide towards implementation (Lima, CEPIS). Foster, I.D.L., Ilbery, B.W., Hinton, M.A. (1989) Agriculture and water quality: a preliminary examination of the Jersey nitrate problem, Applied Geography 9: 95 113. Fowler, S. (1995) Water wise: the RSPBs proposals for using water wisely (Sandy, Royal Society for the Protection of Birds). Foster, S.S.D. (1976) The vulnerability of British groundwaters to pollution by agricultural leachates, MAFF Technical Bulletin, 32: 6891. Foster, S.S.D., Bridge, L.R., Cripps, A.C., Darling, W.G. (1977) The Groundwater Nitrate Problem, London: Institute of Geological Sciences. Frederiksen HD. 1996. Water crisis in developing world: misconceptions about solutions. Journal of Water Resources Planning and Management 122 (2): 7987. Frigot, D. (2001) Jerseys Rural Heritage: a Farming Way of Life, Jersey: Royal Jersey Agricultural & Horticultural Society. Gale, R., Barg, S. & Gillies, A. (1995). Green Budget Reform: an International Casebook of Leading Practices, London: Earthscan. Gass, G.M., Robins, N.S., MacDonald, A.M. (1996) Groundwater Resources Degradation in Jersey: Socio-Economic Impacts and their Mitigation, Keyworth: British Geological Survey. Gazmuri Schleyer R. & Rosegrant MW. (1996) Chilean water policy: the role of water rights, institutions and markets. Water Resources Development 12 (1): 3348. Giddens, A. (1984). The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press. Gleick, P.H., P. Loh, S.V. Gomez, J. Morrison (1995) California water 2020: a sustainable vision (Oakland, Pacific Institute for Studies in Development, Environment and Security). Green, A.R., Feast, N.A., Hiscock, K.M., Dennis, P.F. (1998) Identification of the source and fate of nitrate contamination of the Jersey bedrock aquifer using stable nitrogen isotopes, in Robins (ed.) pp. 2335.

208

Bibliography

Green, C.H. (1990). Economics in a Time of Failure, London, Middlesex Polytechnic, Flood Hazard Research Centre Publication No. 180. Guthrie, J. (2005). Homes will not ruin countryside, says minister. Financial Times 1718 September. Haddad BM. (2000) Rivers of Gold: Designing Markets to Allocate Water in California. Washington DC: Island Press. Hearne RR. & Easter KW. (1995) Water Allocation and Water Markets: an Analysis of Gains-from-Trade in Chile. Washington DC: World Bank. Hills, J.S. (1995). Cutting Water and Effluent Costs, Rugby: Institution of Chemical Engineers. Hodge, C.A.H. & Seale, R.S. (1966). The Soils of the District around Cambridge. Harpenden: Agricultural Research Council. Hodgson, G. M. (1994). Cultural and institutional influences on cognition. Hodgson, G.M., Samuels, W.J. & Tool, M.R. (eds). Institutional and Evolutionary Economics A-K. Cheltenham: Edward Elgar. Hodgson, G., Samuels, W. & Tool, M. (Eds). (1994). The Elgar Companion to Institutional and Evolutionary Economics, Cheltenham: Edward Elgar. Hoekstra, A.Y. and P.Q. Hung. 2002. Virtual water trade: a quantification of virtual water flows between nations in relation to international crop trade, Value of Water Research Report Series No 11. Delft: The Netherlands: IHE. van Hofwegen P. & Svendsen M. 2000. A Vision of Water for Food and Rural Development. The Hague: Netherlands Directorate General for Development Cooperation. Holmes, M.R.J. (1979) The place of applied nitrogen in crop production in Chemistry and Agriculture, London: Royal Society of Chemistry. House of Lords (2003). Water Bill [HL]. London: Stationery Office. Howard, J. (1999) La Rosi` ere Jersey: the Largest Freshwater Reverse Osmosis Plant in the British Isles, St. Helier: JNWWC. Howsam, P. & R.Carter (Eds) (1996) Water policy: allocation and management in practice (London, Spon). H otte, M.H.H., van der Vlies, J. & Hafkamp, W.A. (1995). Levy on surface water in the Netherlands, in Gale et al. op.cit. pp. 220230. Jersey New Waterworks Company Limited (1998) A Description of the Water Supply of Jersey, St. Helier: JNWWC. Jersey New Waterworks Company Limited (2002) Annual Report and Accounts for the Year Ended 31st December 2001, St Helier: The Jersey New Waterworks Company Limited. Kaminsky, G., Lizondo, S., Reinhart, C. (1998). Leading indicators of currency crises. IMF Staff Papers (45) 148. Kay M., Franks T., Smith L. (1997) Water: Economics, Management and Demand. London, E. & F.N. Spon. Kemper, K. (1996) The cost of free water: water resources allocation and use in the Curu valley, Cear a, Northeast Brazil (Link oping, University of Link oping). Kemper KE. (2001) Markets for tradable water rights. In Meinzen-Dick RS. & Rosegrant MW. (2001a). Kemper KE. & Olson D. (2000) Water pricing: the dynamics of institutional change in Mexico and Cear a, Brazil. In Dinar A. (ed) pp: 339357. Kinnersley, D. (1994) Coming clean: the politics of water and the environment (London, Penguin). Koubratova M. 1997. Water pricing for irrigation during the transition to market economy in Bulgaria, in Kay M., Franks T., Smith L. 1997. Water: Economics, Management and Demand. London: Spon Press.

Bibliography

209

Kraemer, A. (1995) Water resources taxes in Germany in: R. Gale & S. Barg with A. Gillies (Eds) Green budget reform: an international casebook of leading practices (London, Earthscan) chapter 16. Kuhn, T.S. (1970) The Structure of Scientific Revolutions, Chicago: University of Chicago Press. Langley, J. & Kett, S. (2005) Quinquennial Review of Jersey Flowing Waters (1998 2004). Lant, C. 2003. Commentary, Water International 28 (1): 113115. Lee, F. (1994). Full cost pricing, in Hodgson et al., op.cit. Vol. 1, 262266. Linn, J.F. (1983). Cities in the Developing World. New York: Oxford University Press. London Assembly Environment Committee (2005.) Down the Drain: Londons Water Usage and Supply. LAEC, London. Lott, J.E., Parkinson, R.J., Fuller, M.P., Beattie, L. (1999) Nitrates and Phosphates in Surface Water on Jersey: Sources and Land Management Strategies, Plymouth: Department of Agriculture and Food Studies of the University of Plymouth. Lott, J.E., Parkinson, R.J., Fuller, M.P., Beattie, L. (1999) Nitrates and Phosphates in Surface Water on Jersey: Sources and Land Use Management Strategies, Plymouth: University of Plymouth Department of Agriculture and Food Studies; and Howard Davis Farm: States of Jersey Department of Agriculture and Fisheries. Lundqvist, J. (1998) Avert looming hydrocide, Ambio 27 (6): 428433. Lundqvist, J., U. Lohm, M. Falkenmark 1985. Strategies for river basin management: environmental integration of land and water in a river basin, Dordrecht, Reidel. McDonald, A.T. & D. Kay (1988) Water resources: issues and strategies (Harlow, Longman). MacRae, D. & Whittington, D. (1988). Assessing preferences in cost-benefit analysis: reflections on rural water supply evaluation in Haiti. Journal of Policy Analysis and Management, 7, (20), 246263. Maslow, A. (1954). Motivation and Personality. New York: Harper and Row. Matthews, A. & Carter, A. (1999) An Assessment of the Persistence and Mobility of Chlorthal Dimethyl and Chlorthal, Keyworth: ADAS. Meinzen-Dick R. & Rosegrant MW. 1997. Water as an economic good: incentives, institutions, and infrastructure, in Kay M., Franks T., Smith L. 1997. Water: Economics, Management and Demand. London: Spon Press. Meinzen-Dick RS. & Rosegrant MW. (2001a) Overview. In Meinzen-Dick, R. & Rosegrant, M.W. (eds) (2001b). Meinzen-Dick RS. & Rosegrant MW. (eds) (2001b) Overcoming Water Scarcity and Quality Constraints. Washington, IFPRI. Merrett, S. 1997. Introduction to the economics of water resources: an international perspective, London, UCL Press. Merrett, S. (1999a) The regional water balance statement: a new tool for water resources planning, Water International, 24 (3): 268274. Merrett, S. (1999b). The political economy of water abstraction charges, Review of Political Economy 11, (4), 431442. Merrett, S. (2001) Behavioural Studies of the Domestic Demand For Water Services In Africa, Water Policy, 4 (1): 6981, 2002. Merrett, S. (2002a) Water for Agriculture: Irrigation Economics in International Perspective, London: Spon Press. Merrett S. 2002b. Twelve theses on the cost and use of irrigation water. Irrigation and Drainage 51 (3): 265268.

210

Bibliography

Merrett, S. 2003. Virtual water and Occams Razor, Water International 28 (1): 103105. Merrett, S. (2007) Behavioural studies of the domestic demand for water services in Africa In The Price of Water (2nd Edition). IWA Publishing, London. Merrett, S. 2008. Farm-level drought management: an irrigation case-study from the UK., The Price of Water: Studies in Water Resource Economics and Management 2nd edition, London: IWA Publishing. Merrett, S. (2005a) Integrated water resources management and the hydrosocial balance, in Merrett, S. (2005c): 2036. Merrett, S (2005a). The Price of Water: Studies in Water Resource Economics and Management. International Water Association, London. Merrett, S. (2005b) Deconstructing households willingness-to-pay for water in lowincome countries, in Merrett, S. (2005c): 118131. Merrett, S. (2005b). Catchment water deficits in the 21st century. Water Policy, 7 (2): 141149. Merrett, S. (2005c) The Price of Water: Studies in Water Resource Management and Economics, London: IWA. Merrett, S. (2005c). Catchment water deficits: an application to Zambias Kafue River Basin. Water Policy (forthcoming). Merrett, S. & C. Sharp (1991) The denarius hypothesis: house price inflation in owneroccupied London, in: M. Satsangi (Ed.) Changing housing finance systems (Glasgow, Centre for Housing Research, pp. 271312). Miloradov, M. & P. Marjanovic 1998. Guidelines for conducting water resources assessment, Studies and reports in hydrology No. 55, Paris, UNESCO. Morris, J. et al. 2003. The sustainability of irrigation in England and the impact of water pricing and regulation policy options. In Maksimovic, C., Butler, D. Memon, F.A. (eds) 2003. Advances in Water Supply Management. Lisse: Swets & Zeitlinger: 623632. Mu, X., Whittington, D. & Briscoe, J. (1990). Modelling village water demand behavior: a discrete choice approach. Water Resources Research, 26, 521529. National Rivers Authority (1993) NRA water resources strategy (Bristol, National Rivers Authority). National Rivers Authority (1995) 199596 annual abstraction charges (Bristol, National Rivers Authority). National Rivers Authority Thames Region 1994. Future water resources in the Thames region: a strategy for sustainable management, Reading, NRATR. Office for National Statistics (2004). Regional Gross Value Added. ONS, London. Office of Water Services (2005), Security of Supply, Leakage and the Efficient Use of Water 200405 Report, Birmingham: Ofwat. Ofwat. (1994). 199394 Report on the Cost of Water Delivered and Sewage Collected, Birmingham: Ofwat. ORiordan, T. (ed.) (1995) Environmental Science for Environmental Management, Harlow: Longman. Owen, D. 1994. A Review of Techniques for Systematic Client Consultation. World Bank: Southern Africa Department. Oxford University Press (1979) The Compact Edition of the Oxford English Dictionary Volume I. Oxford: Oxford University Press. Parkinson, R.J, Beattie, L., Fuller, M., Hillman, J., Lott, J. McCrory, D. (2001) Nitrates and Phosphates in Jersey Surface Waters: Sources and Land Management Strategies. Jersey: Department of Agriculture and Fisheries. Pearce, F. (2003) Replumbing the planet, New Scientist, 7th June: 3034.

Bibliography

211

Perry, C.J. (1996) The IIMI Water Balance Framework: a Model for Project Level Analysis, Colombo: International Irrigation Management Institute. Perry CJ. 1996. Alternative Approaches to Cost Sharing for Water Service to Agriculture in Egypt. Colombo: International Irrigation Management Institute. Pigou, A.C. (1932) The economics of welfare (London, Macmillan). Pigram, J.J., R.J. Delforce, M.L. Coelli, V. Norris, G. Antony, R.L. Anderson, W.F. Musgrave (1992) Transferable water entitlements in Australia (Armidale, Centre for Water Policy Research). Public Services Department (2000) Water Made Clear: Guidance Notes on the Provisions of the Water Pollution (Jersey) Law, 2000, Bellozanne: PSD. Rees, J. with S. Williams (1993) Water for life: strategies for sustainable water resource management (London, Council for the Protection of Rural England). Rees, J., S. Williams, J. Atkins, C. Hammond, S. Trotter (1993) Economics of water resource management (Hull, University of Hull, R&D note 128). Reisner M. (1993) Cadillac Desert: the American West and its Disappearing Water. New York: Penguin Books. Ricardo, D. (1821) On the principles of political economy and taxation (Cambridge, Cambridge University Press, edited by P. Sraffa, 1951). Robins, N.S. (ed.) (1998) Groundwater Pollution, Aquifer Recharge and Vulnerability, London: Geological Society. Robins, N.S. (2000) The Water Resources of Jersey: an Overview, Wallingford: British Geological Survey. Robins, N.S. & Smedley, P.L. (1998) The Jersey Groundwater Study, Keyworth: British Geological Survey. Robinson, J. & J. Eatwell (1974) An introduction to modern economics (London, McGraw-Hill, revised edition). Rogers, P. 1991. What water managers and planners need to know about climate change and water resources management, Proceedings of the first national conference on climate change and water resources management, US Army Corps of Engineers. Rogers, P. 1992. Comprehensive water resources management: a concept paper, Washington DC, World Bank. Rosegrant MW. & Binswanger HP. (1994) Markets in tradable water rights: potential for efficiency gains in developing country water resource allocation. World Development 22 (11): 16131625. Rosegrant MW. & Gazmuri Schleyer R. (1996) Establishing tradable water rights: implementation of the Mexican water law. Irrigation and Drainage Systems: pp. 263 279. Rosegrant MW. & Meinzen-Dick RS. (1996) Water resources in the Asia-Pacific region: managing scarcity. Asian-Pacific Economic Literature 10 (2): 3253. Rosegrant MW. & Ringler C. (1998) Impact on food security and rural development of transferring water out of agriculture. Water Policy 1 (6): 56786. Russell, B. (1996) History of Western Philosophy and its Connection with Political and Social Circumstances from the Earliest Times to the Present Day, London: Routledge. Sandmo, A. (1987). Public goods, in Eatwell, J., Milgate, M. &, Newman, P. (Eds), The New Palgrave: a Dictionary of Economics, pp. 10611067, London: Macmillan. Schiffler M. (1997) Intersectoral water markets: a solution for the water crisis in arid areas? In Kay M., Franks T., Smith L. (1997). Schiffler, M., H. K oppen, R. Lohmann, A. Schmidt, A. W achter, C. Widmann (1994) Water demand management in an arid country: the case of Jordan with special reference to industry (Berlin, German Development Institute).

212

Bibliography

Scott Wilson Pi esold (2003), Integrated Kafue River Basin Environmental Impact Study Main Report, Ashford: Scott Wilson Pi esold. Seckler D. et al. 2000. World Water Supply and Demand: 1995 to 2025. Colombo: International Water Management Institute. Shiklomanov, I.A. 1997. Assessment of water resources and water availability in the world, Stockholm, World Meteorological Organization. Shiklomanov, I.A. (2000). Appraisal and assessment of world water resources. Water International 25(1), 1132. Simon, H.A. (1959). Theories of decision-making in economics and behavioural sciences. American Economic Review, 49, 253283. Simon, H.A. (1982) Models of Bounded Rationality: Volume 1, Economic Analysis and Public Policy; Volume 2, Behavioural Economics and Business Organization. Cambridge, Mass.: MIT. S oderbaum, P. (1994) Environmental policy, in: Hodgson et al, vol. I, pp. 193199. Steppacher, R. (1994) K.William Kapp, in: Hodgson et al, vol. I, pp. 435441. Talbot, R. & Whiteman, R. (1996). East Anglia and the Fens. London: Cassell. Tuddenham, M. (1995) The system of water charges in France in: R.Gale & S. Barg with A. Gillies (Eds) Green budget reform: an international casebook of leading practices (London, Earthscan, chapter 14). Turner, R.K. & W.R. Dubourg (1993) Water resource scarcity: an economic perspective (Norwich, Centre for Social and Economic Research on the Global Environment). Veblen, T. (1919). The Place of Science in Modern Civilisation. New Brunswick: Transaction Books. Watson RT. et al. 2001. Climate Change 2001: Synthesis Report of the Intergovernmental Panel on Climate Change. Cambridge and New York: www.ipcc.ch/pub/syreng.htm Whittington, D. 1998. Administering contingent valuation surveys in developing countries. World Development, 26(1), 2130. Whittington, D., Briscoe, J., Mu, X., & Barron, W. (1990A). Estimating the willingness to pay for water services in developing countries: a case study of the use of contingent valuation surveys in Southern Haiti. Economic Development and Cultural Change, 38(2), 293311. Whittington, D. & Choe, K. (1992). Economic Benefits Available from the Provision of Improved Potable Water Supplies. Washington DC: US Agency for International Development. Whittington, D., Choe, K. & Lauria, D. (1997). The effect of giving respondents time to think on tests of scope: an experiment in Calamba, Philippines, in R.J. Kopp, W.W. Pommerehne, & N. Schwarz (eds), Determining the Value of Non-marketed Goods:: Economic, Psychological and Policy Relevant Aspects of Contingent Valuation Methods, Rotterdam: Kluwer. Whittington, D., Davis, J., & McClelland, E. (1998). Implementing a demand-driven approach to community water supply planning: a case study of Lugazi, Kenya. Water International, 23, (3), 134145. Whittington, D., Lauria, D.T., Choe, K., Hughes, J.A., Swarna, V. & Wright, A.M. (1993). Household sanitation in Kumasi, Ghana: a description of current practices, attitudes, and perceptions. World Development, 21, 733748. Whittington, D., Lauria, D. & Mu, X. (1991). A study of water vending and willingness to pay for water in Onitsha, Nigeria. World Development, 19, 179198. Whittington, D., Lauria, D. & Mu, X. (1991) A study of water vending and willingness to pay for water in Onitsha, Nigeria, World Development 19 (179-198). The material in this chapter originally appeared in the journal Water Policy, Volume 8, Number 4, 2006: pp. 339349.

Bibliography

213

Whittington, D., Lauria, D. Okun, D.A. & Mu, X. (1989). Water vending activities in developing countries: a case study of Ukunda, Kenya. Water Resources Development, 5, 158167. Whittington, D., Lauria, D., Wright, A.M., Choe, K., Hughes, J.A. & Swarna, V. (1993). Household demand for improved sanitation services in Kumasi, Ghana: a contingent valuation study, Water Resources Research, 29, (6), 15391560. Whittington, D., Mu, X. & Roche, R. (1990). Calculating the value of time spent collecting water: some estimates for Ukunda, Kenya. World Development, 18, 269 280. Whittington, D., Okorafor, A., Okore, A. & McPhail, A. (1990). Strategy for cost recovery in the rural water sector: a case study of Nsukka District, Anambra State, Nigeria, Water Resources Research, 26, 18991913. Whittington, D., Smith, V.K., Okorafor, A., Okore, A., Liu, J.L., & McPhail, A. (1992). Giving correspondents time to think in contingent valuation studies: a developing country application. Journal of Environmental Economics and Management, 22, 205 225. Whittington, D. & Swarna, V. (1994). The Economic Benefits of Potable Water Supply Projects to Households in Developing Countries. Manila: Asian Development Bank. Whittington, D., Wu, X. & Sadoff, C. (2005). Water resources management in the Nile basin: the economic value of cooperation, Water Policy 7(3), 227252. Willows RI. & Connell RK. Eds. 2003. Climate Adaptation: Risk, Uncertainty and Decision-making. Oxford: UK Climate Impacts Programme. Winpenny, J. (1994) Managing water as an economic resource (London, Routledge). World Bank (1975). Issues in Village Water Supply. Washington: World Bank. World Bank Water Demand Research Team (1993). The demand for water in rural areas: determinants and policy implications, The World Bank Research Observer, 8(1), 4770. World Health Organisation (1970) International Standards of Drinking Water, 3rd Edition, Geneva: WHO. Young, C.P. & Gray, E.M. (1978) Nitrate in Groundwater, Medmenham: Water Research Centre. Zaman, M. (1996) Sustainable irrigation management through cost recovery system, in: Howsam & Carter, op. cit. pp. 221227.

WEBSITES
http://www.anglianwater.co.uk/ http://www.environment-agency.gov.uk/ www.environment-agency.gov.uk/regions/thames/ www.thames-water.com/

Index

A abstraction charges, scope of developing a taxonomy of charge-setting principles, 9396 impact on users, 9899 and sustainable catchment management, 9698 and theory of rent, 9293 Anglian region, drought management of, 5960 Anglian Water Servicess drought plan, 6364 drought plan and water economy, 6465 Environment Agency Anglian Regions drought plan, 6263 regulatory instruments, 6567 water economy of the region, 6062 water utilities in the region, 60 Anglian Water Services Limited (AWS), 44, 60, 6364 B baseline statements, for regional water balance, 10, 21, 3839 behavioural studies, of African domestic water services discrete choice model, rationale for, 113114 market networks, 106108 methodology, 110113 uses of water, 108110 water demand schools work, on the cities of Kumasi and Onitsha, 106

Bellozanne sewage treatment works (STW), 33 biological oxygen demand (BOD), 136 black reservoir, 14 Bruns, B. R., 83 C catchment area, of water, 2 Dwyer catchment, 160164 general hypothesis of surplus and deficit, 165 global deficit, 160164 hydrosocial balance, 159 water surplus, 56 catchments water economy instream, 38 outstream, 41 rainfed, 38 chalk groundwater, 61 charge-setting taxonomy average total cost charge, 94 environmental regulation charge, 94 incentives charge, 9596 marginal cost charge, 9495 market-clearing charge, 9394 no charge, 93 Pigovian charge, 95 revenue-maximizing charge, 93 chemical oxygen demand (COD), 136 climate change, 58 2004 Code of Good Agricultural Practice for the Protection of Water, 143 Colorado-Big Thompson scheme, 94 compensating non-leakage (CNL), 45

214

Index
concessions, in trading of water-rights, 81 costs, for access to water, 34 D demand, for water, 23 consumption of water, 88 economic demand for water, 8889 need for water, 88 supply-side leakage and evaporation of water, 8990 use of water, 8788 for waste water services, 133135 demand and supply functions, for farmers water-rights, 7879 density drivers, for catchment water deficits and catchment water surpluses, 180, 199 disposal charges environmental regulator, 137138 objectives, 132133 and water utility, 136137 DOE Pollution Paper No. 26 (DoE 1986), 141 domestic water and waste water services, in the low-income countries, 104106 double-entry water accounting, 10, 15 drought defined, 58 management, see Anglian region, drought management of Dubourg, W. R., 96 Dwyer catchment population, productivity and output change over time, 161 redemptive options, 161164 E EC drinking water standards (80/778/EC), 141 91/676/EC Nitrates Directive, 142 economic demand, for water, 8889 ECs Urban Waste Water Treatment Directive (91/271/EC), 149 effluent disposal, 131132 engineering variables, 19 England and Wales British Waterways, 15 EUs Urban Waste Water Treatment Directive, 33 evapotranspiration (ET) losses, of water, 41 F farmers supply function, 80 farm-level drought management strategies Anglian Region, description of, 44 informational strategy, 5156 infrastructural strategy, 4750 review of the activities of the agribusiness, 4547 financial accounting practice and regional water balance statement, 13 G

215

Giddens, Anthony, 110 gold reservoir, 15 green reservoir, 15 groundwater and surface water abstraction, 13 H Haddad, B. M., 81 household water use behaviour, 45 hydrological variables, 19 hydrosocial balance, 2022, 38 of a catchment, 39 for a defined region in 2003, 87 of Island of Jersey, 25 for a specified area in a base year, 45 for a specified region in a base year and a scenario year, 40 of Thames catchment, 183184 hydrosocial cycle, 1011, 13 flow types in, 23 I industrial effluent policy and uses of water demand for waste water services, 133135 disposal charges, 136138 generation and regulation of industrial effluent, 131132 measurement of pollution, 135136 objectives of disposal charges, 132133 instream water, case of, 1 integrated water resources management (IWRM) bridge between quality and quantity, 2124 and hydrosocial balance, 2022 Island of Jersey, case example of, 2434 International Hydrological Programme, 19 Introduction to the Economics of Water Resources: An International Perspective I, 9 Island of Jersey hydrosocial balance of states of, 25 nitrate pollution on, see nitrate pollution and management, on Island of Jersey physical geography and hydrology, 24 water quality, 2834 water supply, 2427 water use, 2728 Itezhi-Tezhi reservoir, 170 J Jerseys farmers and irrigation methods, 29 Jersey Water Code, 150 Jersey Water (1972) Law, 144

216
K

Index
pollution measurements, 135136 pre-treatment of water, 134 R rainwater collection, 13, 26 recycling, of water, 2, 1314 regional water balance statement change statements and its use, 1619 formulation of, 1013 storage capacity of regions, 1415 supply sources, 1214 use behaviour category wise, 12, 15 rent theory and abstarction charges, 9293 reused water, 2 Ricardo, D., 9293 river basins water economy basin water productivity, 41 benefits and its sharing, 3742 Rosegrant, M. W., 81 ruralurban market theory in abstraction rights, limits of concessions, 81 legal context, 8283 partial sale of the farmers water-rights, 8182 sale of land, 8182 third party effects, 8384 time-scale propensities, 81 transaction costs, 8485 water rights markets, 84 Russell, Bertrand, 71 S sale of land and water-rights, 8182 scenario statements, for regional water balance, 10, 21, 3839 Schiffler, M. H., 79 Scott Wilson Pi esold study, of groundwater abstraction for the Kafue River Basin, 170 seawater desalination, by JNWWC, 145146 Silver Birches plc, review of the activities of the agribusiness company information, 45 drought warning chart, 55 hydrosocial balance for, 46 infrastructural planning, 5156 monthly Anglian Water Services, 47 purchases of irrigation water, 54 supply planning at, 4850, 5253 water supply and use, 4647 soakaways and septic tanks, recommendations on, 30 storage capacity, of regions, 1415 supply-side leakage and evaporation, of water, 8990 sustainable catchment management and abstraction charges, 9698

Kafue Gorge Upper (KGU) hydroelectric power station, 170 Kafue River Basin description, 168170 economic demand for water, 171 environmental requirements for, 172174 estimated total surface water abstraction rights and the available resource in cubic metres/second, 171, 173 food source of, 172 groundwater abstraction, 170171 Kemper, Karin, 8384 Kyoto consensus, on water, 7475 L leakage data, of Thames Catchment area, 199203 leather industry and effluent wastes, 134 M maximum admissible concentration (MAC), 3133, 141142 Meinzen-Dick, R. S., 81, 83 Merretts law, on regional water balance, 16 The Middle East Water Question: Hydropolitics and the Global Economy, 68, 73 N nitrate pollution and management, on Island of Jersey benefits of, 149155 the department of agriculture and fisheries, water quality management by, 143144 of groundwater and surface waters, 141143 households, water quality management by, 147148 the Jersey New Waterworks Company Ltd. (JNWWC),water quality management by, 144147 waste water discharges, water quality management of, 148149 O Occams razor, to the virtual water thesis, 71 outstream water, case of, 1 P partial sale, of the farmers water-rights, 8182

Index
T Thames catchment, 14, 41, 91 density analysis of, 186188 economic production, 184185 hydrology of, 181 hydrosocial balance, 183, 183184 increase in number of houses and flats, 194196 leakage losses in, 199203 micro-components of householdwater use in, 194 population, 183184, 192193 range of environmental services, 181 tipping process during deficit, 188189 useful information on, 192 water deficit, 185186 water pricing, 185 theses, on costs and use of irrigation water, 101103 third party effects, in trading of water-rights, 8384 time-scale propensities, in trading of water-rights, 81 transaction costs, in trading of water-rights, 8485 U UK Water Bill, 43 urban actors demand function, 7879 V Val de la Mare catchments nitrogen export coefficients, 30 virtual water concept, 78 flaws with Allans thesis, 70, 7374 importation of, in MENA region, 70 and Kyoto consensus, 7475 as a metaphor, 7273 Occams razor to the virtual water thesis, 71 W

217

water deficit concept, 69 water economy, 59, see also river basins water economy Water Pollution (Jersey) Law 2000, 2930, 148149 water quality, in Island of Jersey household discharges, 30 household usage behaviour, 30 JNWWC supplies, 3033 Public Services Department discharges, 3334 by rainfall and irrigation, 29 Water Resources Act 1991, 63 water resources management balance (WRMB), 19 water-rights, 5 farmers abstraction rights, 80 ruralurban market theory in abstraction rights, limits of, 8085 and urban demands, 7880 water services, in any region, 106 Whittington, Dale, see behavioural studies, of African domestic water services willingness-to-pay concept, in low-income countries anchor prices, 126128 data collection, 118119 private agendas, 124126 problems in the application of demand theory to survey practice, 119120 sanctions for non-payment of the water bill, 123124 and substitutes, 121123 survey methods, 117118 vs. ability-to-pay, 120121 World Meteorological Organization (WMO), 19 World Water Forum, in Kyoto, 7475