Public Disclosure Authorized

Eni ronmental
Accoulnting
Public Disclosure Authorized

for Sustainable
Development
Public Disclosure Authorized

edited by E'
YusufJ. Ahnad
Salahi El Serafy
Ernst Lutz
Public Disclosure Authorized

Bank Symposium
A UJiNEP-World
 4~~~~~~~~~~

i
 Accounting
En4ronmental
for
Sustanable
Development

A UN.EP-WorldBank Symposium
I
I I
EnvironmentalAccounting
for
Sustasinable
Development
editedby
Yusuf J. Ahmad
Salah El Serafy
Ernst Lutz

The WorldBank
Washington,D.C.
1989 The InternationalBankfor Reconstruction
and DevelopmentI THE WORLD BANK
1818H Street,N.W.,Washington,D.C.20433,U.S.A.

Allrightsreserved
Manufactured in the UnitedStatesof America
FirstprintingJune 1989

The findings,interpretations,and conclusions expressed

in this studyare entirelythose ofthe authorsand should
not be attributedin anymannerto the WorldBank,to its
affiliatedorganizations,or to membersof its Boardof
ExecutiveDirectorsor the countriestheyrepresent.

Library
of CongressCataloging-in-Publication
Data

Environmental accountingfor sustainabledevelopment:

selectedpapersfromjoint uNEPfWorld Bankworkshops
I editedbyYusufJ. Ahmad,SalahEl Serafy,Emst Lutz.
p. cm.-(A WorldBanksymposium)
Includesbibliographies
and index.
1. Naturalresources-Accounting.2. Environmental
impactanalysis. 3. Nationalincome-Accounting.
4. Economicdevelopment-Environmental
aspects. L. Ahmad,YusufJ. II. El Serafy,
Salah. Im. Lutz,Ernst IV. UnitedNations
Environment Programme. V. InternationalBankfor
Reconstructionand Development.VI. Series.
HF5686.N3E58 1989
333.7'14-dc2O 89-33458
ISBN0-8213-1224-3 CIP
 Foreword
Internationalorganizationsare making a substantialef-
fort to incorporateenvironmentalconcernsiintheir regu-
laroperations.A great deal of work is nowbeing done to
help clarifythe linkages between developmentand the
environment. Thisworkwillenable us to irntegrateenvi-
ronmentaland resourcemanagementconcernsmore ef-
fectivelyin the economicdecisionmakingprocess.It will
be an essentialcomponentin our effortsto establishthe
basisfor long-termand sustainabledeveloprnent.
Environmentalaccountingis a complexand sometimes
elusivesubject.It is also a tool with great potential,how-
ever,a tool that can help ensure that future calculations
of nationalincomemore accuratelyrepresenttrue, "sus-
tainable"income.The current systemof nationalincome
accountinghas some limitations.Gross domesticproduct
(GDP) figuresare widelyusedby economists,politicians,
and the media. Unfortunately,they are generallyused
without the caveat that they represent an income that
cannotbe sustained.Current calculationsig5nore the deg-
radationof the natural resourcebase and viewthe sales
of nonrenewableresources entirelyas income.A better
v
way must be found to measure the prosperity and pro-
gress of mankind.
We have witnessedincreasingpressure on the environ-
ment and the natural resource base. We have come to
understand that, where te environmentis concerned,
"there is no such thing as a free lunch"and that someone
will eventuallyhaveto bear the "externalcosts" of pro-
duction and consumptionactivities.We must learn to
dis5tinguish
betweentrue incomegenerationand the draw-
ing downof capitalassets by resourcedepletion or deg-
radation.
This volumereflects the attention given during the past
six years by the World Bank and the United Nations
EnvironmentProgramme,as wellas others, to this impor-
tant topic. We hope it will help bring about a needed
changein attitudeand perceptionabout this issue.
BarberB. Conable MostafaTolba
ExecutiveDirector
President UnitedNations
The WorldBank EnvironmentProgramme
I~~~~~~~~~~~~~~~~~~
 Contents

Preface xi
Abbreviations xii
Contributors xiii

1. Environmental and Resource Accounting: An Overview 1

SalahEl Serafy and ErnstLutz
Shortcomingsof the CurrentNationalIncomeMeasures 1
The Necessityof MeasuringSustainableIncome 2
DefensiveExpenditures 2
The Depletionand Degradationof NaturalResources 3
ResourceAccounting 4
LinkingEnvironmentaland ResourceAccountsto the SNA 5
ConstructingEnvironmentaland ResourceAccountsfor DevelopingCountries 6
A Varietyof Approaches,but a CommonTheme 6
Note 7
References 7

2. Toward a Measure of Sustainable Social Net National Product 8

HermanE. Daly
SustainableIncome 8
DefensiveExpenditures 9
Depletionof NaturalCapital 9
Reference 9

3. The Proper Calculation of Income from Depletable Natural Resources 10

Salah El Serafy
ConceptualBackground 11
Weaknessof the DepreciationApproach 12
Conversionto a PermanentIncomeStream 13
Clarificationof the UserCosi,Approach 14
Conclusion 17
Appendix.SplittingReceiptsinto Incomeand Capital 17
Notes 17
References 18

vii
viii Contents

4. Introducing Natural Capital into the SNA 19

Anne Harrison
ExploitableResources 20
PermanentResources 20
Valuationof the Consumptionof NaturalCapital 22
EnvironmentalEnhancementPrograms 22
UnplannedOccurrences 23
Sustainability
and the SNA 23
Summaryof Recommendations24
Note 25
References 25

5. Measuring Pollution within the Framework of the National Accounts 26

Derek W Blades
The Outputof Pollutants 26
PollutionDamage 26
Costsof PollutionAbatement 28
Benefitsof PollutionAbatement 29
Conclusions 31
References 31

6. Correcting National Income for Environmental Losses: Toward a Practical

Solution 32
RoefieHueting
The Needto Use GDPwithProper Caveats 32
CorrectingNationalIncomefor DefensiveEnvironmentalOutlays 33
ComplementingCorrectionsfor DefensiveOutlaysthroughSurveys 34
ComplementingCorrectionsfor DefensiveOutlaysthroughStandardsfor Sustainable
EconomicDevelopment35
Notes 38
References 38

7. Environmental Accounting in DevelopmentPolicy: The French Experience 40

JacquesTheys
The FrenchStratifiedSystemof Information 40
AdaptingEnvironmentalAccountsto DevelopingCountries 45
Proposals 49
Conclusion:Towarda Programof Action 50
Appendix.Nomenclaturesof the FrenchNaturalPatrimonyAccounts 52
Notes 53
References 53

8. Linkages between Environmental and National Income Accounts 54

RichardB. Norgaard
SNA Developed
by Consensus,Not Theory 55
The ValueAggregationIssue 55
The Bounded-Knowledge SynthesisIssue 56
Conclusions 57
Notes 57
References 58
 Contents ix

9. Environmental and Nonmarket Accounting in DevelopingCountries 59

Henry M. Peskin
Usesof ExpandedAccountsin DevelopingCountries 59
ImplementationProblems 60
SuggestedCourseof Action E3
Notes 64
References 64

10. A Proposed Environmental Accounts Framework 65

Henry M. Peskin
Theoryof EnvironmentalAssetServices 65
ValuationConcepts 67
Investmentand Depreciation 67
Sourcesof EnvironmentalAssetDepreciation 69
Overviewof the NationalAccountingStructure 69
Responseof ConventionalAccountsto Nonmarketand EnvironmentalActivity 71
A ModifiedAccountingStructure 72
The Environmentand the ModifiedAccounts:Optionsfor Modifying
GNP 76
Notes 77
References 78

11. Environmental Accounting and the System of National Accounts 79

Peter Bartelmus
EnvironmentStatisticsand PhysicalAccounting 80
Environmentand NaturalResourcesin the SNA 81
MonetaryEnvironmental Accounting 82
Indicatorsof SustainableProductand Expenditure 84
FutureWork 85
Notes 86
References 86

12. Recent Developments and Future Work 88

Ernst Lutz and Salah El Serafy
SNA Expert GroupMeeting(Vienna,March1988) 88
Joint UNEP-WorldBank Expert Meetingon EnvironmentalAccountingand the SNA
(Paris,November1988) 89
SNA Expert GroupMeeting(Luxemburg, January1989) 89
Future Work 90
Annex.The Treatmentof NonreproducibleAssets,Stocks,and Flows
underPresentSNA and "M.60"Standardsfor Reconciliation
Accounts 90
References 91

Appendix. Participants at the uNEP-WorldBank Workshops 93

Index 97
I I I
 Preface
This volume contains selected papers from a series of
workshopsjointlysponsoredby the WorldBank and the
United Nations EnvironmentProgramme(UNEP). UNEP
was created at the StockholmConferencein 1972. Dur-
ing the past seventeenyears, membersof its Governing
Councilhave urged UNEP to help clarifythe linkages
betweendevelopmentand the environment.The council
believedthat such clarificationwouldhelpintegrateissues
of environmentaland resourcemanagementooncerninto
the frameworkof economicdecisionmakingand thiuspro-
vide a basis for long-termand sustainabledevelopment.
In 1982, UNEP'S Session of a Special Character,which
commemoratedthe tenth anniversaryof the Stockholm
Conference,requestedthe executivedirector to develop
methodological guidelinesfor developingcountrieson en-
vironmentalaccountingand its use in developmentpolicy
and planning.
In February1983, UNEP conveneda consultativemeet-
ing under the chairmanshipof YusufJ. Ahmad,assistant
to the executivedirector and director for specialassign-
ments at UNEP. The mainobjectiveof the meetingwasto
ascertainwhetherenvironmentalaccountingcouldbe de-
velopedas a publicpolicytool,based on the presentstate
of the art. Robert Goodlandwas the only 'WorldBank
representativeat that first meeting.He subsequentlypro-
vided much leadership and encouragedeconomiststo
focuson these critical issues. He and YusufJ. Ahmad,
Salah El Serafy,and Jacques Theys chaired the subse-
quent workshops,whichwere sponsoredjointlyby UNEP
and the WorldBank.
xi
Numerousparticipantsat the workshopsmade signifi-
cant contributions,either by presenting papers or by
substantialparticipationin the discussions.They are too
manyto acknowledgeindividually; insteadwe havehad to
content ourselveswith listing the names and affiliations
of all participantsin the appendix to the text. Of the
more than thirty papersor notes that were presented,we
haveselectedthe seventhat appearedto us to have pro-
vided broad insightsor brought out new aspects of a
complexand sometimeselusivesubject.In addition,we
haveincludedthree contributionsthat werewrittenafter-
wardby workshopparticipants(Chapters4, 9, and 10).
In general, the papers indicate current thinkingon the
issueof environmentaland resource accounting.Readers
interestedin examiningpapers not includedin this vol-
ume are encouragedto contactthe particularauthors.
The chapters in this volumereflect different aspects
and approachesto environmentalaccounting.They are
concernedmostlywith financialand economicconsidera-
tions and the prospects of modifyingthe U.N.Systemof
NationalAccounts(SNA) to reflect issues of environmen-
tal and natural resourceconcern.
In publishingthis volumewe havereceivedencourage-
ment, support, or comments from Ramesh Chander,
Vinod Dubey,Stanley Fischer, Robert Goodland,Ken-
neth Piddington,Ibrahim Shihata, MichaelWard, and
JeremyWarford.
YusufJ. Ahmad
Salah El Serafy
Ernst Lutz
 Abbreuiations

CES Conferenceof EuropeanStatisticians isic Internationalstandard industrialclassification

coFOG Classificationof the functions of government IUCN InternationalUnion for Conservationof Nature
(expenditures) and NaturalResources
COIP Classification of outlaysof industriesby purpose MCLA MonetaryCenter forLatin America(also CEMLA)
CPC Centralproductclassification MEB Materials-energybalance
ECA EconomicCommission for Africa NDP Net domesticproduct
ECE EconomicCommissionfor Europe NNP Net nationalproduct
ECLA Economic Commission for Latin America NRA Naturalresourcesaccounting
(alsoCEPAL) OECD Organisationfor Economic Co-operationand
EPP EnvironmentProtectionProgram Development
ESCAP Economicand Social Commission for Asiaand OPEC Organizationof PetroleumExportingCountries
the Pacific PACE Pollutionand abatementcontrolexpenditures
FAO U.N.Food and AgricultureOrganization SGDE Sustainablegross domesticexpenditure
FDES Frameworkforthe Developmentof Environment SGDP Sustainablegrossdomesticproduct
Statistics SNA U.N.Systemof NationalAccounts
GDFCF Grossdomesticfixed capitalformation SNDP Sustainablenet domesticproduct
GDP Grossdomesticproduct SNNE Sustainablenet nationalexpenditure
GNP Grossnationalproduct SSDS Systemof Socialand DemographicStatistics
LARIW International Associationfor Research on In- UNEP U.N.EnvironmentProgramme
comeand Wealth UNESCO U.N. Education Scientific and Cultural
ICOR Incrementalcapital-outputratio Organization
INSEE InstitutNationalde la Statistiqueet des Etudes UNSO U.N.StatisticalOffice
Economiques WHO WorldHealthOrganization

xii
 Contributors

YusufJ. Ahmad RoefieHueting

Senior Adviserto the executivedirector, U.N. Environ- Head, EnvironmentalStatisticsDepartment,Central Bu-
ment Programme,Nairobi,Kenya reau of Statistics,Voorburg,Netherlands
Peter Bartelmus Ernst Lutz
Chief,EnvironmentStatisticsSection,U.N.StatisticalOf- Senior Economist, Environment Department, World
fice, NewYork,U.S.A. Bank,Washington,D.C.,U.S.A.
DerekW. Blades RichardB. Norgaard
PrincipalAdministrator,Economicsand Statistics De- AssociateProfessor,Energyand ResourcesGroup, Uni-
partment, Organisationfor EconomicCo-operationand versityof California-Berkeley,U.S.A.
Development,Paris, France Henry M. Peskin
HermanE. Daly consultant,SilverSpring,Md., U.S.A.(formerlywith Re-
Senior Economist, Environment Department, World sourcesfor the Future)
Bank,Washington,D.C., U.S.A. JacquesTheys
Salah El Serafy Head of Mission,Ministryof the Environment,Paris,
EconomicAdviser,EconomicAdvisoryStaff in the Office France
of the Senior Vice President-Operations,Norld Bank, Jan W. vanTongeren
W.Mashington,
D.C.,U.S.A. Chief,NationalAccountsSection,U.N.StatisticalOffice,
Anne Harrison NewYork,U.S.A.
Consultant,Stockbridge,UnitedKingdom(formerlywith
the WorldBank)

xliii
I I I
 EnvironmentdlandResourceAccounting:
An Overtiew
Salah El Serafy and Ernst Lutz
Most production and consumptionactivAities have some
effect on the physical environment.As economicand
populationgrowthhave occurred,they have increasingly
put pressureon the environmentand the natturalresource
base. Yearsago, whenthe pressure wasstill small, econ-
omistsmayhavebeenjustifiedin makingno referenceto
the role playedby the environment,both as a resource
base and as a "sink" to receivethe residuesof the pro-
ductionand consumptionprocess.But there is little jus-
tificationforthis now.
Economistshaveconsideredthe side-eifectsof produc-
tion and consumptionactivitiesto be "externaleffects."
But such effects are external only if a narrow view is
taken, whichdoes not considerthe effecton the resource
systemas a whole.Althoughthis systemis generallylarge,
it is neverthelessfinite,and in certainrespectsit is subject
to great stress. The realizationis growingthat, wherethe
environmentis concerned,"there is no such thing as a
free lunch"and that someonewilleventuallyhaveto bear
those "externalcosts."If a broaderviewis talken,environ-
mental costs would be internalizedin the production
processes.In this connectionit is essential to reckon
costs and benefitsproperlyand to distinguishclearlybe-
tween true income generationand the drawingdown of
capitalassets by resourcedepletionor degradation.This,
in generalterms,is the subjectof this volume.
Shortcomings of the Current National
IncomeMeasures
Incomeaccounting,if properly done, is a most useful
tool foreconomicanalysisand policyprescriptions.It can
indicatethe levelof economicactivity,its variationsfrom
yearto year,the size of savingsand investment,the limits
society can consume out of its current receipts, factor
productivity,industrial structure, comparativeperform-
1
ance, and many other things. Developmentplanners,
economists,and politiciansthus makefrequentuse of the
nationalincomemeasureof gross nationalproduct (GNP)
and its variants, such as GDP (gross domesticproduct)
and NNP (net nationalproduct) for a varietyof purposes.
GDP, the most commonlyused variant of aggregate in-
come, is essentiallya short-term measure of total eco-
nomic activityfor whichexchange occurs in monetary
terms withina givenyear.It is valuablemostlyfor indicat-
ing short- to medium-termchanges in the level of eco-
nomic activity,and it is particularlyuseful for demand
managementand stabilizationpolicies.As calculatedat
present, however,it is less useful for gauging long-term
sustainablegrowthpartly becausenatural resourcedeple-
tion and degradationare being ignored. Furthermore,
GDP is often used inappropriately as an indicatorof "wel-
fare," frequentlywithoutany caveat about its shortcom-
ings for that purpose. The concept of welfareis much
broader than a monetarymeasure of income. It covers
manydimensionsof subjectivewell-beingother than those
that involvemarkettransactionsor that can be measured
in monetary terms, particularlyfor people whose basic
materialneeds havebeen met.
As most economistsknow,there are severalcontrover-
sial issuesconcerningnationalincomeaccountingas cur-
rentlybeing practiced,such as the treatment of leisure,
householdand subsistenceproduction,other nonmarket
transactions,and the servicesof long-livedconsumerdur-
ables. This chapter willnot deal with any of those issues;
instead it addresses certain environmentaland natural
resourceissuesas they relate to the proper measurement
of incomeand variationsin assets. GDP, as measuredat
present, does not adequatelyrepresent true, sustainable
incomebecauseof two shortcomings.These concernthe
treatment of environmentalprotectioncosts and the deg-
radationand depletionof natural resources.The fact that
2 SalahEl Serafy and Ernst Lutz
these issues are not properly dealt with in the current
U.N.Systemof NationalAccounts(SNA)is a seriousflaw
from an accountingpoint of view.As a result, policy
advicebased on measurementsproducedunder the SNA
can be faultyto the extent that GDP does not adequately
reflectenvironmentaland natural resourceerosion.
The Necessity of Measuring Sustainable
Income
True income is "sustainable"income. This is a key
True~
~~ ~ ~ ~ ~ ~~~
pointstressedbyDaly(Chapter2) and El Serafy(Chapter
3). True income can be thought of as the maximum
amountthat can be consumedin a given period without
redcinofth
reducingthe amount ossbleconumpionin
amuntof
possibleconsumptionin a fuure
future
period.This conceptencompassesnot only current earn-
ings but also changes in assets: capital gains increase
ingsm;
b it.al..soes in assets.capi.Tal gssincrease
ionce;pit lncosehs
rede incoed T . he esenc ot
concpt
hasbee
f icom
sttedby Sr Jhn ick as
the maximumvaluethat a person can consumeduringa
time period and stfi expect to be as well off at the end
of the period as at the beginning(Hicks 1946,
a p. 172).
e mental or natural capital is to be treated conceptually,
ofudnth
Prudentperiodoma
economicmanatghembeginng(uics 1946,
management requires that p.o172).
govern-
ments knowthe maximumamountthat can be consumed
by a nation without causing its eventual impoverishment.
It is important,therefore,that nationalincomebe meas-
ured correctlyto indicate sustainable income. Adjust-
mentsof the SNAappear to be necessaryin the two areas
noted above, since these are currently not dealt with
satisfactorily:the so-called"defensiveexpenditures"to
protector restorethe environmentand the depletionand
degradatio
resourcesused
of natura
degradationof natural resources.
Defensive Expenditures
Actionis often taken to defendthe environmentagainst
encroachmentby economicactivity,and the SNAtreats its
costas generatingincome.Defensiveexpenditurescan be
large or small, dependingon wherethe boundariesare
drawn.Thisvolumeconsidersonlydefensiveexpenditures
againstthe unwantedside-effectsof productionand con-
sumption(such as pollution)but not those relating to
national security,even though similar argumentswould
applyto them.The most obviouscategoryto be included
under defensiveexpendituresis the cost of environmental
protectionactivities.Anotherpossiblecategorywould be
car repair and medicalexpenses as a result of traffic
accidents.Leipert listed other costs that might be in-
cludedand recentlymeasuredthe defensiveexpenditures
for the Federal Republicof Germany(Leipert 1986 and
1987).
It does not make sense to incorporate expenditures
incurred to redress some or all of the negative conse-
quences of production or consumptionactivitiesin the
stream of incomegenerated by economicactivity.It has
thereforebeen proposedthat such outlaysshould not be
countedas finalexpenditure,as is currentlythe case,but
rather as intermediateexpenditure.
A conceptuallydifferentapproachcan be taken by con-
sideringresourcessuch as water,air, soil, and so forth as
natural capital.When such capital is drawndownor de-
graded,this shouldshowup as consumptionin the meas-
urements of national income whether or not defensive
expendituresare actuallybeing incurred to correct for
the negativeeffectsand restore the drawn-downnatural
~cptl Them difrec betweenletheme
defesiv expendi-
captal. The dlifferencebetween the defensiveexpendi-
tures actuallyincurred and the depreciabonof the envi-
ronmentalcapital wouldbe reflected at the level of net
domestic
posed product and
by Harrison (NDP). This approach
outlined has in
in more detail been pro-
Chapter
4. A s arrconcal approach ha been propoed
4. A similar conceptual approach has been proposed by
Peskin (Chapter10), whoproposesthe introductionof a
nature account in additionto the standard accounts for
households,industry,and government.Aside from the
dfulty ing a consen to hoite environ-
difficulty of reaching a consensus as to how the environ-
the greatest difficultylies in actuallyestimatingthe level
of environmentalservicesand damages.
Measuring Pollution within the sNA Framework
Pollution-that is, the dischargeof wastesin waysthat
raise the cost of later activities,harm people, or reduce
the enjoymentthey get from their surroundings-is an
tant area in from
their accountsngs-isan
imp re inwichntental aconts clde
to Improveenvironmental policymaking.Blades
(Chapter 5) distinguishesfour aspects of pollutionand
considersthe extent to whichit is possibleand usefulto
measurethem withinthe frameworkof the national ac-
counts. These aspects are the output of pollutants, the
damagesof pollution, the costs of abatement,and the
benefitsderivedfrom it.
Althoughit maybe feasibleto use nationalaccountsto
measure the output of pollutants,the informationso ob-
tained maybe too generalto be usefulfor environmental
policymaking.Chapter 5 notes that although there are
conceptualand practicaldifficultiesin estimatingthe total
costs of pollutiondamage,it wouldbe possibleand help-
ful to identifysome of the main costs that are already
includedin the nationalaccountsbut are not shownsep-
aratelyat present. The costsof pollutionabatementare a
part of defensiveexpenditures.They havebeen measured
in severalcountriesand have been incorporatedin mac-
roeconomicmodelsto showthe effectof abatementpoli-
cies on prices, output, and employment.In this area the
national accountswould be a valuabletool for environ-
mentalpolicymaking,and Chapter 5 considers in detail
the conceptual and practical problems of measuring
abatementcosts. Finally,althoughit wouldbe interesting

to measure the "market valuation"of the benefits of
pollutionabatement,the practicaldifficultiesinvolvedare
enormous,and it wouldgenerallynot be feasibleto incor-
porate such data in the national accountson a regular
basis.
Measuringthe Differencebetween Environmental
Standardsand Actual Behavior
Hueting(Chapter6) discussesseveralw&s toDdeal with
defensiveexpendituresand lists the pros and ons of the
defensive
andexpenditures
lists the ps ad ccan
variousoptions.Sincehe is skepticalabout the usefulness
of the willingness-to-pay-method,he prefersth't environ-
mental standards be determined by considerations of
health and sustainable development. The costi of achiev-
healtuchstandasutialedev beent. ost ofe
ingld
wouldsuho
standards
show would
how far nthen
b has
etimted.
s These rosts
a country drifted away from
sustainable economic development. Although the ap-
susainab enomic de .t Althoug
the
h wouldap-
poach ,ha inui tivenappa, dterm ingthecost wol
oinkagesbetweenproduction and consumptionactivities
linkges betweent
and the environment.
Generallyspeaking,the issueof the treatmentof defen-
siveexpendituresfor nationalaccountingassumesgreater
significancethe higher the degree of industrializationof
the countryconcerned.The issuesof depletionand deg-
radation, which are considered below, ar-enot directly
raain whc ar cosdee below, ar.o drcl
relatedto the levelof industrialization,however,although
they seem to be particularlyrelevant to countries that
base their economicactivitieson the exploitationof nat-
ural resources.
The Depletion and Degradation
of Natural Resources
There is an evident asymmetryin howthe SNA treats
man-madeassets and natural resources. Man-madeas-
sets-buildings and equipment,for example--arevalued
as productiveassets and are writtenoff againstthe value
of productionas they depreciate.Naturalresourceassets
are not so valuedor adequatelyaccountediforin most
instances,however,and their loss producesno charge in
the national accounts against current income to reflect
the decreasein potential future production.If a country
is exhaustingits renewableor nonrenewableresources,
its current incomewill thus be inflated by the sale of
natural assets that willeventuallydisappear.Differences
in recording under the SNA may arise depending on
whethera resourceis publiclyor privatelyowned.Private
companiestend to take a long viewof the natural assets
they own, and many make provisionfor the decreasein
the capital stock of natural resources, and in certain
countries tax legislationpermits such provisionsto be
excludedfrom taxableincome.No such exclusionis ef-
Environmental
andResource
Accounting:
An Overview 3
fectedin developingcountriesin whichnatural resources
are exploitedby the public sector.
Underlyingthis asymmetryis the implicit,as well as
inappropriate,assumptionthat natural resources are so
abundantthat theyare costlessor haveno marginalvalue.
Historicallythey havebeen regarded as free gifts of na-
ture-a bias that providesfalsesignals for policymakers.
This approachignoresthe depletionof valuableresources
and confusesthe sale of commerciallymarketablenatural
assets with the generationof income.Thus it promotes
and seemsto validatethe idea that rapid economicgrowth
be obtainedby exploitinga resource base that may
be rapidlydiminishing.The growthcan be illusory,and
the prosperityit engenderstransitory,if the apparentgain
i i
in income means permanent loss in wealth, that is, if at
least part of the receipts is not redirectedinto new pro-
Ite
ductive investments.As income is inflated, often con-
smto sas,adtecutycnendgt opa
sumption is also, and the country concerned gets compla-
cent about its economicperformance;as a result the
adjustmentin economicpolicygets delayedby the seem-
ing prosperity.In this regard, proper income accounting
is an aid to better decisionmaking,but, of course, it does
not guarantee that improveddecisionswill actuallybe
made.
Existingnatural capital of geological(nonrenewable)
and biological(renewable)resourcesas wellas of flow
resources(suchas water and air) is needed for industrial
and agroindustrialproduction.Newgeologicaldiscover-
aswa astreyl anduconservation, dosnovers
tes, as wellas recyclkgand conservaton, do not reverse
the depletion of known stocks. The newlydiscovered
stocks themselvescomefrom a finite stock of resources,
and they merelyextend the time span over whichdeple-
tion can continue. Depletion of renewable natural re-
sourcescan haveserious indirect effectsby reducingthe
sustainable flow needed for industrial inputs and eco-
systemservices.
Similarly,
crop productionat the expenseof soil erosion
cannot be sustained.Only carefulhusbandryof environ-
mental capacitiescan ensure sustainableand potentially
larger flowsof incomein the future.The optimisticargu-
ment that human ingenuityis bound to find substitutes
for the naturalresourcesbeingdepletedmaybe generally
valid,but it would be imprudentfor societyto base its
behavioron such optimismand wouldbe wrongfor econ-
omistsand accountantsnot to take rational precautions
in case this does not occur.
Two main conceptualapproachesto deal with the de-
pletion or degradationof natural resources have been
proposed:the depreciationapproachand the "user cost"
approach. The principle of depreciationof man-made
capitalcan be appliedstraightforwardlyto the consump-
tion of capitalembodiedin renewableand nonrenewable
resources (Daly,Chapter 2; Harrison,Chapter 4). Since
geologicaland ecologicalinformation on depletion or
degradationcomesin physicalunits, this must be priced
4 SalahEl Serafy and Ernst Lutz
or valuedin some way before some adjustmentcan be
madeto GDP to arrive at a correctednet product. Valua-
tion couldbe basedeitheron the principleof replacement
cost,wherereplacementis possible,or on (the discounted
valueof) willingnessto pay. Present conventionswould
valuethe depletedor degradedresourcesat currentprices
whereavailable.If such a correctionwere effected,the
grossproductwouldremainunchanged,but the net prod-
uct wouldbe adjustedto reflect the depreciationof envi-
ronmentalcapital that has occurred duringthe account-
ingperiod.
Because the depreciationapproach would leave GDP
unadjustedand becauseit would wipe out from the net
product the entire proceedsfrom natural resourcesales,
the user cost approach (see Chapter 3) has been pro-
posed as a wayof properlytaking accountof depletionof
mineralresources.Possessionof a natural resource con-
veyson its owneran incomeadvantagethat is deniedto
those withouta natural resource,and it is not satisfactory
to arrive at a measurementof zero net income, as pro-
duced by the depreciation method. The user cost ap-
proach avoidsthe difficultiesof putting a valueon the
stock of the resource, but it relies on the consciousas-
sessmentof current extractionrates in relation to the
total availablestock, measuredin physicalterms.Depend-
ing on the rate of depletionand on a discountrate, the
revenuefrom the sales of a depletableresource, net of
extractioncost, can be splitinto a capitalelement,or user
cost,and a valueaddedelement,representingtrue income.
T-hecapitalelementrepresentsasset erosion,and it has
been proposedthat it shouldbe hypothetically (El Serafy
1981) or actually(Ward 1982) reinvestedin alternative
assets so that it continuesto generate income after the
resourcehas been totallyexhausted.Unlikethe deprecia-
tion method,this wouldseek to alter the reckoningof GDP
itself,not just of NDP. This methodwouldbe in harmony
with accountingprinciplesand woulduse current market
pricesfor valuationpurposes,but it wouldrequirea rule-
of-thumbdiscountrate for convertingthe capitalsalesinto
an incomestream.It is alsorootedin a properunderstand-
ing ofthe economicmeaningsof "valueadded"and "rent,"
whichshouldnot be confusedwith assetsales.
How the net revenuecan be split into user costs and
true incomeis explainedin Chapter3. One needs firstto
decideon a discountrate, r, say 5 percent. Second,one
has to determinethe number of periods, n, over which
the resourceis being liquidated.This can simplybe read
from the ratio between total reserves and whatever
amountis extractedin the current period.Then the for-
muladevelopedby El Serafy(1981) is used to calculate
the ratio of true income,X, to net receipts (exclusiveof
extractioncosts),R:
1
X/R 1 -
XIR = I)n+1*for
(1 + r)~~1
R-X would be the "depletionfactor" (or user cost) that
should be set aside and allocatedto capital investment
and excludedfrom GDP, while X would represent true,
that is, sustainable,income.
This method is flexible enough to handle changing
levelsof extraction,movementsin the discount rate, and
alterationsin reserve estimates.Such alterationswould
includenew discoveries,whichwouldchangethe reserve-
to-extractionratio. In the aboveformulathis is denoted
by n, that is, the life expectancyof the reservemeasured
in years at the current period's extraction rate. The
methodis not concernedwith valuingtotal reserves,but
onlywith the fractionof the resourcebeing liquidatedin
the current accountingperiod, whichis valuedat current
prices.That fractionreliesentirelyon physicalquantities,
since the price is the same in the numeratorand denom-
inator.The method can be adapted to deal with mineral
extractionunder conditionsof deterioratingqualityof the
product and rising costs of extraction.Resourceowners
usuallyminethe richer depositsfirst, leavinginferiorde-
positsfor later periods,thus inevitablyraisingthe cost of
futureextraction.
This method, like all accounting methods, does not
indicatean ex ante optimalrate of depletion,but merely
mirrors decisionsalready made by the resource owner
about liquidatinghis natural resource.The ownerusually
determineshis extractionrate in the lightof manyfactors,
includinghis expectationof future price changes. If he
decidesto extract20 percent of his reservesin one year,
then n in the aboveformulationis equalto 5; the income
content of his net receipts, using a 5 percent discount
rate, wouldbe 25 percent; and the user cost to be rein-
vested is 75 percent. If, however,he extracts only 10
percent of his reserves,that is, plans to exhaust his re-
source over ten years, then he needs to set aside for
reinvestment58 percent of his net receipts,and willthus
enjoy 42 percent as current income. The correction
needed to reckon "true" incomeout of natural resource
sales is higher the nearer the resource is to exhaustion,
and lowerthe longerits life expectancyat current extrac-
tion rates. Onlya 1 percentreductionwouldbe necessary
in net receiptsto arrive at "true" incomeusing a 5 per-
cent discountrate ifthe resourceis beingliquidatedover
100 years. The choice of the discount rate materially
affectsthe calculation.A high discount rate, whichde-
pressesfuture againstcurrent valuation,raises the ratio
of "true" incomein current receipts. But alternativein-
vestmentsmust be found in whichto sink the depletion
factor (R-X) so that it can yieldthat much as a return.'
Resource Accounting
For resourceaccounting,data need to be collectedon
renewableand nonrenewablenatural resourcesprimarily
the purposeof planningtheir long-runexploitationin
pursuitof sustainableeconomicactivity.Severalindustrial

countries,includingCanada,France,Japan,Norway,and
the UnitedStates,havedevelopedresourceaccountsthat
are tailoredto their availableresourcesand pollicypriori-
ties. Some of the importantfeaturesof the F'renchap-
proachto resourceaccountingare describedbelow.
Insteadof the term "resourceaccounting"the French
use the expression"patrimonialaccounting,"whichcould
be describedas "accountingof the nationalenvironmen-
tal heritage" (Theys,Chapter 7). This is broader than
resource accounting because it includes, for example,
cultural heritage in addition to natural resources. The
French resource accountingapproach is jintendedulti-
mately to relate economicgrowth to the quantities of
natural resourcesthat haveto be used up or imnported to
make economicgrowth possible.Such a syst.emwould
also help to optimize the economicvalue of available
natural resources, determine the fraction of GDP that
should be set aside for the efficient protection of the
environment,and orient economicgrowthso that it does
not threaten ecosystems.
When fullydeveloped,the systemwould be versatile
and serve various ends. It would optimize the use of
natural resourcesas factors of production (for example,
inversionof a quantitativeinputloutputtable that would
indicatethe intermediateuse of natural resourcesin the
productiveprocess);it would describe the economicas-
pects of resourceuse (such as whichresourceswouldbe
marketedand in what quantitiesand values,how to im-
provethe productivityof processingindustries to opti-
mizethe use of natural resources,and what the opportu-
nity costsare of alternatives);it wouldtreat resourcesas
"environmentalgoods" (taking into account changesin
the qualityof the environment,the costs and benefitsof
environmentalpolicies,and the economicconsequences
of alternativeenvironmentalpolicies);and it wouldtake
stock of the national environmentalheritage and define
the long-termimplicationsof its transformations,so that
it could be preserved for future generatiDns.Since re-
sourcesto developsuch a comprehensive systemare nec-
essarilylimited,stress is placedon satisfyingthe needsof
the policymaker.Althoughit would be easier to collect
environmentaldata in the form of flexiblereports on the
state of the environmentand country profiles,the need
for developinga systemof environmentalaccountsis par-
amount,so that the informationis standardized,exhaus-
tive, summed up in physicaland monetary terms, and
comparablein time and space. The long-termgoal is to
match the standards already reached by national (eco-
nomic)accounting,whichmake the SNA such a powerful
planningtool forshort-termeconomicmanagement.
The French approachis only one amongsseveralbeing
pursuedin industrialcountries(for example,Norwayhas
been using resource accountsfor severalyears; see Alf-
sen, Bye, and Lorentsen1987).The French approachis
to build up balance sheets of resourcesand to monitor
their change from year to year, with emphasison meas-
Environmentaland ResourceAccounting:An Overview 5
urements in physicalterms. Physicalmeasurementsare
clearlyessential,and without them accountingin mone-
tary terms is impossible.Built into the French approach
is the assumptionthat a comprehensive physicalinventory
systemmust be in place beforeany changescan be pro-
posed in nationalaccountingmethodology.This is a point
of viewshared by many,but there are many others who
wouldwantto see nationalaccountingmethodsadjusted
graduallyas measurementsbecomeavailable.
Linking Environmental and Resource
Accounts to the SNA
The current versionof the SNA, whichhas been in effect
since 1968, does not contain an explicitenvironmental
dimension.The ongoing SNA revisionwas mandatedby
the U.N.StatisticalCommissionto simplifyand clarifythe
existingsystem rather than to propose radical changes.
The Commissiondesired to maintain consistency in
time series,evenif the series containedconceptualshort-
comings.
Amongenvironmentalists and economistswithenviron-
mentaland resource concerns,there are severalschools
of thought about the best approach to the accounting
problem (Norgaard,Chapter8). Some advocateenviron-
mental accountingin physicalterms and havelittle inter-
est in establishingany linkagewith the SNA. Their aim is
to use indicatorsof physicalchange to influencepublic
opinionand environmentalpolicies.At the other end of
the spectrumare those who feel that environmentalac-
countingwould not have an adequate effect unless the
accounts are monetized and integrated into the SNA,
whichwouldproducean adjustednationalincomethat is
more sustainable.
We believethat environmentalaccountingin physical
terms is essential,particularlyas this wouldcovercollect-
ing data that indicatethe directionand rate of changein
the quantityor qualityof a resource.At the same time,
werecognizethat "monetization,"to the extentpossible,
is importantas welland that a linkagewith the SNAand
an adjustmentof the current incomeconceptsand meas-
urementsare urgentlyneeded. Giventhe current state of
the art, however,we believethat more conceptual and
empiricalwork is necessarybefore GDP and NDP in the
core of the SNA can be replacedby more sustainableGDP
and NDP. That is whywe, as an interimstep, encourage
the constructionof satelliteaccounts,linkedwith the SNA,
in whichthe adjustmentswould be made (see Chapter
12). In other words,by havingsatelliteaccounts,the user
couldcomputesustainableGDP and NDP (SGDP and SNDP)
in them. This half-waysolution would not represent a
threat to the historicalcontinuityof GDP but has a fair
chance of being adopted. If it is adopted, national ac-
countantsmaytake the issues discussedhere more seri-
ouslyand mighteventuallybe willingto adjustthe core of
the SNA itself.
6 SalahEl Serafy and Ernst Lutz
Bartelmus(Chapter 11) discussesvariousoptions for
linkingor integratingenvironmentaland resource issues
with or into the SNA. His chapter was commissionedby
the workshoporganizers to summarizethe views and
approachesexpressedat the workshops.
Constructing Environmental and Resource
Accounts for Developing Countries
In order for resource concerns to be reflected ulti-
matelyin the SNA and in policymaking,it is necessaryto
makeprogressnowat an operationallevelso that govern-
mentofficials,nationalaccountants,and economistsalike
cantsee howf, natioincluenvironntalt and resucon-mistsali
can see how to include environmental and resource con-'
cerns in the calculations.But it is clear that the develop- argues that income measuresshould exclude all capital
cers i th cacultins.Butit
s cea tht te dvelp-
ment of environmentaland resource accountswill take
time.~~~~~~~~~~~~~~
time. This fact, however,shouldThsfc,hwvr.hudntke
ttsia
not keep statistical n
and
planningofficialsin developingcountriesfrom initiating
relevant work now, especiallyon minerals or forestry,
wheredata to a largeextent are alreadyavailable.In the
caseof Indonesia,Peskin (Chapter9) arguesthat a local
researcheffortshouldstart right away,supportedinitially
withperiodicconsultantinputs. Ideallynot only environ-
mentalhbut also other important nonmarket factors
metl bu als ote imoran nonsarket factorsndgidcooicpliy
should be consideredin an expandedaccountingstruc-
ture. This iea iselabortediChaptrNorgaard (Chapter8) is skepticalabout the economists
Repetto and others (1987) have applied resource ac-
Reetito feand fothers( resourcer ac-
counstinto feliand forvestrreourest
g ionesa. For
radationnet of regrowthand suggestedthat it be treated
rike depreciationof man-madeassets, that is they pro-
likeddereduciaong is,pthep o
oe maN-aDe assets,thmated
posed reducing the NDP by the estimated depletion.A
similarapproachwas followedforvaluingthe depletionof
fulmreservesachtwasee rservationsabout this aetnd an
fuel reserves (but see reservationsabout this and an
alternativeproposal in El Serafy Chapter 3). Another
alternat proposal
sive inE Sersa1y a ter3. Another
empirical stud
e Maona and Aoilren18 esiaTed thea
cost~~ ecnm ofsi rsonJv
to the. Th
. anua
amountestimatedwasUS$350millionto US$415million,
whichis slightlyless than 0.5 percent OfGDP.Morethan
95 percent of this cost is the on-site cost of declining
These two studies have madevaluable
productivity. . conA
productivity.Thnesetwo studies have madevaluablecon-
tributions,but it is lear that further empiricalwork is
nedd(e Catr1)
A Variety of Approaches,but a CommonTheme
A varietyof approachesfor amendingthe SNA are pro-
posed by variousauthors in this volume.This, however,
should not detract from the central theme argued by all
of the authors: in their present form the guidelinesfor
income calculationunder the SNA leave out important
aspectsof economicdevelopmentthat shouldbe brought
into the accounts.Theseguidelinesnowproducereadings
of levelsof activityand growthover time that can lead to
faultypolicyadvice.Such readingsfrequentlyexaggerate
income and thus encourageconsumptionand promote
habitsof behaviorthat cannot be sustainedoverthe longer
term.
An interestingargument over "desirableand practica-
ble" adjustmentis highlightedby the approachesregard-
ing depletableresourcesof Harrison(Chapter 4) and El
Serafy(Chapter3). Both are in fundamentalagreement
about what constitutessustainableincomeand what does
not. Harrison wouldwork withinthe existingframework
ofteNAbprevighedininofialemd
of the sNA by preservingthe definitionof final demand
capital as a parallel entry to consumptionof man-made
capital,withappropriateadjustmentsto NDP. Further,she
consumptionand thereforethat net productsshould be
usdto ind theleve of eoni ctivityuad it
used to indicate
development overthe
time.level of economic
El Serafy, activity
by contrast, and re-
would its
defelthendistinctioet intermeit and finlde-
mand th at theeo n atu and finalde-
bved basgeing sale of natural capitalmust not
of that incomeshouldbe excludedfrom GDP itself,as well
as from the net product. Thus the GDP measurement
couldcontinueto be used extensively,as is nowthe case,
to describeperformanceand guideeconomicpolicy.
and accountantsconcernedabout the environmentever
being able to agree on a set of corrections that would
simplyrectifyand fillin gaps in the existingSNA to pro-
duce one aggregatefigure expressedin monetaryterms.
He claimsthat the existingSNA contains contradictions
becauseit is basedon conventionsand reflects consensus
rather than being built on deductivereasoning.He does
not viewsustainabilityas implicitin the definitionof in-
come, which both Daly and El Serafydo, but as an
"ethical" goal, representing a "separate social objective
of development strategies." He believes the undervalua-
tion of the concerns of future generationsreflects the
nonparticipationby future generationsin the capitalmar-
kets of today.
more conventionalviewwouldascribesuch underval-
uto oteueo o ihadson ae hc
uces the ue of tur nbei almost to nhing
reduces the value of future net benefits almost to nothing
the more distant the future. Since futuregenerationswill
never be able to participate in today's capital markets,
the surest way of reflecting their preferencesis to use
lowerdiscount rates. Norgaard is raising questions re-
garding the economicapproach to the "sustainabilityof
development,"whichis based on accountingthat relies
on market valuations.He espouses "methodologicalplu-
ralism" in the belief that a multiplicityof perspectives
wouldensurethat "all valuesare respected to the extent
possible,"so that decisionmakershaveinformationalert-
ing them to "as many aspects of environmentaland re-
sourcephenomenaas possible."Norgaard,however,does

not spell out how such alternative value systerns can be
established or used; nor does he speculate on the sort of
solutions they would produce.
There are other areas in which the workshop partici-
pants have expressed different points of vievw,but the
main message this volume hopes to convey is ihe urgent
need to recognize the shortcomings of the cun-ent meas-
ures of income and to work toward a more sustainable
concept and measurements-a common thread in all the
contributions included here.
Note
1. J. M. Keynes(1936)first introducedthe conceptof "user
cost" in relation to capital equipment.He defined it as the
"maximumnet valuewhichmighthavebeen conserved. . . if it
tthe equipment]had not been used."He describedthis concept
as constituting"one of the links betweenthe present and the
future." (Seechap. 6 and its appendixof his GeneralTheory.)
Projectanalysisof depletablemineralshas also made use of the
concept of user cost at the micro level (see, fcr example,
Schramm1986).
References
Alfsen,Knut, TorsteinBye, and Lorents Lorentsen.1987.Nat-
ural ResourceAccountingand Analysis:The NorwegianEx-
perience,1978-86. Social and EconomicStudy 65. Oslo:
CentralBureauof Statisticsof Norway.
El Serafy,Salah. 1981. "AbsorptiveCapacity,the Demandfor
Environmentaland ResourceAccounting:An Overview 7
Revenue,and the Supply of Petroleum." The Journal of
Energyand Development7(1, Autumn):73-88.
Hicks,John R. 1946. Valueand CapitaL2nd ed. Oxford:Ox-
ford UniversityPress.
Keynes,John Maynard.1936. The GeneralTheoryof Employ-
ment,Interest,and Money.London:Macmillan.
Leipert,Christian.1986. "SocialCosts of EconomicGrowth."
Journalof EconomicIssues20(1):109-31.
. 1987. "DefensiveAusgaben in der Bundesrepublik
Deutschland,1970bis 1985:AbsoluteWerte und Relations-
zahlenmitdem BSP."InternationalInstitutefor Environment
and Society,Berlin.Processed.
Magrath,William,and Peter L. Arens.1987. "The Costsof Soil
Erosionon Java:A NaturalResourceAccountingApproach."
WorldResourcesInstitute,Washington,D.C.Processed.
Repetto, Robert, MichaelWells,ChristineBeer, and Fabrizio
Rossini.1987. "NaturalResourceAccountingfor Indonesia."
WorldResourcesInstitute,Washington,D.C.Processed.
Schramm,Gunter.1986. "PracticalApproachesfor Estimating
ResourceDepletionCosts."In EdwardMiles,Robert Pealy,
and Robert Stokes, eds.,NaturalResourcesEconomicsand
PolicyApplications:Essays in Honorof James A. Crutch-
field. Seattleand London:Universityof WashingtonPress.
UnitedNations,EconomicCommissionfor Europe. 1987. En-
vironmentStatisticsin Europe and North America:An Ex-
perimentalCompendium.NewYork.
Ward,Michael.1982. Accountingfor the Depletionof Natural
Resources in the NationalAccounts of DevelopingEcono-
mies.DevelopmentCentre PublicationCDIR(82)3010. Paris:
OECD.
 2

Towarda Measureof SustainableSocial

NetNationalProdudt
Herman E. Daly

SustainableIncome ment of NNP overestimatesthe maximum net product

availablefor consumption.Consequently,NNP increasingly
Thecentralcriterionfor definingthe conceptof income failsas a guide to prudent conductby nations.
has been wellstated by Sir John Hicks: Twoadjustmentsto NNP are needed to makeit a closer
approximationto Hicks' concept of income and a better
of incomcalctionsofinepracticl aairs
isthgivepurpose guide to prudent behavior.One adjustmentis simplyto
to giv pople antindicationeofsthegamountewhich extendthe principleof depreciationto coverconsumption
theyoang onu impoverishinth msele,
without, of natural capital stocks depleted through production.
defolling out idea, it woldsem tat weo h
mans The other is to subtractdefensiveexpenditures,or regret-
dfne acan'suduincome
as theek
mandstimuvaluecwhcea tableexpendituresnecessaryto defendourselvesfrom the
can consumeduring a week,and stll expectto beg unwantedside effects of our aggregateproduction and
well.off ahen o te wes he wasat te begin- consumption.Regrettabledefensiveexpendituresare in
ning.~~~~~~~~~~~
esnsvs Thswe. epast ebte the nature of intermediategoods, costs of production
off in the future;when he livesbeyondhis income he
plans to be worse off. Rememberingthat the practical rather than final product availablefor consumption.To
purposeof income is to serve as a guide for prudent correct for havingcounteddefensiveexpendituresin NNP,
pupse of inom is to srv as..... a gud for prudent theirsizemustbe estimatedand subtractedto arriveat
conduct, I think it is fairlyclear that this is what the
meaningmust be. (1946, p' 172).
central an estimateof maximumsustainableconsumption,or true
centralmeaningmust be. (1946, P. 172). income.
Thesame basic ideaof incomeholdsat the nationallevel. To summarize,let us define the correctedincomecon-
Incomeis not a precise theoreticalconceptbut rather a cept, "sustainablesocialnet nationalproduct"(SSNNP), as
practicalguideto the maximumamountthat can be con- net nationalproduct (NNP) minusboth defensiveexpendi-
sumedby a nation withouteventualimpoverishment.WVe tures (DE) and depreciationof naturalcapital(DNc). Thus,
knowthat we cannot consumethe entire gross national
product (GNP) without eventually impoverishingourselves,
so wesubtractdepreciationand get net nationalproduct This definitionentails no interference whatsoeverwith
(NNP),whichis usuallytaken as incomein Hick's sense. the current structure of the U.N.Systemof NationalAc-
The centraldefiningcharacteristicof incomeis sustaina- counts(SNA). There is no loss of historicalcontinuityor
bility.The term "sustainableincome"ought thereforeto comparability.Two additionalaccounts are introduced,
be considereda redundancy.The fact that it is not is a not for frivolousor trendy reasons,but simplyto gain a
measure of how far we have strayed from the central closer approximationof the central and well-established
meaningof income and, consequently,of the need for meaningof income.No attemptis madeto deal with the
correction. controversialissues of national income accounting,such
But could we reallyconsumeevenNNP year after year as inclusionof leisure,disutilityof labor, householdpro-
without impoverishingourselves?No, we could not, be- duction, and services of long-livedconsumer durables.
causethe productionof NNPrequiressupportingactivities The relationof incometo welfareis not addressed.
that are not biophysicallysustainable,and the measure- Since NNP is a familiarconcept, it remains only to

8

discussbrieflythe new accounts:defensiveexpenditures
(DE) and depletionof natural capital (DNC),whichare by
no meansnovelideas,but are not yet includedin as part
of an extendedSNA.
Defensive Expenditures
The explosionof the populationsof humanbodies, of
artifactsof all kinds, and of the populationsof plants and
animalsexploitedfor human use that has happened in
the past fiftyyearsmightbetter be called an "implosion,"
since it has occurred in a finite environment.The term
implosionsuggestsa compressingtogetherrather than an
expandingapart, a processof congestion,mutualinterfer-
ence, and self-canceling collision.Defensiveexpenditures
reflect this increasinglyprevalentphenomenonoi mutu-
allyinterfering,self-canceling activities.
The categoryof defensiveexpenditurescan be largeor
small dependingon where the boundaries are drawn.
ChristianLeipert of the InternationalInstitutefor Envi-
ronment and Societyin Berlinhas suggestedfivebroad
categoriesof defensiveexpenditures.
e Defensiveexpendituresinducedby the overexploita-
tion of environmentalresourcesin the general courseof
economicgrowth,such as the costs of all environmental
protection activitiesand expendituresfor environmental
damagecompensation.
* Defensiveexpendituresinducedby spatialcoricentra-
tion, centralizationof production,and associatedurbani-
zation,such as increasedcommutingcosts,housing,and
recreationcosts.
* Defensiveexpendituresinducedby the increasedrisks
generatedby the maturationof the industrialsystem,such
as increasedexpendituresfor protection against crime,
accident,sabotage,and technicalfailure.
* Defensiveexpendituresinducedby the negativeside
effects of car transport, such as traffic accidentswith
associatedrepair and medicalexpenses.
* Defensiveexpendituresarisingfrom unhealthycon-
sumptionand behavioralpatterns and frompoor working
and livingconditions,such as costs generatedby drug
addiction,smoking(bothactiveand passive),and alcohol.
These categoriesare neither exhaustivenor mutuallyex-
clusiveand are naturallysomewhatarbitrary.Category1
might, in our classification,fit better under DNC.But it
represents expendkeeping
a start at subtracting
representsa start at subtractingexpendituresthlaido not
reflectany increasein the net product availablefor con-
sumptionwithoutsumptionwithouteventualimpovershment.national
eventualimpoverishment.
Depletion of Natural Capital
This is entirelyanalogousto the depreciationof man-
madecapital.In fact, Keynesjustifiedthe conzeptof user
Towarda Measureof SustainableSocialNet NationalProduct 9
costfor man-madecapitalby analogywith the more self-
evidentcase for charginguser cost for natural resources.
The obviouscategoriesof natural capital are geological
(nonrenewable)and biological (renewable). Both, of
course, are depletable.Depletionof geologicalcapitalis
necessaryfor industrial and agroindustrialproduction.
Newgeologicaldiscoveriesdo not reversethe processof
depletion,but they do extend the time span over which
the depletioncan continue.
Depletionof renewablenatural capitalis in some ways
a more serious matter becausereduced stocks or popu-
lationsof plants and animalswilllead to a reduction in
sustainableflow of resource inputs and ecosystemserv-
ices. Onlyby a future investment(reductionin consump-
tion) could the larger sustainableflow be reestablished,
and eventhat is often not possible.Evenin commercially
exploitedpopulationsthat are above the level of maxi-
mum sustainableyield and therefore would yield more
with a smallerstock,the consumptionof the stock dimi-
nution is not a sustainablesource of income,but rather
capitalconsumption.
Geologicaland ecologicalinformation on depletion
comesin physicalunits and must be priced or evaluated
in some waybeforeit can be subtractedfrom NNP.This
willno doubt involvesome arbitraryconventions.Valua-
tion might be based either on the principleof replace-
ment cost or on willingnessto pay, whicheveris less, in
order to be "conservative."It is not clearthat any greater
arbitrarinesswould be involvedthan already exists in
current estimatesof depreciationof man-madecapital,
especiallyif one counts obsolescenceor "moral" depreci-
ation as wellas physical.
Income, as Hicks emphasized,is not a theoretically
precise concept,but rather a practicalguide to prudent
behavior.Surelysomereasonableallowance,howeverim-
precise,for depletionof natural capitaland some correc-
tion for the double countingof defensive(intermediate)
expendituresare required if the income concept is to
remain a guide to prudent behaviorby nations, whichis
its fundamentalreasonto be. The two adjustmentsare in
keepingwith the central idea of income and involveno
disruptionof the existingSNA.The need for such adjust-
mentshas alreadyled independentscholarsto begin work
on measuringdefensiveexpendituresand the depletion
ofntrlcpa.Whtemisstogvofcalttu
andnforal cogi. .tio te to j e aimedat
the incomeconceptas a reliableguideto prudent
bhvo nawrdta a hne infcnl ic
naion in aounting was firsinitionalize
income accounting was first institutionalized.
Reference
Hicks,JohnR. 1946.Valueand Capital.2nd ed. Oxford:Ox-
fordUniversity
Press.
 The ProperCalculationof Income
from DepletableNaturalResources
Salah El Serafy
Recognitionis growingthat income is not being accu-
rately calculated for economies based on natural re-
sources.Somewouldevensay that, for these economies,
nationalaccountingmethodsproducemisleadingcalcula-
tions. They lead to measurementsthat neither faithfully
describeeconomicperformanceex post, nor can theybe
used as a basis for useful policy proposals.For such
economies,current accountingpractices exaggeratein-
come, encourage unsustainablelevels of consumption,
and obscure the necessityto implementgreatly needed
policyadjustment.The problemis relevantto practically
all countries where nonrenewableresources are being
exploitedand where renewableresources are being run
downwithoutbeing restored. But it is most acute where
such resourcesare being exploitedin the public sector,
eitherdirectlyor throughforeigninterests.
In the moreindustrializedcountries,whereexploitation
typicallyoccurs in the private sector, tax allowancesfor
depletiontend to correct the calculationof the "value
added"believedto be generatedby such activities.Such
correctionof courseis effectedas depreciation,reducing
the gross product by an elementto coverdepletion.The
correctionis frequentlynot exact, but it is a step in the
right direction.This process is being helped by the fact
that whenproperties containing marketablenatural re-
sources,such as subsoil deposits,exchangehands, their
marketvaluetends to reflect their natural-resourcecon-
tent. Bycontrast, no such correctionis made in most of
the developingcountries, whose economiesdepend in
varyingdegrees on the exploitation of their natural re-
sources, such as mineral extraction or the commercial
loggingof forests to maketimber and paper. This prob-
lem, therefore, is one of paramountimportancefor the
developingworld.
The practice in these countries followsthe United
NationsSystemof NationalAccounts(SNA),whichtreats
10
revenue derived from the sale of natural resources as
currentincome,or rent, that is availablefor consumption.
If the revenueaccruesto the publicsector, it can be used
just like revenue from any other source, such as the
proceedsfrom income taxes. Giventheir short perspec-
tive,the politiciansin charge of such economiesoften do
not want to be remindedthat the revenuederivedfrom
liquidatingtheir country'snatural assets is neither recur-
rent nor sustainable.And many a developingcountry
rejoicesin havingits leadershippraisedfor illusoryrapid
economicgrowth, apparentlyhigh rates of saving and
investment,and deceptivelystable or near-stable price
levelsbroughtabout by import surpluses.
Such apparentprosperityis bought at the costof asset
erosion-a sure recipefor future economicdecline.Thus
natural resourcesare exportedand used up to prop up a
truly unbalanced, but seeminglycomfortable, external
balance.An overvaluedexchangerate inevitablydevelops,
and relativeprices are upset as a "Dutch Disease"syn-
drome sets in, wherebythe prices of nontradablegoods
and servicesrise in relationto those of tradables.' Con-
sequently,the economy'scapacityis reduced to produce
and exportthe productsof nonnatural-resource-based ac-
tivitiesthat could providebadlyneeded employmentand
(sustainable)future income.Any comparativeadvantage
the countrymay havegets sacrificedduring a period of
ephemeralprosperityand illusorygrowth.This is partic-
ularlytrue wherethe exhaustionof the resourceis immi-
nent. Needlessto say, the citizensof these countriesfind
it onlytoo easyto adjustto a higherlevelof consumption.
When the bonanza ends and the natural resource is
almost exhausted,standards of living have to fall, and
intolerablepressures develop on the external balance.
Quite often the country finds itself saddled with a high
externaldebt,whichit contractedin the prosperousyears
when it had overestimatedits capacityto borrowand its
 The ProperCalculationof IncomefromDepletableNaturalResources 11
creditors had mistakenlyassumed that the prosperity
would continue.The governmentthen finds itseliFin an
impossiblesituationin whichthere are no marginsleft to
providea cushionfor urgentlyneeded policy adjustment
that should havebeen initiatedyearsbefore.The halcyon
period of plenty willhave come to an end, and all the
putativeeconomizingthat had been done during those
years is seen in retrospectto havebeen false and futile.
Defectiveaccountinghad led economicbehaviorand pol-
icy analysisastray.
The fundamentalprinciplethat is flouted by applying
conventionalnational income accountingto depletable
resourcesis the separationthat must be maintainedbe-
tween incomeand capital.This principletells us that if
you liquidateyour assets and use the proceedsfor con-
sumption,you are livingbeyondyourmeans,and in doing
so you are underminingyour abilityto create future in-
come. The accountingprofessionwas born, in the late
MiddleAgesin the city states of the Mediterraneanbasin,
largelyto separate from the proceeds accruingto mer-
chants that part which they could use to finance their
families'current needs. Those merchantshad to guard
againstconsumingtheir capital,whichwas the sourceof
their continuedwell-being.From its infancythe account-
ingprofessionspecifically has addressedthis task. In pres-
ent-daylanguagethe accountantswere asked to define
sustainablelevelsof consumption,and they could do so
onlyby attemptingto keep capitalintact.
The same principlewastaken up by AdamSmith,who
sawcapitalas a means"to increasethe productivepowers
of labor" and as an asset whosemaintenancewas imper-
ative, since it "is alwaysrepaid with great profit, and
increasesthe annual produce by a much greater value
than that of the support whichsuch improvementsre-
quire."The Wealthof Nations(Smith1776) states that:
The gross revenue of all the inhabitants of a great
country, comprehendsthe whole annual produce of
country,ncomprehends the whol annuea produ o
thi
lan an laor th nea reeu,wa .ean
free to them after deductingthe expense of maintain-
ing; first, their fixed;and, second,their circulatingcap-
ital; orencroaching
what, without
un
ital; or what, without encroaching upon their ccapital,
.r ,
they can place in their stock reserved for immediate
consumption,
or spn.pntersbsseieovn
icnsum, amuspmend
ord uoTheir subsiste oo. n
iences, and amusemn. their
r eat tooeisei
proportion,not to theirgross,but to their neat revenue.
The SNA, in failingto distinguishbetweenunsustainable
receipts, derived from the sale of natural assets, and
sustainableincome,producedby the factors of produc-
tion, disregards the fundamental Smithian concept of
"neat revenue,"whichshouldguidethe consumptionand
assessmentof the wealthof the revenuerecipienlt.
The distinction between capital and income has re-
mainedcrucial throughoutthe developmentof econom-
ics. In the present day, Hicks paraphrasedthis principle
into a definitionof incomeas that amountwhicha person
can consumeduringa givenperiod and still be as welloff
at the end of the period as at the beginning. 2 More
specifically,
weare told in no uncertainterms that:
if a person's receiptsare derivedfrom
tation of a wastingasset, liableto giveout the
. .. exploi-
at a future
date, we shallsay that his receiptsare in excessof his
income"(Hicks1946,p. 187).
Natural resources are certainly "wastingassets" if they
are nonrenewable(for example,minerals),or, if they are
renewable(for example,forestsexploitedfor timber,fish-
eries, or agricultural soil), are not actually renewed
through careful maintenance,thus causing the receipts
fromtheir exploitationto giveout in the future.Ignoring
this elementaryfact makes a mockeryof what has been
passingas economicanalysisand policyprescriptionfor
economiesbasedon natural resources(and in particular,
those based on minerals),in which no effort has been
madeto compensatefor drainingthe nationalwealth by
depletingthese resources.Maintainingcapital intact is
not a marginalissue.It is centralto all economicbehavior
and analysis,and it is a poor economistindeed who is
unableto tell capitalfrom income.
Conceptual Background
The confusionof capital and income,whichhas been
the standardapproachto incomecalculationin this area
and which derives support from the SNA, is becoming
increasinglyuntenable.This chapter offersa wayto esti-
mate the true incomecontent of the proceedsfrom min-
eral sales. The treatment of incomefrom renewablere-
sources such as forests, which have to be maintained
through replanting, or fisheries, which have to be re-
stocked,is more straightforward.Where such replanting
or restocking is effected at technologicallyacceptable
rates that wouldkeep capitalintact,these activitiescould
be charged against the gross returns from the natural
resourceto obtain the net valueadded generated;this is
similar to the way capital consumption is treated in na-
ioaacunsSilesonloblngtoheam
conal accounts. Sofl erosion also belongs to the same
category of a natural resource whose depletion can be
offsetby restoration,and the cost of restoration should
be chargedagainstthe gross product of the soilto obtain
a true estimateof the net product.
But quite often, particularlyin poorer countries, the
resourceis not restored to the same levelof activity.As a
result the valueadded that appears to be generatedcon-
tains capital elements that should be removed.In this
casethose whoestimatenationalincomeshouldimpute a
capitalconsumptioncharge based on technicallyaccept-
able criteria against current receipts to obtain the true
incomefromthese activities.For soil erosion,some esti-
mate maybe necessaryof the decliningpowerof the soil
12 SalahEl Serafy
to produce, and this can be based, for example,on de-
cliningland yieldsover time. This chapter, however,ad-
dressesonlythe problemof estimatingincomegenerated
from depletable,nonrenewableresources.
My thinking on this topic began to be shaped by a
senseof discomfortoverwhat I thought to be an inappro-
priate use of economicconcepts when the pricing of
petroleumbeganto attract the attention of economistsin
the early 1970s. To my mind the oil market had long
been an oligopsonisticmarket, dominated by powerful
multinationalconglomerates.Economicanalysishad con-
tributed little to understanding how prices were deter-
minedin that sort of market,beyondthe traditionalmod-
els of oligopoly theory, which concentrated on how
equilibriumwasreachedrather than on the levelof prices
producedby it. Later,when oligopsonyin the petroleum
marketgavewayto an apparentmonopolyallegedlyinsti-
tuted by the Organizationof PetroleumExportingCoun-
tries (OPEC), the price increaseswere too facilelyattrib-
uted to the powersof the exporters'cartel.
It was curious that many analysts overestimatedthe
competitionprevailingin that market before 1973 and
underestimatedit afterward,emphasizingOPEC'S monop-
olypower.It wasevenmore curiousthat manyanalystsin
the 1970sappearedto think that iffree competitionwere
to prevail,competitiveequilibriumwouldindicatea price
equalto the marginalcost of extraction,whichwas-and
still often is-referred to as production,and that it was
only becauseof the allegedcartelizationof supply that
the price wasable to rise abovethat cost.3
This constructionwas later challengedby those who
were aware that the price of an irreplaceablenatural
resource,such as petroleum,should perhaps contain a
user cost or capital element,representingthe erosionof
the resource.Even under free competition,the marginal
cost of extractioncould not possiblyindicatean equilib-
rium price level, since the cost of extraction is tanta-
mountto the cost of asset liquidationand cannot deter-
minethe valueof the veryasset being sold.Hotellinghad
to be resurrected and used with great dexterityby an
importanteconomistlike Solowfor a more convincing
explanationof petroleumprice increasesbeforethe eco-
nomic professioncould be persuaded. 4 But it has not
been completelypersuaded,and doubtersstill abound.
Parallelto the microeconomicconfusionaboutthe pric-
ing of natural resources, other inaccuracieshave also
beenperpetratedand havedistorted thinkingabout mac-
roeconomicsin countries where the exploitation of de-
pletablenatural resources is significant.If the marginal
cost of extractionwas the only cost, then any surplus
accruingto the sellers was pure rent and represented
valueadded to be includedin the gross domesticproduct
(GDP). This certainlyis impliedby the accountingprac-
tices currentlybeingused underthe SNA. Basedon these
practices,the expansionof economicactivityas a conse-
quenceof acceleratingthe liquidationof subsoilassets is
applauded as good economicperformance and is con-
fused with the growth that comes from labor, capital
formation,technologicalprogress,and efficientorganiza-
tion. The revenueaccruingto countriesthat depletetheir
natural resources in this way is reflected in increased
savingrates and investmentcoefficientsand in improved
parameters, such as incremental capital-outputratios
(ICORS),which shed deceptivelyfavorablelight on the
economicperformanceof such economies.Policyadvice
basedon these calculationsbecomesdulled at best-and
downrightwrongat worst.
The conceptof rent in this situationis profoundlymis-
used and totally misapplied.In the perception of the
classicaleconomiststhe rent that qualifiedas valueadded
derivedfrom the indestructiblepowersof nature.5 Such
revenueis clearlysustainablewherethe powersof nature
to reproduce it are not impaired,and it can therefore
legitimatelybe counted as income. The surplus, net of
extractioncosts, emanatingfrom liquidatingnatural re-
sources, however,has little kinship with either rent or
quasi-rentas definedby Marshall(1920).
There seemsto be no alternativeto bringingthe capital
nature of such exploitationinto the open and integrating
this in all economicthinkingand measurements,not just
to gauge welfareadequately,but to savethe disciplineof
economicsfrom disrepute.Even noneconomistshave on
occasionrightly perceivedthat mineral extraction reve-
nues are not whollycurrent income.A smalland under-
developedcountry such as Libyacould thus legislateas
earlyas 1963 (whenit firstbegan to extractpetroleumin
commercialquantities)that at least 70 percent of petro-
leum proceedshad to be allocatedto development.The
perceptionwasstrong in that verypoor countrythat this
uniquewealth truly belongedto future generationsand
should not be squanderedon consumption,as would be
impliedby treating its sales as current income.To recall
Hicks'sstandarddefinition,currentincomeis that part of
receiptswhich, if devotedto consumption,would leave
the earner no worse off at the end of the accounting
period than at its beginning.
Weakness of the Depreciation Approach
Likeother economistsof the same bent, I thought first
of usingthe "capitalconsumption"or "depreciation"ap-
proach to treat income from depletable mineral re-
sources.As the resource is depleted by the quantity of
extractionduringthe year,the amountof depletion,val-
ued at current prices, can be deducted from the gross
proceeds,just as, for example,the depreciationof capital
equipmentused for manufacturingis subtractedfrom the
gross valueadded by manufacturingactivities.The min-
eral extraction earnings can still be reckoned in GDP,
providedthat the valueofthe depletionis deductedfrom
 The ProperCalculationof IncomefromDepletableNaturalResources 13
it for calculatingnet income.The problem of the exact
valuationof capital consumptionin this case appearsto
be of secondaryimportance.Much more impcrtant is to
try to make some adjustment.Variousmethods are al-
ready used to treat inventoriesand other capital assets
used up in the processof production.Shortcuts,approx-
imations,and arbitraryestimationsare used throughout
national income calculations,and no special harm can
come from adding the depreciationof natural.resources
to the list.
On reflection, however,I movedaway from this ap-
proach,both for practicalas wellas conceptualconsider-
ations. First the conceptual.It is wrongto describe as
current productionthat whichis not current production.
GDP is an importantmeasurementand is much more in
use than NDP (net domesticproduct).Even if NDP and its
national parallelNNP are correctlymeasured,the whole
apparatusof GDP withits structure,input-outputrelations,
and changes over time would remain incorrectly calcu-
lated if revenuesfrom depletableresourcesare counted
as valueadded in GDP.
It is not by chance that the gross product, rather than
the net product, is the preferredquantityfor macroeco-
nomic analysis.It is often used as a denorninatorfor
crucialmacroeconomicratios, with the nominatorbeing
moneysupply,exports,imports,externaldebt, debt serv-
ice, savings,capitalformation,and so forth.As Hickshas
suggested,the conceptof net incomeis usuallyeschewed
because it is alwaysarbitrary. It relies on estimatesof
depreciationand inventoryuse that are a miixedbag of
hlistoricalcosts and estimationbased on accountingcon-
ventions, tax lawsand allowances,and insurance com-
ven ration taxlaws
s, andallowasjcves, vansuratonce bcom-
pany practices, as well as subjectivevaluationby eco-
nomic agents who do the reckoning and wvhohave a
variety of expectationsabout the future (Hlicks1981,
chap. 9). If an incomecorrectionis to be made,it should
apply thereforeto the gross product itself,and it is not
enough to effectthe adjustmentat the net product level.
Another reason whyI discarded the depreciation ap-
proach to rectifyingincome accounting for depletable
resource activitiesis the fact that countrieswith market-
able natural resourcesare evidentlybetter ofi than those
withoutsuch resources,and they can enjoya higherand
sustainablestandard of livingthan the latter by virtueof
their resourceendowment.Such an advantageshouldbe
reflectedin calculatingthe incomeof both groups.If we
deduct from the gross receiptsfrom mineralsales in any
oneyear an amountequalto the depletionalongthe lines
describedabove,the valueof net income fromthis activity
becomes zero. Where a countryderives 100 percent of
its receipts from,say, petroleumextraction--anextreme
case of a Saudi Arabia-the depreciationapproach (ig-
noring the multipliereffectof ancillaryactivitiesrelated
to extractionas wellas the contributionof other sectors
to valueadded)wouldgiveus a GDP of 100 and a NDP of
zero-a measurementthat is not particularlyedifying.For
the gross product this approach would not make any
adjustmentand would simplyeliminatethe net product
altogether.Such a measurementof net income would
belie the observablefact that havingsubsoilmineral de-
posits to exploitgives their possessorsan income edge
over thosewho do not havethat advantage.
Conversion to a Permanent Income Stream
Mineraldepositsand other comparablemarketablenat-
ural resourcesare assets. Salesof assets do not generate
valueadded and should not be includedin GoP. They do
vate addedand howeve ich inbe Tto
generateliquidfunds,however,whichcan be put to alter-
nativeuses.A countrymaychooseto spend the proceeds
(net of extractioncosts) on consumptionor investment
or any combinationof the two. But this is neither here
nor there. Froman accountingpoint of view,however,an
incomecontent of the net receiptscan be estimated.This
incomecontent shouldbe part of GDP since it represents
valueadded. The argumentfor this proceedsas follows. 6
If an owner of a wastingasset is to consumeno more
than his income,he must relendsome part of his receipts
so that the interest on it will make up for the eventual
failureof receipts from the wasting asset in the future.
This proposition,which can be found in Hicks (1946
chapter 14), suggestedthe need to convertthe mineral
asset concerned into a perpetual income stream. The
finiteseries of earningsfrom the sale of the resource,say
a ten-year series of annual extractions leading to the
extinctionof the resource, has to be converted to an
infiniteseries of true income, such that the capitalized
valuesof the two series are equal. From the annual earn-
insfosae,nicmepronttcnbepntn
ings from sales, an incomeportion that can be spent on
consumptionshould be identified;the remainder,a capi-
tal element, should be set aside year after year and in-
vestedto create a perpetualstream of incomethat would
providethe same level of true income,both during the
life of the resourceas wellas after the resourcehas been
exhausted.The two constituent portions of current re-
ceipts need to be defined:the income portion and the
capital portion. Under certain a,ssumptions,which are
neither too restricting nor too unrealistic,the ratio of
true incometo total receiptsis:
1
XIR = 1 - ____r_n_
)
where X is true income, R the total receipts (net of
extractioncost), r the rate of discount,and n the number
of periodsduring whichthe resourceis to be liquidated.
R-X would be the user cost or depletion factor that
should be set aside as a capital investmentand totally
excludedfrom GDP. On the expenditureside, this deple-
tion factor would represent a disinvestmentthat should
14 SalahEl Serafy

be set against capital formationin new assets, so that Figure3-1. Income Contentof MineralSalesat Various
total expenditurewouldstill be equalto the true income. Life Expectanciesand DiscountRates
If all of the receiptswere devotedto consumptionand if
new capitalformationfell short of the depletionfactor,
the accounts should show a negative value for capital 1.0 r = 10percent
formation,thus reflectingthe disinvestmentthat had oc-
curredin the accountingperiod.
The ratioXIR dependsonly on two values:the reserves-
to-extractionratio, that is, the life expectancyof the re-
sourcemeasuredin years,and the discountrate. A country 3 percent
that liquidatesits mineralreservesoverfiftyyearsneedsto 0.8/
set asidefor reinvestmenta smallerportionof its receipts r = 2 percent
than anotherthat liquidatesits reservesovertwentyyears,
and thus it can count a larger portion of its receiptsas r4I
income.Similarly, if the receiptsset aside can be invested
at a higherinterestrate, say 10 percent,a higherportion II
wouldbe reckonedto incomethan if the interest was 5
percent.Accordingto this formula,witha discountrate of ' 1
5 percent, a countrythat liquidatesits natural resource r = 1 percent
overten years can consideras incomeonly42 percentof E
its annual receipts,whileanother witha fifty-yearhorizon
can reckonas muchas 92 percentof its annualreceiptsto 0.4
currentincome.At a 10 percentdiscountrate the former's I
current incomewouldbe 65 percent of the receiptsand 1l
the latter's 99 percent, whichwould require almost no /
correctionto GDP estimatesas currentlymade.
Figure3-1 shows the ratio XIR (the portion of total
receipts that is true income) as a function of the life 0.2
expectancyof the resource,N, measuredin years, at ten
alternativediscountrates, decreasingfrom 10 to 1 per-
cent. The same relation is given in Tables3-1 and 3-2.
Table3-1 showsthe income content of mineral sales at
elevenalternative discount rates from 0 to 10, for re- 0...,.l.,.l.,.l.,.Il.,.l...,i.l.,.l.,.l
sourcelife expectanciesof 1 to 100 years.This is shown 0 40 80 120 160 200
as percentageshares of the receipts that are currently Ratio of reserves to extraction (A
being treated under the SNA as if they were whollyin- (Lifeexpectancyof resourcein years)
come. Table 3-2 is the complementof Table 3-1 and Note:r = alternative discount rates.
presents the user cost content of the annual sales, ex-
pressed also as a percentage of total receipts, for the
same discountrates and life expectancies.This percent- the utilityof the resourceto future generations.Current
age representsthe capital elementthat, I believe,should practicesfor calculatingGDP accordingto the injunctions
be excludedfromGDP as a depletionfactor. of the SNA are thus seen to be built on one of two
The calculations show that the present practice of untenablepremisesor a combinationof both: that the
counting mineral sales proceeds as current income im- natural resource being liquidatedwould last foreverand
pliesthat the fraction that the welfareof futuregenerationsdoes not matter.
1
(1 + r)n+l Clarification of the User Cost Approach
equalszero in the previousformula.For onlythen would In defenseof this approach,the followingpointsshould
XIR = 1. This would be obtained, irrespectiveof the be clarifiedand emphasized.
discountrate, by havingn = oo;or alternatively, wheren
is finite,by havinga veryhigh valueof the discountrate * A discountrate must be chosen. This decisionhas
so that r tends to infinity.Such a high rate of discount to be arbitrary,but the arbitrarinessof the discountrate
implies a very strong time preference of the resource is not in principle any differentfrom the arbitraryesti-
ownersand is tantamountto setting a very lowvalueon mation methods used extensivelyunder the SNA. A rate
 The Proper Calculation of Income from Depletable Natural Resources 15

Table 3-1. Income Content of Mineral Sales (XIR)

(percent)
Life expectancy of the - Discount rate (r)
resource(A)(years) 0 i' 2 3 4 5 6 7 8 9 10
1 0 2 4 6 8 9 11 13 14 16 17
2 0 3 6 8 11 14 16 18 21 23 25
3 0 4 8 11 15 18 21 24 27 29 32
4 0 5 9 14 18 22 25 29 32 35 38
5 0 6 11 16 21 25 30 33 37 40 44
6 0 7 13 19 24 27 33 38 42 45 49
7 0 8 15 21 27 32 37 42 46 50 53
8 0 9 16 23 30 36 41 46 50 54 58
9 0 10 18 26 32 39 44 49 54 58 61
10 0 10 20 28 35 42 47 52 57 61 65
15 0 15 27 38 47 54 58 66 71 75 78
20 0 19 34 46 56 64 71 76 80 84 86
25 0 22 40 54 64 72 78 83 86 89 92
30 0 27 46 60 70 78 84 88 91 93 95
35 0 30 51 66 75 83 88 91 94 96 97
40 0 34 56 70 80 86 91 94 96 97 98
50 0 40 64 78 86 92 95 97 98 99 99
60 0 46 70 84 91 95 97 99 99 99 100
80 0 55 80 91 96 98 99 100 100 100 100
100 0 63 86 95 98 99 100 100 100 100 100
Note: Figuresare roundedto the nearest digit.

Table3.2. CapitalContent(or "UserC'ost")of MineralSales (1 - X/R)

(percent)
Life expectancy of the _ Discount rate (r)
resource(IV(years) 0 1 2 3 4 5 6 7 8 9 10
1 100 98 96 94 92 91 89 87 86 84 83
2 100 97 94 92 89 86 84 82 79 77 75
3 100 96 92 89 85 82 79 76 73 71 68
4 100 95 91 86 82 78 75 71 68 65 62
5 100 94 89 84 79 75 70 67 63 60 56
6 100 93 87 81 76 71 67 62 58 55 51
7 100 92 85 79 73 68 63 58 54 50 47
8 100 91 84 77 70 64 59 54 50 46 42
9 100 90 82 74 68 61 56 51 46 42 39
10 100 90 80 72 65 58 53 48 43 39 35
15 100 85 73 62 53 46 42 34 29 25 22
20 100 81 66 54 44 36 29 24 20 16 14
25 100 78 60 46 36 28 22 17 14 11 8
30 100 '73 54 40 30 22 16 12 9 7 5
35 100 '70 49 34 25 17 12 9 6 4 3
40 100 66 43 30 20 14 9 6 4 3 2
50 100 630 36 28 14 8 5 3 2 1 1
60 100 54 30 16 9 5 3 2 1 1 0
80 100 45 20 9 4 2 1 0 0 0 0
100 100 :37 14 5 2 1 0 0 0 0 0
Note: Figuresare roundedto the nearest digit.
16 SalahEl Serafy
of 5 percent or thereaboutscan be chosenas approximat-
ing what the classicaleconomistsused to call a natural
rate of time preference.This could be changed periodi-
cally,say everyfiveyears,guidedby changesin the long-
term marketrates.
* Underthe proposedformula,the settingasideof part
of the proceedsfor reinvestmentis only a metaphor.The
ownermay dispose of his receipts any wayhe chooses.
But he should be made aware of the fact that his true
income is only a fraction of his total receipts. Proper
accountingshould conveythis fundamentalmessage.
* Equallymetaphoricalis the process of calculating
the yie!dsfrominvestingthe set-asidepart of the proceeds
at the choseninterest rate. As stated previously,the rate
should approximatean availablemarket parameter that
wouldindicateprudent behaviorfor the asset liquidator
and wouldguidehis decisionsabout extraction.Thus he
maydecideto delayextractionif the market interestrate,
availablefor financialinvestment,appears lowerthan the
rate at whichhis resourcewould appreciateif left in the
ground. But he need not in practice sink his finds in
physicalor financialassets at that interest rate. However,
he wouldbe wiseto seek such a rate as a minimumyield
on his new investments.The so-calledHotelling Rule
states that if he left his natural resource alone, it would
appreciateat the market discount rate because of its
growingscarcity.
* Likewise,the extractionschedule,assumedto be at
a constant rate over some time horizon, is also a para-
digm and is used only for makingthe calculations.The
owner has a given resource. He may extract it for two
yearsor twenty.Everyperiod he may decideto alter his
plans, dependingon current prices and the expectations
thereof,by increasingor decreasingthe annualextraction
rate. He is at libertyto do so. All the formulaneeds is
the ratio between the total reservesand the amount ex-
tractedin the current period.Supposean owner,whohad
been planningto liquidatehis reservesover a ten-year
period, decides to accelerate extractionbecause of an
expecteddeclinein future prices (reckoningthat, since
his market share is small, he can do this with impunity,
that is, withoutdepressingprices)and now decideson a
five-yearhorizon.All that is necessaryis to use the new
ratio of reservesto extraction,and this can be decided
period by periodand changedeveryyear ifneed be.
* The same appliesto the discoveryof new deposits-
or a downwardadjustmentin reserves-usuallya tough
nut to crack. The new discoverydoes not have to be
counted as income,as some have suggested.All that is
necessary,if this approach is followed,is to alter the
reserves-to-extraction ratio (A)in the calculations,ihat is,
if it is decidedto keep the extractionscheduleas before.
In this case, the discoverywill reflect itself in higher
income than before, as shown by movingfrom left to
righton the x-axisin Figure3-1. However,the ownermay
verywellkeep the reserves-to-extraction ratio unchanged
by raisinghis annualextractionwhenhe realizesthat the
reservesare larger than he had thought. This will also
translateinto higher income.
* It is not necessaryto estimatethe absolutevalueof
the total mineralreservesor to resort to what is known
as "wealthaccounting."Neitheris it necessaryto predict
future prices. The owner of the resource does all the
predictionsnecessary,and these are reflected in his an-
nual extraction,whichthe accountanthas to relate to the
size of the total reservesin order to estimateincome.By
implication,it is assumedthat the unit valueof the total
resourceis the same as the current price. Suchvaluation,
appearingin both the numeratorand denominatorof the
formulagiven earlier,cancels out, and what remains is
the ratio betweentwo physicalquantities:the size of the
reservesand the annual extraction,that is, the number of
yearsremainingbeforethe resourceis exhausted.Specu-
lation about the ftuturecourse of prices, however,does
occur, and this, as mentionedabove,affects the rate at
whichthe resource is liquidated,but this is not the ac-
countant'sproblem.
* The problemsof the terms of trade or of changed
technologythat mightlead to drasticchangesin the valu-
ationofthe resourceare not addressedhere.Suchchanges
haveto be acknowledged by the incomeaccountantswhen
they occur.The focusof this approach is on the volume
of extractionin the accountingperiod as it relatesto the
total volumeof the reserves.In the manner of national
accounting,the market valuationof the product is taken
as givenand is used merelyto weightthe volumein order
to assessthe activity'scontributionto GDP.
* The proposed methodcould be applied immediately
to mineral depositsthat are more or less ascertainable,
such as petroleum,forwhichthe industryestimatesproven
7
reservesand publishesthese estimatesregularly. But even
for petroleum,and certainlyfor metals, ownerstend to
mine richer and more accessibledeposits first, which
means that later extractionsinvolveprogressivelyhigher
extractioncosts. Risingextractioncosts can undermine
the sustainabilityof the activityas much as the physical
exhaustionof the resource.When market prices fall be-
lowextractioncosts,manyprevioussellers,still sitting on
large deposits,find profitableoperation impossible.Esti-
mation of the volumeof reserves therefore should be
adjusted downward by a factor that would reflect the
risingfuturecost of extraction.Shortcutsfor such adjust-
ment need to be devisedcaseby case.
* It is importantto remember that the issue here is
nationalincomeaccounting.Even if the identifiedglobal
reserves of a mineral get adjusted upward,the fact re-
mainsthat the reservesof individualcountriesinevitably
are depleted as they are exploited.Nationalincome ac-
countingshould reflect this individualnational aspect of
the activity.
 TheProperCalculationof Income fromDepletableNaturalResources 17

Conclusion infinite series X. The capitalized value of the finite series

R, accruing in equal amounts over a period of n years,
Although the user cost approach appears radical in that would add up to:
it seeks to alter the calculation of GDPunder the SNAfor r 1 1
certain activities, it is economical and practicable. It is an n | - + n+ld
effective way of impressing on developing countries that ER* (=
depend on the exploitation of subsoil deposits that natu- o1 1
ral resources are being exhausted as they are exploited. I + r
The method proposed is in harmony with standard eco- The infinite series X would add up to:
nomic concepts. The national income accounting prac-
tices set out under the SNAdistort these concepts when X
applied to depletable resources. They falsely call rent that X* 1
which is not rent, and include in value added that which 0 1 - + r
is not value added. A second-best alternative would be to
use the depreciation approach, deducting a depletion fac- Setting E R* = E X* and multiplying by the denomina-
tor from an inflated GDPto reach a corrected NDP.The o o
user cost proposed in this chapter would be the correct tor in both quantities,
measurement of this depletion factor, not the "full-value" r
depreciation, which, as argued above, would wipe out all X= R 1 - I
the activity from the net product. I (1 +r) 1
The correction ought to be made in the flow accounts X/R = 1 -
of the SNAat the GDPlevel. It is not enough to record (1 + r)n+1
depletion in balance sheets, reconciliation, or satellite 1- X/R = 1
accounts. This approach would make it unnecessary to (1 + r)n+l
attempt to show in such accounts absolute values of total
reserves and their annual changes-values that would be In thiifmionit is asm at teripts I
as arbitrary as they would be unedifying. Its adoption alternatively, they accrue at the end, the fraction XIR
would lead to the proper understanding and measure- 1
ment of the special economic activity of depletable natu- would be 1 - . It is also assumed that the rela-
ral resource exploitation, and consequently to better pol- (1 + r)
icy analysis. tiveprices of the resourceand the goods and serviceson
which the stream of income will be spent do not change.
Appendix. Splitting Receipts into Income If there is reason to believe, for instance, that such goods
and Capital and services will appreciate over time relative to the re-
source, the capital element to be set aside has to be larger
In this chapter, receipts from the sales of a depletable (and the income content smaller) to make it possible to
natural resource are net of extraction cost. The extraction maintain a constant income stream in real terms. The
cost contains elements that do not directly generate value converse is true if there is reason to believe that the
added, such as materials used up in the process of extrac- resource would appreciate relative to the goods and serv-
tion, but would normally also contain payrnents to factors ices that would make up future income. But these are
of production, which should be included in GDP in the refinements that could be incorporated in the method
usual way. suggested and would not affect much the results obtained.
A time series of expected net receipts iR from the sale The method proposed, with the implicit assumption of
of a resource that, as a result of exploitation, will come constant relative prices, seems adequate if the direction
to an end in a future year n contains a lrue income in which relative prices will change is uncertain.
element X, where X < R, such that if R -- X (the capital Notes
content) is invested year after year at interest rate r, the
accumulatedinvestmentwouldcontinueto yieldthe same
levelato incomes t d 1. The term "DutchDisease"originatedin the 1960sto refer
leveltof
isnecoessar
X. idntfyXR,thtis te rpotin to the adverseeffects on Dutch manufacturingof natural gas
t is necessary to identify XIR, that is, the. proportion discoveries.Generallyspeaking,increasedrevenuesfrom a nat-
of net receipts that can truly be called income, and its ural resource encouragesspending on nontraded goods and
complement 1 - XIR, the capital element, also as a drawsresourcesout of the traded, nonnaturalresourcesector,
proportion of net receipts. The capitalized value at inter- thus stiflingdiversificationand retardingdevelopmentof non-
est rate r of the finite series of receipts R should equal naturalresourceexports.See, for example,Corden (1984).
the capitalized value at the same interest rate of the 2. See Hicks(1946),p. 172. See also Keynes(1936),chapter
18 SalahEl Serafy
6, on "The Definitionof Income,Savingand Investment,"and
that chapter's"Appendixon UserCost."
3. OPEC admittedlymet regularlyto agree on the prices at
whichits memberswouldsell oil. But such prices clearedthe
marketwithoutany quotasimposedto regulatesupply.Not until
1982, however,did OPEC behave like a cartel, with individual
quotasindicatedfor its members,but likeall cartelsthis attempt
to maintainpricesin a decliningmarketpalpablyfailed.
4. See Solow(1974),in whichhe recalledHotelling'spath-
breakingarticle, "The Economicsof ExhaustibleResources,"
Journalof PoliticalEconomy39 (April1931):137-75.
5. "[T]heoriginaland indestructiblepowersof the soil" as
formulatedby Ricardo(1821).
6. 1had been thinkingalong these linesfor some timeand
first expressedmy viewson this topic in a paper deliveredin
March 1979 to the staff of the OAPEC in Kuwait (El Serafy
1979). I elaborated these viewsand proposed a method for
estimatingincomefrom depletablenaturalresourcesin a later
paper on absorptivecapacity,presentedin 1980 to an energy
conferenceorganizedby the Universityof Colorado.The con-
cern at the time wasthat the so-called"capitalsurplus"econo-
mies, which exported petroleum,had too low an absorptive
capacity.It was felt that if that could be increased,it would,
throughincreasedimports,restoreequilibriumto the petroleum
buyers'balancesof payments.I attackedthis approachbecause
it reflectedthe short-terminterestsof petroleumconsumersand
not that of the ownersof this scarceresourceand of humanity
at large. See El Serafy (1981). An appendix to that paper
entitled, "HowMuchof PetroleumReceiptsCan Be Reckoned
to Income?"proposedthe formulashownin the appendixto
this chapter.
7. Occasionally estimatingreserveswouldraisecontroversy,
but this should not inhibit approximationsthat can be later
revised.In the words of the late Sir Dennis Robertson,"it is
betterto be approximatelyrightthan preciselywrong!"
References
Corden, W. M. 1984. "BoomingSector and Dutch Disease
Economics:Survey and Consolidation."Oxford Economic
Papers36:359-80.
El Serafy,Salah.1979. "The OilPrice Revolutionof 1973-74."
Journalof Energyand Development4(2, Spring):273-90.
1
. 1981. "AbsorptiveCapacity,the Demandfor Revenue,
and the Supplyof Petroleum."Journalof Energyand Devel-
opment 7(1,Autumn):73-88.
Hicks,J. R. 1946. Valueand Capital.2nd ed. Oxford:Oxford
UniversityPress.
. 1981. Wealthand Welfare:CollectedEssays on Eco-
nomic Theory.Cambridge,Mass.:HarvardUniversityPress.
Keynes,John Maynard.1936. The GeneralTheoryof Employ-
ment,Interest,andMoney.London:Macmillan.
Marshall,Alfred.1920. Principlesof Economics.8th ed. Lon-
don:Macmillan.
donMacil.
Ricardo,David.1821. On the Principlesof PoliticalEconomy
and Taxation.3rd ed. London:John Murray.
Smith,Adam.1937. An InquiryInto the Nature and Causesof
the WealthofNations, 1776.Cannan Edition.NewYork:The
ModernLibrary.
Solow,Robert M. 1974. "The Economicsof Resourcesor the
Resourcesof Economics."AmericanEconomicReview64(2,
May):1-14.
 __ _ __ _ _
IntroducingNaturalCapitalinto the SNA
Anne Harrison
Duringthe past ten to fifteen years public reaction to
environmentalissueshas broadenedfrom a concern for
preservingendangeredwildlifeand aestheticallypleasing
landscapesto the realizationthat the wholeprocess of
economicdevelopmentdependson the utilizationof nat-
ural resources.In extremecases,such as in Sub-Saharan
Africa,the physicaland economicsurvivalof millionsof
people dependscriticallyon the managementof both the
quantityand qualityof land and waterresources.
When environmentaleconomistswish tc demonstrate
their concernin quantitativeterms and to illustratethe
effectof alternativescenarios,they find the present Sys-
tem of NationalAccounts(sNA) (UnitedNal:ions1968) to
be inadequatefor a number of reasons.The problemsare
both of omissionand commission.On the one hand,
manyactivitiesundertakenbywomenin developingcoun-
tries, such as findingand carryingwater amdfuelwood,
are excludedfromthe presentmeasuresof grossdomestic
product (GDP). On the other hand, majc,rprojects to
rehabilitatepollutedriversor otherwiserestore degraded
environmentalresources are included in GDP, and in-
creasesin this typeof activitylead to increasesin GDP.
Steps to redress some of these shortcomingsare in
hand. Variousspecializedstatistical systemshave been,
or are being, developedto addressquestionsin specific
areas, and the environmentis no exception.Physicalre-
source accountsare being developedto show,in volume
terms, inputsand outputsof physicalprocesses,including
natural resourcesas wellas manufactured'products.The
"Frameworkfor the Developmentof EnvironmentalSta-
tistics" (UnitedNations 1984), for example,attemptsto
showthe interactionamongenvironmental,economic,and
social statistics. Various proposals have beern,and are
being, examinedto developan aggregatethat includes
19
4
the "desirables"that the SNA omits and excludes the
"undesirables"that the SNA includes.
Such alternativestatisticalsystemscould form satellite
accountsto be used in conjunctionwith the revisedSNA.
To the extent that such systemsare greater elaborations
of specificdetail and rearrangementsof existing items
and inclusionof others, no conflict should arise. How-
ever,giventhe claimof the SNA to providea comprehen-
sive frameworkunder which all such satellite accounts
can be developed,it is importantto establish an appro-
priate interfacewithenvironmentalmatters in the process
of reviewingthe SNA.
From an environmentalperspective,neither exhausti-
ble resources such as mineral deposits nor permanent
resourcessuch as land and water should be treated as
freegiftsof nature. Althoughland and water seem at first
to be free, these resources are not automaticallyself-
renewing,and although humanity may have power to
renewthem,it has a muchgreater powerto destroythem.
Economicactivitynecessarilyinteracts with nature, and
allowancemustbe madefor the existenceof environmen-
tal programs,whichhusband and maintainthese penna-
nent resources,and for the lack of such programs,which
leads to the destructionof natural resources.
The presenceor absenceof such programsis therefore
an economicdecisionthat directly affects the potential
economicactivityin subsequentperiods, just as a deci-
sion on the rate of depletionof nonrenewableresources
does. In both cases, therefore, there is a strict parallel
with the decisionto create and maintainman-madecapi-
tal, and environmentalistsargue that the SNA should rec-
ord these resources as alternativeforms of capital.The
question,therefore,is howmuchpresent practicesunder
the SNA wouldchange if these resourceswere treated as
20 Anne Harrison
natural capital. The main purpose of this chapter is to
answerthis question.
ExploitableResources
Muchof the past discussionbyresourceeconomistshas
concernedthe treatmentof nonrenewableresources,such
as mineraldeposits.In fact, similarargumentsalso apply
to natural forests used for timber and to natural fish
stocks. Since both of these are renewable,even if not
actuallyrenewedon a realistictime scale,it is convenient
to considerall three cases as exploitableresources.
In the production accounts of the present SNA, when
an exploitableresourcesuch as oil is extractedand sold,
onlythe directcosts associatedwithits extraction,includ-
ing labor, are deducted from its market value, and the
wholeof the differenceis treated as gross operatingsur-
plus. Net operating surplusdiffersfrom this only to the
extent that fixed man-made capital is consumed. Al-
though the processof oil extractionis treated as produc-
tion, neither the value of newlydiscoveredreserves nor
changes in the value of reservesbecause of changes in
worldprices is treated as production.
This results in the rather anomalouspresentationof
infdrmationabout stocks. The value of the unexploited
oil at the start of the year appears in a balancesheet for
the industryand the nation, as do the start-of-yearvalues
of man-madecapitalassets, but, since all changesto the
valueof exploitableresourcesbetween the start and end
of the year are explainedin the reconciliationaccount,
this occurrencein the balance sheet is like a memoran-
dum item. If exploitableresourcesweretreated as natural
capital, part of what is now treated as gross operating
surplusin the productionaccount wouldappear as con-
sumptionof natural capital, and net operating surplus
would be decreasedby this amount. A matchingentry
wouldappear in the balance sheet of natural capital as-
sets.
The question arises of what value is appropriate to
attach to the exploitableresourcebefore its exploitation.
Daly(Chapter2) and others suggestthat, in additionto
the direct cost of extraction,an entrepreneurialreturn
(calculatedperhaps as a standard markup on the direct
costs) should also be deductedfrom the valueof the oil
at the wellheadand that the residualwouldrepresentthe
subsoilvalue. This straightforwardapproachhas the ap-
peal of simplicityand economicsense. Variationsin world
priceswouldlead to matchingvariationsin subsoilvalues
and would determinewhen a deposit was commercially
viableand when not; the latter being when the subsoil
valuefell to zero or below.In certain circumstancesex-
ploitationmightcontinueevenif this valuewerenegative,
forexample,becausethe directcostscouldnot be averted.
In such a case, the entrepreneurialreturn wouldhaveto
absorb the operatingloss, and on the balance sheet the
valueof the reserveswouldbe givenas zero.
Morecomplicatedalternativevaluationsfor mineralex-
traction have been proposed. For example, El Serafy
(Chapter3) suggestsdeterminingan income streamwith
the use of a discountedcash flow analysisof expected
earningsfromthe depositover its expectedlifetime.
Permanent Resources
Anothertypeof expenditureof concernto environmen-
talists is related to the preservationof land, water, and.
clear air. This is generallydescribedas defensiveexpen-
diture, althoughfurther categorizationis instructive,and
incomeas wellas expenditureneedsto be considered.In
contrast to the exploitableresources discussed above,
these resources will be referred to as permanent re-
sources.
Beforeconsideringhow to deal with issues related to
permanentresources,it is helpfulto reviewthe SNA guide-
lineson the treatmentof man-madecapital.Two typesof
expendituresassociatedwith capital occur on a yearly
basis.The firstis the consumptionof fixedcapital(depre-
ciation),whichis an allowanceto permitthe replacement
of the capital asset at the end of its usefullife. Overthe
economyas a wholethe valueof the consumptionof fixed
capitalin a year is the valueof all man-madeassets that
were used up in that year and that haveto be replacedif
the level of man-madecapital stock at the start of the
yearis to be kept intact.The second is the currentexpen-
diture necessaryto repair and maintainthe capitalstock.
The SNA distinguishesbetween current and capital re-
pairs: "expenditureson current repair and maintenance
make good breakagesin fixed assets and keep them in
properworkingorder, whileoutlayson capital repair and
alterationlengthenthe expectednormallifetimeof use ol
fixed assets or increase the productivityof these good
significantly"(para. 6.123).
This distinctionmay be blurred in practice, however.
Althoughcurrent maintenancemay not directly extend
the life of an asset, the lack of maintenancemayshorten
it, and, in practice, estimates of the lifetime of assets
assume regular maintenance.There is thus a relation
between the levels of current maintenance and capital
consumption.Indeed,for a specificgroup of assets, in-
cludingroads and dams,the SNA states: "it may be con-
sideredthat expenditureson repairsand maintenanceare
sufficientto maintainthe asset in its originalcondition"
(para. 7.20);this is consideredas justificationfor making
no provisionfor consumptionof these fixed assets. This
provisionhas alreadybeen questionedin the SNA revision
process,and it is likely to be changed in the new SNA
manual.'
In severalpoorerdevelopingcountries,for example,the
retrenchmentof governmentexpenditurehas led to the

neglectof road maintenance.Eventuallynew capitalpro-
jects have been necessaryto replace roads, and these
involvecapitalexpenditurethat wouldnot haveoccurred
if there had been adequatecurrent maintenance.Thus it
is suggestedthat if maintenanceis not adequateto keep
roads in their original condition, then the capital con-
sumptionof these assets should be estimated.Although
not yet discussedin detail, the appropriatevalueof the
capitalconsumptionwould seem to be that of the "miss-
ing" maintenance.
Howdo these practicalitiescarry over to the proposal
to treat naturalresources as capitalassets?Expenditures
associatedwith permanent resources can be of two re-
lated, but rather different,types. The first are expendi-
tures intendedto preventthe degradationof natural re-
sources,and the second are to redress degradationthat
has alreadyoccurred.Preventiveexpenditureclearlypar-
allelscurrent maintenanceand, in particular,the current
maintenanceof roads. If all industriesensure that land,
air, and water are not degraded,these natural resources
remain in their original condition,and Ihe preventive
expenditureis clearlycurrent.
Basically,such preventiveexpendituresmaybe madein
one of three ways:by industryitself(eithervoluntarilyor
in response to governmentlegislation),by government
and funded by taxes leviedon industryaccordingto its
pollutionpotential,or by governmentand fundedby gen-
eral revenue.In the first case, expendituresby industry
to preventpollutionwould be classifiedas intermediate
expenditures.In the second case, since the paymentto
governmentis relatedto the serviceprovided,this activity
of governmentshould be treated as a public enterprise.
Paymentsto governmentshould not be treated as taxes,
but as feesforservices,and thus shouldalso be classified
as intermediateinputs. (Governmentprovisionof waste
disposalservicesto industryis alreadytrealtedas a public
enterprise in some countries.)In the third case, when
governmentfunds pollutioncontrol out of general reve-
nue, this appears as a final expenditure.
Someresource economistsargue that all such preven-
tive expenditureshouldbe excludedfrom GDPand there-
forewishto categorizeevengovernment-funded programs
as intermediateexpenditure,but this proposalis inconsis-
tent with accountingdefinitions.The distinctionbetween
intermediate and final expenditure is determined by
whetherthe product is resoldto anotherec:onomicagent,
not by the nature of the product. For welfareor other
analyses,specificactivitiesmay be excludedbecause of
their nature. In addition to environmentalprotection
costs,such candidatesas defenseexpendituresand crime
and drug "industries"have been put forwardfor exclu-
sion. The derivationof an alternative measure of a re-
stricted set of activitiesby deductingfrom (and possibly
addingto) those includedin GDPusingnormativecriteria
is a legitimateand arguablydesirabledevelopment.But it
IntroducingNaturalCapitalinto the SNA 21
does not make the case either for suppressingGDP as a
measureof all economicactivityregardlessof desirability
or for changingthe accountingbasis for discriminating
betweenintermediateand finalexpenditure.
Eventhe alternativepresentedhere-that whenpreven-
tive expenditureis undertakenby industryor by govern-
ment on a charge-backbasis,it is intermediateexpendi-
ture-is oversimplifiedand misleading.The preventive
expenditureconsistsof the purchaseof labor and of goods
and services,whichin turn have labor input. Therefore
initiatinga preventiveexpenditureprogramdoes increase
the valueadded,and thus GDP, evenif it is first recorded
in the accountsas intermediateexpenditure.Indeed,the
fact that increasingpollutionpreventionwould increase
employmentopportunitiesis often cited by environmen-
talistsas an argumentfor initiatingsuch programs.Under
the present SNA, all alternativeslead to the conclusion
that initiatinga pollutionpreventioncampaign,however
funded,increasesGDP.
It is suggestedhere, however,that no such apparent
increase in GDP should be shown.The consequenceof
treating natural resourcesas capital assets is that degra-
dation of those natural assets is treated as capital con-
sumptionand shouldbe includedin GDP. This treatment
would exactlyparallelthe present situationof the treat-
ment of roads describedabove.The newSNA proposalis
that GDP should include either the cost of maintaining
roads in their original conditionor the consumptionof
man-madecapital of the same amount.This recognizes
that the present practice, in whichthere is no mainte-
nance, leads to GDP being understatedby the amountof
this missingcapital consumption.For the environment,
therefore,it is proposedthat if preventiveexpenditureis
incurredand no pollutionresults,GDP is correctlymeas-
ured. But ifsuch programsare neededbecausepollution
is increasing,but theyare not being initiated,then GDP is
underestimatedby the cost of the programs. A simple
numericalexampleof the alternativesis given in Table
4-1. Thisillustratesthe casein whicha countryat present
has a GDP of 100 with consumptionof man-madecapital
of 10.The effectof environmentaldegradationis 5, which
Table4-1. Effecton GDPof Introducing
an EnvironmentalProtectionProgram
aPreent SNA Proposedrevision
Without With Without With
program program program program
GDP 100 105 105 105
Consumption
of
man-madecapital 10 10 10 10
Consumption
of
naturalcapital - - 5 0
NDP 90 95 90 95
- Not applicable.
22 Anne Harrison
is exactlyredressedby introducingan environmentalpro-
tectionprogram.
This highlysimplifiedexample highlightsthe conse-
quencesof the proposedtreatment and of preservingthe
existingpositionon natural capital.Netproductmeasures
are the same in both systems,whichshowsthat the intro-
ductionof an environmentalprotectionprogramleads to
an increase in NDP. Under the present SNA there is an
increaseof equalsize in GDP, whereasunder the proposal
described here there is none. But this is achievedby
increasingthe present valuationof GDP without the pro-
gram to allowfor the consumptionof natural capital.
At first it may seem counterintuitivethat the degrada-
tion of permanentresourcesleads to an increasein GDP.
This reflects the common lack of awarenessthat the
"gross"in GDP means "before allowancehas been made
for consumptionof capital." Commonusage overlooks
this part of the definitionor assumesthat the allowance
for depreciationis fairlyconstant overtime, so gross and
net product measuresmovein line with one another,and
gross can be used as a proxy for net without undue
distortion. This assumptionhas proved misleadingfor
manypoorercountrieswherenew capitalis not acquired
as fast as the old is exhausted.In environmentalterms
the assumptionis equivalentto assumingthat the environ-
ment is not degraded,and it is the obviousrefutationof
this that is the originof the presentconcern.An objective
of the present reviewof the SNA must be to ensure not
only that GDP and NDP are properlymeasured,but also
that the conceptsare clearlyexplainedto facilitatebetter-
informedand more appropriatecommentaryand analysis.
An importantuse of GDP is for comparisonsovertime
or betweencountries.Althoughthere maybe reluctance
to increasepast estimatesof GDP to allow for the con-
sumptionof natural capital,the alternativeis also disturb-
ing. Withoutthis revision,countriesintroducingenviron-
mental protectionprogramswill show increasesin both
GDP and NDP. The worse the environmentaldegradation
that is being reversed,the greater this increasewillbe.
Underthe proposalabove,onlyNDP willincrease,and the
gap betweengross and net product willnarrow,in accor-
dance with both the economicand accountinginterpre-
tation of gross and net. The entries for this capital con-
sumptionas wellas for man-madecapitalwould also be
includedin the balancesheet and wouldshowthe cumu-
lativeeffect on the resources availableto the nation of
continuedenvironmentaldegradationor of its restitution.
Valuation of the Consumption of Natural Capital
Assumingthat the theoretical proposal above is ac-
cepted, there remainsthe question of howconsumption
of natural capital should be measured.Consumptionof
man-madecapital,althoughnot without its owndifficul-
ties of measurement,can be calculatedat a disaggregated
levelbased on differentlifetimeassumptionsfor different
types of assets in differentindustries.For the consump-
tion of natural capitalsuch an approachis neither possi-
ble nor appropriate.Industryis not the only degrader of
the environment,and it may be impossibleto attribute
the degradationto individualindustries.
Consider,for example,the case of three factories, all
dischargingeffluentinto the same river.It is possiblethat
the river couldabsorb the effluent fromany one but not
all three or that the combinationof differenteffluents
causedthe problem.In such casesattributionof the cause
of the pollutionamongthe three factorieswouldbe diffi-
cult and not especiallyinstructive.Degradationmayalso
be caused by final consumers,for example,by the car
exhaust of householdsor government.It may also occur
not as a result of activitywithinthe country,but as an
involuntaryimport from a neighboringone. For all these
reasons,therefore,it seems appropriatethat a single ad-
justment in moneyterms to the SNA definitionof gross
productmeasureswouldbe adequate.This mightbe seen
as analogousto the adjustmentfor imputedbankingserv-
ices, which is made in total only and not attributed to
individualindustries.
For the national accountsfor a year, it is the changes
during that year in the environmentalendowmentsthat
need to be captured,not the deviationfrom an absolute
state of perfection, and these annual changes can be
measured,for example,by air and water qualityindexes.
The cost of making improvementscould then be esti-
mated. Althoughthe measurementand valuationof the
consumptionof natural capital may have to be approxi-
matedand simplified,these inexactitudesmay be no less
tolerablethan those alreadyencounteredin such areas as
the ownershipof dwellings,the valuationof subsistence
output, and the derivationof holdinggains and losseson
intangibleassets.
Environmental Enhancement Programs
In this discussionit has been assumedthat there is a
prima facie case for includingadjustmentsfor the con-
sumptionof natural capital in all countries,althoughfor
some countrieswhereno environmentaldegradationoc-
curs the appropriateadjustment may be zero. Another
factor to consider,although it may apply irregularlyand
to only a few countries,is how to treat major programs
to reverse the degradationthat has already occurred.
Such programs are most likelyto be funded by govern-
ment,but there is no reason in principlewhytheyshould
not be fundedby industryor by private nonprofitinstitu-
tions. Whoeverundertakes the expenditure, it clearly
wouldenhance the natural resourcesand would increase
the availabilityof such assets in the future. This under-
linesthe case for treating the assets as capital and the
expenditureas capitalexpenditure.

Treatingenvironmentalenhancementprogramsas cap-
ital expenditureis a logicalconsequenceof introducing
the consumptionof natural capital as an adjustmentbe-
tweengross and net product measures,since they are in
effectnegativecapitalconsumption.It was argued above
that the measurement of natural capital consumption
should be based on annual changesin qualityindexesof
environmentalendowments.If such an index showsan
improvementover the year, the value of natural capital
consumptionfor that resource is zero, but the valueof
the improvement,whichis by our definitionthe valueof
the enhancementprogram,should enter the azcountsas
the same broad type of expenditure,that is, as capital
rather than current. Some misattributionof capital ex-
penditure may go undetected if the program does not
make a positiveimprovementbut only mitigal:esthe de-
gree ofdegradationin the environment.If the data sources
were rich enough, adjustmentsfor these situationscould
be made, and separate adjustmentscould be made for
differenttypes of naturalresources.Sincethe proposalis
to adjust for the consumptionof, and enhancementsto,
natural resourcesat the nationallevel,however,aggrega-
tion of the effectsacross projectsand resourcetypeswill
still ensureconsistencyin the flowaccountsand balance
sheets at the nationallevel.
It may be argued that natural resources are not in
themselveseconomicallyproductiveassets and therefore
that expenditureto enhance them should not be treated
as capital. However,not all items includedin gross do-
mesticfixedcapitalformation(GDFCF) in the presentSNA
are economicallyproductive. Schools and universities
produce better educated people, but they will be more
productiveonly if employmentopportunities exist that
utilizethe extra educationgained. It is not unknownfor
such institutionssimplyto increasethe number of edu-
cated unemployed,althoughthe GDFCF is not decreased
to reflect this. Nor shouldit be. Schoolsand universities
are means of improvingthe human capital stock. The
deploymentof that capitalis a separateissue.Someman-
made capital assets are purely defensive,such as the
Thamesbarrage, a mechanicalbarrier across the River
Thamesto hold back waterduringhigh tide and adverse
windsto preventLondonfrom beingflooded.Somne serve
obviouslynoneconomicfunctions,such as a new cathe-
dral or a spy satellite.The range of assetsto be included
in GDFCF in the new SNA is due to be reviewed,and a
distinctionin classificationbetweenproductiveand non-
productiveassets might be very desirable.But even on
presentgroundsa case can be made to treat majorenvi-
ronmentalenhancementprogramsas capitalformation.
Thesesorts of programsare not mentionedin the pres-
ent SNA. It is difficultto be categoricalabout present
practice, but it seems that the implicitrecornmendation
in the SNA to treat such expenditureas current is gener-
allyfollowed.In line with the discussionabove,such gov-
Introducing Natural Capital into the SNA 23
ernmentprograms,eveniftreated as currentexpenditure,
will at present lead to increasesin GDP but willnot be
separatelyidentifiable.The proposalbeing made here is
that majorenvironmentalrehabilitationprogramsshould
be classifiedas capitalexpenditureand identifiedas en-
hancementof natural resources,regardlessof who funds
them. In supplementaryanalysesa more limitedclass of
capitalexpenditure,includingonly "productive"capital,
howeverdefined,maybe used, parallelingthe derivation
of the more restrictedmeasuresof "desirable"activities
describedin the discussionof preventiveexpenditure.
Unplanned Occurrences
The sections above dealt with current and capital ex-
penditures in relation to the environmentand in the
contextof plannedactivities.The environmentand sub-
sequent economicactivity,however,are also greatly af-
fectedby a natural disaster,such as an earthquake,or a
man-madedisaster, such as a major chemicalspillage.
The proposalsdescribedabove could encompassthe ef-
fect of such unplanned occurrences.A disaster would
cause consumptionof natural capitalto increaseand net
product measuresto fall.This wouldbe offsetin part by
anysignificantexpenditureto restorethe situation,which
wouldcount as enhancementof natural capitalfromwhat
it wasafter the disaster.The same is true of the effecton
naturalresourcesin one countryof degradation(planned
or unplanned)originatingin another country.
Sustainability and the SNA
Apartfrom the intrinsicinformationto be containedin
environmentalaccounts, there are two further benefits
that this initiativehas brought to national accounting.
The first has been to focusattention on the difference
betweennet and grossproductmeasures,since evenwhen
the distinctionrelatesonlyto man-madecapital,there are
numerousoccasionswhennationalincomeon a net basis
wouldbe a more appropriateindicatorthan the ubiqui-
tous GDP. The other benefit is more diffusebut arguably
more importantin the longer term. The present SNA has
depersonalizedeconomicsby concentratingon produc-
tion processesand associatedtechnologies.Environmen-
talistshavebroughtattentionback to the questionof how
peoplelive and the qualityof life,whichis associatedwith
an understandingthat respect for life is intimatelyand
inevitablybound up with respect for the environment.In
statisticalterms this gives a new opportunityto bridge
the presentseparationbetweeneconomicand social sta-
tistics by articulating the interaction between environ-
mentaldevelopmentand humandevelopmentthroughthe
economicproduction process. It is this integration of
natural, man-made,and human resources that provides
the frameworkfor definingsustainabledevelopment.
24 Anne Harrison
Environmentaleconomistsare still workingto agreeon
a definitionof sustainableincome.But centralto it is the
point, recognizedby Hicks,that if assets are consumed
withoutbeing replaced,one is worseoffat the end of the
period than at the start, and that consumptionin the
period coversnot just income but also an element of
wealth.
The presentSNA containsthree measuresof GDP-from
the income,output, and expendituresides-all of which
are identicalin value. It may be possibleto detennine
measuresof sustainableincome and sustainableoutput
that are equalin the aggregatebut not in their disaggre-
gatedparts. For example,ifincomefroma miningactivity
is used to fund alternativeproductiveactivitywhen the
mineralresourcesare exhausted,the output frommining
is not sustainableeven if total output and total income
are. But there is no necessityfor the income in a later
period to be at exactlythe same levelas in the previous
period.Which levelof incomeis to be used as the basis
of sustainableincome?Is it the sustainabilityof income
or incomeper headthat is of interest?Is sustainabilityan
absoluteor a relativeconcept?
The answerto these questionsmust dependin part on
the needs of analysts.It may wellvaryfrom one applica-
tion to another and may depend on assumptionsabout
future technologicalchange and other subjectivepredic-
tions. Giventhis uncertainty,such work seemsat present
a candidateforsatelliteaccountsrather than total integra-
tion with the SNA. However,a substantiallink could be
establishedby adopting "sustainabilityfactors." These
factorswould be nothing more than the ratio of capital
stock at the end of the period to capital stock at the
beginning, adjusted for price change when necessary.
Hicks'sdefinitionof income-"what you can consumein
the period and be as welloff at the end as at the begin-
ning"-is equivalentto sayingthat new capitalmustbe at
least as great as capitalconsumedduringthe yearso that
the capitalstock at the end is at least as great as at the
start. In this case a sustainabilityfactor would have a
valueof one or greater.A valueless than one impliesthat
too much capital has been consumedand that the con-
sumptionlevelis not sustainableindefinitely.
Despitethe simplicityof the definition,this conceptof
a sustainabilityfactor has severaladvantages.It can be
applied at detailed levels as well as in the aggregate.
Sustainabilityfactorsfor man-madecapital could be dis-
aggregatedby assetand industry,and, evenat the detailed
micro or project level,sustainabilityfactors for environ-
mentalresourcescould be calculatedseparatelyfor vari-
ous categoriesof land, for water, and for air. Further,
since the factors are independentof price effects,they
can be basedon the same sort of quantifiedinformation
used to calculatethe consumptionof natural capital as
proposed above. Indeed their compilationwould be a
logicalstep in derivingestimatesof consumptionof nat-
ural capital.Suchmeasurescouldalsobe appliedin areas
of humancapitalby measuring,for example,the ratio of
the number of trained doctors in the country at the end
of a period to the numberat the start. Further variations
on the basic ratio are also possible.For long-termplan-
ning, periods longer than a yearmightbe used; in other
cases assets per head of populationmightbe the basisof
the ratio.
Not only would the adoption of a set of sustainability
factorsthat can be calculatedeasilyand withoutambigu-
ity form a bridgebetweenthe main SNA and the satellite
environmentalaccounts,theywouldhighlightdirectlythe
differencein valueand interpretationof measuresof gross
and net product. Provisionof, and referenceto, a set of
such factors can only help to spread the understanding
of the conceptsof the nationalaccountsand raise discus-
sion in the popular mediato a more informedplane. This
consequencewouldbenefitboth nationalaccountantsand
environmentalists and wouldbe an importantstep on the
road to the introductionand acceptanceof more elabo-
rate conceptsof sustainableincomeand product.
Summary of Recommendations
If satelliteaccountsfor the environmentare to comple-
ment the sNA, the basic accountingconventionsin the
two systemsmust agree. In particular, the distinction
betweenfinal and intermediateexpenditureand the ex-
tent of capitalconsumptionthat is the differencebetween
gross and net domesticproduct must be agreedon if the
two systemsare to be harmonized.The proposalsmade
in this chapter to reach this harmonizationimply the
followingchangesfor the SNA.
a. It wouldbe necessaryto introduce a balance sheet
for natural resources.
b. Part of what is at presenttreated as gross operating
surplusfor exploitableresourceswouldbe treated as con-
sumptionof natural capitaland would appear as such in
the productionaccount.
c. An adjustmentfor the consumptionof natural capi-
tal wouldbe calculatedand appliedas a singleadjustment
of the economyas a whole.
d. Both of the adjustmentsin b and c would be de-
ductedfrom net product measuresas nowdefined.
e. Projects designed to enhance environmental re-
sourcessignificantlywouldbe classifiedas natural capital
formation.This attributionwouldbe appropriatewhether
the expenditurewas undertakenby governmentor indus-
try.
f. The estimatesof consumptionof natural capitalfrom
b and c and for enhancementof natural capital from e
wouldboth carry overto the balancesheet.
g. Sustainability
factorsshouldbe introducedthat show
for all capital-man-made,natural, and human-the ratio

of stock at the end of the period to that at the begiining.
This would help develop more comprehensive measures
of sustainable income in the longer term and to empha-
size the practicality and desirability of using net rather
than gross product measures where this is appropriate.
Note
1. The presentSNA was publishedin 1968 and is IIowin the
course of review,which started in the early 1980s and is ex-
pectedto lead to a revisedversionin 1991. The reviewprocess
consistsof a series of three or four meetingsa year,each on a
IntroducingNaturalCapitalinto the SNA 25
specifictopic and attended by experts from both developing
and developedcountriesand representativesof the international
organizationsconcernedwithcollatingand publishingeconomic
statistics.
References
UnitedNations.1968. "A Systemof NationalAccounts."Publi-
cation E.69.XVII.3. NewYork.
. 1984. A Frameworkfor the Developmentof Environ-
ment Statistics. StatisticalPapers, Series M, no. 78. New
York.
 _ ~~~~~~~~5
MeasuringPollutionuithin theFramework
of theNationalAccounts
Derek W Blades
The nationalaccounts could make a usefulcontribution
to environmentstatisticsin the importantarea of meas-
uring certain aspectsof pollution.Pollutionis taklento
mean the disposalof waste solids, liquids,gases, heat,
and noisein waysthat harm people or reducetheir enjoy-
ment of theirsurroundings.It is helpfulto distinguishtwo
sides of pollution:the act of polluting, such as by dis-
charge of noxiousgases or fluids into the environment,
and the abatementof pollution,whichcoversgovernment
monitoringof pollutionlevelsas wellas measurestalken
by industryand householdsto reduce the output of pol-
lutants.This discussioncoverstwo aspectsof the act of
polluting(the output of pollutantsand pollutiondamage)
and two aspectsof the abatementof pollution(the costs
and the benefitsof pollutionabatement).
The Outputof Pollutants
Since pollutantsmay be regarded as a particular (uin-
wanted) type of industrial output, it seems natural to
studythe generationof pollutantsby an extensionof the
conventionalinput-outputtable. Table 5-1 summarizesa
model proposed by Ayres and Kneese (1969), which
showsthe overalleffectof economicactivityon the envi-
ronment. It is based on the materialsbalance principle:
goods that are taken out of the environmentcirculate
through the production-consumptioncycleand are re-
turned to the environmentin a changed,but not dimin-
ished,form.
MatrixA is a conventionalinput-outputtable displaying
the commoditycompositionof intermediateconsumption
by industries. MatrixB shows the amounts of various
natural resources used in production, including "free"
resources such as air and water. Matrix C shows the
amountsof wasteproductseventuallydischargedinto the
environment.
26
To date onlyone attempt has been made to applysuch
a model. This was developedby Victor (1972) for the
Canadianeconomyand has been used to determinethe
output of pollutantsassociatedwith one dollar'sworth of
final demandfor variouscommodities.The data require-
ments for matrixesB and C are clearlyveryheavy,partic-
ularly since they involveinformationthat is not usually
collectedin industrialsurveys.In addition,however,there
are at least two other reasonswhyrelativelylittle interest
has so far been shownin this approach.
First,matrixesB and C containvariousunits of volume,
suchas gallonsof wateror cubicfeet of nitrogenin matrix
B and millicuriesof cesiumor tons of sulfur dioxidein
matrixC. Thus the entries cannot be aggregatedto pro-
duce meaningfultotals, and the relative importance of
variousinputs and outputs can be measuredonly in an
arbitraryfashion.The second problem concerns the op-
erationalvalueof the model.The kind of data that poli-
cymakersneed about the output of pollutantsis highly
site-specific.The informationthat the pulp and paper
industryuses x gallons of water and emits y tons of
chlorineis not in itselfvery helpful.Such data have op-
erational significanceonly when they can be related to
informationabout the particular sites where pulp and
paper millsare located.In short, matrixesB and C are of
limitedinterestfor most policypurposesbecausethe data
they contain are too generalizedto be applied to local
problems.
Pollution Damage
Pollutioncauses damages, such as sickness and pre-
mature death, stuntedgrowthof crops, disappearanceof
wildlife,or corrosionof metalsand masonry.The national
accountsreflectsome of the costs of some of these dam-
ages in, for example,increased health expendituresby

Table 5-1. The Effect of Economic Activitf on the Environment
Industries
Commondities A C
Primary inputs _ _
Total
Environmental
commodities B _
households or government, lower labor productivity,
highercosts of foodproduction,and increasedoutlaysfor
repairing and maintainingbuildings,but they are not
explicitlyshownas costs of pollutiondamage.Wlouldit
be possibleto do so? Table 5-2 showsan input-output
table that could serve as a frameworkfor showingexplic-
itly some of the costs of pollution damage that are in-
cludedin the national accountsbut hiddenat present.
The columnsheaded DC indicatecosts frornpollution
damage. ColumnsA and B are intermediate costs in-
curredby industryin, for example,cleaningpollutedwa-
ter forhuman consumptionor industrialuse and protect-
ing buildingsand equipmentfrom atmosphericpollution.
ColumnC would includethe costs of repairs to govern-
mentbuildingsbecauseof air pollutionand-a largeitem
in many countries-the costs of medicalservices con-
nected with pollution-linkeddiseases.Householdexpen-
ditures on these medical services would be shown in
columnD, whichwouldalso includesuch cosis as insula-
tion to protect houses from traffic noise, air-filtering
equipment,and water filters to removechemicalsfrom
drinkingwater.
Finally,columnE includescapitaloutlaysby producers
and governmentin connection with pollutiondamage.
These would includepart of the cost of replacingbuild-
ings and equipmentwhoseliveshavebeen shortenedby
pollution.
ColumnsA through E containonly a smallpart of the
total cost that could be attributed to polluticn damage.
Table 5-2. Input-Output Table Isolating PollutionDamage Costs
Prodlucers ...__-
1 2 Total consumption
Other
DC DC Other -F
Commodities
1 A BC
2
Total __
Indirect taxes .__._
Factor incomes
Restof world _
Note:DC = pollutiondamagecosts.
MeasuringPollutionwithinthe Frameworkof the NationalAccounts 27
Final Wastedischarged
demands Total to the environment
i
In a study by the United KingdomAtomic Energy Au-
thority (1971) the costs of air pollutionwere deemed to
include, in addition to medicalexpenses, the value of
production lost through illness and death, the loss of
imputedincomeof housewivesand students,and the so-
cialcostsof prematuredeath and illness.The socialcosts,
whichwereby far the largestpart of the total, werebased
on a valueof 1,000 for each year of life lost and on the
ranges of legal compensationawardedfor sickness.The
valuationprocedures used for these various costs are
often very arbitraryand subjectto great uncertainty.Just
howuncertainmaybe judged from the results of studies
on the costs of air pollution in the United States pub-
lishedannuallyby the Councilon EnvironmentalQuality.
For 1968, estimatesof the cost of air pollutiondamage
ranged from $8.1 billionto $20.2 billion,and for 1970
from $7.0 billionto $12.3billion.Saunders(1977, p. 51)
observedthat "the range of uncertaintysurroundingsuch
estimatesis enormous.It maybe suggestedby the skep-
tics that the usefulnessof such figures. .. is in the sphere
of publicrelationsrather than of socialscience."
The costs discussedabove in connectionwith Table
5-2 are less open to criticism on these grounds, since
. they are alreadyincludedin the national accounts;they
only need to be separatelyidentified.Whilethe difficul-
ties of doing this should not be minimized(often the
costs of pollution-relateddamagecannot be easilydistin-
guishedfrom the costs of normalwear and tear), it seems
worthwhileto try to identifysome of the major costs of
P Grossfixed
Government Private I capital
consumption formation Stocks Exports Totals
DC Other DC Other DC Other
D E
28 Derek W Blades
pollutiondamage.These mightinclude,for example,the
costs of treating pollution-linkedrespiratorydiseasesfor
persons livingin large towns and the estimatedcosts of
protecting buildings and machineryfrom atmospheric
pollution.
Costs of Pollution Abatement
The third aspectof pollutionto be consideredconcerns
the costs borne by industries,and to a lesser extentlby
governmentand households,to reduce the output of pol-
lutants. Elsewherein this volume,these costs are de-
scribedas "defensiveexpenditures."The columnsheaded
PA in Table 5-3 show where these costs are currently
includedin the national accounts,althoughthey are not
nowdistinguishedseparately.
ColumnsA and B showpurchasesof intermediateand
primaryinputsfor pollutionabatement.These represent
the current operatingcosts incurred by industryfor this
purpose.ColumnC showsgovernmentspendingfor pol-
lution abatement,mainlyin connectionwith monitoring
pollutionlevels.Householdpurchasesof goods,such as
catalyticconvertersto reduce emissionsof noxiousex-
haust gasesfromvehiclesor less-pollutingequipmentand
suppliesfor home heating, are shownin columnD. Fi-
nally,columnE shows businessand governmentinvest-
ment outlaysthat are mainlyfor pollutionabatement.
Input-outputtables of the kind outlined in Table 5-3
havebeen quitewidelyapplied.Modelsdevelopedfor the
UnitedStatesand the Netherlandsare describedin OECD
(1974), modelsfor Italyand the United States are out-
lined in Pearce (1976), and three differentmodels for
Japanare describedin OECD(1977).Theyhavebeenused
to assess the effect of abatement policieson variables
such as price levels, consumption,investment,produc-
tion, employment,and the trade balance.
Data on pollution abatement outlays by government
could be obtainedfrom a functionalclassificationof gov-
Table 5-3. Input-Output Table Isolating Pollution Abatement
Producers
Producers______________ |Government
1_________
f 2 Total consumption
PA Other PA Other PA Other
Commodities
I A B C
2
Total __ _
indirectices . . .
Rest of world l l l l
Note: PA = pollutionabatementcosts.
emient expenditures.The UnitedNations"Classification
of the Functionsof Government"(COFOG) identifiesout-
layson both monitoringand abatement.Data on pollu-
tion abatement expendituresby households, mainly in
connectionwith vehiclesand home heating, are usually
obtainedfromproducersrather than directlyfrom house-
holds. There are some conceptualproblemsin defining
abatementexpendituresby householdsthat also applyto
expendituresby industries. These problemsare consid-
ered below.
To date the most rigorous, or at least the best-docu-
mented,effortsto compilestatisticson "pollutionabate-
ment and control expenditures"(PACE)by industryhave
been made in the UnitedStates by the Bureau of Eco-
nomic Analysis.The problemshave been discussedby
Cremeans(1974 and 1977).
First, there is what has been describedas the baseline
problem.This is the difficultyof establishinga baseline
fromwhichto measurethe extent of pollutionabatement
activities.Longbeforethe environmentbecamea fashion-
able topic,producers,households,and governmentswere
controllingpollutionto some extent. Shouldantipollution
activitiesbe measuredfrom a zero baseline,so that they
covertraditionalas wellas newerabatementprocedures,
or shouldthe baselinebe brought forwardto includeonly
recentlyintroducedcontrol procedures?The second ap-
proach has been adoptedin the United States; in effect
pollutionabatementis definedas activities,like scrubbing
smokeor coolingwastewater,whichare undertakenbe-
cause of recent antipollutionlegislation.The use of a
mobilebaselinemakesthe definitionof abatementexpen-
ditures somewhatfuzzy,but it is probablythe only prac-
tical approachfor data collection.Respondentswillhave
some notion of the extra outlaysthey have incurred be-
cause of recent antipollutionlegislation,but they would
find it very difficultto estimate their costs if they were
allowedto cause unlimitedpollution,becausethey have
neverbeen in this position.
Costs
i Gross fixed T
Private capital
consumption formation Stocks Exports Totals
PA Other PA Other
D E I
 MeasuringPollutionwithinthe Frameworkof theNationalAccounts
The second difficultyis the problem of joint costs.
Some equipmentand industrialprocessesare both good
for business and less polluting. In such cases it is not
clear how much of the additionalcost of the equipment
or process should be treated as a pollutionabatement
expenditure.In the UnitedStatessurvey,respondentsare
askedfor their "best estimatesof the cost of that portion
of plant and equipmentexpendituresattributableto spe-
cial features for [air and water] pollution abatement"
(Cremeans1974, p. 62). Respondentspresumablybase
their best estimateon the costs of alternativemodelsthat
producethe same output as the equipmentacluallyused
but that lack the model'sspecialantipollutionfeatures.In
the unlikely,but not impossible,eventthat a new pieceof
equipmentis as productiveas an older model but less
pollutingand less costly,it wouldbe reasonableto attrib-
ute a negativecost to the antipollutionactivity.The U.S.
surveydoes not attempt to identifysuch cases, perhaps
because of the difficultyof makingaccurate estimates.
However,new,less-pollutingprocessesmaysometimesbe
cheaperthan older,more-pollutingones.An often-quoted
exampleis the switchby pulp and paper manufacturers
fromsulfiteto sulfateprocesses.Sincethe new methodis
both less pollutingand cheaper,the pollutionabatement
involvedhas a negativecost.
A third difficultyconcernsthe respondent'sawareness
of antipollutioncosts. For example, if new legislation
forbidsopen-castmining,the additionalexpenseof shaft
mining should be counted as an antipollutien cost. In
general, however,respondentswould probablyoverlook
costs of this kind and would in any event have great
difficultyin estimatingthem.Similarproblemsarisewhen
antipollutionpressuresresultin demandshiftsfrommore-
to less-pollutingproducts. Examples here include the
switch to unleaded gasoline,biodegradablezontainers,
and low-phosphatedetergents.These goods are all more
costlythan the more-pollutingitemsthey replace,but the
producergenerallywouldnot regardthe extra investment
and higher operating costs involvedin their production
as a pollutionabatementcost. Indeed,in the IJ.S. survey,
such costs are explicitlyexcluded,and manufacturersare
supposedto report only the outlaysincurred in reducing
pollutionin their ownproductionprocesses.
Becauseof these problems,some criticshavesuggested
that it is futileto try measuringbusinessoutlayson pol-
lution abatement.However,this positionmaybe too ex-
treme. In attempting to compile the PACE series, the
Bureau of Economic Analysisis respondingto a real
concernfelt by policymakersand business people, who
need to know the costs involvedin differentpollution
abatementpolicies.What is open to questionis whether
the approachused by the bureau is the best wayto meet
this need. There are two lessonsto be learned from the
U.S. experience.First, it maynot be usefulto compilea
time series for PACE statistics.In someyearsseveralwide-
29
rangingantipollutionlawsmaybe enacted,and, whenthis
happens,it is certainlyusefulto measure their effecton
industrialcosts.In other years,however,there maybe no
new legislationrequiringoutlayson pollutionabatement,
and there is then no obviousneed for a survey.In short,
the aim should be to measure PACE in response to a
particular piece of legislationin particular years, rather
than to considera continuoustime series.
The second lesson is that the conventionalsurveyap-
proach-mailing questionnairesto a large number of re-
spondents-may not be suitable for PACE statistics. As
noted, the U.S. questionnairerequires respondents to
estimate some importantcomponentsof PACE, but there
are some pollutionabatement costs that haveto be ig-
noredbecauserespondentsare unawareof them.In these
circumstancesit maybe better to investigateintensivelya
smallnumber of establishments.The larger samplinger-
rors probablywouldbe offsetbythe smallernonresponse
errors, and the enumerators could collect information
about abatementcosts that the respondentsdo not rec-
ognize as such. In addition,serious pollutionproblems
are confinedto a relativelysmall number of industries:
home heating, transport, electricitygeneration, chemi-
cals, paper and pulp, and steel. Thus,an intensiveinves-
tigation of relativelyfew establishmentsmay yield accu-
rate data on the bulk of pollutionabatementoutlays.
Benefits of Pollution Abatement
One of severalbenefitsthat may result from pollution
abatementis the recoveryof usablematerials.In practice,
such benefitsare usuallyquite small, and, if they can be
identifiedat all, the most convenientwayto treat them in
a national accounts frameworkis probably as negative
abatementcosts. Far more importantare improvements
in health, enhancedenjoymentfrom the natural environ-
ment, and related benefitsthat the communityat large
derivesfrom pollutionabatement.There are severalways
to valuethese indirectbenefits.
One possible approach is to consider the damages
avoided.Thus, if it could be establishedthat a particular
type of pollutioncaused1,000 dollars'worth of damage,
then it could be argued that abating this particular pol-
lution providesbenefitsworth the same amount.Several
authors have indeed assumed that damage costs and
abatementbenefitsare two sides of the same coin. Jaszi
(1971, p. 224), for example,doubts the possibilityof
measuringthe benefitsof environmentalprotection be-
cause he "cannot formulatea question relating to the
damagesof pollutionthat could be answeredby knowl-
edgeable respondents in a meaningfulway" (emphasis
added).The valuationof goods and services,however,is
not generallybased on the assumptionthat the benefits
receivedequalthe damagesavoided.Clothingis not val-
ued at the cost-in terms of social ostracism or legal
30 Derek W.Blades
penalties-of goingaround naked, and food is not valued
at the cost of starvingto death. In conventionalnational
accounts,goods and servicesare valued in one of two
ways.Goodsand servicesthat are sold on the marketare
valued at their market prices, that is, at prices that are
assumedto measurethe purchaser'svaluationof the ben-
efitsderivedfromthem,whilegoodsand servicesthat are
providedfree or at heavilysubsidizedprices are valuedat
their cost of production.
The outputs ofpollutionabatement-clean air and pure
water,for example-obviouslycannotbe sold on the rnar-
ket. So the natural approachfor the nationalaccountant
is to value the benefits of pollution abatement as being
equal to the costs of pollution abatement. If, as was
suggestedin the previoussection, abatementcosts were
shownseparatelyin the national accounts,users would
have all the informationneeded to valuethe benefitsof
pollutionabatement in preciselythe same way that the
benefitsof other publicgoods-national defenseor public
administration,for example-are valued.
This simple,orthodoxmethodof valuationmaybe crit-
icized on the grounds that the market, or the community
as a whole, may not necessarilyvalue the benefitqof
abatementat the costs of abatement.The same objection
may,of course, be made against valuingother nonmar-
keted goods at cost, and several authors have indeed
argued that all goods and services in the national ac-
counts should be valuedat market prices regardlessof
whetheror not they are actuallymarketed.Peskin and
Peskin(1978)havesuggestedhowthis mightbe done for
pollutionabatement.
The problemmaybe demonstratedin Figure5-1. The
demandcurve (D) shows how much air cleanlinessthe
market would buy at differentprices, while the supply
curve (S) givesthe price of various levelsof air quality,
that is, the cost of abating air pollution.If clean air were
sold on the market,the amountsuppliedwouldbe deter-
mined by the intersectionof D and S. The air supplied
wouldbe the moderatelycleanA2 type,the publicwould
buy it at price P2, and the valueof purchases-the num-
ber that wouldbe recordedin the national accountsun-
der householdconsumptionexpenditures-is represented
by the area 0, P2, S2, A2.
In practice, of course, the level of air pollutionis set
not by the market but by governmentfiat. A government
sensitiveto businesspressuresmightdecideon lowabate-
ment standards that result in air of the rather dirty A,
type, whilea governmentdominatedby consumerinter-
ests mightlegislateto produce very pure A3 type air. In
the first case, the number that goes into the national
accounts(as pollutionabatementcosts) is representedby
the smallarea 0, P1, SI, A,. However,the demandcurve
indicatesthat the marketwouldhavepaid Pi' rather than
P, for the amountof pollutionabatementrepresentedby
Al, so that area P1, P1 ', D1, S, maybe taken as a measure
of the extent to whichthe "at cost" valuationof pollution
Figure5-1. Supply and Demand for CleanAir
Price
Demand(D)
\
PI
P3
l_
P2
SI D3
P3
A, A2 A,
(dirty) Index of air quality - (clean)
abatementunderstatesthe benefitsof abatementasjudged
by the market. In the second case, the "at cost" value
representedby 0, P3 , S3, A3 is much higher than the
benefitsas judged by the market. For most people, type
A3 air is not worth the cost; the market wouldpay only
P3 ', and the area P3 ', P3 , S3, D3 represents the amount
by whichthe "at cost" valuationof abatementoverstates
the market valuationof the benefitsof abatement.
To measurethe marketvaluationof abatementbenefits,
it is necessaryto estimate the demand curve in Figure
5-1. The direct approachof askingpeoplehowmuchthey
would pay for differentamounts of abatementdoes not
appear to have been widelyused. One problem is that
shrewdrespondentsmaydeclareno interest in payingfor
any pollutionabatement because they know that since
clean air and pure water are public goods, they could
never be excludedfrom enjoyingthem. This causes the
marketvaluationof abatementbenefitsto be understated.
Anotherdifficultyis that it is not clear whetherrespon-
dents should be treated as buyersor sellers of clean air
and water (Seneca and Taussig 1974). Should they be
asked howmuch they wouldpayto enjoya givenlevelof
pollution abatement, or how much they would charge
polluters for the use of their clean air and water? The
latter approach, which is based on the view that the
inhabitantsof a countryhavecollectivepropertyrights in
their environment,wouldgenerallylead to a highervalu-
ation of abatementbenefits.To be realistic,however,it is
unlikelythat either question could elicit anything but
fancifulresponses. People are simplynot very good at
puttingprices on thingsthat are never marketed.
 MeasuringPollutionwithinthe Frameworkof the NationalAccounts 31

In these circumstances,variousindirectapproachesmay total damagescaused by pollution, they would indicate

be used.For example,the benefitsof cleaninga lake may the successof abatementprogramsin reducingpollution
be valued by finding out how much people spend at damagesovera period of years.
presentto travelto the nearestunpollutedlake. This and Nationalaccounts have been widelyused to measure
other indirect methods used in connectionwith water the costsof pollutionabatement.This is clearlyan impor-
pollutionare summarizedin Kneese(1971).The benefits tant use, since policymakersneed to know the cost of
of reducingtrafficnoisemaybe valuedby comparingthe alternativeabatementmeasures.Althoughthere are some
prices of housessubjectto differentnoise levels.Several problemsin definingand measuringabatementcosts, in
studies on airport noises in the United Kingdom are generalthe aimshouldbe to measurethe abatementcosts
discussedby Pearce (1976). incurredbyparticularindustriesin responseto particular
The results of such studies have been used in some pieces of legislation.For this purpose, it willnot usually
cost-benefitanalysesrelating to particular problems at be necessaryto constructtime series statisticson abate-
particular times. In viewof the sporadicnature of such ment costs. To collectcost data from industry,in-depth
studies and of the controversythat has sometimessur- investigationof a small sample of firms may be more
roundedtheir findings,however,it is unrealisticto sug- effectivethan conventionalsurveymethods.
gest that market measuresof abatementbenefits could
be includedin annual national accountsstatisticswithin References
the foreseeablefuture. Valuationat cost is all that can
reasonablybe achievedat present. Ayres,R. U., andA.V.Kneese.1969."Production,
Consump-
tion,and Externalities."
American EconomicReview 59(3,
Conclusions June):282-97.
Councilon EnvironmentalQuality.Annual.Environmental
This chapter has reviewedwaysin whichnational ac- Quality.Washington,
D.C.
counts statisticscan be used to assess and manage a Cremeans,J. E. 1974."CapitalExpenditures
by Businessfor
major environmentalconcern-pollution. More specifi- Air and WaterPollutionAbatement,1973 and Planned
cally, it has examined how the national accounts can be 1974."Survey of CurrentBusiness54(7,July):58-64.
used to study four aspects of pollution: the output of . 1977."Conceptual
and StatisticalIssuesin Developing
pollutants,the damagescausedby pollution,the costs of EnvironmentalMeasures:
RecentU.S.Experience." Review
abatingpollution,and the benefitsderiveclthierefrom. of Incomeand Wealth23(2,June):97-116.
The use of national accounts to study the output of Jaszi,George.1971. "Review."
Surveyof CurrentBusiness51(7,
pollutantswillhave a low priority in most countries.In- July)(PartII, 50thAnniversary
Issue):183-227.
formationon quantitiesof pollutantsemittedmust be site- Kneese,A.V.1971."TheBenefit-CostAnalysis
of Environmen-
specificin order to be used to monitorpollutionlevelsor tal Pollution."In Proceedings
and Documentation of the
to designcontroltomeasures. Nationaltotals forte output
Ntionl toals or the United
masurs.
dsigncontol ouput sium onNations Economic
Problems Commission
Relatingto forNew
Environment.Europe Sympo-
York:United
of pollutantsare of limitedvaluefor environmentalanal- Nationso
ysis.It is unlikelythat the nationalaccountscan be used Nations.
OECD (Organisationfor EconomicCo-operationand Develop-
to measure the benefitsof pollutionabatement,at least ment).1974.EconomicImplications of PollutionControL
on a "market" basis. Although it may be theoretically Studiesin ResourceAllocation. Paris.
possibleto attach market valuesto the benefitsof abate- _____ 77.Environmental Policiesin Japan.Paris.
ment, there are enormouspracticaldifficultiesin doing Pearce,D.W.1976.Environmental Economics. Londonand
so. For the foreseeablefuture the only way the national NewYork:Longmans.
accountscan be used to valuethe benefitsof abatement Peskin,H.M.,andJ. Peskin.1978."TheValuation of Nonmar-
is through the cost side. This method is used now by ket Activitiesin IncomeAccounting." Reviewof Incomeand
nationalaccountantsto valueother publicgoods,such as Wealth24(1,March):71-93.
nationaldefenseand public administration. Saunders,C. T. 1977."The Feasibilityof Welfare-Oriented
One can take a more positiveattitude toward other Measuresto Supplementthe NationalAccountsand Bal-
aspectsof pollution.Althoughit is difficultto define and ances:A TechnicalReport."Studiesin Methods,SeriesF,
measurein full the costs of damagescausedby pollution, no.22.NewYork:UnitedNations.
it seems feasibleto identifysome of the costs that are Seneca,J. I., andM. K.Taussig.1974.Environmental Econom-
alreadyimplicitlyincludedin the nationadaccounts. Ex- ics.Englewood Cliffs,N.J.:Prentice-Hall.
ampleshere includepollution-related medicalexpensesin UnitedKingdom AtomicEnergyAuthority. 1971.An Economic
the final expenditures of governments and households as and TechnicalAppraisalof Air Pollutionin the UnitedKing-
well as outlaysfor pollution-causedrepair and mainte- dom.London:HerMajesty's
StationeryOffice.
nance in the intermediate consumption of industries. Al- Victor,Peter A. 1972. Pollution:Economy and Environment.
though these expendituresconstitute only part of the London:AllenandUnwin.
 CorrectingNationalIncome for EnvironmentalLosses:
Towarda PracticalSolution
Roefie Hueting
An increasein nationalincomeis often called economic
growth.It is identifiedwith increased welfareand con-
ceivedof as an indicatorof economicsuccess.However,
this terminologyis incorrect for a number of reasons.
One of these is that the apparent increase in national
incomemaybe accompaniedby the degradationof a vital
scarce good: the environment. The misconceptionof
terms reflectsthe beliefthat thingsgo well,economically
speaking,solelywhen production,as measuredin gross
domesticproduct (GDP), increases. This belief also in-
cludesthe convictionthat an increasein GDP is necessary
to solve the environmentalproblem, since this creates
room for financingthe conservationand restoration of
the environment.However,this propositionis questiona-
ble to the extent that environmentaldeteriorationis a
consequence of the prevailing pattern of production
growth.
The environmentis being endangered because eco-
nomicpolicystressesthe increasein productionas meas-
ured by national income,but it neglectsto account for
the depletion of natural resources and damage to the
environment. This depletion, in turn, endangers sus-
tained economicdevelopment.This gavea great impulse
to the ideas of Kuznets,whosuggestedas early as forty
years ago that national income be correctedto provide
politiciansand the public with a better insightinto the
developmentof welfareovertime (Kuznets1947, 1948a,
1948b).
A bigstep in the right directionwouldbe possibleif all
the factorsthat influencewelfarecould be expressedin
one commondenominator(for example,money).For the
environment, this would involveconstructing shadow
prices for environmentalfunctions(possibleuses of the
environment),which are directly comparable with the
prices of goods and servicesproduced for the market.'
Withthe aid of these, the valueof the losses of environ-
32
mentalfunctionscould then be establishedand entered
in the Systemof NationalAccounts(SNA) as cost items
whenthey occur.
Unfortunately,shadowprices for environmentalfunc-
tions can be constructed only in exceptional circum-
stances.It is possiblein principleto estimatethe cost of
measuresthat lessen the pressure on the environment
and thus preservethe desired uses of the environment.
These so-calledeliminationcosts may be regarded as
productioncosts of an environmentalfunction. With the
aid of calculationsof the costs of various degrees of
elimination,supply curves for environmentalfunctions
can be constructed.Only in a few cases, however,is it
possible to construct demand curves, because the de-
mandforenvironmentalfunctionsusuallycannot manifest
itself in market behavior.Therefore,other options need
to be considered.But as long as the traditional, uncor-
rected concept of GDP is used, many caveatsneed to be
kept firmlyin mind.
The Need to Use GDP with ProperCaveats
Whilework is goingon to correct the measurementof
GDP, one recommendationshouldbe implementedimme-
diately:proper caveatsshouldbe used and additionalfacts
should be presented in the yearly publicationsof the
nationalaccounts.
Thus, it has to be stressedthat an increasein national
incomeis not identicalto economicgrowthand economic
success.Further,production,as measuredby GDP, is only
one factor that influences the level of welfare. Other
factorsare the environment(includingnature and natural
resources), leisure time, income distribution, working
conditions,employment,and the safetyof the future in-
sofar as it depends on our dealings with scarce goods.
Therelationbetweenproductiongrowthand environment

in the past must also be madeclear. The burden on the
environmentis equal to the product of the number of
people and the amountand type of activityper person.
Sustainableeconomicdevelopmentthereforerequiresthat
the nature of our activitiesand the number of our species
be adapted to the carryingcapacity of our planet; such
adaptationsmaydecrease GDP in the short run but will
providesustainedeconomicdevelopmentin the long run.
It shouldalso be kept in mind that the SNA'S mainvalue
is for short-termmacroeconomicanalysis.Finally,market
prices do not indicatethe marginalutilityof goodswhen
their productionand consumptionare at the expenseof
scarce environmentalfunctions.
In additionto such caveats,as much quantitativeinfor-
mation as possibleshould be providedon the losses of
environmentalfunctionsand their consequences.To the
extent possiblethis should be done in rrmonetary terms,
such as compensationand eliminationcositsand financial
damage.This can be supplementedby informationex-
pressed in physicalunits, such as emissions,concentra-
tions, fragmentationof the countryside,and losses of
species.The figures in physicalterms can be related to
standards for environmentalfunctionsand will thus de-
scribe their availabilityor loss. For exarnple,if certain
thresholdsof emissionsand subsequentconcentrations
are exceeded,the function of surface water as a basic
materialfor drinking water is lost. The informationin
physicalunits could be collectedin so-calledsatelliteac-
countsof the nationalaccounts(Theys,Chapter7).
The recommendationsmade above-that proper cave-
ats be used and that GDP be supplementedby monetary
and physicalinformationon the environment-can and
should be implemented immediately.T'he rest of this
chapter reviewsthree optionsfor correctingC.DP.
Correcting National Income for Defensive
Environmental Outlays
Whennational incomeis calculatedaccordingto pres-
ent conventions,severalactivitiesthat have a cost char-
acter, and thereforeought to be entered as intermediate
deliveries,are currentlydesignatedas finadconsumption.
Kuznets,one of the great theoreticiansof the conception
of nationalincome,emphasizesthis point (Kuznets1947,
1948a, 1948b).
To the three classes of expendituresdistinguishedby
Kuznets (expenditureinvoked by an urban pattern of
living,expenditureinherent in participationin the tech-
nicallyand monetarilycomplexcivilizationof industrial
countries,and the majorpart of governmentactivity),the
expenditureon measuresthat compensatefor, redress,or
guard againstlossesof environmentalfunctionscouldbe
added.Theseexpendituresare currentlye:nteredas inter-
mediatedeliveriesinsofaras the measures,are taken, and
directlypaid for, by privatefirms.Whenthe measuresare
CorrectingNationalIncome for EnvironmentalLosses 33
paid for by the governmentor private households, or
whentheyare takenby privatefirmsbut financedthrough
levies imposed by the government, they are currently
treated as finalconsumption.
However,all these outlaysshould be entered as inter-
mediateexpenditurewhen a long time series such as that
for national incomeis composed.For the losses of envi-
ronmentalfunctionsare currentlynot entered as costs at
the time they occur, and thereforetheir restoration or
compensationshouldnot be entered as a finaldeliveryat
the momentit is undertaken. This procedure cannot be
changed because it is impossibleto construct shadow
prices.It is thereforedoublecountingto enter outlaysfor
compensatingthese lossesor for eliminatingthem as final
consumption.Expenditureon preventingand compensat-
ing environmental lossesdoes not contributeto the quan-
tity of consumergoods; these outlaysprotect or replace
scarce environmentalgoods that were already available.
(Someof the disadvantagesattributedby Kuznetsto the
urbanpattern oflife,such as the greater distancebetween
home and work, maybe interpretedas losses of function
resultingfrom spatial competition,and the expenditure
on this as the cost of compensatorymeasures.) 2
Most authorities in this field, including most partici-
pants in the uNEP-WorldBank workshops on environ-
mental accounting,favoreither merely isolatingthe de-
fensive outlays for the environment or isolating and
correctingfor these items in an alternativepresentation
of national income alongsidethe current one. In both
cases, the time series would not be interrupted. The
current figure would remain intact for analytical pur-
poses, but the interactionbetween production and the
environmentwouldbe better understood, and the ability
to provide informationin physicalunits and to collect
these data in satelliteaccountswould,of course, not be
hamperedat all.
The approach of isolating and correcting, however,
seemspreferable.The politiciansand the public,whoare
now givenonly one figure of income,wouldthen realize
that part of the increasein nationalincomeconsistsof an
increase in cost, since defensiveenvironmentaloutlays
are increasing(Leipert 1987). If the income correction
was made, however,severalclarificationswould have to
be made in the introduction of national accounts
publications.
* The correctionis onlypartial, since the greater part
of environmentallossesare neither restored nor compen-
sated.
* The mutations in both the current and corrected
figures of GDP do not represent economicgrowthor the
courseof welfareovertime.
* Reallocationby shiftingthe productionand consump-
tion packageon behalfof the environment(for example,
more bicyclesand publictransport and fewerprivate cars
34 Roefie Hueting
becauseof the pressureof physicalrestrictionsand levies)
constitutesa sacrifice.This is not reflected in the SNA,
insofar as the pattern of consumptionthat is thus pro-
duced differsfrom the first choice,whichwasmadewith-
out regardto environmental considerations.The shiftswill
often lead to a check on productiongrowth.This is auto-
maticallyreflectedin the SNA but is not visible.
* Isolationand correctionformpart of a broader field
of defensiveoutlays(see Hueting1980;Kapp1963; Kuz-
nets 1947, 1948a, 1948b, 1954; Leipert 1986a, 1986b;
Mishan1984;and Tinbergen1985).
ComplementingCorrectionsfor Defensive Outlays
through Surveys
The possibilitiesfor preferencesfor the current and
future use of environmentalfunctionsto manifestthem-
selvesin marketbehaviorare verylimited(Hueting1980).
Thereforeeffortshave been madeto trace these prefer-
ences byaskingpeoplehowmuchtheywouldbe prepared
to pay to whollyor partially restore lost environmental
fts th
being
dfncton willingnessto pay.cHoweverm this method does
neoalwillinnes
not tolpay.eStimateS
alwaysprovidereiable However
forthis
manymethods.
reasons.
Informationon the significanceof environmental
functionsis deficientin manycases.This is especiallyso
for the functionsthat determinethe future qualityof the
environment.For these (life-support)functionsthere is
oftenthe risk that interruptingcomplicatedprocesses,for
instance,ecosystems,maylead to seriousovershootsand
collapse,and the chance that technologies not yet in-
vented or operationalmay cope with those risks. Many
peoplemaynot be able to weighthese risksand chances,
and thus to answerhowmuch they are prepared to pay
for avoidingthem.Accordingto the biologicalliterature,
the possibilitythat overshootsand collapsemay occur if
the growthpattern of production(and population)is not
changedconstitutesthe most importantpart of the envi-
ronmental problem (see, for example,Odum 1971). If
individualsare not awareof the importanceof an environ-
mentalfunction,the surveymethodis pointless.
* There is a considerabledifferencebetween saying
that one is willingto spend money on somethingand
actuallypayingfor it.
* The questioningmethodin fact tries to approachthe
valueof a collectivegood as ifit werea privatemark.etable
good(bytryingto findsomepointson the demandcurve).
In a market the bidder knowsfairlywellwhat qualityand
quantitycan be acquiredby differentbids. In a collective
situation, however,this is not possible,becauseit is not
knownhowmuch other people are goingto bid. Without
a considerableamount of additionalresearch it is also
not knownhow much moneyis requiredto attain differ-
ent qualitystandardsfor the environmentalfunctions.
* In order not to make the questioningunjustifiably
vague,some research on environmentalaccountinghas
to be done beforehand.For clear air, clean water,and so
forth are not homogeneousgoodsfroman economicpoint
of view,as water and air have quite a few differenteco-
nomic functions.If the persons being questionedare to
have a clear picture of the issues, they must be given
informationon the significanceof the differentfunctions,
the consequencesof their loss, and the measures and
costs involvedin their restoration.All together this con-
stitutes a huge amount of information,whichwould not
be easyto survey.Althoughthe willingness-to-pay method
might be justified for one or two factors affectingthe
immediatelivingconditionsof people asked, it is most
probablynot a sound base for correctingnationalincome
for all environmentallosses.
* Peoplemaybe interestedin the effectsof their bids,
together with the (unknown)bids of others, on, for in-
stance,.employmentlevelsand consumptionpatterns. For
answeringlegitimatequestionsabout this, studiessuch as
the scenario study mentioned belowhave to be elabo-
rated, and the resultshaveto be presentedto the persons
questioned.This hardlyseemsfeasible.Again,what might
be justifiedon a micro scaleis most probablynot justified
on a macro scale.
* Askingpeople howmuch they are prepared to pay
suggeststhat conservingthe environmentalwaysrequires
extra provisionsthat must be paid for. In quite a few
cases, however,conservationis a matter of refraining
from doing things rather than of doing them, and this
savesrather than costsmoney.Thus,not buildinga road
through a mountainousarea that is vulnerableto erosion
is cheaper than buildingit, cyclingis cheaperthan driv-
ing, wearing a sweater and using an extra blanket is
cheaper than raisingthe temperature,and confiningthe
consumptionof lettuce to the summerseason is cheaper
than eating it throughout the year. People who realize
this maymodifytheir answersbecauseof such considera-
tions.
* Some peoplewillprobablybe convincedthat it does
not matter what theybid becausetheir bid willnot influ-
ence environmentalpolicyat all, and this convictionwill
influencetheir bid.
* Some people may think they have a "right" to a
healthy and safe environmentand will probably react
accordinglyby not makinga bid at all.
a Some people willprobablyhave their doubts about
the participationof others (the Prisoner'sDilemmafrom
gametheory) or prefer to waitand see (the Free Rider
Principlefromthe theoryof collectivegoods).
* The environmentis an important collectivegood.
But it is not the only one. Dikes, public administration,
and the army are too, while police and educationhave
clear collectiveproperties.To be sure of not exceeding
budget restrictions,people also have to be questioned

about howmuch they are prepared to pay for the other
collectivegoodsof the society.This hardlyseemsfeasible.
Again,what might be justifiedfor one or two separate
environmentalagentson a micro scalemaybe impossible
on a macroscale.
* The willingness-to-paymethod also measures the
consumer'ssurplus.In national incomethe total valueof
the goods is found by multiplyingthe quantity of each
good by its respectiveprice and then addingtogetherthe
resultingamounts.Usingthis procedure,the consumer's
surplusis not expressedin the levelof nalional income.
Thus, a doctor who savesa patient's life creates a value
that, whateverone may think about its exact size, is
certainlyhigherthan the valueaddedrecorcledin national
income.The intra-marginalutility of goods,whichis ig-
nored in nationalincome,willapproachan infinitevalue,
becauseit includes the utility of the first unit of food,
drink, and so forth. For this reasonthe resultsof willing-
ness-to-payresearch are not suitableto be used in con-
junction with the figures of national incc,me.An addi-
tional objectionto incorporatingthe consumer'ssurplus
in the willingness-to-pay approach is that the resultsre-
flect the income distributionmore directlythan do the
prices of market commodities;the differencesbetween
rich and poor in the weightsof the "votes"becomegreater
whenthe consumer'ssurplusis included.The occurrence
of differencesin weightsof "votes" is often defendedby
the argumentthat the contributionto the nationalpack-
age of goods and servicesby the rich is greater than by
the poor: their incomesare higherbecauseof the greater
relativescarcityof their abilities.This argument is not
validwith regard to the environment,because it is not
producedby humanity.
Becauseof the limitationsmentionedabove,the willing-
ness-to-paymethoddoes not presenta firm enoughbasis
for correctingnationalincome for lossesof scarce envi-
ronmentalfunctions.A correctionbased on this method
may
lead to inaccurate estimates of envi.ronmentalde-
mlea.
mine,
Complementing Corrections for Defensive Outlays
through Standards for Sustainable Economic
Development
It can be concludedfromthe abovethat the correction
of nationalincomeis highlydesirable.But shadowprices
for environmentallossesthat are directlyccmparablewith
market prices (on which national income figures are
based)can be constructedonlyexceptionally, becausethe
intensityof the preferencesfor the present and future
availability
of environmentalfunctionscan mostlynot be
measured. For this problem there is a well-defensible
practicalsolution.This consistsof supplementingthe cor-
rections for defensiveenvironmentaloutlayswith esti-
CorrectingNationalIncomefor EnvironmentalLosses 35
mates of the expenditureon the measuresrequired to
meet physicalstandards for the availabilityand qualityof
environmentalfunctions (includingnature and natural
resources).4 The standards can in turn be based on the
prerequisitesof healthand a sustainableeconomicdevel-
opment,that is, developmentwithsustainableyields.This
seemsthe most naturalguidelinefor setting standardsfor
environmentalfunctions. Standards and the measures
based on them mayreferto the occupationof space, the
use of soil, the availabilityof stocks of natural resources,
the compositionof products, the consumptionof raw
materialsand energy,the emissionof pollutants,and the
concentrationof chemicaland other agents.
The standards can be related to environmentalfunc-
tions. Thus it is possibleto formulatethe wayin whicha
forestshould be exploitedto attain a sustainableuse of
its functions.Thesefunctionscouldincludebeinga source
of wood, regulatorof the water management,preventer
of erosion, bufferof carbon dioxideand heat, regulator
of the climate,gene reserve,objectof studyfor ecological
research,supplierof natural productsfor the local popu-
lation,and sourceof incomefromtourism.The estimated
expenditureon the measuresrequiredto meet those stan-
dards then indicatesin monetaryterms howfar a nation
has drifted awayfrom its (supposed)end or standard of
sustainableuse of its forestresources.
Likewiseit is possibleto formulatethe way in which
surfaceand groundwatershould be exploitedto arrive at
a sustainableuse of its functions.These functionscould
include water for drinking,for agriculture,for cooling,
for flushing and transport, for industrial processes,for
recreation,for navigation,in the natural environment,
and forwastedumping.The estimatedexpenditureon the
measuresrequired to meet those standards would indi-
cate in monetaryterms howfar a nation has drifted away
from a sustainableuse of its water resources.The same
holdstrue for the use of air, soil, and space.
The measuresmayrange fromselectivelycuttingtrees,
reforesting,buildingterraces,drainingroads, maintaining
buffersin the landscape,and selectivelyusing pesticides
and fertilizersto buildingtreatment plants, recirculating
materials, introducingflow energy, altering industrial
processes,makingmore use of public transport and bi-
cyclesinstead of private cars, and using space so that
plant and animalspecieshave enough room to survive.
Whenput into practice,these measureswilllead, in one
wayor another,to different-sustainable-productionand
consumptionpatterns. It followsfrom the earliersections
in this chapter that the levelof welfarewillnot decrease
if people prefer to maintain a sustainabledevelopment
rather than to risk the overshootsand collapseinherent
in the present production and consumption patterns.
Whetherthis is the case cannot be provedor disproved,
becausein manycasesit is impossibleto constructshadow
pricesfor environmentalfunctions.
36 RoefieHueting

Table6-1. Comparisonof Traditionaland EnvironmentalScenarios

(1980= 100)
Traditional Environmental
Aspects growthscenario scenario
Plant and animalspecies Further deteriorationby, among Checkeddeteriorationor even
other things,increasedproductivity improvement
in agriculture,more use of space,
and more use of persistentcom-
pounds;onlymeasuresagainstpol-
lutionhavebeen taken
Emissions
Air pollution
Sulfurdioxide(SO2) 154 24
Nitrogenoxide(NOr) 130 76
Particulatematter 310 58
Hydrocarbons 266 48
Carbonmonoxide(CO)(total) 290 31
Processes 75
Stationary - 48
Transport 36 21
Carbondioxide(CO 2) - 83
Lead 44 0
Waterpollution
Mercury 53 29
Cadmium 79 36
Otherheavymetals 109 57
Persistentorganiccompounds - 7 0b

Phosphates 81 29
Oxygendemand 35 17
Thermalload 139 44
Solidwaste 152 108
Harborsludge - 100
Energyrelated 1,400 467
Gypsum 210 90
Industrialslag 358 136
Chemicals 91 64
Radioactivewaste
Laboratorywaste 100 100
Nuclearenergywaste 300 0
Reprocessingwaste 598 0
Fissionablematerial 633 0
Energyinput 123 81
Coal 505(605)c 292
Uranium 463(0)' 0
Landscape Further deterioration(seeplant and Checkeddeteriorationor even
animalspecies) improvement
Livabilityof cities - Improvement
Production
CDP 164 127
Agriculture 181 135
Industry 321 177
Services 181 145
Availableincomeper capita 116 104
Unemployment 82 83
Employment 110 111
Tax burden 98 103
-= Not available.
a. Thisfigureis notconsistentbecausethecarbonmonoxide
emissions
of processes
andstationary
sourcesaremissing.
b. Veryroughestimate.
c. If nuclearenergyis notused,coalinputincreases.
Source:Hueting1987.

The expenditurenecessaryto carryout thesie(and sim-
ilar) measurescan be estimated.The sum ofthe amounts
found wouldindicatein monetaryterms howfar the na-
tion has drifted awayfrom a sustainableeconomicdevel-
opment. This indicatorcan be used in conjunctionwith
the figuresof the nationalincome,which,after correction
for defensiveoutlays,showshow far the nation has pro-
ceeded in raisingthe levelof production.
No measurescan be formulatedfor irreversiblelosses.
If plant and animalspeciesbecomeextinct,no restoration
measuresare possible.The same probablyholdsfor the
total loss of the topsoilof a mountainousarea. An arbi-
trary value then has to be assignedto these losses, of
whichone can say for certainonlythat the valueis higher
than zero.
A more or less similar problem seems to arise for
nonrenewableresources.Onceoil has been exploitedand
exportedor used, it is gone forever.This loss constitutes
a cost for the country,since oil reservesare finite.If no
alternativesare developed,the generation that experi-
ences the depletion (or a sharp rise in the operating
costs) of the resource willsuffer a severeeconomicloss.
Since the prospect of a safe and prosperous future for
one's childrenand one's children'schildranis a normal
humanneed, the diminutionof the reservesof nonrenew-
able resources such as oil constitutesa cost here and
now.
The depreciation of nonrenewableresources can be
valued by estimatingthe costs involvedin the develop-
ment and practical introductionof alternativessuch as
solar energy, substitutes for minerals, and recycling
methods.
An example of a practical applicationis a scenario
studyelaboratedfor the Netherlands(Hueting1987).This
study scans the effects on the level of production and
employmentof the transition to a sustainableeconomic
developmentbetween 1980 and 2000. The conditionsor
standards for such a developmentand the measuresto
meet these conditionsor standards were formulatedby
specializedscientificinstitutes. The estimated costs of
these measureswere enteredinto the econometricmodel,
whichthen scanned the effects on production and em-
ployment.The resultswere contrastedwith the resultsof
a scenario based on a traditional growthpolicy,elabo-
rated alongsidethe environmentalscenario. Thus, com-
parison of the outcomeof the two scenarioscan indicate
the differencesbetween a policy aimingat sustainability
and one aimingat increasingthe nationalincomeaccord-
ing to the present pattern. Because the environmental
scenario makes an importantstep in the directionof a
sustainableeconomicdevelopmentover the wholefieldof
productionand consumption,comparisonolfthe results
of the two scenarios also gives a rough impressionof
what may be expectedif GDP is correctedaccordingto
standardsfor a sustainableeconomicdevelopment.
CorrectingNationalIncomefor EnvironmentalLosses 37
The mainresults of the studyare shownin Table 6-1.
The total net expenditure(or cost of forgoneproduction)
to meet the conditionsor standards set in the environ-
mental scenario is equal to the difference in gross na-
tional product (GNP) betweenthe two scenariosin 2000.
The benefits(in regainedand safeguardedenvironment)
are equal to the differencesin the environmentalindica-
tors in 2000. A similarstudy is in progress for Taiwan
(Hueting1986a).
An example of a practical applicationof correcting
nationalincomefigures,albeitnot based on standardsfor
attaininga sustainableeconomicdevelopment, is the study
by the Economic Councilof Japan, which developsa
measure of net national welfare.In this approach the
nationalincomefiguresare correctednot only for actual
expenditureon compensationand restorationof environ-
mental functions,but also for the estimatedcost of the
measuresnecessaryto attain the level of environmental
quality prevailingin the base year (in this case 1955).
Thus, de facto, a correction of national income figures
on the base of standardswaselaboratedthere (Economic
Councilof Japan 1974,pp. 99ff).
De Groot(1986) can alsobe mentionedin this respect.
In that studythe ecologicalparametersare formulatedfor
the performanceand sustainableuse of, among others,
the functionsof water purification,fishery,gameanimals,
and nursery.The StimulationProgrammefor Researchin
HumidTropical Forest Lands (Tropenbos)is now con-
ductingresearch on the conditionsfor the conservation
and rationaluse of forests and forestland resourcesand
on howto bring these conditionsinto practice(Bax and
others 1986; Ross and Donovan1986; Grainger 1987;
Hueting1986b).Apart from these and other, more inte-
gral, approaches,numerousstudies notablyon the meas-
ures and costsinvolvedin reducingemissionsof harmful
agents have been carried out in quite a few countries.
Somehealthstandards havebeen establishedfor the con-
centrationof harmfulagentsallowedfor variousfunctions
of water and air, such as drinking,swimming,and physi-
ologicalfunctioning.
From these experiencesit seems feasible to correct
national income figures on the basis of standards for
health and a sustainable economic development. The
methodhas three advantagesand four drawbacksor im-
perfections.
The first drawbackis that the results of the approach
do not represent individualvaluationsin the true sense,
as has beenextensivelyexplainedabove.For,amongother
things, the intensityof the preference for a sustainable
economicdevelopmentcannot be measured.This simul-
taneouslyimplies,however,that the intensityof the pref-
erences for the acceptanceof the adverse effects and
future risks involvedin the present growth pattern of
productionand consumption,and thus for the growthof
GNP, is equallyunknown.Both of these aspectsshouldbe
38 RoefieHueting
clearly mentioned when the results of the method are
presented in the publications of the national accounts.
Second, the method ignores the loss of welfaresuffered
by people who have a strong preference for the survival
of plant and animal species apart from their role in main-
taining the life-support functions of our planet (which is
tainreruingthe
lfore-susofutaionabed lourpmenet).(whee ps
a prerequisite for a sustainable development). These pref-
erences could be compared with the preferences for cre-
ating and maintaining art or churches. This might not be
considered indispensable for sustainable development and
yields, but their loss would constitute a decrease in wel-
fare for those who appreciate them. The same holds true
for species that might not be considered indispensable
for a sustainable economic development. This too should
be mentioned when presenting the results.
Third, no measures can be formulated for irreversible
losses. This holds true for any method. Fourth, the
method is laborious.
The first advantage to the method is that it seems to
provide an acceptable supplement to national income fig-
ures by indicating the losses of environmental functions
in monetary terms, since shadow prices cannot be con-
structed.
Second, the method compels the definition of an exact
content of the term "sustainable economic development."
Without such a content the term remains vague and not
operational for economic policy regarding the environ-
ment.
Third, the physical data required for comparison with
the standards come down to basic environmental statistics
that have to be produced anyhow if we are to get a grip
on the state of the environment. The formulation of the
measures to meet the standards and the estimates of the
expenditure involved are indispensable for policy deci-
sions. In other words, the work necessary to supplement
national income figures might be laborious, but it has to
be done anywayif one wants to formulate a policy on the
environment,
Based on the arguments given above, the national in-
comefigures shudbecreceoinsufficient
come figures should be corrected for the losses of envi-
ronmental functions (including resources) on the basis of
standards for health and a sustainable economic develop-
ment. This would provide alternative national income fig-
ures alongside the existing ones. The differences between
the two will show politicians and the public in monetary
terms how far the country has drifted away fiom the
course of sustainable economic development.
Notes
1. Thischapter uses the conceptof environmentalfinctions,
which was designed to describe the economicaspect of the
environment.For an economicapproach,the environmentcan
best be interpretedas the physicalsurroundingsof humans,on
whichthey dependcompletelyin all their activities.Withinthe
environmentseveralpossibleuses can be distinguished.These
are calledenvironmentalfunctions.Whenthe use ofan environ-
mental functionby an activityis at the expense of the use of
another (or the same)functionby another activityor threatens
to be so in the future, loss of functionoccurs;lossesof function
constitutecosts. In this situation,which is called competition
between functions,the environmenthas an economicaspect.
Competitionbetween functionscan be qualitative,spatial, or
quantitative.Qualitativecompetitionamountsbroadlyspeaking
to pollution:the use of the environmentalfunction "dumping
ground for waste"is at the expenseof other functions.It is as
if there were an intermediatestep. An activityintroduces an
agent (forinstance,a chemical,heat, or noise)into the environ-
ment, as a result of whichthe qualitychanges;this maydisturb
other uses or render them impossible.In spatialand quantita-
tive competitionthere is not enough space or matter to meet
the existingdemandfor it. The use of a functionalso includes
the passiveuse of the function"natural environment"to con-
serve the actual and potentialutilitiesof ecosystems,nowand
in the future,and to retainthe diversityof speciesof our planet.
Competitionbetweenfunctionscan take all sorts of forms.But
in most casesit is a questionof the environmentbeingused for
productionand consumptionactivitiesat the expenseof another
desireduse (Hueting1980).
2. This chapter followsthe usual procedureof classifyingas
compensationcosts expendituressuch as extra provisionsfor
the drinkingwater supply,buildingswimmingpools when the
waterhas been polluted,or travelcosts to natural parks when
naturalareas in the vicinityhavedisappearedand of classifying
as financialdamagesuch effectsas diminishedproductivitybe-
cause of damageto crops and acceleratedcorrosionor devalu-
ation of properties caused by pollution. At the same time,
asking people how much they would be prepared to pay to
conserveand restorethe environmentis classifiedunderwilling-
nessto pay.Someauthors,however,placeall of these categories
underwillingnessto pay.
3. An overviewof the methodsused, includingquite a few
results(albeitnot in the contextof correctingnationalincome
figures) can be found in Kneese (1984), Ewers and Schulz
(1982),and Schulz (1985).For criticismof the method in the
environmentalfield,see Kapp(1972,pp. 17 ff).
4. Whenthe applicationof all availabletechnicalmeasuresis
to meet a standard, the estimatesmust be supple-
mentedwiththe marketvalueof the reductionin activities(in
additionto the measures)necessaryto meet the standard.
References
Bax, J., R. Hooghiemstra,and H. deWitte,eds. 1986.Initiating
a GlobalResearchProgrammeto Conserveand Develop
TropicalRainForests.TropenbosInformationSeries 1. Ede,
Netherlands:Tropenbos.
Economic Council of Japan. 1974. MeasuringNet National
Welfareof Japan. Tokyo.
Ewers,H. J., and WernerSchulz.1982. Die monetarenNutzen
gewasserguteverbessernder Massnahmen-dargestellt am
BeispieldesTegelerseesin Berlin.Berlin:Erich SchmidtVer-
lag.
Grainger,Alan. 1987. Report of the 7iel Consultation27-31

October1986. TropenbosInformationSeries 3. Ede, Neth-
erlands:Tropenbos.
Groot,R. S. de. 1986. "Functionsand SocialEconomicImpor-
tance of the Dutch Waddensea."Agricultural University,
Wageningen.Processed.
Hueting,Roefie. 1980. New Scarcityand EconomicCrowth.
Amsterdam:North-Holland.
. 1986a."Reporton MyVisitto Taipei as an Adviserfor
the Taiwan2000 Study."CentralBureauof Statistics,Voor-
burg,Netherlands.Processed.
. 1986b."Frameworkfor a Cost-BenefitAnalysisfor Dif-
ferent Uses of a HumidTropicalForest Area." Paper pre-
sented for the Tropenbosseminar,September22-24. Tro-
penbos,Ede, Netherlands.Processed.
. 1987. "An EconomicScenarioThat GivesTop Priority
to Savingthe Environment."EcologicalModeling38:123-
40.
Kapp, K. W. 1963. Social Costs of BusinessEnterprise.Lon-
don:AsiaPublishingHouse.
. 1972. "SocialCosts,Neo-Classical
Economics,Environ-
mentalPlanning:A Reply."SocialScienceInformation11(1):
17-28.
Kneese,A. V. 1984. Measuringthe Benefits of CleanAir and
Water.Washington,D.C.:Resourcesfor the Future.
Kuznets,Simon. 1947. "NationalIncomeand IndustrialStruc-
ture." Proceedingsof the InternationalStatistical Confer-
ences,Vol.V (TheEconometricSocietyMeeting,September
6-18, 1947,Washington,D.C.).Calcutta.
. 1948a."Onthe Valuationof SocialIncome."Economica
(February/May):1-16, 116-31.
CorrectingNationalIncomefor EnvironmentalLosses 39
. 1948b. "Discussionof the NewDepartment of Com-
merce IncomeSeries (NationalIncome:A New Version)."
Reviewof Economicsand Statistics30:151-79.
. 1954. "NationalIncomeand EconomicWelfare."In
SimonKuznets,ed., EconomicChange.London.
Leipert,Christian.1986a."SocialCostsof EconomicGrowth."
Journalof EconomicIssues20(1): 109-31.
. 1986b. "Whatthe GNP Does Not Accountfor." Devel-
opment:Seedsof Change(3):58-65.
_ und Volk-
. 1987. "Folgekostendes Wirtschaftsprocesses
swirtschaftlichteGesamtrechnung;Zur Identifikationvon
steigendenkompesatorischen Ausgabenin der Sozialproduk-
trechnung."Projektbericht.InternationalInstitutefor Envi-
ronmentand Society,Berlin.Processed.
Mishan,E. J. 1984. "GNP: Measurementor Mirage?"National
WestminsterBank QuarterlyReview(November):2-13.
Odum,E. P. 1971. Fundamentalsof Ecology.3rd edition.Phi-
ladelphia:W.B. Saunders.
Ross,M.S., and D.G.Donovan.1986. "WorldTropicalForestry
ActionPlan: Can It Save the TropicalForests?"Journalof
WorldForestResourcesManagement2:119-36.
Schulz,Werner.1985.BessereLuft, was ist sie uns wert?Eine
gesellschaftlicheBedarfsanalyzeauf der Basis individueller
Zahlungsbereitschaften.Berlin:Erich SchmidtVerlag.
Tinbergen,Jan. 1985. Production,Income and Welfare:The
Searchfor an OptimalSocial Order.Whitstable:Harvester
Press.

Enironmental Accountingin DevelopmentPolicy:
The FrenchExperience
Jacques Theys
The French experience might be a valuableguide in
evaluatingthe outlook for environmentalaccounting,be-
cause originalconceptshavebeen developedand a wide
range of statistical tools tested there during the past
decade. The French examplethus makes it possibleto
take rapid stock of the manydifferentapproachestaken
by the developedcountries,for whichit has provento be
an exemplarytesting ground. In a period of economic
crisis,the concernto incorporatethe environmentmore
effectivelyin economicpolicy has in recent years led
France to promote the establishmentof "natural patri-
mony accounts," which are similar to Norwegian"re-
source accounts."' However,it would be givinga clis-
torted viewof realityto isolatethese patrimonyaccounts
fromthe overallenvironmentalinformationsystem.Their
part in the planningprocess, and the very methods by
whichthey are drawn up, are meaningfulonly in this
overallframework.
The French Stratified System of Information
Informationcurrentlyavailableon the environmentin
France is formallyorganized into a hierarchicalsystem
that comprisessix levels,ranging from field data (which
are extremelyfragmented)to global indicators of the
qualityof life and pollution(Figure7-1).
* LevelI includes a large number of heterogeneous
data, either specificto the environmentalfield(results of
scientificinventoriesor monitoringnetworks)or nonspe-
cific(socioeconomic data).The qualityof this information
can vary considerablybecausethere is no adequate sys-
tem of standardization;for example,attempts to institu-
tionalizeecologicalbalance sheetsin businessfirmshave
failedthus far.
40
7
* Level11comprisescompositeyet sectoral statistics
on water,atmosphere,land, noise,and so forth; these are
generallyby-productsof administrativeactivity.During
the past severalyears, statisticaland documentarydata
banks havebeen addedthat coversuch areas as industrial
risk, marine pollution,regionalflora and fauna, and the
cost and efficiencyof nonpollutingtechnologies.
* Level III features comprehensivestudies published
periodicallyor compiledfrom a wide range of sources:
national (and, more recently,regional) reports on the
state of the environment,indicatorsof progress in meet-
ing the nationaleconomicplan, or a compilationof eco-
nomicdata. These studiesbring together in a singledoc-
ument all existing information on environmental
expenditure,damagecosts,and the macroeconomiceffect
of environmentalpolicy.
* LevelIV contains actual environmentalaccounting.
On the one hand, there are patrimonyaccounts, which
willbe discussedlater and whichare usuallyexpressedin
physicalterms. On the other hand, there are satellite
accounts,whichare meant to organizeprimarilymone-
tary data. Peripheralto the national accountingsystem
(whose accountingnomenclaturesit borrows), satellite
accountingis designedto assessnational environmental
expenditure,its funding,and its beneficiaries.As yet, only
a fewcategoriesof this systemhaveactuallybeen set up:
water,waste,nationalparks, and hunting.
* LevelV includestwo models,since the accounts in
LevelIV alone are not sufficientto be used as a basisfor
forecastingor simulations.The first model (POLLEN) is
designedto estimatethe effecton such issuesas produc-
tion, prices, and foreigntrade of variousenvironmental
policies. The second (SPIRE) operates in the opposite
directionand is used to assessthe pollutionor resource
extractionlevelslinked to variousdevelopmentstrategies
or alternateproductionand consumptionprocesses.
 EnvironmentalAccountingin DevelopmentPolicy:The FrenchExperience 41

Figure7-1. The FrenchEnvironmentalData S'ystem

LevelVI \ Aggregate welfare

National indicators;
aggregates GNP modified

| Simulation model . | Forecasting model |National

of the effects of of macroeconomic effects / accounts
Level development scenari'os
t - <_ of environmental
1 policies<
Models on the environmentMareomi

_/ ~~~~~~ ~ ~~(SPIRE)
(POLLEN) models

of theEeffects
of of macroeconomic ffects acconts
_. patrimony
Level IV [\P Environmental accounts
Accounts Patrimony account satellite account National
accounts

t LevelLState
\ of the Regionalreports Pl roEss conomic data on Plan data
Statistical environment on the state of _ lain progres the environment /
syntheses (national report) the environment indicators (annual report) Damage
assessment

ev Periodic sectoral statistics

Sectoral
r, i . Data banks
statistics ~Water
Air Soil Flora/ Others
fauna

Level I Specific data Nonspecific data 1

DisLevel I |(results of monitoring (production, consumption, |
tDisaggregated} | networks and population, energy, I
basic datay | scientific inventories) agriculture, and land use)

- - - - Prospective relation; not yet implemented.

indicatoirs.As yet,
* LevelVI concerns quality-of-life
France has made no attempt-unlike the United States
and Japan-to supplementgross national product (GNP)
with one or severaladditionalrepresentativeindexes of
national welfareor wealth.Indeed,using a multicriteria
methodology,some work has been undertakento assess
the overalldevelopmentof local livingconditions.How-
ever,these effortsare still experimental.
Naturally,all of these levels are linked in varyingde-
grees to external informationsystems,such as research
resultsin publicationson the cost of damagescausedby
42 JacquesTheys
pollution,generalmacroeconomictheories on whichthe
operationof the SPIRE and POLLEN modelsdepends,and
demographicand economicstatistics that are used to
supplementreports on the state of the environmentand
to build up basic data. Morerecently,a connectionwas
establishedbetweenthe naturalpatrimonyaccountsdrawn
up by the Ministryof the Environmentand economic
patrimonyaccountsestablishedby the NationalInstitute
2 Theseinterfacesallrepresentareas in which
of Statistics.
the environmentcould potentiallybe incorporated in a
generalpolicyframework.
The structure described above is not the result of a
preestablishedplan, but of an empiricalcombinationof
sets of data compiledto meet specificneeds over time.
The variouslevelshavedevelopedin a quasi-independent
manner.Initially,modelswere constructedand indicators
weredevelopedveryrapidlyand reliedheavilyon nonspe-
cificdata (on productionactivities)and experts'statistics.
Overthe years,statisticscollectedin the field(for exam-
ple, the resultsof monitoringnetworks)began gradually
to replaceestimates,thus facilitatinga broader dissemi-
nation of comprehensivedocuments,such as reports on
the state of the environment.In the late 1970s a partial
reorganizationof data based on accountingprinciples
and data banks was envisaged.Subsequently,statistics
were compiledon a regionalbasis and field data were
standardized.
The developmentof this data base explainsin part why
todaythese variouslevelsare not regarded as formingan
integratedsystem.The drawbacksof this type of step-by-
stepprocessare a certain amountof duplication,wastage,
and a lack of connectionsbetween differentbodies of
data. The advantageis that each componentis adjusted
to the specificrequirementsfor whichit was designed.In
this connection it would be interesting to determine
whetherempiricism-andthe associatedwastage-are un-
avoidable,or if the developingcountries can be spared
thisstage of a trial-and-errorapproachin order to be able
to proceeddirectlyto the accountingstage.
An AccountingSystem for NaturalAssets:
FrenchPatrimony Accounts
Despite the diversityof availabledata, there was no
instrumentthrough whichthe conditionsof the long-ter-m
regenerationof nature couldbe taken into accountwithin
the overallframeworkof economicdevelopment.For this
reason, in 1978France decidedto designan accounting
systemto assess both quantitativelyand qualitativelythe
state and the developmentof the natural patrimony,as
wellas the causesand effectsof its evolution.The author-
itiesopted to establishaccountsexpressedin both physi-
cal and monetaryunits, whichwouldbe independentof
nationalaccountsbut connectedto them byvariouslinks.
Underlyingthe conceptof an environmentalaccounting
systemis a long-standingcriticismof the inadequaciesof
GNP and economicaggregatesas welfareindicators;this
also existsin severalother developedcountries.Bertrand
de Jouvenel(1966) officiallyproposed that the national
accountsincludeenvironmentalservices,the cost of pol-
lution, and resourceextraction.He did not want to abol-
ishexistingstructures,but to appendsupplementarycom-
ponentssuitablefor long-termuse.
It became clear, however,that a simple adjustmentof
the existingsystemwouldnot be sufficient.That is why,
after the existingenvironmentaldata systemswere first
critiqued,it took ten yearsbefore a decisionwasmadeto
establishthe patrimonyaccounts. Muchdata on the en-
vironmentis available,but it is poorlyintegratedinto the
planningsystembecauseit is fragmented.It is alsodiffi-
cult to competewith a systemof economicdata shaped
by decadesor centuriesof accountingpractice.
What is neededto developenvironmentalaccountingis
a processsimilarto the one forty years ago, whichpro-
duced the Systemof NationalAccounts (sNA). Such a
process would developa consistent frameworkthat re-
flectsthe complexityof the relationshipbetweenhuman-
ity and nature as clearlyas possible;wouldcreate a stan-
dardizedlanguage,whichmaybe adoptedby economists,
environmentalmanagers,and scientistsalike; and would
devisea streamlinedset of indicatorsto serve as a basis
for decisionmaking.
Assessingnaturein all of its dimensions.The French
accountingsystemattempts to includeall componentsof
nature that can be quantitativelyor qualitativelychanged
by humanactivity.Componentsthat lie outsidethe realm
of human influence(such as geomorphology,solar rays,
or magneticfields) are thus excluded.Definedthis way,
naturalpatrimonycan be consideredas:
* A collectionof isolated components:nonrenewable
components(such as metallicore or fossilfuel), environ-
mentalmedia(soil,water,and the atmosphere),and living
organisms(such as animalsor plants)
* A set of ecosystems,that is, relationshipsbetween
components within specific systems (such as forests,
groves,heaths,grasslands,wet areas, and coastal ecosys-
tems)
* A set of territories, that is, spacesdelineatedeither
geographically(such as mountainsor coastalland), insti-
tutionally,or abstractly(suchas gridsor observationnet-
works).
Naturally,these groupsare not analyzedseparately,but
in terms of their relationshipswith other natural compo-
nents (hence the importanceof the conceptsof ecosys-
tems and cycles)and, most important,with humanactiv-
ities. The abovecategoriesare henceforthsupplemented
by the categoryof "agent" (individualsor institutions
usingthe natural environment)and a summaryclassifica-
 EnvironmentalAccountingin DevelopmentPolicy:TheFrenchExperience 43

Table 7-1. Structure of a Central Account for a Resource

Component Initial Potential Development Available Harvestingand Othereffectsof human Final
category stock Increase resource restitution resource extraction activities(pollution) stock
1
2
3
4

N
Total

tion of the relationship between people and nature, such
as extraction, pollution, nondestructive use cf the envi-
ronment, access, land use planning, development, and
selection. For obvious reasons, the categories chosen for
agent accounts are derived from those used in national
accounting systems (such as business firms, households,
government services, and foreign firms or entities).
Compared with similar experiments, the most original
feature of French patrimony accounts is that the natural
environment is analyzed according to each of its basic
functions: economic, ecological, and social. This func-
tional approach is in fact essential when it comes to
revealing natural potential and qualitative damages result-
ing from market and nonmarket activities. It explains to
a large extent why patrimony accounts have been devel-
oped independently from national accounts; the main ob-
jective is to demonstrate not a net profit or loss from the
exploitation of nature, but tradeoffs between the eco-
nomic, ecological, and social functions of natural re-
sources. Nomenclatures of the French Natural Patrimony
Accounts are given in the appendix to this chapter.
The rigor of an accounting approach. Patrimony ac-
counts do not purport to be an organized and periodical
inventory of information alone, as are, for example, re-
ports on the state of the environment. As already indi-
cated, they represent a commitment to the notion that
the relationship between people and their environment
can be effectivelyportrayed in accounting terms.
This concept first appears in the systematic construc-
tion of an account for each component, such as raw
materials, the physical environment, living organisms, and
ecosystems. This account describes the available stock of
the component at the beginning and end of a given period
(initial and final stock), as well as the factors that may
have brought about any changes: increases or decreases
in stock from natural causes and the destruction or de-
velopment caused by humanity. Such an account, called
the central account, is an itemized balance sheet (see
Table 7-1 and a practical example in Table 7-2).
One of the advantages of the accounting approach is
that the coherent picture it offers accurately reflects the
loops and relations. In theory, each change in one part of

Table 7-2. Example of a Central Account: Growing Stock of a Commercial Forest, 1969 to 1979
(thousandsof cubic meters)
Resource/asset Broadleaf Coniferous Total Use Broadleaf Coniferous Total
Volume of growing stock 980.1 6,526.5 7,506.6 Natural reduction 5.6 21.0 26.6
in 1969 (mortality)
Natural growth of initial 401.9 2,583.5 2,985.4 Accidental reduction 9.7 481.2 490.9
stock (breakage and windfall)
Natural growth by Resource extraction 92.0 1,474.0 1,566.0
reproduction (commercialfelling)
(recruitment) Self-consumption 13.6 395.0 408.6
Adjustment -29.4 +1,239.2 1,209.8
Volumeof growing stock 1,330.7 5,758.0 7,088.7
in 1979
Total 1,422.2 9,368.4 10,790.6 Total 1,422.2 9,368.4 10,790.6
44 JacquesTheys
the accountingsystemshouldproducean equivalentchange
in accountingterms.This interactionis most clearlyilus-
trated in the ecologicalsector accounts and linkage ac-
counts,whichanalyzethe relationshipbetweennature and
the components,and agent accounts,whichsum up the
linkages with the natural environment of a given agent
An objectiveof the central accounts is to aggregate
data to the extent possible.At first glance,this objective
wouldseem to contradictthe use of highlydifferentiated
physicalunits of measurement(suchas surfacearea, cu-
bic metersof wood, numberof species,and tons of coal)
or of qualitativeindicators.This obstacleis partiallyover-
come by creating equivalentclasses at the assessment
level accordingto three main functions:economic,eco-
logical,and social. This method can be applied using
monetaryunits wherejustified,but also energyand bio-
massequivalentswhenpossible.Althoughthis systemhas
attempted to produce aggregatedata, the concept of a
single natural patrimonydevelopmentindicator-similar
to the majoraggregatesused in nationalaccounting(such
as GNP)-has been abandoned.
Flexible applicationof the framework. The develop-
ment of central accounts must take into account the
specificityand complexityof natural phenomena. The
limitationsof an aggregateapproach havealready been
shown:in practice it is neither possibleto combinethe
indicatorsrepresentingeach of the three main functions
nor to comparethe measurementsspecificto extremely
dissimilarcategoriesof resources(for example,forestar-
eas and watervolume).Thus, differentaccountsfor each
of these three functionswere created, and data were
aggregatedonlyfor the mainresourcecategories(suchas
water,fauna,flora, and atmosphere).Consequently,vec-
tors, not indicators,must be used. Similarly,the objec-
tives of thoroughness and systems impact accounting
wouldbe too costlyto use if they were applied to the
letter. For this reason it was initiallydecidedto confine
letter.For this reason it was initiallydecidedto confine
the analysisto a fewprioritysectors (forests,water,soil,
landuse,land
and wildlife)
anduse, and aa few
wildlife)and basicinteractions.
fewbasicinteractions.very
Becausethe phenomenato be measuredare so heter-
ogeneous,the principleof standardizationhas been par-
tiallyreassessedin favor of differentiatingbetween com-
ponent,ecozone,and agent accounts.
* Componentaccounts deal with subsoil assets, the
soil, the atmosphere,water, and the flora and fauna.
Materialor energybalancesare presented at the begin-
ningor the end of a periodand indicatechangesresulting
fromnature or humanactivity.
* Ecozone accounts register the changes in land use
and
Ecoze
stacusfcoutsyster. R othehangesingla oduse
and the status of ecosystems.Remote sensing methosds
are
etablih
neessar
theeto accunts.particularly
Arenecssary establishthpes
nto ple
acunts humaninstit
Agent,which
ac cuntsidealwinthpeopleandgor f thua
istiu-
tions,
whichare classifiedin the categoriesofthestan-
dard nationalaccountsand deal with economicand non-
economicactivities.
In the future these accountswillmost likelybe supple-
mented by,and connectedwith,social statistics.
The flexibilityof the French system makes patrimony
accounts resemble more of a general framework than a
rigidsystemof accounts.This flexibilityenablesresource
managersto remaindirectlyinvolvedin constructingsuch
accounts. This has been the case in France, where re-
search into methodologyat the central level has been
combinedwith highlydecentralizedpracticalinitiatives.
The roleof economicassessment. Thosefor whomthe
word "accounts"first signifiesa monetarybalance sheet
(or a cost-benefitanalysis)could be disappointedby the
French approachto patrimonyaccounting,whichconcen-
trates on physicaldata. In fact, these economicconsider-
ations appearon at least three levels.
* The systemis designedto lay the groundworkfor a
genuinephysicalor "bio-economy."The questionat hand
is to revealthe relationbetween the economicand non-
economicfunctions of nature, to analyze the circum-
stances in whichthese functionsare reproduced,and to
determinehow certain underdevelopednatural potential
maybe transformedinto productionfactors. In a word,
patrimonyaccountingis a key tool in assessingthe eco-
logicalbuildingblocksof developmentand is a represen-
tation of the conceptof sustainablegrowth.
* Patrimonyaccountsalso havemany links or bridges
with the national accounting system. Agent accounts,
whichuse categoriesborrowedfrom patrimonyaccounts,
serve as a system of passagewaysthrough which it is
possibleto movefrom classicproduction and consump-
tion activitiesto changesin resource stocks or environ-
mentalquality(or viceversa), as shownin Figure 7-2. In
..
adionacutsrodexniuesmetoepr
and developthe natural environment(satelliteaccounts).
* Last, but not least, it must be rememberedthat the
purpose of such accounts is to estimate simultane-
ously the ecological,social, and economicvalue of the
natural patrimony.This involves,on the one hand, not
only assessingthe environmentalstock, but also tracing
the damagingeffectof its conversion.On the other hand,
it involvesevaluatingthe extensionof services rendered
by nature. The entire systemis organizedwith reference
to this objective: priority is given to a functional ap-
proach, qualitativeassessmentsare introducedfollowing
physicalanalysis,and aggregateindexesare sought.
Thisbeing the case,manyproblemsremainunresolved,
patcual in th il fmntr sesetpno
in the field of monetary assessment. Opinion
is divided,especiallywith respect to the most suitable
methodof assigningan economicvalueto resourcestocks:
usingmarket prices, consideringthe discountedvalueof

Figure 7-2. Links between Economic, Agent,
and Patrimony Accounts
Economic
activities
National Input-outputaccounts
economic
accounts Production Consumption
Uses , t Resources
Typology of relations ____
Agent beTvpology
of relations _ Noneconomic
accounts Social 4
Extractio Accretion uses,
Pollution and and such
depletion development as
access'
Impacts t
PatrimonyChangesin resourcestocksseeaplnigccrsadobctv.
accounts and environmentalquality
total anticipatedrevenuesand services,or even rneasur-
ing opportunity or replacementcosts. All of these ap-
proaches yielddifferentresults, and it is clear that this
subjecthas not yet been fullythought out.
Despite the seeming complexityof the F'ren,chap-
proach, tangibleresultshavebeen obtainedover a fairly
short period and at a relativelylow cost (three staff-years
and $100,000 in research grants) by focusingattention
on a limited number of areas (water,forestry,soil, and
wildlife)and by decentralizingthe data gatheringprocess.
Additionalyearsof studywillbe requiredto translatethe
resultsof patrimonyaccountinginto changesin the rela-
tive price structure of natural resourcesand into mone-
tary values.Therefore,in the immediatefuture, the pri-
ority has been given to using componentaccounts in
long-termmodelingand to buildingecologicalsector ac-
countsthroughremotesensing.
The Experience of Franceand OtherDeveloped
Countries:A Comparisonof Planning Objectives
As a conclusionto the firstpart of thischaptsr, it would
be interestingto situatethe Frenchexperiencein relation
to the general trend of establishingenvironmentalinfor-
mation systemsand to compare the variousapproaches
adoptedin the lightof planningobjectives.On examining
some of the more representativeprojects, one is struck
by their diversity.Each countryseemsto haveadaptedits
environmentalaccountsto its own national needs, insti-
tutions,and culture.
Americanand Japaneseproposalsare clearlybased on
the theory of welfareeconomics:their aim is to modify
GNPby assessingin monetaryterms the servicesderived
Environmental
Accounting
in Development
Policy:TheFrenchExperience 45
fromthe environmentand the damagesborne by it. Nor-
wegianresource accounting,however,is based on an
analysisof eco-energyand materialbalance.This system
translates economic activity into flows of matter to energy
and viceversa.The Canadianapproachdrawsheavilyon
the concept of impact studies and applies them at a
macroeconomic level.This approachis concernedprimar-
ily withthe dynamicof ecologicalsystemsthat havebeen
disturbed by the economic system.
_ The experience of the Netherlands hinges on the neces-
sity of harmonizingthe economic,social, and ecological
activities functionsof nature. The French patrimonyaccounts,al-
though based on this conceptof function,share with the
Norwegianaccountsthe aim of attaininga genuineecon-
omyof wealthbasedon an assessmentof the potentialities
(bothmonetaryand nonmonetary)of usingnaturalassets.
The approachespursued by differentcountries reflect
several plannng concems and objectives.
* To diagnosethe current (or foreseeable)state of the
environment(Canada,France,and Norway)
* To analyzethe effect of environmentalpolicies (or
the absencethereof)on economicperformanceand wel-
fare (Japanand UnitedStates)
* To assess the value of resources that can be ex-
ploited for developmentas wellas the conditionsof their
regeneration(Franceand Norway)
* To identifythe possibleoptions among the social,
economic,and ecologicalobjectivesof development,with
a viewto globalnegotiation(France and Netherlands).
In conclusion,no natural resourceaccountingeffortso
far has madeit possibleto attain all of these goalssimul-
taneouslyor to coverthe entire range of planningpro-
cesses(as illustratedin Figure 7-3).Moreover,it is diffi-
cult to imaginea systemcapable of meetingthe needs of
economists,ecologists,the public, decisionmakers,and
environmentalmanagersall at the same time. Developing
countriesmay thus haveto choose from the variouspos-
sible approachesby referringto their own needs,levelof
development,and environmentalmanagementproblems.
This is discussedbelow.
Adapting Environmental Accounts
to Developing Countries
It is no coincidencethat proposals currentlyput for-
ward about environmentalaccounting are varied. This
heterogeneitysimplyreflects the varied political con-
cerns, institutionalstructures,and problemsin the devel-
oped countries.It also indicatesthe difficultyof findinga
consensuson a theoreticalframeworkthat wouldpresent
clearlythe complexrelation between the economyand
the environment.Therefore,a case-by-case,country-by-
country, and perhaps even region-by-regionanalysis is
calledfor, based on a simplifiedaccountingframework.
46 JacquesTheys

Figure 7-3. Environmental Accounts and Economic Planning

Potentialities Activities Impacts

Accumulation or
depletion Environmental Direct use of
goodsand goodsand

Useof gods
Outputan d services Impactof
Mobilizable Inpt Nonmarket Market service (consumption)o
potentials production produiction activities
* Human * Economic
* Nature * Ecological
* Movableassets Imports . _ Social
and currency Waste, Absorption
* Realand pollution of pollutants On
capitalgoods * People
9 Nature
Resourcefunctions * Movableand
* Economic I \ real goods
* Ecological _ . Available Useof * Institutions
* Social * publicservices, publicservices (positiveand negative)
including
Public environmental
management services
1 Policies
* Plans
Projects o
r 9~~~~~~~
Incentives Criea
and standards evu

Obetie
mxmie0Output O b j 'ective:maximize
weath-*a goals social,economic,and

|Long-termchoie|| Social |
discountrate values l

iVote:Someaccountsassessthe potentialand stockof resources(suchas in Franceand Norway),whileothers measurethe effectsof environmental

damageor improvement(suchas in Canadaand the UnitedStates).

Problems Specific to the Developing World
The first question that naturally comes to mind is the
applicabilityof environmental accounting in the context
of the physical problems experienced by the developing
countries. It is clear, however,that such accounts will not
be created unless they correspond to a perceived eco-
nomic need.
Economic issues and their implications. Despite the
enormous diversity of structures and economic situations
in the developing world, there are a few common features
that directly affect environmental accounts. First, most
people are aware of the critical role that natural resources
continue to play in the economic activity of developing
countries at all levels (employment, production, export,
and consumption). Most developing countries-particu-
larly the poorest-depend on the export of primary com-
modities or one or two other products for their growth.
Resource regeneration and improvement is thus a vital
factor in these economies.
Second, the importance of natural resources is made
even more apparent if nonmarket activities are consid-
ered. Another feature of developing countries is the exis-
 Environmental
Accounting
tence of a traditionalsector, dominatedby nonremuner-
ated activitiesincludingself-production,self-consumption,
and barter, alongsidea modern sector, whichis highly
integratedinto the internationaltrade system.To date,
populationgrowth and the frequent choice of capital-
intensivetechnologyhave limitedthe capacity of these
countriesto absorb the modern sector.
It is thereforethe properfunctioningof this traditional
sector (for example,accessibilityof local,nonreplaceable
resources)that ensures the basic survivalof hundredsof
millionsof people. The clearest exampleis noncommer-
cialenergysources(such as firewoodand refise), which
accountfor 50 percent of the developingworld'senergy
consumption.In the People's Republicof the Congo,for
example,firewoodrepresents 80 percent of all energy
sources and takes up 25 percent of householdbudgets.
During the past decade, its price for villageinhabitants
has increasedtenfold.
Third,this superpositionin of heterogeneouseconomic
systemstranslates directlyinto an enormoussocial and
geographicalinequality.Public works projects, recently
constructedmodernindustrialcomplexes,and rapid urban
expansionare so many enclavesof modernity,whichat
times stand in strong contrastto the rest of the country.
The growingdisparitiesin income,consumptionpatterns,
and lifestylescall into question the usefulnessof tradi-
tionalstatisticalassessmentmethodsbasedon averages.
A finalfactor specificto developing-country economies
is their dependenceon the outsideworld. This has also
had considerableinfluenceon the introdLiction of envi-
ronmentalaccounts.On the one hand, the importanceof
export goods, whosevalue is often set by the interna-
tional market, and, on the other hand, the weaknessof
sectors producing directly for the domestic market,
sharplyreduce the effectivenessof national rnechanisms
reuatn reouc supl ademn. FutemoeX
reuatn reouc sppl an deman. F .rhroe
given the spread of the nonmarket sector, it is obvious
that, in this situation,prices cannotfulfilltheir traditional
role as indicatorsof scarcity.
Theeas arseveral
Thnereare s ipslcar s oof tthese economic Issues
implications
for environmentalaccounting.
* There is no doubt that the assessmentand develop-
ment of natural resource potential can have both an eco-
nomic and ecologicalinterest for developingcountries,
providedthat it leads to the formulationof aggregate
indicatorsand coversboth commercialand noncommer-
cial resources(see also Peskin,Chapter10).
* Giventhe poor representativenessof monetary as-
sessments,it is necessary,at least initially,to concentrate
on physicalaccounting.
* The existence of extreme inequalitiesor different
situationsrequiresthe creation of separate accountsfor
each homogeneousgeographic region or population
group.
in Development
Policy:TheFrenchExperience 47
* Because of the highlypolarized nature of develop-
ment,many cases maywellwarrantlimitingthe environ-
mentalaccountingsystemto an analysisof a fewproduc-
tion and consumptionchannelsor primarycommodities.
Physicalproblemsand their implications.It wouldbe
inappropriatehere to providea detailedinventoryof the
environmentalproblemsin the developingworld. It is
more importantto summarizehowthese problemsdiffer
fromthose experiencedin the industrialcountries.Envi-
ronmental concernsin the developedcountrieshave fo-
cused largelyon pollutionrisks and the conservationof
the naturalenvironment.But despiteextensivelocaldam-
age caused by uncheckedindustrializationor urban ex-
pansion,the greatest environmentalthreat in developing
countries is the disappearanceof the resource base on
which the survivalof millionsof people depends. The
problemsrelated to resources include shortages of fire-
wood and deforestation,deteriorationof arableland and
desertification,depletionof the quality and quantity of
water and the prevalenceof waterbornediseases,reduc-
tion in geneticdiversityand wildlifestock (wildand do-
mesticated, including fishing resources), and damage
causedby natural disasters.
These problemstend to overlapinto a chain of cause
and effect, which makes partial approaches of limited
value. Disturbancesin the hydrauliccycle,for example,
may be related to erosion, which is in turn linked to
deforestation.The destabilizationof ecologicalsystems
mayalso be linkedto povertyand inequality,the popula-
tion explosion,legal systems,dependencyon the outside
world,and adversenatural conditions.
This situationhas severalimplicationsfor establishing
accounts.
The stock of natural resources and its development
take precedence over pollution problems in developing
uteriec vr
countries.
* An accounting system will be even more advanta-
geous if it reflects the systems' effects and cyclical phe-
nomena.
* Although the nomenclature of natural resources
seemseasilyadaptableto the specificcontextof develop-
ing countries, the categorization of agents should be care-
fullyreviewedand, if necessary,revised.
AvailableInformation
The high cost of compilingnew data and the chronic
lack of qualifiedpersonnelin the combinedfieldsof sta-
tisticsand the environmentmake it difficultin the short
term to prepare resource accounts in developingcoun-
tries. Gatheringinformationon the environmentis gen-
erallynot yet a priorityconcern for statisticiansin these
countries. In these circumstances,it is thereforelikely
48 JacquesTheys
that environmentalaccountingwillin the short term ei-
ther be based on existingdata alone or depend on statis-
tical programs already envisagedin other fields. This
restrictionmakesit strategicto establisha preciseinven-
tory of all currentlyavailabledata (mapping,scientific,
and statistical).
In 1977 the United Nations Office of Statistics ad-
dresseditselfto this taskby undertakinga generalsurvey
of 122 countriesand case studieson Kenya,the Domini-
can Republic,and the FijiIslands(UnitedNations1982).
The main problemsfound were dispersedand compart-
mentalized data, insufficientspecific information,ex-
tremely uneven quality of nonspecificdata, and over-
general statistics prepared by international agencies.
Theseare discussedbelowin turn.
Dispersedand compartmentalizeddata. Somedata on
the environmentalready exist in the developingworld.
These data, however,havebeen compiledby a widevari-
ety of differentagencies,ministries,scientificinstitutes,
business firms,and internationalorganizationsand are
scattered and diversified.Usuallyappearingas by-prod-
ucts of research studies, sectorial surveys,and project
evaluations,they are presented in diverseforms, such as
individualobservationson specificsituations,statistical
series, files, and maps. Most often used directlyby the
institutionsproducingthem, these data are neither stan-
dardized, centralized,nor even widelydisseminatedor
published.Thus in Thailand,more than eighty different
bodiesproduceenvironmentaldata, whichuse frequently
incompatiblenomenclaturesand formats.Forty-fivesuch
institutionsexistin Kenyaand thirty-eightin Fiji.
Althoughconsiderablepotential does exist, it remains
largelyundeveloped.In this connection,the creation of
an environmentalor resource account could representa
decisivestep forward,evenif it werelimitedto creatinga
minimumof harmonization-and thus economies of
scale-betweencompetingdata systemsand to servingas
a meetinggroundwheredata producersand users-who
otherwisehavelittle contact-could meet.
Insufficientspecificdata. Some progressin collecting
data has been made duringthe past years.Severalcoun-
tries nowpublishreports on the state of the environment.
Monitoringnetworkshavebeen establishedto analyzeair
and waterpollutionfroma fewstandpoints.Remotesens-
ing is now being used systematicallyto observechanges
in soiluse. At the instigationof the Economicand Social
Commissionfor Asiaand the Pacific,pertinent data have
been collectedin most Asian countrieson deforestation
and water resources (ESCAP 1982). These efforts, how-
ever,remain the exception.In Kenya,for example,which
is one of the more advancedAfricancountries,statistical
data are still inadequateon such importantphenomena
as deforestation,air and waterpollution,soilerosion,and
marginalsettlementsin rural or urban areas. Generally
speaking,existingdata are often not sufficientlylocalized
and are discontinuousor isolated.Once again, they are
gatheredby a host of independentagencies,whichfail to
circulatetheir findingsamongthemselves.
In general,there is little data availableon stocks and
even less on their growthor depletion.Assessmentsof
natural resource potential are generallydealt with in ad
hoc studiesor research,such as those carried out before
a major project is implemented.Purely economicesti-
mates most often remain confinedto the university(at
least at a global level),and studiessummingup environ-
mental expenditureat the national level are extremely
rare. Finally-and most significant-qualitativedata are
virtuallynonexistent.
Unevenquality of nonspecificdata: abundant but not
directlyusable. The observationaboveleads to a simple
conclusion:the feasibilityof environmentalaccountingin
developingcountrieswilldependprimarilyon the ability
to adapt data originallycompiledfor other purposes or
to use existingstatisticalinfrastructuresto gather infor-
mation from other sources.In most countries periodical
statistics are availableon population;employmentand
output, by major sector of activity;land area under agri-
culture,agriculturalyields,and production;livestock,fer-
tilizer and pesticides;quantityof timber and forestprod-
ucts for export; catches of fish; and construction of
permanentdwellings.
Through a few relativelysimple operations, such as
geographiccoding of data or the use of other data (for
example,pollution coefficientsor resource-use coeffi-
cients per unit of output), statistics directly usable for
environmentalaccounts mightbe obtained, and an envi-
ronmentalsectioncould graduallybe added to some sur-
veys.
Naturally,the successof patrimonyaccounts depends
on the qualityof their connectionwith the SNA, whichhas
the tremendousadvantageof havinga homogeneousand
consistentframeworkand of being appraisedannuallyin
most countries.Nationalaccountingis a necessarytran-
sitionpoint, particularlyif environmentalaccountsare to
be used in national economicplanning. There are two
limitationsto this approach,however.
First, it is clear that considerablegaps exist in the
statistical material compiled-even at the purely eco-
nomiclevel.For example,verylittleaggregateddata exist
on householdspendingand consumptionor on income
distribution,and informationrelatingto volume(deflated
accordingto a price index) is often contestable(IARIW-
ECA 1982).
Second, it is obviousthat the significanceof accounts
and the use of aggregatesas overallindexes of activity
are considerablyreduced by the fact that they do not
include nonmarketgoods and services. Gunnar Myrdal
 EnvironmentalAccountingin DevelopmentPolicy:TheFrenchExperience
(1970) pointed out long ago the limitations of applying
the notion of GNPto developing countries whose econo-
mies include sizable nonmonetary agricultural sectors.
These two limitationsexplain why the quality of aggre-
gate data collected can vary considerably. In Africa, for
example, estimates of GNP growth during 1960 to 1970
ranged, according to the sources and countries concerned,
from 1 to 11 percent for Chad, from 1 to 13 percent for
Mali, from 1 to 6 percent for Burkina Faso, and from 1 to
2 percent for Senegal (World Bank 1981). These gaps,
which may at first seem to represent insurmountableobsta-
cles to creating environmentalaccounts, can also, paradox-
ically, be viewed as significant opportunities. As pointed
out on numerous occasions, improvingthe quality of con-
ventional economic data systems (national accounting, in
particular) will obviously benefit enviromnental accounts
by emphasizing the nonmonetary sector.
Two examples-among many others-exist of this con-
vergence. First, by excluding wood-based fuels, current
statistics obscure the specific nature of the energy crisis in
the developingcountries and may lead to irrational invest-
ment decisions. In Zaire, for instance, if noncommercial
energy sources were included, energy consumption would
jump from 3 to 10 percent of gross domestic product
(GDP). Second, because of the quality of data compiled on
nonmarket food crops, Kenya was able to lorecast harvest
yields that proved to be invaluable during the drought in
the late 1970s. A project specificallyaimed at collecting
data on natural resources not intended for export markets
could thus easily become economicallyprofitable.
Abundant but overgeneral data from inzternationalor-
ganizations. Beginning in the mid-1970s, many interna-
tional organizations-including the Food and Agriculture
Organization (FAO); Unesco Man and the Biosphere Pro-
gram; the World Health Organization (WHO); and the
U.N. Environment Programme (uNEP)--have involved
themselves in far-reaching programs to compile data on
the environment and natural resources.
For example, the FAO recently completed a soil map on
a scale of 1:1.5 million, as well as a map of areas in the
process of desertification. It has also launched a project
to partition Africa and Asia into homogerneousecological
regions and has analyzed soil potential for twelve types of
crops (UNEP 1981). It periodically dissenninates data on
soil use (in five categories), pesticide use, area of land
under irrigation, and fishing yields. Two years ago, it
completed an inventory of tropical forests and their
growth, in conjunction with Unesco and UtNEP.
For its part, UNEP has coordinated a monitoring net-
work to oversee climate, long-range transport of pollut-
ants, environmentalhealth, oceans, and terrestrial natural
resources (GEMS program). In many fields, these data are
often the only available at the moment. The pollution
monitoring network coordinated by the GEMS program
49
comprises only 400 stations for the entire world, which
represents an average of only 2 to 4 in each country. Last
but not least, entire sections of environmental concerns
have poor or inadequate data coverage; this is especially
true of wildlifeand flora and of environmental economics.
Environmental Accounting as a Tool
for Policymaking
Natural resource accounting could act as a catalyst for
framing environmental policy in developing countries.
Many of these countries have complained that it is diffi-
cult to put the concept of integrated development into
practice without a tool to facilitate the distribution of
investment to various areas of the environment or to
identify beforehand the payers and beneficiaries of the
policies envisaged. To a large extent, environmental ac-
counting answers this dual need. In this respect, it can
fulfill, at the national level, the same role of initiator or
driving force that impact studies have at the project level.
It can supply environmental institutions with the means
to open dialogue with economic administrators and mem-
bers of the scientific community.It can also provide these
same institutions with effective management tools in their
particular field of competence, such as priority criteria
and minimal data on anticipated expenditure and its im-
pact. Environmental accounting cannot, of course, substi-
tute for political will, but, for environmental administra-
tions, which are often weak, it can constitute a powerful
means of internal structuring and recognition with re-
spect to other ministerial offices.
Proposals
The preceding survey shows that the development of
environmental accounting in the developing world will
inevitablyencounter a set of serious obstacles. The prob-
lem of insufficient and poorly dispersed basic data is
compounded by several factors.
* The shortage of financial and human resources that
can be mobilized in the light of other, more urgent,
priorities for statistical data that have not yet been filled.
* The lack of a coordinating structure capable of
bringing together multidisciplinaryteams.
* The unsuitability of a cost-benefit approach in sec-
tors dominated by nonmarket activities.
* The present lack of political demand for this type of
tool, particularly by economic planning bodies.
Although these obstacles may seem sizable, they are
not insurmountable. There are a few factors that might
help in establishing environmental accounts, and results
from the few projects recently undertaken in developing
countries have been encouraging.
50 Jacques Theys
Severalfactorscan possiblyfacilitatethe setting up of
environmentalaccountsin the developingcountries:their
environmentalproblemsmaybe more apparentthan in
the industrialcountries,the convergenceof requirements
for economicand resource statisticsmay be more obvi-
ous, and the intermediatematerials and results of the
developedcountriescan be used.
In addition,remote sensing is an effective,well-devel-
oped tool. It is particularlywell adapted to evaluating
naturalresources,and, havingdemonstratedits reliability,
remotesensingcan be an extremelyusefultool in creat-
ing environmentalaccounts.
* It is relativelyinexpensivecomparedwithother meth-
3
ods of statisticaland cartographicalanalysis.
* Its globaland compositenature allowit to analyzea
large numberof variablessimultaneously, thus furnishing
qualitativeas wellas quantitativeresults.
* It makes it possible to process rapidly large and
inaccessibleareas.4
* It is particularlysuitablefor repeated use, whichis
importantin takingperiodicmeasurementsof differences
in the state of the environment.
To date, there are no examples of natural r esource
accountingprojectscarried through to completionin de-
velopingcountries.However,significantprogresshas been
madewiththe launchingof a fewpartialeffortsunder the
auspicesof the U.N.StatisticalOffice(uNso),UNEP, and
5
EScAP, and on the initiativeof individualcountries.
Cote d'Ivoireis a good example,as it began research
in 1983 with a view to setting up a quantitativeand
qualitativesoilaccount (Dravieand others 1983).Within
a few months,it establisheda methodologicalframework
and appliedit to a smallregion,whichpermittedfull-scale
activitiesto be carried out in three directions.
First, a dual soil classificationsystem(functionaland
qualitative)was created. Second, a central account was
set up, whichdescribedboth the changesin land alloca-
tion and the evolutionof its potentialuse by phenomena
such as erosion, desertification,and certain agricultural
practices.Third, an economicaccount was established,
whichcoveredboth the investmentcost involvedin soil
protectionand improvementand the monetaryeffect of
the growthof potentiallyexploitableassets (taking into
account theories on production techniques).Full-scale
compilation of statistics for these accounts must, of
course, be started before any conclusionsmaybe drawn
fromthis project.
At the National Level
The preparationof accounts cannot be contemplated
withoutstrong motivationon the part of the interested
countries.
* In the poorest countries,priorityshouldbe givento
preparing reports on the state of the environment(add-
ing, where appropriate,a survey of investmentexpendi-
ture).
* Patrimonyaccountsshouldbe developedonly in the
economicallymost advanced countries and should em-
phasizeaccountingin physicalterms before contemplat-
ing linkageswiththe SNA.
* Data collectionand processingshould initiallyfocus
on one or two resources regarded as economicallyand
ecologicallystrategicby each country.
* The national (macro)and regional(major projects)
approachesshouldbe combinedas far as possiblewithin
the frameworkof flexibleand decentralizedinstitutional
machinery,whichwilldrawmaximumbenefit from exist-
ing skillsand information.
At the InternationalLevel
As was the case whenthe SNA wasestablished,interna-
tional cooperationwillundoubtedlybe necessaryto help
initiate the process of creating environmentalaccounts.
The followingsix proposalsfall withinthat context.
* A detailedsurveyshouldbe madeof the cost, logis-
tics, human resource requirements,and performanceof
accountingsystemsbeing set up in the developedcoun-
tries. At the same time, an audit on the reliabilityof
remotesensingmethodscould be carried out.
* Full-scaleexperimentsshould be funded in a few
countries(one per continent),whichfocuseach time on
a differentaspect that is locallyconsideredto haveprior-
ity.
* Specificimprovementsin existingstatisticalsystems
should be promotedunder the auspicesof the U.N. Sta-
tistical Office,such as reorganizingdata according to
ecologicalzoningpatterns, compilingeconomicdata per-
taining to environmentalmanagementin a single com-
pendium, or adding a section on the environmentor
natural resourcesto industrialand agriculturalsurveys.
* An internationalnetwork should be establishedto
facilitatethe disseminationof the results of successful
experiments,the exchangeof viewsand intermediatema-
terials, access to data collectedby internationalorgani-
zations,and the creationof a file of cost-benefitstudies.
* Trainingprogramsshouldbe developedfor the stat-
isticians,economists,and scientists involvedin the ac-
countingsystemsprocess,or trainingmodulesshould be
createdto make national accountantsand planning offi-
cialsmore aware of the importanceof environmentalac-
counting.
Conclusion: Towarda Programof Action
Actually,the greatestproblem connectedwith environ-
mental accountingis not to define once and for all a
 EnvironmentalAccountinginDevelopmentPolicy:TheFrenchExperience 51

model transposableto all the developingzountries,but willbe possibleto attain all these objectiveswithin two
to launch a process of implementationthat takes into years.
accountthe assets and constraintsspecifict:oeach one of
them. It is thus necessaryto go beyonda list of desirable SecondStage: Development
activitiesand to propose a program with very specific
stagesand time scales. The second stage should havetwo objectives:on the
The followingprogram proposes a series of steps in one hand, to extend the precedingsteps so as to include
three stages(short-,medium-,and long-term),whosemain fewereconomicaspectsand, on the other hand, to set up
aim is graduallyto arouse in the developingcountries the essentialinstitutionalstructuresin each countrycon-
concerned sufficientmotivationto proceed to the final cemed. The expertsshouldthus be graduallyreplacedby
stage of creatingaccounts. nationalbodies.
The followingactivitiescould thus be carried out dur-
First Stage:Awareness and Experimenta,tion n this
ing hsseodsae
second stage.
The aim of the firststage is to interestplanningofficials * The setting up, in each of the countries in whicha
(and beyondthem, politicaldecisionmakers)in creating pilot studywasdone, of institutionalstructuresresponsi-
accounts as soon as possible.The stress is thus deliber- ble for the accounts(Figure7-4).
ately placed on producing results rapidlyand on using
accountsfor economicforecasting.
For that purpose four or five pilot studies should be Figure 7-4. Using Accounts to Compare Projects
launched, with the participationof experts from these or AlternativeDevelopmentScenarios
countries and from the developedcountries who have
experiencewith environmentalaccounting.Initiallyeach Valueandsustainability
study would prepare a central account (stock account) ofthe project
for one of the resourcesconsideredto be strategicby the
volunteer countries (such as water, forest:,soil, humid ,
zones,fisheries,or genotype).
The first drafts of the accounts thus created are used Ecological Social Economic
in the second stage to analyzeresource processingnet- evaluation evaluation evaluation
works throughout the production system and to assess
the modificationsin the economicvalueof the stock at
the beginning and the end of the period under consider- Ecological Social
ation.It willbe possibleto test severalmethodsof mone- functionof function
tary evaluationconcurrently. natural (nonmarket)
The best wayof turning the precedinginformationto resources
account would be to construct in a third stage, together
with planning officials, two contrasting resource manage- Gainor loss Economic
ment scenarios. One of these would prolong past trends, of function function
while the second would be a conservation scenario. This of stocks Consumption
prospective analysis would illustrate by simple indicators Amountof stocks * Imports
the relevance-both from the economic and physical view- in t, t + 1, t + n * Exports Expenditures
points-of a strategy of sustainable development.The case (quantityand quality) Jobs protection
studies would also localize existing information and as- I J p/
sess the cost and technical conditions involved in obtain- Resourceflows /
ing currently unavailable data. By the end of the first Macroeconomic
stage, the experts should thus be able tD propose an model
initial program for organizing the existing environmental /
data in each country.
Finally,at the same time as this fieldworkis being done, - Usesand _ t_
it would appear to be essential to maintain or to set up Patrimony extraction National Satellite
an ad hoc assessment team. This team would be respon- accounts of resource accounts
sible for periodically following the progress of the case
studies and for assessing the results of this first stage, Scenarios
together with the national leaders. The few rare experi- and projects
ments that have already been highlighted indicate that it (Prospectivemodels)
52 Jacques Theys
* The preparationof a completesystemof accountson
the resourcesstudiedduringthe firststage. This is equiv-
alent to addingto the preceding analysisa satelliteac-
count summingup the expenditureon the management
of this resource,a set of agent accounts,and, in particu-
lar, the bulk of the data on the ecologicalfunctionof the
resources(accountsand indicatorsof defectsin ecosys-
tems). This wouldbe done by the institutionalstructures
describedabove.
* The launchingof internationaltrainingand research
activities.
At the end of this period four or five complete ac-
counts, each dealing with a specificresource (such as
water,soil, or forest),wouldthus be available.This should
sufficeto test the validityof the indicators,ratios, or units
of economicand ecologicalmeasurementsuggestedin
eachcase.
Concurrentlywith this detailedanalysisof a resource,
it would probablybe desirableto use the availabledata
to set up in four or fivevolunteercountriesa provisional
systemof indicatorsof situationor of resultsfor the other
nationalresources.Allthis materialshould finallyenable
an increasingnumber of countriesto become aware of
the relevanceof accounts before the third stage, which
should consist of applying the pilot experimentson a
largescale.
Third Stage:Dissemination
Untilthe programis about to enter its fifthyear,it will
not be possibleto appreciatethe relevanceand feasibility
of environmental accountingin the developingworldand
to arrive at seriousconclusionsregardingthe accounting
structurebest adaptedto the situationin these countries.
The third stage should thus logicallybegin with an inter-
nationalconferenceto assessand adjustthe programof
action.The followinglong-termactionscouldeventually be
launched,dependingupon the resultsof this conference.
In countriesthat initiatedpilot experimentsduringthe
firsttwo stages
* Progressiveextensionof accounts establishedfor a
single resource to the totality of the natural resources
(patrimonyaccounts) and of the expenditureon their
management(satelliteaccounts)
* Simultaneousorganizationof existing information
systems(such as data banks, monitoringnetworks,sur-
veys,and inventories)
* Introductionin planning of a limitednumber of ag-
gregate indicators of the state and developmentof re-
sources, whichare followedup periodically(and estab-
lishedon the basis of the accounts)
* Discriminatinganalysisof the possible linkups be-
tweenresourceaccountsand nationalaccounts
* Development,where possible, of prospective and
macroeconomicmodelsfor assessing,on the one hand,
the economiceffects of environmentalpolicies and, on
the other hand, the consequencesfor the mainecological
indicatorsof differentgrowthscenarios
* Launching,on the basis of the preceding informa-
tion, of a limitednumber of cost-benefitstudies on poli-
cies for combatingconditionssuch as desertification,de-
forestation, and water pollution (in accordance with
nationalpriorities).
In the othercountriesor on an internationallevel
* Preparationof an international-level typicalaccount
structure adaptedto the developingcountriesand of the
correspondingmethodologicaldocuments,such as guide-
linesand intermediatematerial
* Adoptionof an internationalclassification systemfor
resources (such as water, air, soil, and ecosystems)and
effortsto harmonizeexistingnomenclatures
* Improvement of the methodsof storing and dissemi-
nating the data collectedby the internationalagenciesto
make them more coherent and directlyaccessibleto the
developingcountries
* Launchingof a majorinternationalresearchprogram
on indicatorsof defectsin ecosystems(analysisof thresh-
olds) and on resource economics(economicassessment
of damage,calculationof shadowprices, and criteria for
the interperiodmanagementof renewableand nonrenew-
able resources).This researchfocuseson the problemsof
the developingworld and willhave to take explicitlyinto
accountthe nature of the availableinformation.
In conclusion,it is necessaryto revert to another as-
pect, the institutionaldimension.It seems obviousthat
an operation as demandingas the creation of accounts
has no chance of succeedingunless it corresponds,at
leastto beginwith,to specificpoliticalneeds withrespect
to management,and unless it givesbirth to a permanent
frameworkfor dialoguebetweenthe responsiblegovern-
ment departments,scientists,economists,planners, and
statisticians.Methodologicalproblemsare of secondary
importancecomparedwith these necessities.
Appendix. Nomenclaturesof the French Natural
Patrimony Accounts
A. Nomenclatureof components
1. Nonrenewablecomponents
Fossilfuels
Uranium
Metalliferousore
Nonmetallicore
Quarry(includingsand)
Other
 EnvironmentalAccountingin DevelopmentPolicy:TheFrenchExperience 53

2. Environmental media in monetaryterms all stock or capital goods held by eco-

Soils (exposed and immerged) nomicagents(such as housing,capitalequipment,land, sub-
Continental waters (surface water, groundwater, soil resources, livestock,and forests) as well as financial
snow, and glaciers) assets (such as gold, currency,and stock).
Ocean waters (coastal waters and open sea) 3. For Landsat,the taking, processing,interpretation,and dis-
Atmosphere (air, upper atmosphere, and play of imageson a scale of 1:100,000costs roughlyone
climate) tenth of photomappingusingcolor infrared.
3. Living organisms 4. For example,only40 Landsatimageswere neededto estab-
Animal species (wild and domestic)-according lish a 1982-83 inventoryof terrestrial resourcesfor half of
to standard scientific nomenclature Mali,whereasthe same area wouldhaverequired5,000 aerial
Plant species (wild and domestic)--according to photos on a scale of 1:50,000(excludingground surveys).
standardtscientific nomenclature 5. ESCAPplans to supplementits report on the state of the
standard scientific nomenclature environmentwitha sectionon resourceaccounting,basedon
one or two case studies.
B. Nomenclature of ecosystems
1. Water ecosystems References
Open sea
Coastal ecosystems de Jouvenel,Bertrand. 1966."Propositiona la Commissiondes
Inland ecosystems Comptesde la Nation."Ministeredes Finances,Paris. Pro-
2. Terrestrial ecosystems cessed.
Forest Dravie,ChristopheA., GeorgesHainnaux,and Jean LouisWe-
Woodland and pasture land ber. 1983. "Essaid'Etablissementd'un Comptedes Sols en
Health C6te d'Ivoire." Regional African Seminar on Natural Re-
Meadow and land under cultivation sourceStatistics.November.Abidjan.Processed.
Turf ESCAP (Economicand SocialCommission for Asiaand the Pa-
Pioneer ecosystems cific).1982. "Reviewand Appraisalof EnvironmentalSitua-
3. Other ecosystems tion in the ESCAP Region."Bangkok.Processed.
JARIW-ECA(InternationalAssociationfor Researchon Income
C. Nomenclature of agents and Wealth-EconomicCommission for Africa).1982. Record
According to the traditional distinction of the SNA of the regionalAfricanconference,November15-20. Vol.11.
among: Douala,Cameroon.Processed.
Households InstitutNationalde la Statistiqueet des Etudes lconomiques
Business firms (INSEE)et Ministerede l'Environnement.1986.LesComptes
Administrations du PatrimoineNaturel-La Documentation Francaise.Paris.
Other (such as foreign entities) Myrdal,Gunnar.1970. The Challengeof WorldPoverty.New
Other subcategories: York:RandomHouse.
Business: environmental firms or managing bodies Rapport,David,and AnthonyFriend. 1979."Towarda Compre-
(agricultural and nonagricultural) hensive Frameworkfor EnvironmentalStatistics:A Stress
Administrations and environmental public bodies ResponseApproach."StatisticsCanada,Ottawa.Processed.
(associations, local administrations, and central UnitedNations.1982.Surveyof EnvironmentStatistics:Frame-
administrations work, Approaches,and StatisticalPublications.Statistical
Papers,seriesM, no. 73. NewYork.
Notes UNEP (United Nations Environmental Programme). 1981.
"Earthwatch: An In-depth Review.EnvironmentalData."
1. The term "patrimonyaccounts"is a directtranslationfrom eport 1. Nairobi.Processed.
the French language.The Frenchsystemis verybroad and Weber,J. L. 1983. "The FrenchNaturalPatrimonyAccounts."
includes,in additionto naturalassets,those witha historical StatisticalJournalof the UnitedNationsEconomicCommis-
or culturalvalue. sion for Europe1:419-44.
2. Economicpatrimony accounts have been prepared since World Bank. 1981. AcceleratedDevelopmentin Sub-Saharan
1986withinthe frameworkof the expandedSNA. Theyassess Africa:An Agenda for Action.Washington,D.C.

LinkagesbetweenEnvironmental
andNationalIncomeAccounts
Richard B. Norgaard
Almostfour decadesago K. WilliamKapp(1950) argued
that both our conceptionof the developmentprocessand
our nationalincomeaccountswere inadequateior devel-
opmentplanning.Neither our theories nor our primary
indicatorsof economicdevelopmentincorporatethe serv-
ices of the environmentand of resources.This conceptual
problembecamea popularconcernduringthe 1.960s,as
awarenessof pollutionand environmentalmar,agement
problemsincreasedin the developedcountries, By this
time the shortcomingsof the grossnationalproduct (GNP)
becamemore and more apparent. GNP, as conventionally
measured, could be increased by using up limited re-
sources faster at the expense of future generations. It
could also be increasedboth by degradingthe environ-
ment and by correctingenvironmentaldegradation.
Duringthe 1970seconomistsand othersbeganto make
serioussuggestionsfor changesin the Systemof National
Accounts(sNA) to correct the shortcomings.In spite of
the logic behind many recommendations,however,the
onlysuggestionthat manycountrieshaveadoptedin their
national accounts is to keep separate track of environ-
mentalcorrectionexpenditures(nameddefensiveexpen-
ditureselsewherein this volume).
Since the early 1970s an increasingnumber of prob-
lems in developingcountries also have been linked to
inadequateconsiderationof environmentaland resource
factors.Both specificdevelopmentprojectsand general
strategies failed to reflect environmentalopportunities
and constraints.In the early 1970s the politicalleaders,
governmentadvisers,and academicsin developingcoun-
tries rejectedthe concept of an environmentalcrisis as
perceivedby their counterparts in developedcountries.
Duringthe past decade,however,someof them havebeen
active in reformulatingdevelopmentthinkingto include
the relationshipsbetween well-being,environment,and
resources.
The internationaldevelopmentorganizations-as major
54
8
users of informationfrom the SNA, as advisersto devel-
opingcountries, and as coordinatorsof systematicdata
collectionand exchange-have initiatedvariouseffortsto
improvethe usefulnessof environmentaldata. The Con-
ferenceof EuropeanStatisticiansof the EconomicCom-
missionfor Europebegan to investigatethe organization
and use of environmentaldata in 1973.
The U.N. StatisticalOffice(uNso) began working on
the problemsof organizingenvironmentaldata in 1974.
In 1983 the first of several EnvironmentalAccounting
Workshopswas held. In late 1984 UNSO presented "A
Frameworkfor the Developmentof EnvironmentStatis-
tics."
The EnvironmentalAccountingWorkshopshavestrug-
gledwith the complexmethodologicalissues of howvari-
ous heterogeneous environmental data should be se-
lected, organized,aggregated,and incorporatedinto or
linked with the informationprovidedby the SNA. Strong
arguments,presentedpredominantlyby economists,have
been made that the SNA is already an effectivetool. Its
defendersbelievethat it simplyneeds to be improvedby
includingboth the valueof environmentaland resource
servicesin the current account and the variationin the
value of resource stocks and environmentalsystemsin
the capitalaccount Someeconomistsbelievethat appro-
priate techniques exist to fullyincorporate critical re-
source and environmentalvaluesin the SNA. Others be-
lieve that environmental accounting should be
independentof economicaccountingand that linkages
should be made betweenthe accountsas appropriatefor
differenttypes of questions.
Economistshave developednumeroustechniquesdur-
ing the past three decadesfor estimatingthe monetary
valuesof resourceand environmentalservices.The values
of the environmentsor environmentalfactors and of the
resourcesfrom whichthese servicesderive are the pres-
ent valueof the services.Economistshavetried to ground
 LinkagesbetweenEnvironmentaland NationalIncomeAccounts
these techniquesin economictheory,have obtainedade-
quate data in some cases, and have establishedeffective
econometricprocedures.' To most economists,environ-
mentaland resourcevaluationis a provenarea, onewhere
success appears consistentlypossibleafter sufficientef-
fort. Thus most economistsconcernedwith environmen-
tal accountingargue that the greatest problemsof the
SNA with resources and environmentalsystemscan be
correctedby developinga good systemof accounts,with
everythingmeasuredand aggregatedin monetaryterms.
Furthermore,with everythingin monetary terms, plan-
ners could readilycompareoptions.
The skepticsare concernedwiththe limitsof economic
valuationor of any other single technique. They argue
that the environmentalaccountsshouldbe maintainedin
physicalunits to the extent possible and thait multiple
approachesto linkingaccounts with the SNA should be
developedand used as appropriate,given the question
under consideration.They have been unified by their
skepticismof the economists'arguments,and they agree
that questionsrelatedto appropriatedevelopmentare too
important, diverse, and complicated to be pursued
through a single approach. But they have so far not
reacheda consensuson an alternativeapproachor com-
binationof approaches.2
The methodologicaldebate has circledaround the ec-
onomicapproach versusthe use of diverse,less formal,
methods.Manyof these other methodshavebeen exper-
imentedwith in the French systemof envirornmental ac-
3 A consensus,however,has hardly developed
counting.
on whichmethods might be conceptuallymc,resuitable
or practically more effective under different circum-
stances.
sNA Developed by Consensus, Not Theory
Most lay people and many planners assume that in-
creasing the GNP means that the economicwelfareof
peoplehas improved.But GNP only measuresthe levelof
economicactivitythat goes through the market mecha-
nism.Welfareis best thought of as the sum of producers'
and consumers'surpluses-the difference between the
areas under demand curves and above supply curves.
When these areas increase,other thingsheld equal,wel-
fare has improved.Expendituresare an awkwardcombi-
nation of the cost of supplyinggoods and servicesand
the producers'surplus. And other things are rarely held
equal. GNP increaseswhen subsistencefarmersjoin the
labor force and stop producing their own food, when
womenjoin the labor force and hire domestichelp and
childcare, when the stock of housing and factories is
destroyedby war and then replaced,and when the envi-
ronmentdeterioratesand the governmentinitiatescorrec-
tive actions.
The SNA has these internal inconsistenciesbecauseof
the difficultiesof buildingaccounts that are consistent
55
with economictheory.The derivationof the SNA stems
from heated debate as to what should be includedand
how.Thus, the SNA is based on conventionsestablished
through a process of acquiring consensus rather than
basedon proceduresdeduced fromeconomictheory.
The Value Aggregation Issue
Muchof the interestin environmentalaccountingstems
from a concern with the sustainabilityof development.
Economistsargue that sustainability,to the extent that it
is important to people, is incorporated in individuals'
values and reflected in the decisions they make. For
economists,one goal of environmentalaccountingis to
provideappropriateindicatorsof people's willingnessto
maintain or improvethe quality of the environmentas
developmentoccurs.Valuationtechniquesfor nonmarket
goods and services, however,rely on individualprefer-
ences about the future as expressedfor goods and serv-
ices in those marketsthat are working.
Environmentalistsargue, however,that they are ethi-
cally compelledto seek sustainable development. For
these people,sustainabilityshouldbe a constrainton, or
separatesocialobjectiveof, developmentstrategies.Even
if a societydoes not select sustainabilityas a constraint
or objective,it should explicitlymonitor the extent to
whichcurrent preferencesjeopardizefuture generations.
Economictechniques, however,aggregatekey environ-
mentalinformationof importanceto socialdecisionmak-
ers and considersocialvaluesby initiallyweightingthem
with respect to individualpreferences as they are ex-
pressedin privatemarkets.
Economistsmaketradeoffsbetweengenerationson the
assumptionthat interest rates reflect optimal exchanges
between present and future generations. The fact that
future generationscannot possiblyparticipate in capital
marketsprovidesa clear rationalefor social intervention
in capitalmarketsand independentsocialdecisionsabout
conservingresources and environmentalsystems.Socie-
ties could,and to someextent do, protect future genera-
tions through interest rate policy.Societies frequently
makedecisionsto directlyreserveresourcesand environ-
mental systemsfor future generations. The information
for makingthese socialdecisionsshouldnot be presented
solelyin a formreflectingthe currentgeneration'svalues,
as expressedindividuallyby the rate of interest generated
in capitalmarkets.
Althoughthe issue of sustainabledevelopmentis rela-
tivelynew,the philosophicaldivisionbetweenthinkingof
socialgoalsas the sum of the goalsof a society'sindivid-
ualsand thinkingof societyas havingits owngoals above
and beyond those of its individualshas deep historical
roots. Economictheory is based on the formerpremise,
but all societiesoperate on the second premiseto some
extent, mostto a considerableextent. In the publicpolicy
arena, economistsare repeatedlyconfrontedby the con-
56 RichardB. Norgaard
tradictionsbetweenthe premiseupon whichtheir indica-
tors are basedand the premiseuponwhichthey are used.
The ethicaldebate has theoreticalanalogs.The contra-
dictionsbetweeneconomicindicatorsbasedon one prem-
ise about societythat are used in publicdecisionsinvolv-
ing anotherpremiseunderlie,for example,the theoretical
controversyover the use of benefit-costanalysisto effi-
cientlydesignpublic projects. If one of the objectivesof
a project is to redistribute income, what is the signifi-
cance of an efficiencyanalysisderived from valuations
basedon prices generatedby the existingincomedistri-
bution?The weightingof benefitsand costsdependingon
their incidenceby incomegroups is currentlypracticed,
but it is quite controversialamongeconomistsbecauseof
the valuejudgmentsit involves.
Withrespect to sustainability,the parallelsare clear.If
one of the objectivesof public decisionmakingis to en-
hance the likelihoodof sustainabledevelopment,what is
the significanceof environmentalvaluationsderivedfrom
informationgeneratedby an economythat is less sustain-
able than desired?Withinenvironmentaleconomicssimi-
lar questionsare now being theoreticallyconsideredby
comparingtotal systemswithand withouta policychange.
Usingthis approach,the valuesof environmentalservices
are derivedfromthe differencesin utilitygainedwith and
without the policy change. This theoretical "solution"
reversesthe processadvocatedby economistsforenviron-
mentalaccounting:environmentalvalueschange as a re-
sult of the policyinput rather than environmentalvalues
providinginput to the policychange.
The determinationand use of interestrates reflectsthis
theoreticalquandary.Economistsargue that the valueof
resourcessavedfor the next generation,say twentyyears
hence, is equal to their valueat that time discountedby
dividingthe valueby one plus the rate of interest to the
twentieth power. Giventhe economist'stheory of how
interest rates are determinedin intergenerationalcapital
markets,however,it would be just as appropriateto ask
how future generationswould interact differentlyin the
capitalmarket and affectthe rate of interest if they were
wealthierbecause this generation had redistributed re-
sourcesin their favor.
The conventionaleconomicanswer to this issue has
been to argue that market values with and without a
policychangewillbe veryclose for small changes.This,
however,wouldbe incorrectif the ultimateobjectivewere
to significantlychange the direction of development.In
short, if the currenteconomicdevelopmentpath wassuf-
ficientlyinappropriatethat significantcorrectionswould
be calledfor, current market signals would not provide
appropriate indicators of how the path could best be
corrected.Signalsgenerated by economicmethodology
are best for fine-tuningon the margin.Quiteconsistently,
supporters of, and detractors from, the economicap-
proachtend to hold differentopinionsas to whetherfine-
tuningsor significantchangesare needed.
Economictechniquesfor valuingnonmarketgoods and
servicesare alsobased on the assumptionthat nonmarket
goods are a marginalproblem. If nonmarketgoods and
servicescomprisea smallportion of the total economy,
then the inclusion of these goods and services in the
market wouldnot significantlychange the relativeprices
of goods and servicesor individualbehavior.In this case,
existingmarket prices and behaviorcan be used to esti-
mate the valueof nonmarketgoods and services.If non-
market goods and servicescomprise a large portion of
the total economy,however,their inclusionwould signifi-
cantlychangerelativepricesand the valuationof nonmar-
ket goods and services. Again we are faced with the
dilemmathat economictechniquesare more appropriate
the less significantthe problem. And again, advocates
and opponentsof the economicapproachare dividedas
to whetherour problemsare smallor large.
The economicapproachsuffersfromseveraltheoretical
shortcomings.One mightconcludethereforethat another
approach is needed. But all approaches contain similar
methodologicaldrawbacks.Any method of aggregating
environmentaland resource data, for example, implies
valueweightingswhich"should"be made by society.Yet
societyas a wholecannot be involvedin everydecision.
Some methodof weightingis necessaryso that those to
whomsocial decisionmakinghas been delegatedcan as-
sess current conditions,foreseethe consequencesof al-
ternativecoursesof actions,and make decisions.Indeed,
evenif all decisionswere madecollectively,aggregation-
with implicitvaluations-wouldbe necessaryto convey
informationin a usableform to the public. This issue is
beingreferredto as the valueaggregationproblem.
The valueaggregationissue can be used to reduce the
methodologicaldebate to an ethical debate, with each
side leaving the proponents of the other methodology
with little more than their valuessupportingtheir argu-
ment.But one can look at the issuealso in the following
constructive way. Each methodology gives different
weightsto the valuesaround whichinterestgroupsorgan-
ize. With multiple methodologies giving alternative
weightsto differentinterests,the interests of all could be
protectedand a Pareto optimal solutionmightbe possi-
ble.4 The valueaggregationproblemin this light becomes
an argumentfor methodologicalpluralism-for the use of
multiplemethodologiesto help insure that all valuesare
respectedto the extentpossible.
The Bounded-KnowledgeSynthesis Issue
All modelsof social and environmentalsystemssuffer
from another problem. Scientificknowledgestems from
specificmodelsthat simplifyreality,in part, by artificially
boundingthe field of study.Sciencehas not provideda
wayto synthesizeboundedknowledgeinto knowledgeof
the whole.Yet such a synthesisis exactlywhat environ-
mental accountingneeds in order to derive better indi-
 LinkagesbetweenEnvironmentalandNationalIncomeAccounts
cators for planners.This problem willbe referredto as
the bounded-knowledge synthesisissue.
Eachmethodologyfor aggregatingboundedknowledge
worksbetter for some types of knowledge,less wellfor
others. Preferencefor one methodologyover anotherfre-
quentlystems from an individual'sspecial understanding
of a particularaspectof a problemand relativenaiveteof
other aspects. Dependingon the question being asked,
somelinkingmethodologieswillbe ableto addressknowl-
edge more closelyrelated to the question than others.
One procedure,for example,might effectivelyaggregate
soil fertility with economicproductivity,while another
might aggregatethe abilityof the soil to hold moisture
duringdroughtsto the variationsin the need lbr foreign
exchange.Each methodologyfor aggregatinghas advan-
tages dependingon the questionbeing asked.
The bounded-knowledge synthesisissue also supports
a pluralisticapproach to aggregatingenvironmentaland
economicinformation.The rule of "safetyfirst" would
indicatethe use of multiplemethodologiesto insurethat
decisionmakers haveinformationalertingthemto as many
aspects of environmentaland resource phenomena as
possible.Decisionmakers shouldnot be limitedlto knowl-
edgeboundedby the particularmodelsbest addressedby
one linkingmethodology.Plannersand policymakersdo
not alwaysknow whichquestions should be asked and
whichmethodologiesare most appropriate.
The valuationof environmentalservicesand resources
and of their respectivesystemsand stocks is a methodol-
ogy that facilitatesnothing more sophisticatedthan addi-
tion. Marketgoods and servicescan be combinedwith
nonmarketenvironmentalfactorsafter valuationto obtain
better measuresof the level of economicactivity.Simi-
larly,the valueof natural resourcestocks and of environ-
mentalsystemscan be added to that of man-madecapital.
If better measuresof the levelsof economicactivityand
the valueof assets are needed,valuationis an appropriate
approach.
Policymakersand plannerswantto knownot only how
wellthe economyis doingbut howit mightbe improved.
For this they need to be able to predict the cutcomesof
alternativedecisions.And predictionsrequire modelsof
cause and effect-of how differentmixes of economic
activity affect, and are affected by, resources and the
environment over time. GNP is like a speedometer, and
improvements in a speedometer's accuracy could prove
quite important. But a speedometer does riot indicate
whether the wheels are spinning on ice or even on the
road. To address the questions of long-term economic
and environmenta
a mapise planning,V
. n
candenvironmentl planninga m is neededt . Valuation
can still be helpful when combinea with explanatory mod-
els, but for most questions valuation itself is inadequate.
In summary, the theoretical component of the debate
over the appropriate methodologyfor combining environ-
mental and economic infornation has been sustained by
the belief that there is a correct method for understanding
57
the complexinteractionsbetweeneconomicdevelopment
and the environment. 5 The value aggregation and
bounded-knowledge synthesisproblemsindicatethat this
beliefis naive.A consensusfor one methodologycan only
be reached by "nonlogical"means.Logic,however,indi-
cates that multiple methodologies-conceptual plural-
ism-providethe keyto a safer and more pragmaticstrat-
egyfor linkingenvironmentaland economicaccounting.
Sustainabilityis too important,too multidimensional,and
too poorlyunderstoodfor societiesto rely on one meth-
odology.
Conclusions
The approachto combiningenvironmentalinformation
with the SNAfacescertain difficultiesbecauseof the deri-
vationof the SNA(by consensusrather than a deductive
methodology)and its attendant shortcomings;the value-
aggregationissue;and the boundedknowledge-synthesis
issue.
The existenceof these problems means that broad,
aggregateindicatorsfor planners willhaveto be derived
by consensusrather than by a deductivemethodology.
Each specificapproach advocatedby one group can be
shownto containshortcomingsor inconsistencies.There
is thereforea need to pragmaticallywork toward a con-
sensus. Althoughno one methodologycan be logically
correct,the use of multiplemethodologieswillreducethe
likelihoodof makinga significanterror.
In conclusion,less effort shouldbe spent on debating
and more on experimenting,learningthrough doing,and
sharingexperiences.It seemsappropriateat this stage to
try two or three differentapproachesto environmental
accountingin each of four or fivecountrieswith different
types of environmental problems. Future workshops
shouldbe directed towarddiscussingthe practicalissues
encounteredand sharingthe results.
Notes
1. For a generalreviewof and guideto the use of these
techniques,see Hufschmidtand others (1983). Much of the
theoreticaldevelopment
ofthesetechniqueshas beenpublished
in the Journalof EnvironmentalEconomicsand Management.
Peskin (1981) providesthe clearest description of the use of
these techniquesto modifythe SNA.
2. This is welldocumentedin the rapporteurs' summaryof
the meetingin Washington,D.C.,November1984.
3. The Frenchsystemof environmentallinking,for example,
has manyattractivefeatures,but those opposedto the econo-
mists' view cannot agree on its strengths and weaknesses. See
Chapter7.
4. A Pareto optimalsolutionis one in whichsome or many
individualswouldbe better off and, at the same time, nobody
wouldbe worseoff.
5. "Criticalrationality"-the beliefthat rational methodsex-
ist for applyingscientificknowledgeto socialquestions-is the
58 RichardB. Norgaard
dominantbeliefamong scientists,economists,and policymak-
ers. The belief is supported, althoughnot "logically,"by the
mainstreamepistemologyof Popper (1979). During the past
decadethe dominantbeliefsystemhas been attackedbyphilos-
ophers (Feyerabend 1974 and 1978), by systems theorists
(Churchman1979),and by planners(Ulrich1983).
References
Churchman,C. West. 1979. The Systems Approach and Its
Enemies.NewYork:BasicBooks.
Feyerabend,Paul. 1974. Against Method. London:NewLeft
Books.
. 1978. Science in a Free Society. London:New Left
Books.
Hufschmidt,M.,D.E. James,A. D. Meister,B. T. Bower,and J.
A. Dixon. 1983. Environment, Natural Systems, and-
Development: An EconomicValuationGuide.Baltimore,Md.:
Johns HopkinsUniversityPress.
Knapp, K. William.1950. The Social Costs of PrivateEnter-
prise. Cambridge,Mass.:HarvardUniversityPress.
Peskin, Henry M. 1981. "NationalIncomeAccountsand the
Environment."NaturalResourcesJournal21(uly):511-37.
Popper,Karl. 1979.ObjectiveKnowledge:An EvolutionaryAp-
proach. Revisededition. Oxford, U. K.: OxfordUniversity
Press.
Ulrich,Werner.1983.CriticalHeuristicsof SocialPlanning:A
NewApproach.Bern: Paul Haupt.
UnitedNations. 1984. A Frameworkfor the Developmentof
EnvironmentStatistics.StatisticalPapers, series M, no. 78.
NewYork.
 __ _ _ _ _ _ 9
EnvironmentalandNonmarketAccounting
in Developing
Countries
Uenry M. Peskin
National accountingsystems, in one form cr another,
exist in nearly every nation, regardless of the nation's
stage of development.The United Nations encourages
standardizationboth in accountingproceduresand in the
layoutof final tables-presumably to make it easier to
compare economic performance across nations. One
could argue, however,that there is a special need to
depart fromthe standardin developingnations. In partic-
ular,they maybenefitmore than developednations from
modificationsthat better accountfor the coniributionof
environmentaland other nonmarketassets.
An importantcharacteristicof these nations,and a key
indicator of their stage of development, is the large
amount of economicactivitythat occurs outside organ-
ized markets. This activityhas not been ignored by na-
tional accountantsin these countries. Acccrdingto a
surveyconductedby the OECDDevelopmentCentre,most
of the sixty-fiverespondingnations accountl:o some de-
gree for nonmarket,subsistenceactivities(Blades1975).
There is no evidence,however,that any of these nations
accountfor the servicesof the environment,the principal
nonmarketasset that is the concern of ministriesof the
environment.
This chapter points out howa system of national ac-
counts, possiblydevelopedalong the lines of the frame-
work described in the next chapter, could better serve
the needs of developingnations. Certainimplementation
problemsspecificto these nations are discussedand a
courseof actionsuggested.
Uses of Expanded Accounts
in Developing Countries
In developingas wellas developedcountries,there are
two principaluses of the accounts:to generate measures
of economicperformanceand to organizedata on eco-
59
nomicactivity.The OECD surveydid not ask the respon-
dents whythey engaged in nonmarketaccounting.How-
ever,since only a dozen of the 150 nations classifiedas
developingdisplaythe nonmarketinformationseparately,
it can be inferredthat more importanceis being placed
on providingbetter measuresof economicperformance
than on buildingbetter data systems(Blades 1975, p.
406).The concernthat neglectingsubsistenceeconomic
activitycauses total economicactivityto be underesti-
mated seems well justified.Of the forty-eightsurveyed
countriesthat were able to determinethe share of gross
domesticproduct (GDP)accountedfor by subsistenceac-
tivities,nineteenestimatedthat these activitiesaccounted
for more than 20 percent of total GDP. Moreover,this
percentage must surely be an underestimatesince, as
Bladespoints out, many nonmarketproductionactivities
are not accountedfor at all. For example,most countries
surveyeddid not make an estimate for food processing,
handicraft,construction,water portage, and transporta-
tion activitiesin the householdsector.
The OECDsurveydoes not indicatewhetherany of the
respondents attempted to account for any nonsubsist-
ence, nonmarket activities-in particular,the consump-
tion ofthe servicesof the environment.To myknowledge,
no nation, developedor developing,has adoptedformal
accountingfor the servicesof the environment,although
in recent yearsthe U.N.EnvironmentProgramme(UNEP)
has beenattemptingto determinewhatrecommendations
should be made to member nations along these lines.
UNEP and the World Bank are concerned, with some
justification,that neglectof environmentaldeterioration,
and especiallydeterioration of the stock of natural re-
sources,willgive a false picture of economicgrowthas
measuredconventionallyby national account aggregates
such as the net national product (NNP).'Thus, although
the neglect of many subsistence nonmarket activities
60 Henry M Peskin
causes conventionalnational accountingprocedures to
generatedeficientindicatorsof economicperformancein
developingnations, perhaps a more importantreason is
their neglectof the deteriorationof the stock of natural
resources and environmentalcapital. This neglect has
receivedmuch recent attention becauseof the failureof
developmentpoliciesto generatesustainedgrowth.
Historically, overlookingdeclinesin the stock of natural
resourcesin the conventionalaccountswasthoughtto be
of littlesignificance.This viewwas basedon the fact that
discoveriesof new natural resourceswere also neglected,
as were other resource-increasingfactors such as re-
charge and the natural growth of renewableresources.
Therefore,as long as discoveries,recharge, and growth
equaledor exceededdepletions,this mutual neglectwas
thought not to have any adverse social consequences
(Landefeldand Hines 1982).2The current perception,
however,is that discoveries,recharge,and naturalgrowth
are not keepingup with depletion.Moreover,for stocks
of such environmentalcapital as clean water, clear air,
and natural forests,the potentialfor new discoveriesand
rapid increasesin growthappear slimindeed. Therefore,
as long as maintainingthe stock of these resources is
consideredessential to sustainedgrowth in developing
nations(an assumptionexaminedin the next section),the
neglectof their deteriorationin the accountsis especially
serious.
With respect to the second traditionaluse of national
accounts-a frameworkfor the consistent assemblysf
economicdata-conventional accounting practice that
neglectsnonmarketactivityis alsodeficientfor the needs
of developingnations. Althoughthis deficiencyhas not
receivedmuch attention in the literature, it should be
clear that if nonmarketactivityis widespreadin an econ-
omy and if such activityis ignored in national data sys-
tems, then these systemswill not be able to support
accurateanalysesof economicbehavior.
Althoughcynicsmayfeel that such analyticaldeficieri-
cies affectthe communityof analystsfar more than the
communityof developingnations,the lackof a framework
for assemblingnonmarketdata can havemore than mere
academicconsequences.Developmentplannershave be-
comeincreasinglyawarethat a lack of data on nonmarket
activities,especiallythose that lead to negativeexternali-
ties such as pollution, mayproduce a distorted viewof
the likelybenefits of actual and proposed development
projects.I do not believethat they are as equallyaware
of the contributionthat a completeaccountingof market
and nonmarketactivitiescan maketo the analysisof the
opportunitycosts of actual and proposed development
projects.Indeed,for estimatingtrue opportunitycostsas
wellas benefits,it is the accountingframework,and noi:
merelythe assemblyof nonmarketdata, that can make a
significantanalyticalcontribution.
This assertionderivesfromthe fact that, in the absence
of perfectlycompetitivemarkets,observedmonetarycosts
will not measure true opportunitycosts, and observed
output prices willnot measure true socialbenefits.Given
thiswell-known resultfromeconomictheory,development
economistshaveattempted to construct shadow cost or
price measures,whicheither supplementor replace the
conventionalmonetarymeasures. 3 It seems, however,
that
much more attention has been directed towardthe bene-
fit side than the cost side, even though theory suggests
that observedcosts and pricesfail to adequatelymeasure
valuesfor reasonsthat are just as applicableto costs as
they are to benefits. Based on the relative number of
books and articles devotedto each subject, one would
haveto concludethat plannersare far more satisftedwith
their measuresof cost than with their measuresof bene-
fits.
Clearly,however,whenthere is a large degree of non-
market activityin an economy,monetarymeasures can
be just as deceivingfor estimatesof true opportunitycosts
as they are for estimates of social values or benefits.
Moreover,rectifyingthe problem can be just as difficult.
There is, in fact, no easy quickfix.Althoughthe literature
impliesthat correct cost estimatescan be madeby adjust-
ing the observedvalues,determiningthe appropriatead-
justment factors is no easy matter (Little and Mirrlees
1969). For many years,the planning professionhas rec-
ognizedthat the analysisof complexeconomicinterde-
pendenciesrequiressome typeof interindustryor activity
analysis.One of the most importantcontributionsof con-
ventionalnationaleconomicaccountinghas been its abil-
ity to generate the data sets necessaryto support such
analyses.An expandedaccountingsystem that includes
nonmarket,environmentalactivitiescan serve this tradi-
tional role for those developingeconomiesfor whichthe
nonmarketedeconomicactivitiesare as importantas the
marketedones.
Implementation Problems
Severalproblems,uniquein the developingnations,can
arise if the attempt is made to implementan expanded,
nonmarketaccountingframework.Two of these are dis-
cussedin this section.These problemsare not critical in
the sense that their solution is a necessaryprecondition
for implementation.In particular,the valuationdifficulties
discussedbelowmaynot be fullysolvable.However,the
failureto find acceptablevaluationsfor nonmarketgoods
and serviceswill not diminishthe usefulnessof the ac-
countingstructure as a data systemto support economic
planning,particularlyin its abilityto shed light on the
true opportunitycosts of alternativedevelopmentstrate-
gies.
Accounting for EnvironmentalAsset Depreciation
The followingchapter emphasizesthe important dis-
tinctionbetweenphysicaldeteriorationand economicde-
 EnvironmentalandNonmarketAccountingin DevelopingCountries 61
preciation.Onlythe latter is properlyaccounteclfor in an
economicaccountingframework. 4 As noted in that chap-
ter, it is quite possible for a nonmarket (or a market)
asset to remain physicallyintact but to depreciate eco-
nomicallyin the sense that the present valueof the ser-
vicesgeneratedby the asset declinesover time. The op-
posite is also possible.An asset, such as the stock of
standingIndonesianhardwoods,can suffer sorne degree
of physicaldeteriorationand still haveits present value
increaseover time, whichimpliesnegativedepreciation,
or capitalgain.
The effort to introducethe concept of true economic
depreciationinto nonmarket accounting in developing
nations raises a number of issues, some of whichare
amenableto technicalsolutions.Other issues are funda-
mentallymore serious,however,in that they addressthe
wholepurposeof the accountingexerciseand 'the poten-
tial conflictsbetweeneconomicand conservationobjec-
tives.Thus, these latter issuesare addressedfirst.
Environmentalists and, increasingly,economicplanners
are concerned that developmentprograms may deplete
certain vital natural resources. As a result, short-term
economicprogress could be followedby long-termeco-
nomicdecline-that is, the growthmaynot be sustaina-
ble. A related concernis that evenif growthmaybe able
to continueunabated,its compositionmaynot,be suited
to the tastes of future generations.Unlessthe full menu
of assets now availableis conserved,future generations
will be deprivedof the options availableto the present
generation.Can and should the economicaccounts,ex-
pandedto includethe servicesof the environment,reflect
these concerns?
The resources in question need not be confined to
nonmarket,environmentalassets. In fact, often the ex-
amplescited as such are marketedassets,such as miner-
als and farmland.The argumentis that the privatemarket
mechanismis not sufficientto ensure that the stock of
these assets willbe adequatelypreservedfor future gen-
erations.(Presumablythe problemis just as severe,if not
worse,for assets that do not trade in the marketplace.)It
is quite possiblethat a private economyactingoptimally
(to maximizethe presentsocial valueof its assets)could
bring itself to a point of extinctionby exhaustingeven
one of its vital natural resources(Page 1977; Lind and
others 1982).There is no inherent mechanismin a free
market economy to prevent possible self-destruction.
Therefore,the value of vital assets as revealedby the
market underestimatestheir true valueto the very exis-
tence of the society.Accordingto this line of thinking,
the socialvalueof such assets should not be determined
by the present value of income generatedin the future,
but rather by an appealto a conservationethic, an ethic
that rejects economicvaluationas inadequateto reflect
the needs of future generations.
Before examiningthis argument, it should be noted
that ifthe depletionof someasset doeseventuallylead to
the demiseof an economy,there does not appear to be
any convenient-way to reflect this fact in a single set of
economicaccounts. One year's income accounts could
reflect the shrinking of the asset with a depreciation
entry,and a singleset of wealthor asset accounts could
displaythe remainingstock. Only by analyzingsuch ac-
counts overseveralyears could inferencesbe drawncon-
cerning whenthe stock willbe exhausted.Whetherthe
exhaustionof the stock willmeanthe demiseof economic
activitycannot be revealedby the accounts, except, of
course, in the year when the demise occurs. Thus, ac-
countingfor the depreciationof natural resourcesmight
make only a minor contributionto preventingthe even-
tual collapseof the economybecauseof the exhaustion
of a natural resource.
To speak of eventualeconomiccollapseis certainlya
dramatic characterizationof the sustainabilityproblem.
The dismalresult relieson one or more of the following
assumptions:that there is no technicalprogress,that the
exhaustibleresourcehas no effectivesubstitutes,and that
there are, indeed, identifiableassets whose contribution
to the economyis so crucialthat theirphysicalexhaustion
indeedwouldmean economiccollapse. 5 Page apparently
feelsthat it is fairlysimpleto identifynatural resources
that meet these characteristics;he cites, for example,the
establishmentof a strategicstockpileof critical materials
for national defense (Page 1977, p. 184). However,a
strategicstockpileis intendedonly to fulfilldemandsdur-
ingfairlyshort-termeventssuch as droughtor war.Decid-
ing what assets are critical to the sustainabilityof eco-
nomicgrowthcertainlyrequiresa longertime frame.
In the context of developingnations, one naturally
thinksof such resourcesas agriculturalland, forests,and
watersupplyas potential candidatesfor long-runconser-
vation regardless of any calculation of present value.
Those countrieshighlydependenton the export of min-
erals would certainlyadd such mineralsto the list. Yet,
upon closer examination,it is not so clear which,if any,
of these assets are crucial to long-runsustainability(in
Page's sense),and, more important,it is not clear to what
extent their calculatedpresent valuesdistort their true
socialvalues.
Of the assets mentioned above,water, and especially
the supplyof potable water,appears to be the most cru-
cial. Because of the lack of substitutes, its demand is
inelastic(assumingthat desalinizedsalt wateror imported
freshwater are not practicalalternatives).Therefore,it is
quite possiblethat as a nation's supply of fresh water
diminishes,its unit valueincreasesat a greater rate, lead-
ing to an ever-increasingimputed or measuredvalue of
fresh water wealth. Clearly,in this case, the apparent
increase in the nation's water wealth would give a dis-
torted picture of social well-being.Can the same case be
made forthe criticalityof other, so-calledvital assets?To
answerthis question,it is usefulto considerone poten-
tiallycriticalasset, agriculturalland.
62 Henry M. Peskin
Whetheragriculturalland is as criticalas potablewater
depends on what one means by "agricultural land."
Clearly,the elasticityof demand for the aggregateof all
agriculturalland is quite differentfrom the elasticityof
demandforthe individualparcelsthat makeup the aggre-
gate. If these componentsare highly substitutablefor
each other, taken individuallynone could be considered
vital.Moreover,becausethe substitutabilityimplieshigh
elasticitiesof demand, physicalreductionsin individual
parcelswillimplyreductionsin their values.That is, the
reductionsin quantitiesare less likelyto be offsetby high
increasesin prices. Therefore,a present-valuewealthac-
countingwould not falselyindicate an increasein land
wealth.
Whatif someof the parcelswerehighlyspecializedand
the elasticityof demandwerehighlyinelasticfor the prod-
ucts producedon these parcels?In this case,it is possible
(but by no means assured) that a present-valuewealth
accountingcould indicatean increase in wealtheven as
the physicalstock of such land declined.Althoughthis
could possiblybe misleading,there would not be any
serious social consequencesunless the products gener-
ated by this land were so crucialthat the entire economy
dependedon them.
Certainlyit wouldbe disastrousin a developingcountry
if the wealthaccountsindicatedthat the amountof land
(in valueterms)devotedto, say,a crucialexport crop was
increasingwhen,in fact, it was decreasing.However,the
existenceof such a possibilitydoes not justifydeclaring
all agriculturalland as crucialand, therefore,not subject
to present-valuevaluation.Moreover,whatmayappear to
be crucialin the short run may be quite expandablein
the longrun. The stock of silk wormsis hardlycrucialto
the present-day Japanese economy, but it may have
seemedso in the sixteenthcentury.Indeed,it could be
arguedthat, givenhumaningenuity,virtuallyall assetsare
substitutablein the long run. If so, rather than searching
for sustainablelevelsof particularvital assets,the appro-
priate conservationprinciple mightbe to maximizethe
levelof all assets in the aggregate.A natural measure of
this aggregatelevel is its monetary value, that is, the
present value of gross national product (GNP). In other
words, accordingto this argument,the best legacyfor
future generationsis a high levelof (properlymeasured)
incomeand capital formationin the present generation
(Baumol1969).
The accountingstructure put forth in the following
chapterimplicitlyadoptsthis viewwithtwo crucialprovi-
sos (alsoendorsedby Baumol):that GNP (and the under-
lying asset base) be defined to include nonmarket ser-
vices, particularlythe services of the environment,and
that certainassets, such as air, potable water,plant and
animalspecies,beautifulvistas,and the cultural heritage
be recognizedas vital and irreplaceableand therefore
beyondeconomicevaluationand beyondthe scope of the
expandedaccounts.
ValuationCriteriaand Social Organization
If an attempt is made to implementthe expanded ac-
counting structure described in the followingchapter,
severaldifficultiesmust be overcome.Clearly,the lack of
data is one of the more serious problems.There are,
however,evenmore seriousimpediments.Theseare more
serious in the sense that they involveconceptual issues
that, unlike the paucityof data, are not easilycorrected
evenif a developingnation had more resourcesto work
with.
One issue concerns the possibilityof determiningim-
puted nonmarketvaluesthat are independentof a coun-
try's economicand social structure. If, for example,the
valuationprocedurerelies on society'swillingnessto pay
for the nonmarketedservices of a natural resource or
some other environmentalasset, then clearlyit wouldbe
impossibleto devisean economicvaluationwithoutknow-
ing the society'spreferences.
If one acceptsthe principlethat a societyshouldevalu-
ate its ownassets,then the lack of a sociallyindependent
valuationmaynot provetoo troubling.It does create two
problems,however.First,it becomesdifficultto makeany
comparisonsof expanded national income and product
amongcountriesif the countries haveverydifferenteco-
nomic and social structures. Second, tracking economic
performanceover time can also be difficultif there have
been substantialsocial and economicchanges.This sec-
ond problempresents a special conundrumif one of the
purposes of economicaccounting is to determine the
extentof economicchanges.
Thesepoints couldbe used to argue againstnonmarket
accountingin general. However,market prices also de-
pend on a country'seconomicand socialstructure. Thus,
as is wellknownby nationalaccountants,comparisonsof
nationalincomeaggregatescan alsocontainlimitationsif
the countriesbeing comparedare very different.Country
A could have the higher GNP if its relativeprices were
used for the comparison,whileCountryB could havethe
higherGNP if its relativepriceswere used. Althoughthere
is no solutionto this problem,the responsiblecourse of
action would be to use both sets of prices for making
two-countrycomparisonsif the differentprice relatives
wouldmateriallyaffectthe results. Alongthe same lines,
for makingcomparisonsof nonmarketGNP betweencoun-
tries, the prudent course of actionwould be to use valu-
ations that would reflect differenteconomicand social
structures. Of course, in viewof the difficultyof produc-
ing evenone set of nonmarketvaluations,this procedure
is unlikelyto be followed.
Valuationmethods used to implementthe expanded
accountingstructure can be based on methodscommonly
used for benefit-costanalyses.These methodsfrequently
use the conceptof willingnessto pay-the ideabeing that
the valueof a nonmarketedgood or service should be
measuredby what societywouldbe willingto pay for the
 Environmentaland NonmarketAccountingin DevelopingCountries 63
good or servicewere it marketed.In practicethis usually
meanswhat individualsin societywouldbe willingto pay.
This focuson the individualis consistentwith the central
propositionof capitalistic,free market economies:that
the consumeris sovereign.
Relatedto the problem of makingcomparisonsamong
countrieswhen economicand social struzturesdiffer is
the problem of whetherthe willingness-to-pay conceptis
appropriatefor economicsystemsthat have aimore lim-
ited role for consumersovereignty.6 And if it is deemed
inappropriate,what principlesshould take its place, and
what valuationmethodsare consistentwith thleseprinci-
ples? Answersto these questionsare importantcompo-
nents of the task of finding nonmarket valuationsfor
expandednational accountingframeworksin developing
countries.
In a developingcountry, it is important to develop
valuationsand valuationmethods that reflect both the
principlesof the country'ssocialphilosophyand its social
objectives.Sustainableeconomicgrowth-one of the more
importantsocialobjectives-must be translatedinto con-
cepts that lendthemselvesto measurementand valuation
if this objectiveis to be reflected in the expanded ac-
counting system.Indeed, it may not be possibleto de-
velopvaluationschemesthat fullymeasurethiedegree of
sustainability.Yet, it is importantto poinitout that the
accountingframework,evenwithoutcompleteor perfect
monetaryvaluations,can providea data systemthat can
be of tremendoususe to those responsiblefor making
sustainabledevelopmentpolicy.
Suggested Course of Action
Althoughthere is some experiencein implementingthe
expandednational accountingstructure describedin the
followingchapter, the effort to developthis framework
reliedtotallyon data from,and condition;in, the United
States (Peskin1987).No case is knownwherethissystem
or some other nonmarketaccountinghas been fullyim-
plementedin a developingcountry,althoughthere have
been some partial efforts (Repetto and others 1987).
Thus, implementationor, more precisely,initialsteps at
implementationshould be viewedas a pilot project.
Basedon the experiencein the UnitedStates,the start-
ing point is to assemblea completebodyof physicaldata
coveringthe stock of environmentalassels, the demands
for the servicesof these assets (forexample,the demands
for waste disposalservices by industry,the demand for
recreationservicesby householdsand tourists, and the
demandfor potablewater),and the negativephysicalcon-
sequenceson final and intermediatedernandfrom con-
sumptionof these services (for example,adversehealth
effectsfrom water pollution,damageto crop production
from soil loss,and damageto electricityproductionfrom
sedimentation).
After or, better yet, while the physical data are col-
lected,effortsshould be made at monetaryvaluation.In
the U.S. contextbenefit-costtechniqueswere used.Thus,
for example,the valueto firmsof wastedisposalservices
providedby receivingwaterswas estimatedby the oppor-
tunity costs to the firms if they were deprived of these
services.These costs,in turn, were estimatedby various
techniquesthat attempt to measuresocial willingnessto
pay to avoid pollution (for example, the travel cost
method, the property valuation method, and the dis-
countedproductivitymethod).
As noted above,it cannot be assumedthat any of these
methodswillwork in a developingcountry.Thus, a dy-
namicresearch effortis calledfor-dynamic in that it will
have more of the characteristics of developmentalre-
search than of appliedresearch, althoughthe vastlitera-
ture on benefit-costtechniquesmayprovidea usefulpoint
of departure. Becausethis initial effort willbe explora-
tory, the qualityof much of the assembledinformation
and its valuationwill not be very good. This does not
mean that the informationwill not be useful, however.
Indeed,the U.S. experiencehas indicatedthat much of
the preliminaryinformation,especiallyon the physicaluse
of the environmentfor wastedisposal,is of great use to
policymakers.Nevertheless,becauseso much of the infor-
mation necessarilyhas to rely on estimationrather than
on direct observation,the pilot project is probablybest
done by a nongovernmentalagency,such as a university,
in cooperationwith variousministriesthat could provide
some of the supportingdata.
It is difficultat this time to be too precise about the
appropriatestaffingof such a pilot effort in a developing
country.Regardlessof the level of effort, however,it is
important,for two reasons,that the bulk of the work be
done by nationalsand not by foreignconsultants.First,
becausea long-term,multiyeareffort is necessary,heavy
staffingby consultantsis likelyto makethe costs prohib-
itivelyhigh. Second, unless the bulk of the work is done
by nationals,there is a good chance that the accounting
effort will die when the consultants depart. Thus, an
importantgoal is not simplyto constructa resource ac-
counting system, but also-and perhaps more impor-
tant-to developresource accountingskills in the coun-
try.At a minimum,a viableprogram,whichwouldachieve
these goals, would require two full-time nationalsand
about three to six staff-monthsof outside consultantsa
year.
One other considerationin determiningthe level of
effort concerns the scope of the accounting function.
Should the expanded accounts be confined to environ-
mentalassets or shouldthe effort attempt to cover a full
spectrum of nonmarket assets? In principle, the latter
approachis preferable,basicallybecausea fullyexpanded
accountingsystemwillbe able to show important rela-
tions betweenenvironmentalasset use and other nonmar-
ket activity.For example,the amountof labor devotedto
householdfuelwoodproduction,a significantnonmarket
64 HenryM. Peskin
activity in many rural societies, crucially depends on the
location and availabilityof natural forests and, of course,
on competing market and nonmarket sources of energy.
Analysisof such projects as the development of plantation
forests would be greaty facilitated if the accounting sys-
tern reflected these interdependencies.
Realistically, limitations of available staff and money
may preclude development of an accounting system that
fully covers all nonmarket assets. It may be necessary to
confine the effort to environmental assets. Indeed, it may
be necessary at first to confine the effort to one or two
particular assets, such as water and natural forests. Such
initial, modest steps should be encouraged.
Notes
1. In the accountingschemeoutlinedin the followingchap-
ter, depreciationof natural resourceswill not affectthe gross
productaggregates,such as GNP.
2. Of course, such a viewneglectsthe importantdistinction
made in the followingchapter between changes in physical
stocks and changes in the value of those stocks. A country
could be sufferinga real loss in wealtheven if physicalstocks
of certainnaturalresourcesincreased,if the resources:n ques-
tion provedto have decliningvalue. The shift from tropical
hardwoodforests to softwood plantationsin Indonesiamay
providean example.
3. Althoughthe terms "shadowcost" and "shadowprice" are
widelyused, they are often defineddifferentlyby differentau-
thors. Here the terms are used rather looselyto refer to the
unit values, necessarilyimputed, of nonmarketedgoods and
services.Dependingon the imputationprocedures,the term
could be equivalentto severalshadowconcepts.Examplesare
the shadowpricesgeneratedby a linearprogrammingproblem
or the shadowwages,as definedby Little and Mirrlees(1969),
used to measurenet consumptionchangesassociatedwiththe
movementof labor from agricultureto industryas a result of
developmentprojects.
4. The followingchapter shows,however,that the two con-
cepts can be related.Valuedepreciationcan be definedto equal
physicaldeteriorationminus capital gain or plus capital loss.
Nevertheless, the two conceptsare quitedistinct.
5. In effect,the resultpostulatesa verysimpleeconornywide
productionfunction,whichis not separablewithrespectto any
of its inputs.A simpleCobb-Douglas productionfunctionwould
havethe desiredcharacteristics,even though all the inputs in
such a functionare, by definition,substitutes.
6. Followingthe analysisof Marglin(1984),if an economic
systemwere based on neo-Marxistprinciples,consumersover-
eigntywouldbe limited,in theory,to the short-rundetermina-
tion of output mix. In contrast,if an economywere structured
accordingto neoclassicalprinciples,consumersovereignty would
have a much larger role to play.Of particularimportancefor
the justificationof cost- benefit techniquesis the role of con-
sumersovereigntyin determiningrelativeprices.
References
Baumol,WilliamJ. 1969. "On the DiscountRate for Public
Projects."In U.S.Congress,TheAnalysis and Evaluationof
PublicExpenditures.The PPB System. Papers submittedto
the Subcommitteeon Economyin Governmentof Joint Eco-
nomicCommittee,vol. 1. 91st Congress,1st session.
Blades,DerekW. 1975. "SubsistenceActivitiesin the National
Accountsof DevelopingCountrieswithSpecialReferenceto
LatinAmerica."Reviewof Incomeand Wealth21(4,Decem-
ber):391-410.
Landefeld,J. S., and J. M. Hines.1982. "ValuingNonrenewable
NaturalResources:The MiningIndustries,"In U.S. Depart-
ment of Commerce,MeasuringNonmarketEconomicActiv-
ity. Bureauof EconomicAnalysisWorkingPaper. Washing-
ton, D.C.
Lind, Robert C., KennethJ. Arrow,Gordon R. Corey,Partha
Dasgupta,AmartyaK. Sen, ThomasStauffer,Joseph E. Stig-
litz, 3. A. Stockfish,and Robert Wilson. 1982. Discounting
for Timeand Risk in EnergyPolicy. Washington,D.C.:Re-
sourcesfor the Future.
Little,1.M. D., and James A. Mirrlees.1969.Manualof Indus-
trialProjectAnalysisin DevelopingCountries.Vol.11,Social
Cost-BenefitAnalysis.Paris: OECDDevelopmentCentre.
Marglin,Stephen A. 1984. Growth,Distribution,and Prices.
Cambridge,Mass.:HarvardUniversityPress.
Page, Talbot. 1977. Conservationand EconomicEfficiency.
Baltimore, Md.: Johns Hopkins UniversityPress for Re-
sourcesfor the Future.
Peskin,HenryM. 1987. "NationalAccountingand the Environ-
ment." Resources for the Future, Washington,D.C. Proc-
essed.
Repetto, Robert, MichaelWells,Christine Beer, and Fabrizio
Rossini.1987. "NaturalResourceAccountingfor Indonesia."
WorldResourcesInstitute,Washington,D.C.Processed.

AccountsFramework
A ProposedEntronmental
Henry M. Peskin
The objectiveof expandingthe nationalincomeand prod-
uct accountsto encompassthe environmentrequiresthe
developmentof a theory linkingenvironmentaldata with
the economicinformationalready covered by conven-
tional national accounts. Such a theory is presented in
this chapter.
The cost of the proposed expansionis minimizedby
adheringas much as possibleto the accountingstructure
of existingnational income and product accounts,such
as the U.N. Systemof NationalAccounts(SNA). Indeed,
the suggestedchanges are introducedas separableaddi-
tions to the existingaccounts. In this way,the modified
accountspreserve all existinginformation,and the sug-
gested changes can be ignored by users of the conven-
tional accounts who do not require the additionalenvi-
ronmentalinformation.Because of the close linkage to
conventionalaccounting structure, this discussionpro-
videsan overviewof the conventionalincomeand product
accounts,presents a modifiedaccountingstructure, and
emphasizeshowcloselythe proposedmodiFications fit the
existingnationalaccountingconcepts.
Theory of Environmental Asset Services
The firstissuethat ariseswhenone sets out to account
for the environmentis to decidewhat aspect of the envi-
ronmentwillbe measuredor accountedfor. One possibil-
ity, reflecting the popular concern with air and water
pollution,is to accountfor the generationamddisposition
of pollutingmaterials.Thus, followingthe lead of Ayres
and Kneese(1969)and Leontief(1970),a systemcan be
devisedto trace all physicalflowsof materialsthroughout
Note: Someofthe material in thischapter,including
a graph-
icalanalysis,is givenin Peskin (1989).
65
10
the economy,whether or not these materialsare valued
with a positivemarketprice. Usingtechniquesof interin-
dustry or input-outputanalysis,these material balance
approachesare easilylinked to the economicdata drawn
fromnationalaccountingsystems.
Both an advantageand disadvantageof this approach
is its apparent neutrality: only materials are being ac-
countedfor;there is no indicationwhetherthese materi-
als are good or bad. This is in contrastto other environ-
mentalaccountingschemesthat attempt to measure or
otherwiseaccountfor the negativeeffectof pollutionand
human activitieson the environment.These approaches
are more difficultto implement,since a metric must be
devisedto valuethe damagedone by pollutionand envi-
ronmental degradation. Implementinga scheme based
only on the measurementof physicalflows is far easier.
However,even though a material balance accounting
scheme may be useful to organize and displaycertain
environmentalinformation,by itselfit is not sufficientto
augmentthe coverageof the accounts.
The problem is that the elementsincludedin the ac-
counting schemeshould reflect economicwelfare.If, at
the outset, an accountingscheme is set up that, as a
matter of principle, precludes placing a value on the
elementsbeing measured,then such a schemeby defini-
tion saysnothing about economicwelfare.Materialsac-
counting, with its neutral valuationis similarlyneutral
with respect to economicwelfare.Of course, inferences
can be drawn about economicwelfarefrom material ac-
countsjust as from indexesof production.But to do so
requires the aggregationof dissimilarentities measured
in differentunits (for example,tons of cement, barrelsof
oil, and gallonsof water).The material accounts,them-
selves, offer no guidance on how to do this aggregation.
For its full implementation,
the accountingschemeout-
lined below requires the measurement of all entries in
66 HenryM.Peskin
valueterms to allowfor aggregationand the direct com-
parison of environmentalwith nonenvironmentalentries.
Oftenit is necessaryto measuresuch physicalinformation
as the quantity of air and water pollutants emitted by
industrialcategory.Effortsto build materialbalance ac-
countsand effortsto build the expandedaccountscan be
very complementaryactivities.The objectivesof the two
exercisesare, in principle,however,verydifferent.
Rather than materialflows,this expandedschemeat-
tempts to cover flows of servicesgeneratedby environ-
mental capital.It is based on two concepts:all income
originatesfrom capital,and the physicalenvironmentis a
form of capital fundamentallysimilarto the other forms
of tangibleand human capitalthat generate the income
flowsalreadyaccountedfor in the conventionalaccounts.
The first concept,whichexpressesthe fundamentalrela-
tion betweenincomeand capital,has been wellaccepted
in modemeconomicthought ever since it was originally
stated (Fisher1906).The secondconcept,whilehardlyin
the same centralpositionin economicthought,wouldnot
be found to be objectionableby most economists.The
environment"is a form of social capital which society
must somehowmanage if it is to maximizeits welfare"
(Johnson1975, p. 321). Why this type of capitalneeds
social managementwhileother types of capital,such as
factoriesand machines,can rely on private management
is central to the theory of environmentalmanagement.
This theoryunderliesboth the proposedaccountsframe-
workand the proposedmethodsof valuation.'
Although,as arguedabove,materialflowaccountingby
itselfdoesnot providesufficientinformationto add to the
accountingof the elementscomprisingeconomicwelfare,
the material flow and the proposed asset service flow
approachare highlycomplementary.This is becausethe
productionof anymaterial,whetherit is a product to be
sold or a wastematerialto be disposedof, can be linked
to the servicesprovidedby some specificasset. Just as
the productionof goods sold in the marketplacereflects
the capitalservicesof the plant and machinesusecdin the
manufactureof the goods,likewisethe productionof waste
products,such as smokeand spent processand cooling
water,reflectsa demandfor the capitalservicesprovided
by the environment-servicesof waste disposal,cooling,
and the provisionof importantinputs, such as oxygenfor
burningor nutrientsfor crops.
To understandthe relationbetweenthe demandfor the
servicesof environmentalassets and the generation of
pollutants,however,an importantprinciplemust be es-
tablished:the polluteralwayshas some choiceabout how
much pollutantwillbe generated.Thus,whilesome pol-
lutionmaybe an inevitableconsequenceof mostproduc-
tion processes,the amount generated for each unit of
product can vary substantially,dependingon a multitude
of decisionsaffectingproductionlevels,productiontech-
nique, and product mix.2 The model underlyingthis ac-
counting frameworkassumes that these decisions are
madewisely.Morespecifically, the chosen amountof pol-
lutiongeneratedis assumedto be that whichmaximizes
profitsforpollutingproducersor whichmaximizesutility
for pollutingconsumers.
Generally,the less pollution is generated or (what
amounts to the same thing in this simpleexample)the
less use is made of the disposalservicesof the environ-
ment,the more costlyit is for the producer.Therefore,it
is assumedthat there are positivebenefitsto the polluter
associatedwithhigherpollutionlevels,which,however,at
somepoint begin to becomeless positive.Indeed,ifthese
levelsget too high, the pollutionmaybeginto impedethe
productionprocess(say,by affectingthe workers'health),
and thus the benefitsof pollutingmayevendecline.(Since
the producer is assumedto be rational, however,such a
situationwouldnot existfor very long.)
Up to now,the focushas been on the benefitsof using
environmentalassets and on the associatedpollutionby
the polluter.The other side of the coinis nowconsidered,
namely,the "disbenefits"or damagesthat maybe associ-
ated with the use of environmentalassets.There are two
sorts of disbenefits.One is the familiardirectdamagethat
pollutionby one party can inflict on another. Air pollu-
tion's effecton a person's health is a clear exampleof
this sort of direct effect.By measuringthese direct dis-
benefitsas negativebenefits,one can imaginea relation
betweenpollutionand disbenefitsthat, in general, indi-
catesvirtuallyno disbenefitat verylow levelsof pollution,
a rapid increasein disbenefitat moderatelevelsof pollu-
tion, and a satiation of disbenefitat very high levels of
pollution.Althoughthis pattern is realisticfor many pol-
lutantsthat cause adversedirect healtheffects,other pat-
terns are possible.This pattern, however,along with the
assumedform of the relation betweenpollutionand ben-
efits,illustratesthe frequentsituationwherethe net social
benefitsof pollution(the sum of the benefitsand disbene-
fits)reachesa maximumat some positivepollutionlevel,
(wherethe marginalbenefitsand disbenefitsof pollution
are equal)but wherepollutionis at a lowerlevelthan the
polluter finds optimal. Assuminga known relation be-
tweenenvironmental assetuse andpollution,optimalunits
of environmentalservicerather than pollutioncould also
be determined.
The secondtypeof disbenefitoccurswhenthe pollution
or environmentalasset use denies someone else use of
the environmentalasset. For example,using a lake for
waste disposalmaydeny the use of the lake to someone
elsefor gettingpotablewater or fishing.This second type
of disbenefitis referred to as an indirect disbenefitto
distinguishit fromthe direct disbenefitsdiscussedabove. 3
The key feature of this type of disbenefitis that it will
exist only if there is a finite amountof availableenviron-
mentalasset service.
For the proposed accountingframework,three key el-

ementsare relevant.First, and most important,both pos-
itiveand negativebenefitsare associatedwithany amount
of pollutingactivity.Thus a singleaccountingentrywill,
in general, not be adequate to capture the effects of
environmentalasset use. Second,with a finiteavailability
of environmentalasset services,the identificationof one
party (for example,industry)as the polluterand another
party (for example,households)as the injured party is
not alwaysappropriate.The accounting system should
recognizethat any actor in the economymaybe using
environmentalasset servicesto the detrimentof another
actor. A third element concerns the effect of the total
availabilityof environmentalasset services.If the availa-
bility of such services is very great in relation to the
demandsplaced on the environment,and dLepending on
the methodchosento placevalueson these demands(see
below),it may be appropriateto ignore certain environ-
mental asset uses in the accountingsystern.Of course
treatingall environmentalservicesas free goods and not
assigningany valueto them has in part contributedto
the current environmentalproblems. The point being
madehere, however,is that a judgmentmust be madeon
a case-by-case basis on whetherassigninga zero valuefor
accountingpurposesis justified.
Valuation Concepts
For valuationmethods,a distinctionshould be made
betweenthe cost of pollutioncontrolalreadyincurredby
pollutersand the potential cost to pollutersequal to the
forgonebenefitif theyweredenied accesstcithe environ-
ment (whichequals the value to the polluitersof being
allowedto pollute).The former cost is, in principle,al-
ready accounted for in the conventionaleconomicac-
counts (althoughit may not be clearlyidentifiedas a
pollutioncontrol cost). The latter cost is not currently
accounted for. Yet its estimationis important for the
expanded accountingsystem, since it providesan esti-
mate of the value to polluters of disposalservicespro-
videdby the environment.This cost, that is, the prospec-
tive cost to the polluter of being denied any use of the
environmentalasset, measures the value of a nonmar-
keted factor input (the environmentalasset service)just
as, say, the wage bill measuresthe value of a marketed
factor input (labor).
As environmentalassets are used and pollutionpro-
duced, disbenefits are also produced; this is also ne-
glectedin conventionalaccountingsystems.Most availa-
ble measuresof these disbenefitsrely on total valuations:
the total valuesocietywould be willingto pay to avoid
such disbenefits.But these total values are not entirely
consistentwith the valuationconcept used for marketed
goods in the conventionalnational accounis. The usual
valuationfor marketed goods is a price-times-quantity
valuation,the price in questionbeingthat for the last or
A Proposed Environmental Accounts Framework 67
marginalunit sold.Allother so-calledinfra-marginal units
are valued at this same price (assumingthat the seller
cannot practice price discrimination).To measure the
disbenefitsassociatedwith the use of environmentalas-
sets, marginal valuations should be used.4 One important
advantageof marginalvaluations(besidesconsistency)is
that they provide a way to estimate whether observed
environmentalassetuse is greater or less than the optimal
use, whichmaximizesnet benefits.At the optimal level,
the marginalbenefitand disbenefitvaluationsare exactly
equal;at levelsgreater than that, the marginalvaluation
of disbenefitexceeds the marginalvaluationof benefit;
and, at levelsless than that, the marginalvaluationof
benefitexceedsthe marginalvaluationof disbenefit.These
latter situationscan be interpreted as conditionsfor too
much and too little pollution,respectively. Althoughmar-
ginalvaluationsallowfor this interpretation,total valua-
tions are usuallyeasier to estimate.
Up to now, cost and valuationprinciples have been
used for the situationwhere a polluter creates a direct
disbenefit.For indirectdisbenefits,in additionto the total
benefit to party 1 (formerlythe polluter),the fact that
party 2 (formerlythe party sufferingdirectpollutiondam-
age)enjoysa benefitfromthe remainingunits of environ-
mentalassets must be accountedfor. Similarly,the indi-
rect damage party 1 causes party 2 needs to be
considered.To this value must be added the indirect
damageparty 2 causesparty 1 by denyingparty 1 various
unitsof asset service.Againat the optimallevel,marginal
valuationsof benefitsand disbenefitsare equal. As op-
posedto the analysisof directpollutiondamage,however,
the differencebetween benefit and disbenefit estimates
cannot be interpretedas conditionsfor too much or too
little pollutionor asset use. It can be said only that one
party'suse at the expense of the other party's use is too
excessiveto be sociallyoptimal.
Investment and Depreciation
Conventionaleconomicaccounting distinguishesbe-
tween the purchase of goods for consumptionand for
investment.A further distinctionis made between gross
and net investment,the latter being equalto gross invest-
ment less capitaldepreciation.In expandingthe conven-
tional accounts to include environmentalasset services,
these distinctionsshould be preserved. In particular,in-
vestmentsthat increase the environmentalasset stock
should be accountedfor if this is not alreadydone in the
conventionalaccounts. Similarly,there will be an entry
measuringany depreciationin environmentalassets. The
definitionsof investmentand depreciationmay not ex-
actly be the same as those used in conventionalaccount-
ing,however,althoughthe definitionsused are consistent
with economictheory.
Specifically,followingAlfredMarshall(1920), income
68 HenryM. Peskin
(regardlessof whether it originates from marketed or
nonmarketedcapital)is definedas the sumof currentand
potentialfuture additionsto well-being.The formationof
capitalthat allowsfor both additionsto future income
and the maintenanceof current income is gross invest-
ment. If current additionsto well-beingare identifiedas
consumption,gross incomecan be definedas consump-
tion plus gross investment.Net income is simplycon-
sumptionplus net investment,where net investmentis
gross investmentless that portion of capital investment
just necessaryto maintaincurrent levelsof consumption.
Althoughdepreciationis defined conventionallyas the
differencebetweengross and net investment,one should
also distinguishbetween valuedepreciationand physical
depreciation.
One reason that depreciation exists is that the physical
abiit
t geerte
ofcaita onumaleserics dclne
ability*of
captal to generate consumae s decli
aeps
over time. This loss in physicalability-physical epreci-
ation-may also lead to a loss in the valueof the capital
stock-value depreciation;that is, valuedepreciationmay
reflectphysicaldepreciation.However,valuedepreciation
can alsoarise for other reasons.The valueof capitalcan
fall becauseof a changein tastes for consumptionItems
producedby the capitalor simplybecauseof a changein
interestrates.
The relation betweenincome and the change in value
depreciation can be shown as follows. A society's capital
has value presumablybecause it generates a stream of
goods and services,that is, income.Let V0representthis
valueat the beginningof the year and Ql, Q2, and so
forth represent the servicesat the beginningof the next
and subsequentyears.AlsoQ, is definedas grossincome
in year 1.
Vocan be relatedto the Q's as follows:
Q __Q_ Q physicaldepreciationand the portionthat is due to other
(1) VO= (1 + *--+ (1+ +y
wherei is the rate of interest.Since V, the valueof Voat
the end of year 1, is simply
Q2(+ + + Qn+'Il
(1+t) (1+02 (1+1')+*
equation1 can also be written:
Q, V1 Q,+ V1
(2) VO (1+0 + T1 (1+i
fromwhichit followsthat
(3) Qi = ivo + (VO VI).
The term (Vo- V,), representingthe loss in valueof
the initialcapitalstock, can be identifiedas valuedepre-
ciationoccurringin year 1, or D,. As before, gross in-
come, Q, (using the Marshalldefinition),is defined as
consumptionplus gross investment.Since net income
equals consumptionplus net investmentand net invest-
ment equalsgross investmentless depreciation,however,
net incomethus equalsgross incomeless depreciation.It
followsfrom equation3 that net incomeequalsiVo.Thus
equation3 can be rewrittenas:
(4) = Cl + I1 + D,
where C and II are consumptionand net investmentin
year 1. In other words,
Grossincome = net income + (value)depreciation
= consumption+ net investment
+ (value)depreciation
= consumption+ gross investment.
This relation between income and depreciation was
dvlpdwtotayrfrnet h hscldsrc
developedwithout any referenceto the physicaldestruc-
tion of capital. It is true that as capital wears out physi-
cally,future Q's may fall, which explains why VI may
exceed V'. However,future Q's mayfallfor other reasons,
such as a fallin the demandforthe capital'sservices.The
Q's may also decline because of an inabilityto employ
thcailfuy.Ttisifapalomeensabra
declinein labor serviceswouldalsobring about a decline
in future Q's.
Moreover,in contrast to physical depreciation,value
depreciation,V. - VI, need not necessarilybe positive;
that is V c
. ouv in prince ecedion. Conven al
tive and to refer to the case where V, exceeds Vnas a
capitalgain. This chapter does not followthis practice
since it implicitlyassumes that the term "depreciation"
can onlymeanphysicaldepreciation.Rather,valuedepre-
ciationis decomposedinto two components:the portion
of the differencebetween V0 and V,that is due to actual
causes.The latter,if positive,willbe termed "capitalgain"
and, if negative,will be termed "capital loss." Thus by
definition:
(5) Valuedepreciation= physicaldepreciationminus
capitalgain or plus capitalloss.5
Of course, the physicaldepreciationmust be valuedto
makethis computationpossible.Howthe unitsof physical
loss should be valued-using the original price of the
capital,its currentprice,or some other price-is a matter
of somecontroversy.Since the choiceaffectsthe measure
of physicaldepreciation,it has importantimplicationsfor
both corporateaccounting(whichis not of concernhere)
and for national accounting(whichis of concern).If the
focusis on valuedepreciation,however,the implications
of choosing among alternative capital prices can be
avoided by estimatingvalue depreciationby successive
applicationof equation 1 one year apart rather than by
applicationof equation5.

Sources of Environmental Asset Depreciation
Like any other capital asset, the depreciationof the
valueof an environmentalasset is due both to physical
depreciationand capitalloss (or gain).To illustratethese
two sourcesof valuedepreciation,two assets are consid-
ered: one experiencesonly physicalreductionwhile the
valueof the other depreciatessolelybecause each user
perceivesa reductionin the qualityof servicesprovided.
The formerasset mightbe a lake whoseonlyserviceis to
provide drinking water. It is assumedthat as drinking
wateris withdrawn,neithermanmadenor natural replen-
ishmenttakes place, so that the availabilityof drinking
water is steadilyreduced.6 The second examplemightbe
a stretch of seashoreof a givenlength.It is assumedthat
the only serviceprovidedby this asset is recreationand,
further,that as the demandfor this serviceincreases,the
qualityof serviceperceivedby eachdemanderdiminishes.
In the first example-the demand for the services of
the first asset, drinkingwater,overtime by two users-in
order to clearlydistinguishthe two sourcesof valuede-
preciation,each user's demandis assumedto be constant
overtime. Also,to keepmatters verysimple,the demands
are assumedto be identical,and the capitalgain compo-
nent in equation 5 is neglected. Each demander is as-
sumed to have a maximumdemand of, say, X. When
capacityis reducedbelow2X, the remainingasset is as-
sumed to be equallyshared. Under the assumptionthat
the demandfunctionsare identical,such amevenalloca-
tion is Pareto optimal:no other allocationcan makeone
user better off withoutmakingthe other worseoff.
The implicationfor the lake's valueover time is quite
simple. Since the availabilityof drinking water is more
than adequateforboth users, for a time there is no value
reductionand hence no valuedepreciation,eventhough
in real terms there is physicaldepreciation.However,the
valueof this physicaldepreciationis initiallyzero. Yet,as
soon as the lake's capacity is reduced t:) 2X drinking
water units, the valueof the lake beginsto decline,and
value depreciationincreases. That is, the value of the
physicaldepreciationbecomespositive.
The second polar exampleis representedby the sea-
shore. It is assumedthat there is an increasingnumberof
identical individuals,each demanding the same fixed
length of seashore.To showthe effectof this increasing
demandin a two-dimensional diagram,the individualsare
dividedinto two groups of equal size. Three basic situa-
tions arise: (1) when the number of individualsis small
enough that their maximumdemand is less than the
seashorecapacity,(2)when the number ef individualsis
moderate so that their maximumdemands just exceed
seashorecapacity,and (3)whenthe numberof individuals
is so large that their summedmaximumdemandsgreatly
exceedseashorecapacity.For the latter two casesa Par-
eto optimal distributionis assumed:if capacityis z and
A ProposedEnvironmentalAccountsFramework 69
the number of individualsis n, each of the identicalindi-
vidualsis allottedzin of seashore.
The implicationsfor the asset's valuedepreciationde-
scribedin the secondexampleare more complicatedthan
those illustratedin the first example.Even though the
amountof externaldamageeach user inflictson the other
increasesas the numberof individualsincreases,the total
utilityfor all individualsas a group increases.Algebrai-
cally,at the time the numberof individualsn is moderate,
the assumedutilityof each individual(for a Pareto opti-
mal distribution)is u = u(zln).This utilitydecreasesas n
increases.However,the total utilityfor all individualsis
nu(z/n);this utilityincreaseswith increasesin n, although
at a decreasingrate. In effect,previousseashoreusers are
worse off as n increases, but, because there are new
users, total utilitymayincreaseeventhough averageutil-
ity falls.Total utilitymayalsodecrease,however,depend-
ing on the relativesize of an individual'stotal and mar-
ginal utilitywith respect to n. The theoreticalchange in
total utilityas n increasesis u(zln) - u'(zln)Jn,and this
can be either positiveor negative.
In sum, it is not self-evidentthat the valueof an asset
that does not deterioratephysicallydeclinesas it becomes
more heavilyused. A recreationalasset mayexperiencea
capital gain even though it becomes heavilycongested
and even though each user becomes less satisfiedwith
the servicesprovided.Whetherthe valuedepreciationof
such an asset is positiveor negativedepends on a com-
parisonbetweenthe decliningutilityof each user and the
increasingnumber of users. One importantbut perhaps
disquietingimplicationis that if marginalvaluationsare
used, a pristinelake has no valueuntil a point of conges-
tion is reached, that is, the point wheredemand curves
start crossing.Actualphysicalcongestionis not required;
a demandcan existwithoutactualuse if the demandis in
the form of a reservationfor future use. The existenceof
such an optiondemand maysufficeto givevalueto the
pristinelake.
Nevertheless, manywouldfeel that the pristinelake has
valueevenin the absenceof optiondemandor any other
sourceof demand.Even if everyoneacceptedthis view,it
wouldnot vitiatethe theoreticalstructurejust presented.
The keypoint is that the accountingsystemdiscussedin
this chapteris intendedto coveronlyelementsthat affect
economicwelfare.Consistentwith this objective,valua-
tions are confinedonly to economicvaluations-not to
valuationsthat mayhavereligious,ethical,or philosophi-
cal foundations.Unlikeother sourcesof value,the basis
for economicvaluationis the interactionof demandand
supply.
Overviewof the National AccountingStructure
To analyzethe implicationsof this theory of environ-
mentalassets for the proposed modificationto the con-
70 HenryM. Peskin
ventionalnationaleconomicaccounts,it is usefulto dis-
cuss briefly national accounting concepts.7 For
illustration,this discussionfocuseson the U.S. national
accounts and concentratesprimarilyon the irrethod of
accountingfor the inputsand outputs of marketedenti-
ties in production processes. The expanded accounting
frameworkplans to treat the inputs and outputs of envi-
ronmentaland other nonmarketentitiesin an analogous
manner.
Since 1958 the U.S. national accounts haveconsisted
of fivesub accounts:national incomeand product, per-
sonal incomeand outlay,governmentreceiptsand expen-
ditures,foreigntransactions,and gross savingand invest-
ment.The nationalincomeand productaccountincludes
gross nationalproduct (GNP),net nationalproduct (NNP),
and national income (NI)-the indexes that receivethe
most publicattention as indicatorsof well-being.There-
fore, the other accounts willbe ignored in the ensuing
discussion.Although these accounts are not insulated
fromproposedchangesin the treatmentof environmental
asset services,any effectson these accounts willbe re-
flected by changes in the incomeand product account.
Thus, few essentials are lost by focusingon just one
account,and the discussionis greatlysimplified.
The incomeand product account is a consolidationof
many individualaccountingsof productiveactivities,the
bulk of whichare those of businessenterprisesand fed-
eral, state, and local governments. The consolidation
process can be illustratedby consideringa hypothetical
economywhere productionis assumedto take place in
only two businessenterprises,A and B, and one govern-
ment.This economydoes not engagein foreigntrade nor
are there anyinterest chargesfor borrowingmoney.Also,
it willbe usefulfor later discussionif firmnB is assumed
to be a producerof pollutioncontrolequipment.
The dollarvalueof the outputof eachfirmin the period
coveredby the account (usuallyone year) willbe repre-
sented by X, with subscripts indicatingthe destination
and the type of output. Actually,a firm is likelyto pro-
duce and purchasetwo types of output: goods that will
be used up or consumedwithinthe current accounting
periodand goodsthat are used up overmanyaccounting
periods. The formerare termed current goods, and the
latter are termed capitalor investmentgoods.
Firm A's output of current goods is designatedby XAB,
XAC,and XAG,representinggoods that willbe used up in
the accountingperiod by firm B, by consumers,and by
the government,respectively. Expendituresfor such goods
are called current account outlays. Firm A's oultputof
investmentgoods is designatedby XA. Expendituresfor
these goodsare calledcapitalaccountoutlays.Sincethese
goods are not used up in the accountingperiod, the
inputs of such goods to sectors A and B are not shown.
Firm B's output-XBA,XBC,X,,, and XBG-are defined
analogously.
Table 10-1. Input and Output Accounts
for a HypotheticalEconomy
IndustryB
IndustryA (Pollution
controlindustry) Government
Input Output Input Output Input Output
XBA XAB XAB XBA LC X
LA XAG LB XBG XAC
DA X, D, X. XBG
TA XAG TB XBG
-SA -SB
PA PB
Note:X is the outputof goods and services,L is labor,D is deprecia-
tion, T is taxes,S is subsidies,and P is profitor loss.The firstsubscript
refers to the generatingindustrialsector (or government)and the sec-
ond, if any,refersto the receivingindustrialsector.
The right-handside of the account for firm A in Table
10-1 shows the sales of A's output in the accounting
period.The proceedsfrom these sales are dividedamong
severalinputsshownon the left-handside of the account.
These includeX,,, the amount of firm B's output pur-
chasedfor use in the accountingperiod; LA, the amount
spent on labor; TA, the amount spent for indirect taxes
(that is, all taxes excepttaxes on income);SA, any subsi-
dies receivedby firm A (which,as shown,are valuedas
negativeinputs);and P4 , profit or loss.SincePA is deter-
mined as a residual,the left-handand right-handsidesof
A's accountalwaysbalance.The inputside of the account
does not show any capital account outlays. It accounts
only for the inputs used up in the accountingperiods,
whereas capital goods by definition are used up over
severalaccountingperiods. The portion of capitalthat is
used up in the accountingperiod or depreciation(includ-
ing other capital consumptionallowancessuch as acci-
dental damageto capital)is designatedas DA.Againthe
itemsthat makeup firmnB's input, shownon the left-hand
side of B's account in Table 10-1, are defined analo-
gously.
The governmentproductionof goods and services in
the accountingperiodis XGG.The notationindicatesthat
the governmentis viewedas producinggoods and serv-
ices for its ownuse and not for other sectors. In conven-
tional national accounting,because of the difficultyof
valuinggovernmentoutput,XGGis set equal to the wages
and salaries paid to governmentemployees,LG,plus its
purchases from businesses,XAG and X,,. The govern-
mentalproduction account is thus verysimple. Govern-
ment transferpaymentsare not shown,since such trans-
fers do not reflect any new production in the economy.
Similarly, taxes are not shown,since these are not consid-
ered as inputs to governmentproduction.Such financial
itemsas taxes and transfersshowup only in the govern-
ment receipts and expendituresaccount.The consolida-
tion or mergingof the three sector accounts where, by

Table 10-2. ConsolidatedProductionAccount
for a HypotheticalEconomy
Input Output
LA + LB+ LG XAC+ XBC(consumption)
PA + PI
Nationalincome XAI+ XBI
(investmrrent')
TA+ TB
-(SA + SBl)
Netnationalproduct XGG
(government)
DA+ Di
ChArgesagainstgross Grossnationalproduct
nationalproduct
Note:See the note to Table 10-1 for definitionsof the notation.
Entriesin boldrefer to the sumof all entriesabovethem.
assumption,all conventional(thatis, marketed) produc-
tion takes place yieldsthe nationalincome and product
accountfor this simpleeconomy.This account,in effect,
picturesall the economy'sproductionas taking place in
one huge sector.The consolidationis performedbyadd-
ing A'sinputsto B's and the government'sinputs,adding
A's output to B's and the government'soutputs, and
eliminatingitemsthat appear in both the input and out-
put totals. In this simpleeconomy,the itemsso eliminated
are XAB, XB,I XAG, and XBG. Items that represent the
production of one industryfor use by another in the
accountingperiod are termed intermediategoods.In the
United States and in the accountingsystemsof most
westernnations, such intermediategoods are not shown
in the consolidatedincome and product account, since
their value is already reflected in the valueof the con-
sumption,investment,and the governmentalproduction
itemsincluded.
Table 10-2showsthe resultsof the consolidation.The
principal national income and product aggregatesare
definedby the groupingsindicatedby the entries shown
in bold. These entries are the sum of the items listed
abovethem. Thus, GNP is definedas total productionof
goods and servicesfor consumption,investment,and the
government. 8 Since the consolidatedaccount must bal-
ance, GNP mayalso be definedas the sum of lalborpay-
ments, profits, indirecttaxes less subsidies,and depreci-
ation. The sum of labor paymentsand profits alone is
callednationalincome(Nl). It represents the sum of pay-
mentsto factorsof production.The fact that there may
be other factors of production-such as the servicesof
the environment-forwhichpaymentsare not madeis, of
course,the subjectof this chapter.
NI plus indirecttaxes equalsNNP. NNP also equals GNP
less depreciation.It thus providesa measure of output
that accountsfor the fact that a certain portion of gross
output must be set aside to replace deteriorationof the
nation'sstock of capital.The fact that there mayalso be
deteriorationin environmentalcapitalfor whichthere is
A Proposed Environmental Accounts Framework 71
no capitalconsumptionaccountingis alsoa theme of this
chapter.
Unfortunately,this simplifiedexpositionfailsto address
severalcontroversialissues concerningthe ability of the
accountsto measurewell-being.For example,the classi-
fication of many items as investment,consumption,or
intermediategoods is not alwaysself-evident.Manyitems
such as consumerdurablesare classifiedas consumption
items when they could just as easily be classifiedas in-
vestmentitems. Another issue concerns the failure to
consideranyportion ofgovernmentalproductionas inter-
mediate,eventhough such governmentservicesas polic-
ing and inspectionare often essentialinputs to business
operations.A third issue concerns the fact that the ac-
countingstructure and its aggregatesdo not revealhow
any of the componentsare distributedamong members
of the population.A societyin which 10 percent of the
populationenjoys90 percent of the consumptiongoods
producedcould havethe same total GNPas another soci-
ety whereconsumptiongoods are much more evenlydis-
tributed.Fewwouldbelievethat both societiessharedthe
same levelof well-being.
One feature of the national accounts that is reflected
in the above simplified exposition is the limited role
householdsplayas a producingsector. This limitedrole,
in turn, reflects the preoccupationof the accountswith
activitiesthat take place in well-definedmarkets and
where productionis alwaysreflected in monetarytrans-
actions. This preoccupationneglects, for example, the
high portion of productiveactivitythat occurswithinthe
household. 9 The fact that GNPfallswhen"a man marries
his housekeeper"is a curiositythat is highlyrelevantto
the treatmentof all nonmarketproduction,including,as
shallbe shown,the prospectivetreatment of the services
of the environment.
Response of Conventional Accounts
to Nonmarket and Environmental Activity
In recent yearsthe abilityof conventionalaccountsto
reflect social well-beinghas been scrutinizedmore and
more. One can identifymuch of the problem as a failure
of the accountsto reflect adequatelyeconomicyet non-
marketedentities.
Usingthe simpleaccountingframeworkjust presented,
it is possibleto be far more specificabout these inade-
quacies.Since the focus of these efforts to modifythe
accounts has been toward improvingtheir response to
changesin the natural environment,the followingdiscus-
sion of the inadequaciesignoressuch failuresas the con-
ventionalaccount'sinabilitiesto measure fullythe effect
of changes in householdproductive activityor in the
services of government capital. However,the analysis
could easily be modifiedto cover these other nonmar-
keted entities.
72 HenryM.Peskin
As manyhavenoted, a declinein environmentalquality
maybe accompaniedby an increasein GNP. In this simple
economy,increasesin, say,XACcould require industryto
increaseits loadingsof air and water pollutants.WNhat
many find even more bothersome is that the pollution
itselfcould lead to an increasein both the leveland cost
of certain services,such as medicalservices,whichalso
lead to a further increase in conventionallymeasured
GNP. Of course, if the pollutionwere bad enough,work
abilitymightdecline,leading to a fall in LA,LB,ancILC
and a concomitantfall in GNP. The fact that under such
dire circumstancesGNP will movein the "right" direction
will,of course,do little to satisfythe critics.
Althoughdeclinesin environmentalqualityare often
accompaniedby increasesin GNP, it does not necessarily
followthat improvementsin environmentalquality must
be accompaniedby reductionsin GNP. The directionof
the movementof GNP or other account aggregates de-
pends on howthe environmentalimprovementis brought
about. In the simple economydepicted in Table 10-1,
firm B, which was assumed to be in the business of
manufacturingpollutioncontrol equipment,providesXBA
units of antipollutionequipmentto firm A, XBGunits of
equipmentto government,and XBCunits to consumers.
Suppose,in responseto further environmentaldegra-
dation, consumersor the governmentdecideto increase
purchasesof antipollutionequipment.In a fullyemployed
economythe postulated increase in XBCby consumers
must comeat the expenseof a reductionin XAC(thai:is,
a fall in consumptionof goods other than for pollution
abatement),a reductionin XA,or X,, (that is, a reduction
in investment,whichis equivalentto a reduction in sav-
ings),or a reductionin XGG (that is, a reductionin govern-
ment activity).Similarly,a postulatedincreasein XBGby
governmentmust, in a fullyemployedeconomy,be either
at the expenseof XAC,XBC', XAIXBI,or at the expenseof
other componentsofXGG,such as XAG or LG.
Althoughsuch reductionsmay changethe composition
of GNP and NI, it willnot changethe levelof GNP or Nl.10
To the extent that investmentis reduced, NNP may rise
slightly,since depreciationmayfall.(The remainingcapi-
tal maybe usedmore intensively, however,causingcapital
consumptionallowancesto rise.) Of course, if there were
unemployedlabor and capital resources,the increasein
XBCneed not engender any reduction in other outputs,
thus allowingall the accounting aggregatesto rise. In-
deed, an expansionin employmentand output has been
associated with pollution control expenditures in the
UnitedStates (DataResources1979).
The situationis much more complicatedwhen, as is
often the case, businesses,rather than consumers,pur-
chase antipollutionequipment(Denison1979b).The ef-
fect on the accounts depends on whetherthe pollution
abatementexpendituresare current or capital account
outlays.Also, if the outlaysare for investmentgoods, it
willbe necessaryto distinguishthe effecton the accounts
in the year the investmentis made from the cumulative
effectover time.
An increase in current account outlays for pollution
abatementby business and government,that is, X. or
XBG,does not showup on the consolidatedaccount.(They
are canceled out, since they are on both sides of the
consolidatedledger.)However,the fact that such transac-
tions betweenbusinessesare not showndoes not mean
that there willbe no effecton the account aggregates.
The more labor that firmB devotesto producepollution
controlequipmentto be sold to A, that is, XB, the less
labor willbe availableto produce other goods, such as
XB4and XBG.Therefore,barringa sudden changein tech-
nology,the increasein XA must be at the expenseof one
or more of the elementsof GNP.
The same result must occur iffirmA uses its ownlabor
for pollutioncleanup.If A attemptedto maintainits pro-
duction of goods that show up in the GNP (XAc,X ,, and
XA) by reducing its production of intermediategoods,
XAB, firmB wouldhaveto reduce its output,since XABis
B's raw material.The pollutioncontrolfirm B might try
to maintainits productionof goodscounted in GNP (XBC,
X,,, and XBG)by reducing its intermediateproduction,
XBA.This possibilitycan be ruled out, since such a re-
sponsewouldmean an increasein pollutionby A rather
than the postulateddecrease.Thus,divertinglabor either
for the productionof current account,intermediatepol-
lution control equipment or for in-housecleanup must
reduce the GNP."1 If, as is usuallythe case,however,most
of the pollutioncontrol depends on the purchases of
capital equipment,XB,,the short-term effecton the ac-
counts is the same as it was above when the increased
pollution control outlays of consumerswere analyzed.
The only short-termeffectwillbe on the compositionof
GNP, not on its level.
Yet,overthe longterm the situationmightbe different,
since changesin investmentare likelyto cause changesin
other typesof output.Thus,investmentthat is divertedfor
environmental purposeswillresult in fewermachinesand
less equipment to support ordinary production. Some
economistshaveestimatedthat for each 1 percent of in-
vestmentso diverted,output, as measuredby NI, fallsby
approximately0.3 percent (Denison1979a, 1979b).Al-
thoughinvestigators havedisagreedaboutthe exactamount
of the declinecausedby divertedinvestment,there is gen-
eral agreementthat increasesin pollutioncontrol invest-
ment will have some dampeningeffect on the account
aggregates,againmakingthese aggregatespoor indicators
of environmental improvement 12
and overallwell-being.
A Modified Accounting Structure
The abovetheoreticalstructure suggestsboth a frame-
workfor introducingthe servicesof environmentalassets

into the conventionalaccountsand the principle;for eval-
uating these services.Basically,the frameworkassumes
that these servicescan be treated as unpaidfacl:orinputs
to either production or consumptionactivities.The ad-
verseeffectsof pollutionor of the denialof environrnental
servicesto others caused by the use of the servicesas
13
inputscan be treated as negativeoutput or damages. A ing for environmentaldepreciation.The implicitassump-
balancingterm is requiredto assure that the converntional
equalitybetweentotal inputsand outputs is maintained.
The followingset of accounts,modifiedto reflect these
ideas, is rather similar to the conventionalaccounting
structure outlinedin Tables 10-1 and 10-2.14Although
Table 10-2 showsthe consolidatedaccount as a combi-
nation of industrial and governmentalproduction ac-
counts, here the consolidatedaccount is a coimbination
of productionaccountsin four sectors: industry,govern-
ment,household,and the environment(nature).To make
them more understandableto those familiarwith the con-
ventionalaccounts,additionalentries are shown,which,
to simplifythe exposition,were neglectedin the accounts
in Tables 10-1 and 10-2. There are no new entries for
current pollutioncontrol outlays,since these are already
includedin the conventionalaccountingof costs.Asnoted
above, however,it probablywould be usefulto identify
these costsseparately,as has been done by the Bureauof
EconomicAnalysis(U.S.Departmentof Commerce)since
1972.
Industrial Sector
The typical industrial account shown in Table 10-3
containsthree new entries that would ordinarilybe ab-
sent from a conventionalaccountof the industry'sinputs
and outputs.In addition,capitalconsumptionallowances
includenatural resource and environmentaldepreciation
as wellas ordinarydepreciation.Item10 is an accounting
of the environmentalservices,and item 15 is an account-
ing of damages-not just to consumers,but to any agents
in the economythat are damaged,includingother busi-
nesses. Becausethey are "free," environmentalservices
are like a subsidy to the industry.Therefore,they are
entered as a negativeinput. Item 11 is the arithmetic
differencebetween items 15 and 10. It ensuresthat the
modifiedaccounts balance. Since it is defined as the
differencebetweenthe servicebenefitof the environment
and the disbenefitof environmentaldamage,it has been
named net environmentalbenefit." 5 The modifiedindus-
trial account input and output totals equal the conven-
tional input and output totals less the absolute valueof
environmentaldamage.
Government
As is suggestedby the abovediscussionof tne standard
national accounts, the conventionalgovernmentalprod-
A ProposedEnvironmentalAccountsFramework 73
uct account is rather simple.Likewise,the modifiedac-
count shownin Table 104 is very simple. It contains
those additionalentries that accountfor the government's
use of the environment,the resulting damage, and the
necessarybalancingentry. However,it also differsfrom
the industrialaccountin that it displaysan entry account-
tion is that all environmentalassets are "owned"by the
government.
Households
The conventionalnational accounts assume that very
little productiontakes place in households(primarilyac-
countedfor by nonprofitinstitutionsand the servicesof
domestics)(Table 10-5). The focus of the conventional
accounts on activitiesthat reflect market transactions
precludesconsiderationof the outputs relatedto keeping
up a house,preparing meals,raisingchildren,and do-it-
16 Householdsare far more importantin
yourselfprojects.
the modifiedaccounts.Especiallyin developingcountries,
householdsaccount for a substantialportion of environ-
mental damageto water becauseof direct dischargesof
sanitarywastes."7
Householdsare uniqueamongthe producingsectors in
that most of the environmentaldamagecausedby house-
holds(as a result of their consumptionof environmental
services)are inflictedwithin the householdsector itself.
In contrast, industries and governmentstend to inflict
damageoutsidetheir ownsectors.
Nature
The modifiedaccountingsystemdiffersmost markedly
fromthe conventionalsystemin its inclusionof nature as
a separate sector (Table 10-6). Nature is shown as the
primarysourceof all environmentalasset servicesand as
the final consumer of environmentaldamages. Nature
also must be includedbecause it generates a substantial
portion of environmentaldamage.For example,a large
portionof dissolvedsolidsin water havea natural origin,
and, on average,naturallygeneratedparticulatesand ni-
trogen oxides(other than NO2) greatly exceed the man-
madeproductionof these air pollutants.
Althoughsome may havea philosophicalobjectionto
the idea of nature as a polluter,the concept is required
for practical reasons. Availableestimates of damages
caused by air and water pollutants cannot distinguish
between damages from offensiveresiduals that have a
human origin and those with a natural origin. Rather
than attribute all the damageto nonnaturalcauses, it is
more accurate to prorate the damagesbetween the two
sources.
74 HenryM. Peskin

Table 10-3. Industrial ProductAccount for a Typical Sector (I)

Input Output
1. Purchasesfromother industrialsectors 12. Salesto privatesector (currentaccount)
2. Compensationof employeesand proprietors(incl. a. To other industries
rental income) b. To households
3. Profitsafter inventoryvaluationand capitalcon- c. Exports
sumptionadjustment 13. Salesto government
4. Net interest 14. Salesfor grossinvestment
5. Imports
6. Transferpayments
7. Indirecttaxes
8. Subsidiesreceived(-)
9. Capitalconsumptionallowances
Gross industrial sector input Gross industrial sector output
10. Environmentalservices(-) 15. Environmentaldamages(-)
a. Air a. Air
b. Water b. Water
c. Land c. Land
11. Net environmentalbenefit(1.15-1.10)a
Modified gross industrial sector input Modified gross industrial sector output
a. 1.15,L10,andso forth,meansitem15,item10,andsoforth,accountI.

Table 10-4. GovernmentalProductAccount (II)

input Output
1. Purchasesfrom industry(1.13) 7. Governmental
goods and services
2. Compensationof employees
3. Imports
Governmental input Governmental output
4. Environmentaldepreciation 8. Environmentaldamages(-)
5. Environmentalservices(-) a. Air
a. Air b. Water
b. Water c. Land
c. Land
6. Net environmentalbenefit(11.8-11.5)
Modified governmental input Modified governmental output

Table 10-5. HouseholdProduct Account (III)

Input Output
1. Purchasesof intermediategoodsfromindustry 8. Servicesto households
(1.12.b) a. Nonprofitinstitutions
2. Compensationof empLoyees and proprietors b. Domestics
3. Imports
4. Surplusof nonprofitinstitutions
5. Capitalconsumptionallowances
Gross household input Gross household output
6. Environmentalservices(-) 9. Environmentaldamages(-)
a. Air a. Air
b. Water b. Water
c. Land c. Land
7. Net environmentalbenefit(111.9-111.6)
Modified gross household input Modified gross household output

Table 10-6. Natural Account (IV)
Input
1. Environmentaldamages(includingthose naturally
generated)
a. Air
b. Water
c. Land
Natural sector input
Consolidated Income and Product Account
The modified accounts shown above can be consoli-
dated into a modified income and product account, as in
Table 10-7. The procedure is the same as that used to
consolidate the accounts shown in Table 10-2. As before,
all flows within sectors have been eliminated. That is, all
items appearing on both sides of the consolidated ac-
count are eliminated. Also as before, househo'ldsare not
included in the consolidation; nor is the nature account
(except for natural sources of environmental damage).
Although nature is viewedas the source of environmental
services and the consumer of environmentaldamages, the
proposed accounting framework does not vievwnature as
undertaking production activities. It is thus excluded from
the gross production account.
The new entries have been arranged to preserve the
conventional account entries, which will enable those who
Table 10-7. ConsolidatedNational Income and'ProductAccount (V)
input
1. Compensationof employeesand proprietors(incl.rental
income)(1.2+ 11.2+ 111.2)
2. Profits after inventory valuation and capital consumption
adjustment(1.3)
3. Net interest(1.4)
National income
5. Transferpayments(1.6)
6. Indirecttaxes (1.7)
7. Subsidies (-) (1.8)
8. Statisticaldiscrepancy
Net national product
9. Environmental depreciation (-)
Modified net national product
10. Capitalconsumption(1.9+ 111.5)
11. Environmentaldepreciation(+)
Charges against gross national product
12. Environmentalservices(-) (1.10+ 11.5+ I11.6)
a. Air
b. Water
c. Land
13. Net environmentalbenefit(V.19-V.12)
Charges against modified gross national product
A ProposedEnvironmentalAccountsFramework 75
Output
2. Environmentalservices
3. Net environmentalbenefit
(IV.1-lV.2)
Natural sector output
may not be interested in the modifications to simply ig-
nore them. Thus, for example, while the modified ac-
counts show environmental depreciation as a negative
adjustment (item 9) to NNP, it is added back (item 11) to
leave the conventional measure of GNPunchanged. Ar-
ranging the entries in this way should alleviate the fears
of those who object to modifying the conventional ac-
counts on the grounds that such modifications destroy
the integrity of the accounts.
The Modified Account Aggregates
Inspection of the modified consolidated account indi-
cates that modified GNP equals conventional GNP less en-
vironmental damage. Actually this relation is an identity:
it is necessarily true because of the way the entries are
arranged in the accounting structure. Several other ar-
rangements are possible, however,each leading to its own
formula relating the conventional GNP to a modified GNP.
Output
14. Personalconsumption(1.12.b+ 111.8)
15. Grossprivatedomesticinvestment(1.14)
16. Exports (1.12.c)
17. Imports(-) (1.5+ 11.3+ 111.3)
18. Governmental goodsand services(11.7)
Gross national product
19.Environmentaldamages(-)(1.15
+ 11.8+ 111.9
+ IV.1)
a. Air
b. Water
c. Land
Modified gross national product
76 HenryM. Peskin

To showthis, the followingnotation is first defined: the first three, since the relation between the environ-
ment and conventionalGNP has alreadybeen discussed.
VA = chargesagainstconventionalGNP or valueadded
GNP = conventionalGNP Option 1: GNP' = GNP - ED
GNPi= modified GNP, definition i (i = 1, 2, 3, or 4)
ES = environmentalservices It was demonstrated above that the conventionalac-
NEB = net environmentalbenefit count aggregatesdo not alwaysrespond to changes in
ED = environmentaldamage. environmentalqualityin a manner that wouldmakethese
aggregatesacceptableindicatorsof well-being.
GNP tended
Sincethe left- and right-handsides of the consolidated to increasewith environmentaldeterioration,and efforts
accountsmust balance,the followingidentityholds: to improvethe environmentare often reflectedby reduc-
(6) VA - ES + NEB = GNP - ED. tions in GNP, particularlyif these effortswereundertaken
Asno,this
identityimpliesthe followingdefinition byreallocatingbusinesscurrentaccountinputsor shifting
As noted,GNP investment.Clearly,the abovedefinitionof modifiedGNP
of modifiedGNP: seems to perform much better as an indicator of well-
(7) GNPI = GNP - ED. (Definition1) being. GNP' appears to move correctly with respect to
changesin ordinaryGNP and to changesin environmental
Accountingarrangementsare somewhatarbitrary,how- quality.Perhapsthis is whythis modificationto conven-
ever,and other arrangementsare possibleas long as the tional GNP was recommendedby Olson (1977).Yet, this
accountsbalance.For example,by addingES and ED to first definitioncoversonly part of the environmentalre-
both sides of equation6 and noting that NEB = ES - lation.The seconddefinitioncoversanotherpart.
ED, a newidentitycan be formed:

(8) VA + ES = GNP + ES Option2:GNP'=GNP+ES

whichleads to a newdefinition: The theoreticalanalysisbehindthe suggestedmodified

Similarly,addingEs and EDto both sidesof equation6
and
again notingthat NEB= ES -lED producesa t1iird
adagin n of
definitionnotingta
modifiedGNEr
GNP:
(10) GNP3 = GNP + NEB. (Definition3)
Finally,first addingES to both sides of equation6 and
then subtractingNEB from both sides producesa fourth
(ll) GNP4 = GNP. (Definition4)
Thus,modifiedGNPcan be definedalternativelyas con-
ventionalGNP less environmentaldamage,as conventional
GNP plus environmental services,as conventionalGNP pIus
net environmentalbenefit,or simplyas equalto conven-
tionalGNP. These definitionsare by no meansequivalent,
but they are all consistentwith the modifiedaccounting
structure discussedabove.The pros and cons of these
alternativesare discussedin the next section.
The Environment and the Modified Accounts:
Options for ModifyingGNP
The relation between the modifiedaccounts and the
environmentcan be discussedin terms of the modified
GNPconceptsdefinedin the previoussection.Since the
fourth definitionof modifiedGNP is exactly the same as
conventional GNP, however,the discussion need only cover
accounting frameworkdemonstrated that there was a
beneficial environmental service associated with any ob-
served environmentaldamage. This service, being un-
priced and apparently"free,"does not showup directly
in conventionalGNP. The second definitiondirectly ac-
countsfor this unpricedinput. Its use as an indexof well-
being would have interesting and perhaps controversial
implications.For example,the differencein incomebe-
tween an unindustrial,developingsociety in a tropical
climatethat generatesenvironmentalservicesin the form
of warmth and abundant, freely availablefood and an
industrialsocietyin a cold climatethat requiresa highly
sophisticatedagricultural
2
systemmaybe far less if income
is measured by GNP , rather than ordinary GNP. (The
differencemightalso be less under the first definitionif
the industrialsocietywere alsothe more polluted.)
2
However,GNP is prone to possibledouble countingof
environmental servicesconsumedby business.Whilethese
servicesmaynot be accountedfor directly,they may be
reflectedin profits, whichare capturedby ordinaryGNP.
For example,a businessthat can disposeof its wastesin
the ocean has a distinct advantage over a competing
businessthat must treat its wastes.The opportunityto
use the ocean's disposalservicemay show up as an in-
creased profit rate for the business. For the national
accountantto add in an amountequalingthe valueof the
ocean servicewould,in this example,be superfluous.
One interesting aspect of GNP2 concerns its behavior
with respect to pollution control expenditures by busi-
ness. As noted, conventional GNP either is unaltered or

declinesdependingon whetherthe expenditureby busi-
ness is on capital or current account and on whether
capital outays for pollutioncontrol divert capital from
more productiveuses. Assumingfull employment,how-
ever,GNP2 willalwaysdeclineas pollutionconbtolexpen-
dituresincrease.Accordingto the theorybehindthe mod-
ifiedframework,pollutioncontrolexpenditures'meamthat
marketed goods and services are being substitutedfor
environmentalservices.Thus, since ES willdecline,GNP2
will also decline. Some may feel that this result, along
with the potentialfor doublecounting,makesCNP2 a less
desirableindicator of well-beingthan GNP'. However,its
focuson the benefitsof Esis a strongpoint in its favor.
Option 3: GNP3 = GNP + NEB = GNP + ES - ED
This definitionof modifiedGNP appears to be a com-
promisebetweenGNP' and GNP2 and shares the strengths
and weaknessesof both measures.As an indicatorof well-
being it appears to move in the correct direction: in-
creasesin ES and decreasesin ED implyincreasesin GNP3.
Yet there are some circumstances under which GNP3 has
difficultyin movingin any direction. In the absenceof
technologicalchange,decreasesin environmentaldamage
by business, ED, must invariably be accompanied by de-
creasesin environmentalservicesto business,ES. Thus,
under policies of pollution control, NEB may remain es-
sentially unchanged. For this reason, it may not be very
effective as an indicator of well-being after all. If D and
ES are valued using marginal valuations (as described
under "Valuation Concepts," above,) however, GNP3 can
convey important information beyond that of the other
two measures of modified GNP. For considerations of
whether any of the modified GNP concepts are good indi-
cators of social well-being,total valuations of environmen-
tal services and damages are of more interest that mar-
ginal valuations. However, marginal valuations provide a
piece of valuable policy information. As described above,
if the marginal ifvluatins
th margnal
valuations OfEs an ED ae thesame7.
of ES and ED are the same
(neglectingthe minus signs), the allocation of the services
of environmental capital is optimal in the sense that any
other allocation lowers social well-being.Thus, if marginal
valuations are used (and if environmental depreciation is
neglected), an optimum allocation occurs wherl ES = ED
or, equivalently,when NEB equals zero or when GNP3 =
GNP. If NEB were negative or GNP3 < GNP, well-being
would improve if ED were decreased. Such a policy rec-
ommendation to reduce environmentaldamage would not
be very controversial.However,the analysis also indicates
that ifNEBwerepositive,orGNP'
> GNP,theremay10.
that lif NEB were positlve, or GNP > GNP, there may be
too little environmental damage for social optirnality.The
concept of too little environmental damage is a natural
consequence of the fact that environmental asset services
contribute to well-being just as does the reduction in
environmental damages. GNP3 explicitly recognizes this
A ProposedEnvironmentalAccountsFramework 77
dualityand the fact that the benefitsof pollutionreduc-
tion are rarely gained withoutcost: some portion of the
benefitsof environmentalservicesusuallyhaveto be for-
gone.
As noted from Table 10-7, in this frameworkenviron-
mentaldepreciationmodifiesonly NNP, but not GNP. For
this reason environmentaldepreciationwas neglectedin
the foregoingdiscussionof modifiedGNPconcepts.Since
many prefer NNP to GNP as an aggregate measure of
economicperformance,however,it is possibleto recast
all the previousdiscussionin terms of NNP. By simply
subtractingcapital consumptionallowancesand environ-
mentaldepreciation(items10 and 11 in Table10-7)from
both sidesof equations6 through 11, all the conclusions
can be restated for NNP.Thus, for example, NNPI = NNP
- ED, NNP2 = NNP + ES, and so forth. The above
inequalitieshold as well.
Notes
1. Environmental
management
theoryhasfoundits wayinto
severaltexts, but one ofthe best expositionsis the seminalbook
byKneeseandBower(1968).
2. Simpleinput-outputmodels,whichviewpollutionas a
by-productproduced in fixed proportion to salableproduct,
neglectthis importantprinciple.The earliestexampleof such a
modelis Leontief(1970).
3. Both directand indirectdisbenefitscan occursimultane-
os
ously.
4. Marginalvaluationsequal totalvaluationsless consumer's
surplus.
5. That is, if physicaldepreciationis Dp and capital gain is
G, the decompositionof value depreciationrequiresthat V, =
V, - Dp + G or V/,- V1 = Dp - G.
6. Of course, in practice the availabilityof drinking water
may be reduced for other reasons, notablyby poisoningfrom
waste products. To keep the analysis simple, however,it is
assumedthat the lake providesonlyone service-drinking wa-
ter-and not a wastedisposalservice.
For a more detailed discussion
seRglsadRgls(96 90 ofn accounting concepts,
eateto
see Ruggles
Commerce and Ruggles(1956;1970)and U.S.Departmentof
(1954).
8. This simple economydoes not includea foreignsector,
nor are there any interestpayments.If these were included,net
exports (exportsless imports)wouldalso appear on the right-
hand side of the consolidatedaccount and net interest on the
left-handside.
9. One implicationof this householdproductionis that many
consumergoods are really intermediatein nature and thus
contributeonlyindirectlyto well-being.See Hirsch (1976)for a
gooddiscussionof this issue.
This statement assumesthat the economywill remain
fullyemployedin spite of these shifts in output. If factors of
productiondo not shift easily-as maybe the case in the real
world-GNP mayfall.
11. The reductionin the right-handside of the consolidated
accountin Table 10-2must be offsetby an equal reductionin
somecomponentson the left-handside. Sincethe total amount
78 HenryM.Peskin
of labor is fixed,either profits, indirect taxes, or depreciation
must be reduced.
12. See Christiansenand others (1980) for a summaryof
estimatesof the effectof pollutionregulationon the economy.
13. To avoid double counting, these damages should be
measurednet of damagesalreadyreflectedin any reduced pro-
ductioncausedby pollution-relatedproductivitylosses(for ex-
ample,effectsof air pollutionon workerhealth).(RobertRep-
petto calledattentionto this point.)
14. This sectionis basedon Peskin (1976)and Peskin and
Peskin(1978).
15. Net environmentalbenefitis shownas the differencebe-
tweenthe environmentaldamagesentry and the servicesentry.
Sincedamagesand servicesare entered negatively,however,it
actuallyis equal to the absolutevalue of the servicesless the
absolutevalueof the damages.
16. The one majornonmarketactivityassociatedwithhouse-
holds that is includedin the accounts-the imputationfor the
implicitrents from owner-occupied housing-is includedin the
businesssector.
17. In the modifiedsystem,seweredhouseholdsare not con-
sideredto pollutethe environment.Seweredwastesare consid-
ered inputs to municipaltreatmentworks,an industrialsector,
whichis creditedwithany resultingenvironmentaldamage.
References
Ayres, Robert U., and Allen V. Kneese. 1969. "Production,
Consumption,and Externalities."American EconomicRe-
view 59(3,June):282-97.
Christainsen,Gregory,Frank Gollup,and Robert H. Havemnan.
1980. "Environmentaland Health-SafetyRegulations,Pro-
ductivityGrowth,and EconomicPerformance."Assessment
report preparedfor the Officeof TechnologyAssessmentand
the Joint EconomicCommitteeof the U.S. Congress.Froc-
essed.
Data Resources, Inc. 1979. "The MacroeconomicImpact of
FederalPollutionControlPrograms:1978Assessment."Re-
port submittedto the EnvironmentalProtectionAgencyand
the Council on EnvironmentalQuality,Washington,I).C.
Processed.
Denison,EdwardF. 1971. "WelfareMeasuresand the GNP."
Survey of CurrentBusiness51(1,January):1-8.
. 1979a.Accountingfor SlowerEconomicGrowth.Wash-
ington,D.C.:The BrookingsInstitution.
. 1979b."Explanations of DecliningProductivityGrowth."
Survey of CurrentBusines59(8,part 2, August):1-24.
Fisher, Irving.1906. The Natureof Capitaland Income. New
York:Macmillan.
Hirsch,Fred. 1976.SocialLimitsto Growth.Cambridge,Mass.:
HarvardUniversityPress.
Johnson,Harry G. 1975. On Economicsand Society.Chicago,
Ill.:Universityof ChicagoPress.
Kneese,AllenV., and Blair T. Bower.1968. ManagingWater
Quality:Economics,Technology, Institutions.Baltimore,Md.:
Johns HopkinsPress for Resourcesfor the Future.
Leontief,Wassily.1970. "EnvironmentalRepercussionsand the
EconomicStructure: An Input-OutputApproach." The Re-
view of Economicsand Statistics52(3,August):262-71.
Marshall,Alfred.1920. Principlesof Economics.8th ed. Lon-
don:Macmillan.
Olson,Mancur.1977. "The Treatmentof Externalitiesin Na-
tionalIncomeStatistics."In LowdonWingoand AlanEvans,
eds., PublicEconomicsand the Qualityof Life. Baltimore,
Md.:Johns HopkinsUniversityPress for Resourcesfor the
Future.
Peskin,HenryM. 1976. "ANationalAccountingFrameworkfor
EnvironmentalAssessments."Journal of EnvironmentalEc-
onomicsandManagement2(4, April):255-62.
- . 1989. "Accountingfor NaturalResourceDepletionand
Degradationin DevelopingCountries."Workingpaper 13.
EnvironmentDepartment,World Bank, Washington,D.C.
Processed.
Peskin,HenryM.,and Janice Peskin. 1978. "The Valuationof
NonmarketActivitiesin IncomeAccounting."Review of In-
comeand Wealth24(1,March):71-91.
Ruggles,NancyD., and RichardRuggles.1956. The Designof
EconomicAccounts. NewYork: ColumbiaUniversityPress
for the NationalBureauof EconomicResearch.
Ruggles,Richard, and Nancy D. Ruggles. 1970. NationalIn-
come Accounts and Income Analysis. NewYork: McGraw-
Hill.
U.S. Departmentof Commerce.1954. NationalIncome,1954
Edition. Supplementto the Survey of Current Business.
Washington,D.C.:U.S.GovernmentPrintingOffice.
 EnvironmentalAccounting
and the Systemof NationalAccounts
Peter Bartelmus
An importantoutcomeof the 1972 U.N. Conferenceon
the HumanEnvironmentin Stockholmwas to affirmthat
environmentalproblemsare causedby highlyaccelerated
economicgrowthas well as by activitiesinduced by an
actual lack of development.A broader viewof environ-
mental damage was also advanced,which included, be-
sides the dischargeof toxic substances,the degradation
and depletionof natural resources. An integrated and
coordinatedapproach to developmentplanning, which
includedsafeguardingthe natural resources of the earth
for the benefit of present and future generations,was
proclaimedin the "principles"of the conferencereport
(UnitedNations1973).
Eight years later the WorldConservationStrategyad-
vocated the "achievementof sustainable development
throughthe conservationof livingresources"as its basic
aim (IUCN1980).The InternationalDevelopmentStrategy
for the Third U.N.DevelopmentDecadealso referredto
the "need to ensure an economicdevelopmentprocess
whichis environmentally sustainableover the long run"
among its goals and objectives(United Nations 1981,
para. 41). Duringits 38th sessionthe U.N. GeneralAs-
semblycalled upon the World Commissionon Environ-
ment andDevelopmentto proposelong-termenvironmen-
tal strategiesforachievingsustainabledevelopmentbythe
year 2000 (UnitedNations 1984b).Accordingly, sustaina-
ble developmentis the all-pervadingthemreof the Com-
mission'sreport (WorldCommission1987).
Althoughthere is generalagreementon the desirability
of sustainableeconomicgrowth and development,it is
Note: The authoris a staffmemberof the UnitedNations.
Theviewsexpressedherearehisownandnotnecessarilythose
ofthe UnitedNations.
79
less clear how "sustainability"can be definedoperation-
allyso that indicatorscan be developedthat signalsignif-
icant deviationsfrom the sustainable path of develop-
ment.' It appears that politicians in most countries
continueto base their decisionson the assumptionthat
overalleconomicand social welfarerises and falls with
the popularmacroindicatorsof gross domesticor national
product, GDP or GNP. In fact, these indicatorsare usually
taken to moveparallel with aggregateincome,which is
obtainedbymakingan allowancefor the wear and tear of
capitalassets used to produce output and corresponding
income.As far as the producedcapital is concerned, an
elementof sustainabilityis thus introducedin the defini-
tion of nationalincome.This approachreflects the Hick-
sian view of income as a practical guide for prudent
conducttowardconsumption,whichshould not exceed a
certainlimitbeyondwhicha person or a populationwould
be impoverished(Hicks1946,p. 172).
The concept of capital maintenancehas not been ex-
tended to natural and human capital assets in national
accounting.As a result,in the case of natural assets,the
degradationand depletionof the naturalenvironmentand
its resourcescontribute to illusionaryincome flows that
cannot be maintainedin the long run. Moreover,expen-
ditures by governmentor householdsfor "regrettable"
measuresto counteractdeterioratingenvironmentalcon-
ditions actuallyincrease national product and income.
Measuresof economicwelfarewere therefore advanced
that purported to correct the misclassificationof regret-
table expendituresas final and that madeallowancesfor
importantnonmarket activities,environmentaldegrada-
tion, and the depletionof natural resources.Best known
are perhaps the pioneeringeffortsby Nordhausand To-
bin (1972),made popular by Samuelson'stextbook as
"net economicwelfare" or NEW (Samuelson1980, p.
183).
80 Peter Bartelmus
At the requestof the U.N.Committeefor Development
Planning,the U.N.StatisticalOffice(UNSO) criticallyre-
viewedthe concepts,methodologies,and empiricalappli-
cations of these measures.The reviewconcludedthat
considerableconceptualand measurementproblemsmake
estimatesof the valueof environmentaldamage "essen-
tially exercisesfor multidisciplinary
research and experi-
ments, not for routine collectionof statistics,"and that
difficultiesof agreeingon the scope, concept,and meas-
urementof welfarewouldrender such approaches"inap-
propriate for officialand especiallyinternational use"
(UnitedNations1977b,pp. 54 and 66). The environment
statisticsprogramof the UNSO focused,therefore,on de-
velopingphysical statistics and related indicators (see
"EnvironmentStatistics,"below).A significantlimitation
of such data is that they lacka commonnum6raire,which
would permit aggregationacross broad sectors of the
environment.Despiteits limitations,GDP thus continues
to be the mainindicatorof economicsuccessor failure.
The joint UNEPIWorld Bank workshopsset out, there-
fore,to reexaminethe feasibilityof physicaland monetary
accountingand concentratedon the fieldof environment
and natural resources. This field appeared to be better
researchedand easier to assessthan the wholegamut of
welfareeffectsoriginallyaddressedby economicwelfare
measures.A consensusemergedin the workshopsthat
enoughprogresshad beenachievedto link environmental
accountingto the standard Systemof NationalAccounts
(SNA)(UnitedNations 1968) and to include certain as-
pectsof environmentalaccountingin the ongoingrevision
of the SNA. This chapter examinesthe treatmentof envi-
ronmental issues in the SNA and reviewsthe main ap-
proachesto resource and environmentalaccountingfor
possiblelinkageor integrationwith the SNA.
Environment Statistics and Physical Accounlting
The close interactionsamongenvironment,population,
naturalresources,and developmentrequirean integrated
approachto planningand policyformulation,which in
turn need to be supported by comparable social, eco-
nomic,and environmentaldata. On the one hand, envi-
ronment statisticssystemsattempt to compileraw data
from a multitudeof sources and to present them in a
from
of surce
mulitudandto prsentthemin a
coordinatedmannerin statisticalcompendiasuch as year-
books or bulletins. Further aggregationand more rigid
presentation of the stocks and flows of natural resources
in physicalterms are, on the other hand, the objectiveof
resourceaccountsand balances.
Environment Statistics
At the internationallevel,the Conferencefor European
Statisticians(CES 1973)of the EconomicCommissionfor
Europetook the initiativeto developa systemof environ-
ment statisticsthat would complementthe already exist-
ing systemsof economicstatistics (the SNA) and social
and demographic statistics (the System of Social and
DemographicStatistics,SSDS) (United Nations 1975). It
soon becameapparent, however,that the currentstate of
knowledgeabout the environmentand its statisticalmeas-
urement, as well as widelydifferingenvironmentalcon-
cerns and priorities, did not permit an internationally
applicablestatisticalsystemto be established.Therefore,
under the guidanceof the U.N. Statistical Commission
and with the financialsupport of the U.N. Environment
Programme(UNEP), the UNSO developeda flexibleFrame-
work for the Developmentof Environment Statistics
(FDES) (UnitedNations1984a).Since the FDES itselfdoes
not recommendstatisticaldefinitionsand classifications,
no direct linkagethrough commonconceptsand classifi-
cationswas establishedwith the SNA or other statistical
systems.2
The FDESis currentlybeing expandedinto a series of
technicalreports on "Conceptsand Methodsof Environ-
ment Statistics,"whichwillproposedefinitionsand clas-
sificationsfor statisticalvariablesand indicators of high
priority.Becausethese statisticsincludeindicatorsof re-
source depletion and environmentaldegradation,their
use for establishingresourceand environmentalaccounts
in physicalterms willbe addressedspecificallyin a report
on "Statisticsof the NaturalEnvironment."These meth-
odologieswill be applied at the regional and national
levelsin a GlobalProgramof EnvironmentStatistics(U.N.
StatisticalCommission1987, p. 29).
Materials/EnergyBalances
Closelinkageto the SNAhas been attempted in more
selectiveapproachesto accountingfor flowsof materials
and natural resources.The uNso developeddraft guide-
lines for statistics on materialslenergybalances (MEB),
whichwere seen as a moduleof a larger systemof envi-
ronment statistics (United Nations 1976). The primary
purposeof the MEB is to trace the extractionand trans-
formationof materialsand energyfrom natural resources,
through varioussuccessivestages of processing,to final
use, and thence back to the environmentas waste or,
alternatively,
to secondaryuse. Compatibilitywith the SNA
iseacived b usindits ustandrdpatebolessuch asdo
is achievedby using its standard categories,such as do-
mestic output and consumption, imports, and exports,
and its classificationof economicactivity.Contraryto the
conceptsof market transactionsand monetaryvaluation
used in national accounting,physical stocks and flows
and materials/energytransformationprocesses are pre-
sented in such balances.
The U.N.StatisticalCommissionconsideredthe MEB to
be an interestinglong-termapproach,but too ambitious
forshort-termimplementationbecausecountriesstilllack
 Environmental Accounting and the System of National Accounts
the necessarystatisticalcapabilities.As a consequence,
only the energypart of this approachwas further devel-
oped in balancesthat showenergyflowsfromthe produc-
tion of primaryenergy,through conversionprocesses,to
the stage of finalconsumption(UnitedNations1988).
A work programto standardizedefinitionsand termi-
nology for mineral resources was initiated by the U.N.
Committeeon NaturalResourcesand later transferredto
the Statistical Commission.The commissionrequested
the UNSO to further developmineralconsumptionstatis-
tics and raw materialbalances (includingthe secondary
recovery of metals), provided the necessary resources
could be made available.Such resources could not be
found, however,and work on this program has been
suspended,
In focusingon processesof resourceflowsthroughthe
economicsystem,the MEB can also be consideredas a
physicalextensionof the input-outputtables of national
accounting.These modelscan incorporateresource re-
quirementsand wasteresidualoutputs,whichreflectsthe
fact that the materialsienergyinputsand outputs for the
economymust alwaysbalance.However,detailedclassifi-
cationsof productionprocessesthat providefor a further
breakdownof the relativelyheterogeneousclassification
by industrieswould be needed to obtain more stable
technical coefficientsof resource use, production, and
residuals.
Although the SNA providesfor the full integrationof
input-outputtables, country tables were oftenicompiled
separatelyfrom national accounts becausethey focused
on specificanalyticaluses. A future handbookon input-
output statisticsand the current reviewof the SNA as a
wholewilladdressproblemsof better integrationof input-
output practicesinto the SNA framework.For the time
being, however,there are no plans to deal specifically
with natural resourceand energyflowsor pollutionproc-
esses in the input-outputsystem.
ResourceAccounting
A few countries with well-developedmonitoringand
data collectioncapabilitieshave turned to a somewhat
simplifiedphysicalaccountingapproach.The basic objec-
tive of resourceaccountingsystemsis to pirovidea coher-
ent picture of resource use and depletionor increase,
whichcan be linked to, or integratedwith, the national
accounts.Norway'sresourceaccountsconcentrateon se-
lected natural resources and extend the SNA balance
sheets to cover the total stock (includingnonmarketed
stocks) of the respectivenatural resources in physical
units (StatistiskSentralbyra1987). The French "patri-
81
mony"approach is more ambitiousin its attempt to in-
clude ecologicalprocessesthat are not directly affected
by human intervention,that is, interactions among the
physicalcomponentsof the environmentand its biota
(INSEE 1986a;see also Theys, Chapter7).
Environment and Natural Resources in the SNA
Commonto all the above approaches is the use of
physical units, which permits aggregationonly in very
limited areas as, for example,in the energy sector by
means of conversionfactors.Physicalaccounts and the
indicatorspresented in their frameworkassess resource
availability,
use, and overuse,and they are thus important
tools for managingnatural resources. However,they are
only an intermediatestep in the overallassessmentof the
national resource base, which is required to formulate
and evaluateresource policiesthat are compatiblewith
general developmentpolicies(Repetto 1986). To calcu-
late overallmeasuresof economicperformance,a com-
mon numeraire,such as the monetaryunit of national
accounting,is needed.Therefore,aspectsof the environ-
ment that are alreadycoveredby the SNA are examined
belowbefore any further methodologiesof monetaryen-
vironmentalaccountingare described.
Flow Accounts
The basic objectiveof the SNA is to provide"a compre-
hensiveand detailed frameworkfor the systematicand
integratedrecordingof the flowsand stocks of an econ-
omy"(UnitedNations 1968,p. iii).No explicitreference
to the environmentor natural resources is made in any
of the core flowaccountsdescribedin the "blue book" of
the SNA. Variousactivitieslistedin the internationalstan-
dard industrialclassification(isic) refer, however,to the
use of natural resources such as agriculture, hunting,
forestry,logging,fishing,mining,water supply,and trans-
portationas wellas to the environmentallyrelated serv-
ices of housingand sanitation. Similarservicesare also
reflectedin the classificationof governmentexpenditures
by purpose. Since the cost of land improvementis in-
cludedin grosscapitalformation,this conceptalso covers
transactionsdealing with selected natural resources. In
none of these conceptsand classifications,however,are
environmentaltransactions clearly identified. They are
typicallylumped together with other nonenvironmental
transactions.Moreover,certain environmentalexpendi-
tures are treated as finalwherethey actuallyrepresent a
cost to society(see "EnvironmentalCosts and Expendi-
tures," below).
BalanceSheets and ReconciliationAccounts
To date national efforts have largely focused on the
accountingof flowsor transactionsto the neglectof the
82 PeterBartelmus
stocks presented in balance sheets. This occursdespite
the fact that balancesheets form an integral part of the
SNA, linked to its flowaccounts through capitaltransac-
tion and reconciliationaccounts.There has been growing
interest, however,in measuringnational wealt&and, in
particular,its tangibleaspects,a significantpart of which
is madeup of assets of the manmadeand naturalenviron-
ment.
Followinga recommendationof the "blue book," the
UNSO issued detailed guidelineson balance sheet and
reconciliationaccountswithinthe frameworkof the SNA
(UnitedNations 1977a). The guidelinesdistinguishbe-
tween reproducibletangible assets, includingthe fixed
assets and stocks of enterprises, financial institutions,
government,and private nonprofit institutions,and the
nonreproducibletangibleassets of natural resourcesand
historicalmonuments.The classificationof stocks and
fixed assets presents importantaspects of the manmade
environmentand listsbuildingsand infrastructuralworks
as part of fixed assets. The measurementof the availa-
bilityand use of natural resourcesand the assessmentof
relatedenvironmental problemsare consideredto be some
of the main uses of the classificationof nonreproducible
tangible assets. Land, timber tracts and forests,subsoil
assets and extractionsites, and fisheriesare the principal
categoriesof this classification.
In line with the basic principleof the SNA to limitits
scope mainlyto market activities(exceptfor imputations
for directlycompetitivenonmarketactivities),natural re-
sourcesin the publicdomainare excluded.This category
includesrivers,lakes, parklands,unusedwilderness,and
the atmosphere,whichare not subjectto ownershiprights
and are thus generallynot sold or purchased.Outlayson
the permanentimprovementand extensionof these items
are coveredin capital formationand thus increase the
reproducible tangible assets (as "other construction
works"or "landimprovement") of the balancesheets. But
no capital consumptionis chargedagainstthese outlays.
However,the total initialexpenditurefor naturallygrow-
ing but commerciallytraded assets, such as the stock of
animals, timber tracts, plantations, and fisheries, is
countedas capitalformation.Further growthand deple-
tion are not recorded as capital formation but are in-
cludedin "reconciliationaccounts"(seebelow).
Balance sheet accounts are linked to the SNA flow
accounts by capital finance accounts, which portray a
significantpart of the changesbetweenopeningand clos-
ingassetsand liabilities.There are, however,severalother
changes in assets and liabilities,such as revaluations,
unforeseenevents, and, in particular, increases or de-
creasesin nonreproducibleassets,whichare not included
in the flowaccountsof the SNA. To coverall differences
betweenthe balancesheet accountsat openingand clos-
ing dates, these changes are formallyintroducedin so-
calledreconciliationaccounts.
The followingitemsof the reconciliationaccounts are
particularlyrelevantfor environmentalanalysis:natural
growthless depletionof timbertracts, forests,and fisher-
ies; new findsless depletionof subsoilassets;and losses
in land and timber tracts in catastrophesand natural
events.Changesin land qualitydue to erosion,waterlog-
ging, desertification,or upgradingand increased availa-
bilityof mineralresourcesdue to technologicaladvances
are not shownseparately.They are recordedas changes
in the market valueof land and subsoilassets as part of
the revaluationitem of the reconciliationaccounts. All
these items are excludedfrom the capital transaction
accounts and thus do not affect the level of national
incomeor product.
Monetary Environmental Accounting
As alreadydiscussed,physicalindicatorsand accounts
cannot be aggregatedinto overallmeasuresof resource
depletionand environmentaldegradation.Consequently,
they cannotbe compareddirectlywiththe standardmeas-
ures of economicperformanceto assessthe sustainability
of economicgrowth or development.For this reason,
physicalmeasureshavenot becomestandards in national
planning and policymaking,but have been largely re-
stricted to use in environmentaland resource manage-
ment. Becauseof the prominentrole of nationalaccount-
ing and its mainaggregatesin formulatingand evaluating
national plans and policies, various environmentalac-
counting approaches attempted to tackle the issue of
sustainabilityand environmentaldegradationwithin,or as
close as possibleto, the establishedsystemsof national
accounting.
Mostof these approachesfavorthe direct adjustmentof
SNA concepts,classifications,and tabulations.It is hoped
that this would cause politiciansand administratorsto
base their plansand decisionson a conceptof sustainable
income,whichaccountsfor the total cost of destructive
and wastefuluses of the environment and natural re-
sources.Duringthe current revisionof the SNA, however,
it was generally agreed that no significantconceptual
changesshould be introducedin the central framework
of the SNA. Rather,the emphasisshould be on clarifying,
simplifying,and updatingthe SNA for both producersand
users of the systemand on improvingits consistencywith
other internationalstandardsof statistics.A possiblecom-
promise(suggestedbelowin "Future Work")is to elabo-
rate the differentmethodsof environmentalaccounting
in satelliteaccountsand tabulationsof the SNA.3
The followingreviewof the mainapproachesdiscussed
in the uNEP/World Bank workshopsillustratesthe possi-
ble contentsof satelliteaccounts.Three basic categories
of environmentalaccountsare distinguishedaccordingto
the mainenvironmentalconcernsaddressedin the work-
shops, namely accounts of environmentalcosts and ex-
penditures, natural resource capital, and environmental
services.
 EnvironmentalAccountingand the System of NationalAccounts
Environmental Costs and Expenditures
The measurementof environmentaloutlaysand, as far
as possible,of resourcesor "incomings"of the principal
agents of the national accountingsystemhas been pio-
neered by France in "environmentalsatelliteaccounts"
(INSEE 1986b).Operating(running)costs or current out-
laysare distinguishedfrom investmentand other capital
formationexpenditures.One of the mainproblemsin this
approach is to identifythe environmentalactivitiesfor
which outlayshave been incurred. This problem exists
becausethe standard classificationsof the SNA were not
designedto generate statisticson environmentalexpen-
diture.
For governmentexpenditures,the classificationof the
functionsof government(COFOG) does not yet contain a
categoryrelating to the protection of the environment,
but it indicatesthat such a groupingmaybe feasiblein
the future (United Nations 1980). A list of functional
classes that include informationon research and other
aspectsof environmentalprotectionis advance(dfor "fur-
ther examination."A draft classificationof outlaysof in-
dustriesby purpose(coip)wasdevelopedby the UNSO in
1975 but was not further discussedor promoted.It lists
nine categories,one of whichspecifiesoutlayson pollu-
tion abatementand control.For this category,outlaysfor
"end-of-line"measuresare said to be relativelyeasilyal-
located.However,the incrementalcosts of measuresthat
are incorporatedin the productionprocess can be only
estimated. For private nonprofit bodies serving house-
holds, the SNA containsa classificationof purposescon-
sistingof eightcategoriesthat do not specifyenvironmen-
tal purposes.For households,the SNA's classificationof
householdgoods and servicesmighteventuallybe linked
to a generaland more detailedcentralproduct classifica-
tion (cpc), whichhas been prepared by the tJNSO. One
divisionof this classificationreflectsenvironmentalfunc-
tions in categoriessuch as "sewageand reiuse disposal"
and "sanitationand similarservices."
The developmentof functionalclassifications in the SNA
that cut across sectoralboundariesand identifyexpendi-
tures for environmentalpurposeswouldpermitthese ex-
pendituresto be aggregatedfor the wholefield of the
environmentand broken downby environrmental protec-
tion category.The aggregateof the total (finaland inter-
mediate)environmentalexpendituredescribeshowmuch
a nation has been willingto spend on environmental
protectionduring the accountingperiod. T'hisaggregate
also facilitatesinternationalcomparisons,since it is not
affectedby differentinstitutionalarrangementsfor envi-
ronmentalpolicymakingand budgeting.
Functional classificationsof environmentalexpendi-
tures can also be used to measure so-calleddefensive
expendituresincurredby the government,households,or
industryto counteractenvironmentaldegradation.When
such expendituresare counted as final consumptionor
83
investment,they increasedirectlynational product or in-
come. Governmentalexpenditurestypicallyinclude out-
lays for environmental protection and remedy, while
householdsmight bear the cost of medicaltreatment for
pollution-relateddiseases or the avoidance of environ-
mentaldisamenities,for example,by resettlement.If en-
vironmentalexpendituresare incurred by industry,how-
ever, their current cost component is counted as
intermediateconsumption,whichpotentiallydecreases(in
the case of full employmentof the factors of production)
industrialvalueadded,that is, industry'scontributionto
national product. Capitalexpenditurefor environmental
protectionby industryis treated,however,as finalexpen-
diture and is thus fullycountedas an increaseof GNP.
All these expendituresserve only to maintaina certain
levelof environmentalqualityor, in other words,to "de-
fend"societyagainstthe unwantedsideeffectsof produc-
tion and consumption.They could, therefore,be consid-
ered as a cost to society,to be deducted from gross
productionand finalconsumption(see "Indicatorsof Sus-
tainable Product and Expenditure,"below).Apart from
expendituresfor environmentalprotection and compen-
sation for environmentaldamage,defensiveexpenditures
may also include other social costs of urbanizationand
industrialization,such as commutingcostsand provisions
for environmentalhazards in industriesand the working
environment(Leipert1985).Asalreadymentioned,it may
be difficultto identifyseparatelydefensivecosts in the
case of combined measuresthat restore environmental
conditionsand raise productivityor utility at the same
time.
The social restoration cost of meeting or maintaining
environmental standards for safeguardingsustainable
economicdevelopmenthas been advancedas an indicator
that mightusefullysupplementnationalproduct (Hueting,
Chapter6). The indicatordescribeshowfar a nation has
driftedfrom its desiredlevelof environmentalqualityand
can thus be compared with the extent the nation has
progressedin raisingits level of production. Standards
for sustainabledevelopmentcouldbe set by decisionmak-
ers or by science, and appropriate measures to meet
these standardscouldthen be formulatedand costed.For
irreversiblelossesof renewableresources(extinction)or
for nonrenewableresources,the cost of developingalter-
native resources or facilitiescould be estimated. The
methodof usingstandardshas been proposedbecauseof
the difficultiesof devisingshadowprices for environmen-
tal functions(and their losses)that are directlycompara-
ble with marketprices of producedgoodsand services.
Natural Resource Capital
Most changesin the resourceassets of a country have
been excludedfrom the standard flow accounts of the
SNAbecauseof the lack of reliabledata and uncertainties
in estimationand valuationprocedures,particularlyfor
84 PeterBartelmus

assets whosereturns are typicallyspread over many ac- services (counted as a subsidy or negative input) ancl
countingperiods.As describedabove,some of the more deliverenvironmental"bads" (as negativeoutputs)to na-
commercialchanges are recorded separatelyin reconcili- ture's productionaccount(Peskin1981).
ation accounts,but these accounts haverarelybeen pre- The model claimsto be fullyconsistentwith the prin-
pared and thus are largely neglected in policymaking. ciples of neoclassicaleconomicsin measuringthe envi-
Now,however,awarenessof the unsustainability of income ronmentalaspectsof socialwelfare.There are, however,
resulting from the consumptionof finite natural assets considerableproblemsin measuringand valuingnonmar
has created new interest in accountingfor selected re- ket contributions from and to nature and in double
sourcesin physicalaccountsand has also revitalizedat- counting the external economies and diseconomiesof
temptsto assignmonetaryvaluesto the stocks and flows these contributionsthat are already reflectedas produc-
of these accounts. tivitychangesin the productionaccounts.
In principle,such accounts can be developedin har-
monywithnationalaccountingconceptsby extendingthe Indicators of Sustainable Product and Expenditure
SNA's balancesheetsto coverfullynaturalresourcessuch
as water,forest,or wildlife,evenif they havenol:been the The environmentalaccounts described aboveproduce
objectof commercialtransactions.Contraryto the prac- aggregatesthat can be used to modifythe valueadded or
tice of reconciliationaccounts, the separate listing of the valueof gross productioncreated,and the amountof
additions(discoveries,net revisions,extensions,growth, expendituresallocatedfor consumptionand investment
or reproductions),reductions(extraction,deforestation, by those transactors of the economythat use or protect
or degradation),and revaluations
' in "national income ~~~~~~~~~~natural
resources and environmentalservices. Accord-
accounts for natural resources" has been proposed to ingly,the followingnew conceptsof gross and net domes-
record changes in the values of the stocks of natural tic product, incomeand expenditurehavebeen proposed
resourcesduring the accountingperiod (Repetto1986). .. .'
Net changesin the valueof the resourcepositioncan be
estimatedfrom these accountsfor the calculationof sus-
tainableincomeor product indicators.Other procedures SustainableNet Product and Expenditure
to estimatethe depletionfactor in the calculationof sus-
tainableincomeinclude applyingdepreciationprinciples Correctingnational incomefor changes in the net re-
to the stock of natural resourcesor calculatinguser cost source position,which is obtained from accountingfor
allowancesfor their intermediateconsumption(see "Indi- depletion and the increase in the stock of natural re-
cators of SustainableProductand Expenditure,"below). sources,has been proposed for calculatingindicators of
sustainableincome or product (see "Natural Resource
EnvironmentalServices Capital,"above).It has been argued,however,that geolog-
ical discoveriesof natural resources can be neglected,
The amount of environmentalexpendituresor costs since they do not reversethe process of depletion but
described above does not reflect the correct value of extend only the period over which depletion can con-
actualenvironmentaldamage.On the one hand, the value tinue. Consequently,if the depreciationconceptwere ex-
of damagemaybe higheror lowerthan the costof repair- tended to the consumptionof natural capitalby resource-
ing it. On the other hand, accountingfor environmental exploitingindustries,it wouldaccountsufficientlyfor sus-
expendituresneglectsthe fact that pollutionis connected tainability in a measure of sustainable product (Daly,
with both the benefits of using nature's disposal and Chapter2).
absorptionservicesand the "disbenefits"resultingfrom The valuationof naturalresourcestocks couldbe based
lossesof these and other servicesthat benefitconsumers on the principle of replacementcosts or willingnessto
directly.Contraryto defensiveexpenditures,these serv- pay. However,changesin the value of environmentalas-
ices are not priced by the market system and are thus sets can result from both physicaldepreciationand capi-
excludedfrom standard national accounts (see Peskin, tal loss or gain, as, for example,with increasinglyheavy
Chapter10). use of a naturalresource,such as the use of a waterbody
The introductionof a natural productionaccountinto for fishingand recreation.In theory,valuationof capital
the nationalaccountingsystemwas thereforesuggested. loss or gain should be based thereforeon the marginal
Such an account would record environmentalservices benefitsof uses (Peskin,Chapter10).
(for example,waste disposal,cooling, and provisionof The extensionof the depreciationprincipleto the con-
oxygenand nutrients)as output from,and environmental sumptionof natural capitalaffectsthe net valueadded of
damage(in particularpollutioneffects)as input into, the industriesthat exploitnatural resourcesand thus the NDP.
environment,withthe balancerepresentinga net environ- In addition,defensiveexpenditureshavebeen considered
mentalloss or benefit.Industry,government,and house- as intermediaterather than final consumption,and the
hold productionaccountsreceivenature'soutputas "free" furtherdeductionof such expendituresfrom net national
 EnvironmentalAccountingand the System of NationalAccounts
product has been proposed.This procedurewouldyield
a new indicator,whichhas been termed "sustainableso-
cialnet nationalproduct"(Daly,Chapter2).
Such sweepingtreatment, however,loses track of the
accountingidentity betweennational income(the sum of
valueadded createdin production)and nationalproduct
(the value of final demand)because defensiveexpendi-
tures contain elements of both intermediateand final
consumptionand capitalformation.A new,imoresystem-
atic approachto dealingwith(deducting)defensiveexpen-
ditureshas thereforebeenproposedin a draftSNA Frame-
work for EnvironmentalSatelliteAccounting(Bartelmus
and van Tongeren1988).The idea is to reduce the pro-
ductionboundaryof the SNA by coveringonlynonenviron-
mental activitiesand thus excludingall production of
environmentalgoods (filters,chemicals,or wastedisposal
plants)and services(protectingor restoringthe qualityof
the environment)that are used in intermediateand final
demand.In this manner, valueadded and finaldemand
are adjustedto obtain identicaltotal values.Further al-
lowancefor the depreciationof natural resource assets
then obtainsa measurethat couldbe termed "sustainable
5
net national expenditure" (SNNE).
As an alternative,the joint UNEP/World Bank Expert
GroupMeetingin Paris (November1988) suggestedap-
plyinga widerconceptof depreciationto calculate"envi-
ronmentallyadjustednet domesticproduct."This concept
would include both the depletion of the environmental
"stock" assets of natural resources and the degradation
in the qualityof "permanent"environmentad assets.The
depreciationof environmentalqualitycould be estimated
as the differencebetween actual defensiveexpenditure
and the cost of restoring environmentalqlualityto the
levelat the beginningof the accountingperiod (in other
words,"potential"defensiveexpenditures).However,such
a procedureneither accounts for the "inflation"of NDP
by defensiveexpendituresnor reflectsactualienvironmen-
tal degradation or damage. Clearly, monr research is
needed to clarify the relations between damage costs
(expenditures)and actual damage as well as their impli-
cationsfor environmentalaccounting.
SustainableGrossProduct and Expenditure
Unfortunately,net product is infrequenty calculated
and rarelyused in quantitativeeconomicanalysisbecause
of the well-knowndifficultiesof estimatingdepreciation.
Therefore,GDP has usuallyreceivedmore attention in
economicplanningand policy,under the assumptionthat
the grossand net indicatorsmovesimilarlycivertime. The
adjustmentof gross rather than net product or income
has therefore been suggested to obtain a measure of
"sustainablegross domesticproduct" (SGDP') or "sustain-
able gross domesticexpenditure"(SGDE).
The draft SNA Frameworkfor EnvironmentalSatellite
Accountingproposesto calculateSGDP by accountingfor
85
the exploitationof natural resourcesas intermediatecon-
sumptionof differentproductionactivitiesand deducting
a correspondingamount from final demandas a change
in "environmentalassets." 6 For estimatingthe amountof
intermediateconsumptionto be deductedfrom the gross
output of resource-exploitingindustries,the calculation
of a so-calleduser cost allowancehas been advocated(El
Serafy,Chapter3). User costs calculationsavoidthe diffi-
culties of estimatingphysical or value depreciationby
allocatinga certainamountof the receiptsfromextractive
industriesto an actual (Ward1982)or hypotheticalfund,
which, if invested,would "create a perpetual stream of
incomethat wouldprovidethe same levelof true income,
both during the life of the resource as well as after the
resource has been exhausted" (El Serafy, Chapter 3).
Discoveriesare simplydealt with by changingthe para-
metersof the life expectancyof the resource,that is, the
resource-extractionratio.7 SGDE is derivedfrom SGDP by
deductingadditionallydefensiveexpendituresof environ-
mental protection.An approach for deducting defensive
expenditureswithoutviolatingaccountingidentities was
described above(see "SustainableNet Product and Ex-
penditure").
ModifiedGrossor Net Product
The accountingfor environmentalservices and dam-
ages (or their difference,that is, the net environmental
loss or benefit)has been describedas the result of intro-
ducinga natural production accountthat modifiesgross
output in the productionaccounts.This procedurewould
yielda "modifiedgross nationalproduct" (Peskin,Chap-
ter 10). Deductingenvironmentaland other economic
depreciationfrom this aggregateproduces a "modified
net domesticproduct,"whichwouldaccountfor environ-
mentaldegradationas wellas resourcedepletion.Consid-
erableproblemsof measuringand valuingenvironmental
servicesand damageswillprobablymake the estimation
of these indicatorsthe object of research studies rather
than of routine compilationin environmentalaccounts.
Future Work
As indicatedat the beginningof this chapter,worldwide
concernwith the sustainabilityof current patterns of pro-
duction and consumptionhas raised doubts about the
suitabilityof standardaccountingand its main aggregates
for integratedsocioeconomicand environmentally sound
planningand policies.In this context,the WorldCommis-
sion on Environmentand Developmentdrewattention to
the need to shift emphasisfromthe standardenvironmen-
tal policyof dealingwith environmentaleffectsafter they
occurredto "an approach concentratingon the policies
that are the sourcesof those effects"(WorldCommission
86 PeterBartelmus
1987, p. 310). Such policies should be assessed by indi-
cators that present socioeconomic progress net of all
environmental costs. Up to now these costs have generally
been neglected or wrongly accounted for in the standard
macroeconomic aggregates.
The obvious place to develop and standardize concepts
and procedures of environmental accounting is the exist-
ing standard System of National Accounts, the SNA. The
the SNA provides a unique opportunity
ofrevision
ongoin
to examine how the concepts definitions classifications
how the concepts, definitions, classifications,
t examine
and tabulations recently developed in the fields of envi-
ronmental and resource accounting can be related to, or
integrated in, the SNA. At the current stage of develop-
ment it might be premature, however,to radically change
a well-establishedsystem of data collection and account-
ing that serves many different short-, medium-, and long-
term socioeconomic analyses.8 Therefore, standards for
environmentaland resource accounting should be further
elaborated in a satellite system of environmental accounts
of the SNA. Efficient linkage will require some amend-
ofnts the
to
S eicien clnkagewio rsmexand-
ll
ments to the existing classifications as well as explana-
tions of, or cross-references to, alternative concepts and
tabulations in the core recommendations of the SNA.
A first draft of the SNA Framework for Environmental
Satellite Accounting was discussed by the joint UNEP/
World Bank Expert Group Meeting on Environmental
Accounting and the SNA, and several recommendations
were made.
* The revised SNA should provide a place for environ-
mencstal faccntswthinathe overtallsymemand shoduld
discuss the fact that conventional income anc, product
measures do not reflect environmental degradation and
depletion of natural resources.
* Environmental accounting should be presented as
satellite accounts and not in the SNA itself.
* Functional classifications should be developed that
identify separately environmental goods and services, in
particular in COFOG, cpc, coip, and isic.
* Environmental resources that cannot be valued as
stocks should be listed in the form of an inventory that
may include indicators of physical quantity and quality.
* A first step toward measuring sustainable income
and product should be to deduct the depletion and deg-
radation of environmental assets; further research on de-
ducting defensive expenditures is required.
The meeting also suggested that further work be un-
dertaken toward a draft manual through pilot-l:ypecase
studies and the evaluation of existing resource accounting
systems in industrial countries. The uNso has included
the preparation of a handbook on environmental satellite
accounting in its work program and will collaborate closely
with the World Bank, UNEP,and other organizations in
this effort.
Notes
1. For a further discussion of the concept of sustainable
development,see Bartelmus(1987a).
2. See, however,Bartelmus(1987b),who proposed to apply
the generic principlesof the FDES to the SNA and the SSDSto
developan "overallframeworkfor statisticalintegration."
3. The term "satelliteaccounts"is used here in a broadsense
to include,besidesoutlaysand financialresourcesfor environ-
mental protection (as presented in environmentalsatellite ac-
counts in the narrowsense; see, for example,INSEE 1986b),
costsand benefitsof environmentalimpactsand resourcedeple-
tion or discovery.
4. Of course,these modificationsapplyalsoto nationalprod-
uct and income(addingnet factor incomefrom abroad);these
indicatorsare less frequentlyused, however,and, unlessother-
wiseindicated,referenceis made to domesticproduct only.
5. Even though accountingidentitiesare maintainedby this
procedureof deductingdefensiveexpenditures,severalassump-
tions make the meaningof SNNEsomewhatambiguous,espe-
ciallyfor ex-anteanalysis.The first assumptionis that environ-
mental production is not replaced by nonenvironmental
production.Thisis realisticonlyif SNNEis to reflect the aggre-
gate value of nonenvironmentalincome ex post. If it is to
approximatepotentialincome,however,and referto a situation
free of environmentaldistortions,one could assumethat factor
inputsreleasedfromenvironmentalproductionwouldhavebeen
used to generate additional income. In addition, the above
definitionof defensiveexpendituresdoes not account for indi-
rect effects,that is, the employmentof nonenvironmental goods
and services in the supply of goods and services related to
defensiveexpenditures.
6. Changesin environmentalassets introducea new element
into final demand,which replacesthe conceptof capital accu-
mulationwitha newconceptof "changesin tangiblewealth."
7. Anotherpossibilityis to consider discoveriesof new re-
sourcesas the creationof an asset and to record this activityas
own-accountcapital formation of the mining industry (OECD
1985).
8. At least,this appearsto be the viewof the expertsworking
on the current revisionof the SNA,who concludedthat "no
major conceptual changes in the SNA should be introduced"
(UnitedNations1986,p. 6).
References
Bartelmus,Peter. 1987a.Accountingfor Sustainabl Develop-
ment. WorkingPaper 8. NewYork:UnitedNations,Depart-
ment of InternationalEconomicand SocialAffairs.
. 1987b."BeyondGDP: NewApproachesto AppliedSta-
tistics."TheReviewof Incomeand Wealth33(4)(December):
347-58.
Bartelmus,Peter, and Jan van Tongeren. 1988. "SNA Frame-
workfor Environmental SatelliteAccounting."UnitedNations
StatisticalOffice,NewYork.Processed.
CES(Conferenceof European Statisticians).1973. "Report of
the Meetingon Statisticsfor EnvironmentalStudiesand Pol-
icies,"cEsIAC.4015),March19-23, Geneva.Processed.
 EnvironmentalAccountingand the Systemof NationalAccounts
Hicks,J. R. 1946. Valueand Capital.2nd ed. Oxford,U.K.:
OxfordUniversityPress.
INSEE(InstitutNationalde la Statistiqueet des EtudesEcono-
miques). 1986a. Les comptes du patrimoinenaturel. Les
Collectionsde l'INSEE137-138C.Paris.
. 1986b.Les comptessatellitesde l'environnement,meth-
odeset resultats.LesCollectionsde l'INSEE130C.Paris.
IUCN(InternationalUnionfor Conservationof Nature and Nat-
ural Resources).1980. WorldConservationStrategy.
Leipert,Christian.1985. "SocialCostsas a Factor of Economic
Growth."Paper for the Conferencefor a NewEconomics:
The Other EconomicSummit,April 16-19, London;abbre-
viatedversion:1986."FromGrossto AdjustedNationalProd-
uct." In Paul Ekins, ed., The Living Economy:A New Eco-
nomicsin the Making.London:Routledgeand Kegan.
Nordhaus,W.D., and J. Tobin.1972. "Is GrowthObsolete?"In
NationalBureauof EconomicResearch,EconomicGrowth.
GeneralSeries 96. NewYork.
OECD(Organizationfor EconomicCooperationand Develop-
ment). 1985. "Treatmentof MiningActivitiesin the System
of NationalAccounts."Restricteddraft, DESNI/85.4.Paris.
Processed.
Peskin, Henry M. 1981. "NationalIncomeAccountsand the
Environment."NaturalResourcesJournal21:511-37.
Repetto, Robert. 1986. "Natural Resource Accounting for
Countrieswith Natural Resource-BasedEconomies." Paper
for the UNEPIWorld Bank Workshopon EnvironmentalAc-
counting.WorldBank,Washington,D.C.Processed.
Samuelson, Paul A. 1980. Economics. 11th ed. New York:
McGraw-Hill.
StatistiskSentralbyra.1987.NaturalResourceAccountingand
Analysis:TheNorwegianExperience,1978-198f. Oslo.
United Nations. 1968. A System of National Accounts.
E.69.XVII.3. NewYork.
. 1973.Report of the UnitedNationsConfermnce
on the
Human Environment, Stockholm, June 5-16, 1972.
E.73.11.A.14. NewYork.
-. 1975. Towarda System of Social and Demographic
Statistics.E.74.XVII.8.NewYork.
87
- . 1976. "DraftGuidelineson Materials/EnergyBalances."
Report of the SecretaryGeneralon EnvironmentStatistics.
EICN.3/492. NewYork.Processed.
- . 1977a.ProvisionalInternationalGuidelineson the Na-
tional and Sectoral Balance-Sheetand ReconciliationAc-
counts of the System of NationalAccounts. E.77.XV11.10.
NewYork.
- . 1977b. TheFeasibilityof Welfare-Oriented
Measuresto
Supplementthe NationalAccountsand Balances:A Techni-
cal Report.E.77.XV11.12. NewYork.
. 1980. Classificationof the Functionsof Government.
E.80.XVI1.17. NewYork.
. 1981. Resolutionsand DecisionsAdopted by the Gen-
eral Assembly during Its Thirty-FifthSession. General As-
semblyOfficialRecords,supplement48 (Al35148). Resolution
35156:International DevelopmentStrategy for the Third
UnitedNationsDevelopmentDecade.NewYork.
. 1984a.A Frameworkfor the Developmentof Environ-
ment Statistics.E.84.XVII.12. NewYork.
- . 1984b.Resolutionsand DecisionsAdopted by the Gen-
eral AssemblyduringIts Thirty-EighthSession.GeneralAs-
semblyOfficialRecords, supplement 47 (A138147).Resolution
381161.NewYork.
. 1986. "NationalAccountsand Balances:Systemof Na-
tionalAccounts(SNA),Progressin the Revisionof the System
of NationalAccounts(SNA)."Report of the Secretary-General.
EICN.3/1978/5. NewYork.Processed.
. 1988. Energy Balancesand ElectricityProfiles 1986.
E.86.XVII.14. NewYork.
U.N.StatisticalCommission.1987.Reporton the Twenty-Fourth
Session.E/1987119.NewYork.
Ward,Michael.1982. "Accountingfor the Depletionof Natural
Resourcesin the NationalAccountsof DevelopingEcono-
mies." CD/R(82)3010. OECD,Paris. Processed.
World Commissionon Environmentand Development.1987.
Our CommonFuture.Oxfordand NewYork:OxfordUniver-
sity Press.

RecentDelopments and FuhureWork
Ernst Lutz and Salah El Serafy
The Systemof NationalAccounts(SNA) that is currently
in use is based on a document published more than
twentyyearsago (UnitedNations1968).A revisionof this
systemwas started in the early 1980s,with a number of
expert groups lookingat variousissues and considering
howimprovementscould be made with as little distur-'
bance tmprovthemcored systemadewit
posbleithe
as rvisi
bance to the sNA core system as possible.The revisions
are scheduledto be completedby 1991,at whichtime a
newversionwouldbe issued.
Giventhe shortcomingsof the current systemwith re-
spect to the environmentand natural resources,the ques-
tion is how to reach the long-termgoal of makingthe
necessarychanges so that national income accounting
more accurately represents true, sustainable income.
Some,at leastin the past, havearguedforradicalchanges,
whileothers have supported more gradual changesthat
wouldimproveincome estimatesby movingthem in the
right direction. Although it is clearly desirable to make
improvementsas soon as possible,some outstandingcon-
ceptualissuesare yet to be resolved,and much empirical
workmuststilbedonebefre
secifc rcommndatons
work muststill be done beforespecificrecommendations
for changesin the core of the SNA can be made. In the
meantime,the creation of satelliteaccounts appears to
be the only realisticoptionfor progressin this area.
sNA Expert GroupMeeting(Vienna,March 1988)
In the contextof deliberationsoverwhetherand in what
waythe SNA shouldbe revised,the WorldBank presented
a discussionnote to the SNA expert group meetingheld
in March1988 in Viennaunder the agendaitem "Links
between SNA and EnvironmentStatistics."The note fo-
cusedon the two main concernselaboratedin Chapter1:
_ _
Note:The annexwaswrittenbyJan W.vanTongeren.
12
defensiveexpendituresand the depletionand degradation
of natural resources.
For defensiveexpenditures,the Bank recommended
that the experts:
* Discussand, as soon as feasible,decidewhichexpen-
dituresare to be includedunder "defensiveexpenditures"
* Agreethat these are to be shownas a separate line
item
* Decidewhere the adjustmentsare to be made and
wherethe adjustedaggregatesare to be shown(in satellite
accountsor i the main systemof producton, consump-
tion, and accumulationaccounts).
On the issueof resourcedepletionand degradation,the
Bank recommendedthat the expert group:
* Adopt the principleof sNA-linkedsatelliteaccounts,
whereadjustmentsfor resource depletion and degrada-
tion can be made
o Advisecountrieswherea significantportion of gross
nationalproduct (GNP) is based on the depletionor degra-
dation of naturalresourcesthat current income,as calcu-
lated at present,overestimates
sustainableincomeand that
policymakers shouldexplicitly
be alertedto this fact
* Encourageworkon estimatingthe costs and benefits
of resourcedepletionand degradation.
The membersof the group agreedthat these two issues
are of great importance and indicatedthat they shared
the Bank'sconcernabout the environmentand supported
the constructionof satelliteaccountslinked withthe SNA,
as wellas the developmentwithinthe central SNA frame-
work of the functionalclassificationof governmentexpen-
ditures(COFOG) and enterpriseexpenditures(coip)in or-
der to identifyexpendituresrelevant to environmental
analysis.However,they did not agree to changingthe
_ core accounts of the
_ __ SNA in any fundamental way. Three
reasonsweregivenfor this.
88

* Since there are variousconceptualapproachesthat
could be taken for revisingthe SNA, a consensusshould
be reachedbeforethe core SNA is revised.
* In the environmentalfieldmany valuationissuesare
difficult,and more work is required beforestandard val-
uationtechniquescan be applied.
* The group waspleasedthat a list of paperspresented
at the joint uNEPIWorldBank workshopswas attachedto
the Bank's position paper. But since they had not seen
thembeforebecauseof the limitedcommunicationso far
betweenenvironmentalists and nationalaccountants,they
did not knowwhat the issuesand options were (and this
supported the urgent need for producingthis volume).
As a concessionto the environmentalcommunity,the
chairman ~ if an.hnaeuaeeprec
stte tha
chaian stated withsatei andawhen t rienc
hasceenlgainedwithlsateite accounts
hand wen variousd
t deal withvariousassets such as air, water,and land sep-
nationalaccountantsshould be open to an re revision,
enationthaccountan shoulner on to antSNa .reion, * A classificationof intermediateconsumptionby func-
eveneinthatwastmu the twenty-yearipei.o
Duringthe discussionsome other points w
betwengthe lasond the urre reinso
weremadeby
vaiu
pripat.
variousparticipants.
* Usersshouldbe cautionedabout the gross,domestic
product (GDP) as calculatedat present; the revisedSNA
manual ("bluebook")should includea discussionof the
shortcomingsof the GDP and state necessarywarnings
againstits indiscriminateuse.
- Communicationbetween national accountants and
environmentaleconomistsshouldbe encouragedso that
problemscan be discussedjointly.
* Approximationsthat maybe needed in erivironmen-
tal accountingmaynot be worse than some currentarbi-
traryestimationsalreadysanctionedby the SNA,.
* The SNA shouldbe linkednot only withenvironmen-
tal statisticsbut with socialstatisticsas well.
At the end of the meetingthe group agreed that these
issueswouldhaveto be discussedagainand that prefer-
ablynationalaccountantsand resourceeconom,ists should
address them together. A meeting combiningthe two
groups was held in November 1988, as described below.
Joint Unmentald Ac nk th S
nting
EnvironmentalAccountingand the SNAprdcND
(Paris, November 1988)
National accounts experts and environmentalecono-
mists met in Paris to discuss a draft frameworkfor a
proposedset of environmentalsatelliteaccounts(Bartel-
mus and van Tongeren 1988). Both groups were willing
to learn, the meetingwasvery productive,and a consen-
sus wasreachedon severalpoints,
* Natural resources and the environmentare impor-
tant and are likelyto becomemore so in the future, so
theyshould be properlyaccountedfor.
RecentDevelopmentsand Future Work 89
* As presentlycalculated,GDP is not a comprehensive
indicatorthat can satisfyadequatelya diversityof needs,
and its shortcomingsshould be kept firmlyin mind.
* ReplacingGDP with a more sustainablemeasure of
incomeis not yet feasible.For the time being, therefore,
satelliteaccountslinked to the SNA should be created in
whichadjustmentsand alternativecomputationscan be
made.
* The revisedversionof the "blue book" (the volume
describingthe core accountsof the SNA), to be issued in
1991, should have a section in whichsatellite accounts
for the environmentare outlined.
* For the environment,net measures,suchas net domes-
tic product (NDP), maybe more meaningful and shouldbe
emhszdortangsseaumns,ucasGP
emphasizedmore than gross measurements,such as GDP.
* Further developmentsof the draft frameworkmight
arately,since a uniformtreatment maynot be possible.
tion (coIP) and government expenditure by function
should be developed where "environmental goods
(COFOG)
ansevc"aridtfe.
and
e services"
Accountingarefor
identified.
stocks of naturalresourcesshould be
done in physicalterms, not in monetaryterms; monetary
estimations,however,shouldbe attemptedfor changesin
the stock of resourceswithinthe period under consider-
ation.
* For exploitable,commercially salableresourcessuch
as minerals,a user cost approach,such as the one pro-
posed by El Serafy,shouldbe followed(Chapter3).
* For "permanent"resources,such as air, water, and
soil, two options should continueto be pursuedsimulta-
neously:treating restorationcosts as intermediateexpen-
dituresrather than as final ones, as is generallydone in
the SNA, and treating damagesto "permanent"resources
as consumptionof natural capital.
* Two (or possiblythree) NDPsshouldbe computed:
GDP - consumptionldepreciation of man-madecapital
= NDP1
NDPI - consumption/depreciationof natural capital
2
= NDP .
* Further research is necessaryto decidewhat deduc-
tions for defensiveexpenditurewouldbe appropriateto
produce NDP3.
These points of agreementwere communicatedto the
SNA expert meeting later held in Luxemburg,which is
deciedblw
describedbelow.
This was the last expert meetingof many held during
the past few years on the revisionof the SNAbefore an
actual completedraft of the new "blue book" is consid-
ered. Manyoutstandingissueshad to be decided.On the
issueof the environment,the meetingendorsedthe crea-
90 Ernst Lutz and Salah El Serafy
tion of environmentalsatelliteaccounts linked with the
SNA. It wasdecidedthat the new"blue book"willdescribe
these, and further research in this important area was
encouraged.In another context,as far as the concept of
incomeis concerned,the meetingadoptedthe Hicksian
definition(See Daly,Chapter2 and El Serafy,Chapter3),
witha minoraddition,as a basisfor definingthe coverage
and treatmentof capitalflows.Accordingly, incomeis the
maximuman individual(or nation)can consumeduringa
period while keeping intact the stock of capital at the
beginningof the period as well as any net additionsto
capital during the period. The explicitadoption of this
concept does not imply,however,that the central SNA
frameworkhas fullyadopted the concept of sustainable
income.
As a matter of fact:
* GDP will continueto be defined under the revised
sNA without ajtefWorking
SNA without adjustmentfor the degradationof natural
capitalfromthe processof productionand consumption.
* The use of nonrenewablesubsoilassets will not be
taken defning
nto acountin
alue dded.sulting
taken intoccoutinefinngvaueadedexpertmeetingand afterwardbe developedinto a hand-
An alternative sustainable income concept to GDP
thereforewillhaveto be elaboratedin a separate hand-
book on environmentalaccounting.
A descriptionof the current treatmentof nonreproduc-
ible assets, stocks, and flowsin the SNA and a proposal
for its modificationwas circulatedas a background(locu-
ment;it is reproducedas the annex to this chapter.The
proposalis very importantfrom an environmentalpoint
of view,but there wasno time to discussit at the meeting. changesin quantityand quality,and its linkageto mone-
Future Work
Inducingproduction and consumptionpractices with
less negativeeffectson the environmentworldwideis a
formidabletask. Changesare needed in policies, pro-
grams, projects, institutions,the distributionof income
and wealth,and human behavior.But necessarychanges
are often slowin coming.Improvementsin any one area
can make only a partial contributionto an approach
wheremanyparallel activitiesare needed to achieve'de-
velopmentthat is both equitable and environmentally
sound. Environmentalaccountingis only one of many
tools that are needed,and, giventhe lack of past effort,
it willtake time to developit into an effectivetocl to
improvepolicymaking.Nevertheless,seriouswork in this
area should nowbe pursuedactivelyand vigorously.
After several years with little visible progress, some
steps forwardhave been made recently in the fieldsteps fah bn blne
of
environmentaland resourceaccounting.But much niore
remainsto be done both at the conceptualas wellas the
practicallevels.
* Draft manual/handbook. The Bartelmus-vanTon-
geren "DraftFrameworkfor EnvironmentalSatelliteAc-
counting"needs to be developedinto a draft manualand
eventuallyinto a handbookin the handbookseries of the
SNA.
0 Evaluationof existingschemes.Currently,severalin-
dustrialcountries (Canada,France, Federal Republicof
Germany,Japan, Netherlands,Norway,and the United
States) have developedor are developingresource ac-
counting systems.Each of those approaches should be
evaluatedand anygeneral lessonsidentified.
* Pilot case studies in developingcountries.These pi-
lot projectswould: try out the Bartelmus-vanTongeren
frameworkin severalcountrieswith differentnatural and
environmentalresourceproblemsand perhaps in one or
two industrial countries with the most availabledata,
providelessonsfor integrationinto the draftmanual,and
determinethe policyimplicationsof the improvedincome
accounting.
groupnandhexpert
tedatmna esnmeetings.
novdiThe authors
h ioof
studies should meet to sevuentle
sutuinevsioouldm reviewthe draft manual. The
presenTedtoare-
revision would subsequently be presented to an
book.
* In-depthcase studieson developingcountries. Con-
trary to the pilot studies,the in-depthcase studieswould
involvenew data generation and longer-terminstitution
building.
* Physicalresourceaccounting.UNEP has alreadydone
considerable work in this area,but furthereffortis needed
to assessthe environmentand natural resource base, its
tary accounting.
It is foreseenthat these tasks, as wellas other possible
ones, would be undertaken cooperativelyby the World
Bank, the U.N.StatisticalOffice(UNSO), the U.N.Devel-
opment Programme(UNDP), and UNEP, along with inter-
ested foundations(includingthe Ford Foundationand the
NoyesFoundation),bilateraldonors,nongovernmentalor-
ganizations,and developing-country governments.
Annex. The Treatment of Nonreproducible Assets,
Stocks, and Flows under Present SNA and "M.60
Standards" for Reconciliation Accounts
M.60is an importantdocumentfromthe point of view
of environmentalanalysisbecauseit includesa framework
and supplementaryguidelines for developingresource
balances(stocksand flows)that are compatiblewith the
dSNA.Sincesok n developed
M.60was lw)ta much r optbewt h
later than the SNA,
however,it reflectssome advancesin the field that were
not availablewhenthe 1968 SNA was drafted.Therefore
there are severalincompatibilitiesbetween the SNA and
M.60 that should be addressed during the present SNA
revision.

The general rule in the SNA and M.60 is that changes
in nonreproducible assetsare recordedin the capital(flow)
accountswhena transaction(a purchaseor sale)occurs
and in the present M.60 reconciliationaccotnts when
other changesoccur in the asset balances.In additionto
price changes,the latter includechangesfrom newfinds,
growth,destruction,and depletionof assets (see Chapter
11). The valueof nonreproducibleassets in the balance
sheet and reconciliationaccountswhenthere is,no sale is
based on actualvaluesof similarassetsthat are sold(such
as for land) or on the sum of discountedfuture values
(suchas for timbertracts and mineraldeposits'.
The generalrules are not applieduniformlyto all assets
and are unclear for some. The actual treatment distin-
guishesamongland; historicalmonuments;tirnbertracts
and forests; plantationsand orchards;cultivatedagricul-
tural products; livestockfor dairy, breeding,and draft;
and livestockfor meat production.
SpecificApplicationsof the GeneralRule
* Naturalresourcesin the public domain are entirely
excluded.
* Land and historicalmonumentsare include(iin the
balancesheets only after an actual sale.This impliesthat
only entries for changesin these assets are reccrdedin
the M.60reconciliationaccountsif sales take place.
* For mineralresources,newfinds as wellas depletion
are recordedin the presentM.60 reconciliationaccounts.
It is not clear, however,what the exact coverageof the
newfindsis; does it includeall newfinds,does it include
only those mineraldepositsthat are actuallymined,or is
there some intermediatetypeof coveragefor newfinds?
- Natural growthminus depletion of timber, planta-
tions, orchards,and so forth is recordedin the reconcili-
ation accounts.There is no link between thi.streatment
in the balance sheets and production,whichis recorded
only whentimber is loggedand whenorchard and plan-
tation productionis harvested.
* The growthof livestockis included,on the one hand,
as an item in the M.60reconciliationaccountsi(together
with the other natural growthitemsmentionabove)and,
on the other hand, as either grossfixedcapitalformation
(for animalsfor dairy,breeding, draft, and so forth) or
changesin stocks(foranimalsformeat production).These
two treatmentsseem contradictory.
* Agriculturalproduction is not reflected in the bal-
ance sheets or reconciliation accounts and is not regis-
tered as changesin stocks during the growthprocess.It
is recorded as man-made production after it is harvested.
ProposedModifications
The following modifications would make the above
treatmentsmore consistent.
RecentDevelopments
andFutureWork 91
* Nonreproducibleassets should be included in the
balancesheets whenthey are marketedand their use in
human production activitiesgenerates product and in-
come. This impliesthat land should be includedwhenit
is used for agriculture,human settlement,and so forth
and that income should be reflected in GDP, GNP, or
national income, either as a result of marketed or im-
putedproduction,includingown-accountagriculturalout-
put and imputed dwellingsservices to owners. For the
samereason, historicalmonumentsshouldbe includedin
the balance sheets as soon as the assets generate identi-
fiable income.In other words, assets should not be in-
cludedwhentheyare marketedbut whenthey are actually
used for production and other activitiesthat produce
flowsregisteredelsewherein the SNA. It is importantto
establishthis link between income and production, on
the one hand, and the assets used in production,on the
other, to arriveat correct capital-outputratios and other
measuresof productivity.
* The natural growth of agriculturalproducts, prod-
ucts of plantationsand orchards, timber, and livestock
for dairy, breeding,and meat production should be re-
cordedas gross output at the momentof growthand not
at the moment of harvest (for products of agriculture,
plantations,and orchards),logging(for timber), or birth
(for livestock).During growth,the production shouldbe
reflected in changes in stocks, and, once allocatedto
uses,the stocks shouldbe reducedagain. This treatment
wouldeliminatefromthe present M.60 reconciliationac-
countsthe entriesfor naturalgrowthminusdepletionand
wouldchannelthese entriesthroughall changesin stocks,
exceptfor livestockgrowthused for dairy,breeding,and
similarpurposes,whichwouldbe allocatedto gross fixed
capital formation.This proposed treatment is an exten-
sion of the treatment for agriculturalproducts suggested
by the November1988 meeting in BuenosAires organ-
izedjointlyby the EconomicCommissionfor Latin Amer-
ica (ECLAC/CEPAL) and the MonetaryCenter for Latin
America(MCLA/CEMLA).That suggestionwasmadeto bet-
ter link gross output and cost in agriculturalproduction,
particularlyunderhigh inflation.Similarargumentswould
hold for all the other products of natural growth men-
tioned above.
References
Bartelmus,Peter, and Jan W. van Tongeren.1988. "Draft
Frameworkfor EnvironmentalSatelliteAccounting."U.N.
StatisticalOffice,NewYork.Processed.
UnitedNations.1968. "A Systemof NationalAccounts."Publi-
cationE.69.XVI1.3.
NewYork.
 Appendix
Participantsat thetNEP-WorldBan Workshops
The affiliationsgiven were those at the time the work-
shops wereheld.
First Workshop,Geneva,February23-25, 1983
Chairman:YusufJ. Ahmad,UNEP,Nairobi,Kenya
Contributors
Anthony Friend, Statistics Canada, Ottawa, Canada.
"Natural Resource Accounting:International Experi-
ence."
Per Garnasjordet,CentralBureauof Statistics,Dslo,Nor-
way."ResourceAccounts:The NorwegianExperience."
Roefie Hueting,Central Bureau of Statistics,Voorburg,
Netherlands."Use of EnvironmentalDatain Economic
Decisions:Netherlands."
HenryM. Peskin, Resourcesfor the Future, Washington,
D.C., U.S.A."NationalIncomeAccountsand the Envi-
ronment."
MichelPoitier, EnvironmentDirectorate, OECD, Paris,
France. "NationalResourceAccounts:OECD."
WalterSpofford,Resourcesfor the Future, Washington,
D.C.,U.S.A."Environmentand Development."
Jacques Theys,Ministryof Environment,Paris, France.
"NaturalResourceAccounts:France."
AtsushiYoshikawa,EnvironmentAgency,Tokyo,Japan.
"Macroeconomic Accountingfor the Environmentand
the Qualityof Life."
Other Participants
Partha Dasgupta,LondonSchool of Economics,United
Kingdom
RobertGoodland,WorldBank,WashingtonD.C.,U.S.A.
93
Second Workshop, Washington, D.C.,
November 5-7, 1984
Chairman:YusufJ. Ahmad,UNEP, Nairobi,Kenya
Rapporteurs:John Dixon, East-WestCenter, University
of Hawaii,Honolulu,U.S.A.;and RichardB. Norgaard,
JUniversity
of California,Berkeley,U.S.A.
Contributors
Robert Goodlandand GeorgeLedec,WorldBank, Wash-
ington,D.C.,U.S.A."NeoclassicalEconomicsand Prin-
ciplesof SustainableDevelopment."
Roefie Hueting,Central Bureau of Statistics, Voorburg,
Netherlands."EconomicAspectsof EnvironmentalAc-
counting."
RichardB. Norgaardand John Dixon,Universityof Cali-
fomia, Berkeley,and Universityof Hawaii, Honolulu
(respectively),U.S.A. "Project Design and Evaluation
UsingEconomicand CoevolutionaryCriteria."
Jacques Theys, Ministry of the Environment, Paris,
France."EnvironmentalAccountingand Its Use in De-
velopmentPolicy."
OtherParticipants
Ali Al-Damkhi,EnvironmentProtection Council,Safat,
Kuwait
Ali Al-Juwaiser,Saudi Fund for Development,Riyadh,
SaudiArabia
Lester Brown,WorldwatchInstitute, Washington,D.C.,
U.S.A.
DavidBurns,InternationalInstitutefor Environmentand
Development,London,UnitedKingdom
HermanE. Daly,LouisianaState University,
Baton Rouge,
U.S.A.
94 Participantsof the UNEP-WorldBank Workshops
RemediosDe Leon,NationalEconomicand Development
Authority,Manila,Philippines
RobertGoodland,WorldBank,Washington,D.C.,U.S.A.
Norman Holly,Resources for the Future, Washington,
D.C.,U.S.A.
D.C., U.S.A.
G. L. Kamukala,Ministryof Lands, Dar es Salaam,
Tanzania
Charles M. Kasaba, NationalEnvironmentand Human
SettlementSecretariat,Nairobi,Kenya
ArabindaKundu,WorldBank,Washington,D.C.,U.S.A.
JamesLee, WolakahnDavid Burns,Institutefor Environmentand Development,
JamesLee, WorldBank, Washington,D.C.,U.S.A.
RagnarOverby,WorldBank,Washington,D.C.,U.S.A.
A. Padmanabhan,Commissioner for RevenueAdministra-
tion, Madras,India
DavidPearce,UniversityCollege,London,UnitedKingdom
WilbernePersaud, Universityof the West Indies, K;ings- G. L. Kamukala,Ministryof Lands, Dar es Salaam,
ton, Jamaica
HenryM. Peskin, Resourcesfor the Future, Washington,
D.C.,U.S.A.
ThomasRobertson,consultant,Washington,D.C.,U.S.A.
WalterSpofford,Resourcesfor the Future, Washington,
D.C.,U.S.A.
Pedro Vivas,NationalPlanningInstitute,Lima,Penr
JeremyWarford,WorldBank,Washington,D.C.,U.S.A.
Ted Wolf,WorldwatchInstitute,Washington,D.C.,U.S.A.
Third Workshop,Paris, September30-
October 2, 1985
Co-chairmen:YusufJ. Ahmad,UNEP, Nairobi,Kenya,and
Robert Goodland, World Bank, Washington, D).C.,
U.S.A.
Rapporteur:Salah El Serafy,WorldBank, Washington,
D.C.,U.S.A.
Contributors
HermanE. Daly,LouisianaState University,
BatonRouge,
U.S.A."A ModestProposalfor a ModestStart in Mod-
estlyRelatingEnvironmentaland EconomicAccounts."
John Dixon(withRichardCarpenter,LouiseFallon,Paul
Sherman,and SupachitManopimoke),East-WestCen-
ter, Universityof Hawaii,U.S.A.."EconomicMeasure-
ment of EnvironmentalEffects."
HuibJansen and J. B. Vos,Free University,Amstercdam,
Netherlands."Indicatorsfor EnvironmentalProblems
in DevelopingCountries."
RichardB. Norgaard,Universityof California,Berkeley,
U.S.A. "LinkingEnvironmentaland NationalIncome
Accounts."
JacquesTheys,Ministryof the Environment,ParisFrance.
"EnvironmentalAccountsand DevelopmentPlanning:
Proposalsfor a Programmeof Action."
Other Participants
Francis Anokwu, Federal Ministryof Finance, Lagos,
Nigeria
Serge Antoine,Ministryof the Environment,Neuillysur-
Sene Frnc
Seme, France
IwanJayaAzis,Universityof Indonesia,Jakarta
London,England
AlisonA. Gilbert, Free University,Amsterdam,Nether-
lands
RoefieHueting,Central Bureau of Statistics, Voorburg,
Netherlands
Tanzania
Charles Karaba, NationalEnvironmentand Human Set-
tlementsSecretariat,Nairobi,Kenya
Arabinda A. Kundu, World Bank, Washington, D.C.,
U.S.A.
ArmandoLosada,Colombia.
KositPanpiemras,Officeof the NationalEconomicand
SocialDevelopmentBoard, Bangkok,Thailand
WilbernePersaud, Departmentof Economics,University
of the West Indies,Jamaica
A. Pisserides,EnvironmentalManagementUnit, Ministry
of Agriculture,Nicosia,Cyprus
MichelPoitier,OECD, Paris, France
Robert Repetto,WorldResourcesInstitute,Washington,
D.C.,U.S.A.
Juan Pablo Ruiz, EnvironmentOffice,Bogota, Colombia
Harlendos Santos,Ministryof UrbanDevelopment,Bras-
ilia,Brazil
SophieSchlingemann,UNEP, Nairobi,Kenya
Fourth Workshop,Washington,D.C.,
November 17-18, 1986
Chairman:Salah El Serafy, World Bank, Washington,
D.C.,U.S.A.
Rapporteur:DerekW.Blades,OECD, Paris, France
Contributors
Peter Bartelmus,UNSO, NewYork,U.S.A.."Environmen-
tal Accountingand the SNA."
Derek W. Blades,OECD, Paris, France. "Environmental
Statisticsand the NationalAccounts."
HermanE. Daly,LouisianaState University,BatonRouge,
La., U.S.A. "Towarda Measureof SustainableSocial
Net NationalProduct."
 Participantsof the UNEP-WorldBank Workshops 95

Salah El Serafy,WorldBank, Washington,D.C., U.S.A. J. L. Weber,Ministryof Research,Neuilly,France. "Les

"The Proper Calculationof Income from Depletable Comptesdu PatrimoineNaturel(PresentationSuccinte
NaturalResources." d'ExperienceFrancaise."
Anthony Friend, Statistics Canada, Ottawa, Canada.
"Natural ResourceAccounting:A Canadian Perspec- OtherParticipants
tive."
Roefie Hueting,Central Bureau of Statistics,Voorburg,C. Averos,OECD, Paris, France
Netherlands."Possibilitiesfor IncorporatingChanges Mr. Frias, Ministryof Economyand Finance, Madrid,
in the Environment,IncludingNatural Resources,in Spain
the NationalIncome." G.Guignabel,Ministryof Environment,Paris, France
ChristianLeipert,InternationalInstitutefor Environment
Other Participants and Society,Berlin,FederalRepublicof Germany
0. Lone,Ministryof Environment,Oslo,Norway
YusufJ. Ahmad,UNEP, Nairobi,Kenya Lorents Lorentsen,Ministryof Environment,Oslo, Nor-
RobertGoodland,WorldBank,Washington,D.C.,U.S.A. way
Anne Harrison,WorldBank,Washington,D.C.,U.S.A. J. Margat, Research Institute for Geologyand Mines,
RagnarOverby,WorldBank,Washington,D.C.,U.S.A. France
HenryM. Peskin,Resourcesfor the Future, Wlashington, N. Mensah,UNEP, Nairobi,Kenya
D.C.,U.S.A. G. Panella,Ministryof Environment,Rome,Italy
Robert Repetto,WorldResourcesInstitute,WAashington, Sophie Schlingemann,UNEP, Nairobi,Kenya
D.C.,U.S.A.
Sophie.SCh.ingemann,
U.ES,.Nairobi,AKenya Expert Meeting, Paris, November 21-22, 1988
Sophie Schlingemannl, UNEP, Nairobi, Keny'a

Chairman:Salah El Serafy, World Bank, Washington,

Fifth Workshop, Paris, November 20-21, 1986
Chairmen:JacquesTheys,Ministryof Environrnent,Paris;
and YusufJ. Ahmad,UNEP, Nairobi,Kenya
Rapporteur:Per Garnasjordet,Central Bureauof Statis-
tics, Oslo,Norway
Contributors
A. Gilbertand W.Hufkamp,Free University,Amsterdam,
Netherlands."NaturalResourceAccountingin a Multi-
ObjectiveContext."
Roefie Huetmng, Central Bureau of Statishics,Voorburg,
Netherlands,in cooperationwith N. Yunus,University
of Indonesia,Jakarta. "A Note on the Constructionof
an EnvironmentalIndicator in MonetaryTerms as a
Supplementto NationalIncomewith the Aid of Basic
EnvironmentalStatistics."
H. Jansenand B. Vos,Free University,Amsterdam,Neth-
erlands."IndicatorsforAssessmentsof SustainableDe-
velopment."
T. Loikkanen,Ministryof the Environmenil, Helsinki,Fin-
land."EnvironmentalAccountingDevelopmrents in Fin-
land."
I. M. Naredo,Madrid,Spain. "Caracteristiquesdes Ap-
proches EconomiquesNecessairepour Tenir Compte
des ResourcesNaturelles(Comparisonavecle Systeme
Standard)."
D.C.,U.S.A.
Rapporteur:DerekW.Blades,OECD, Paris, France
Participants
YusufJ. Ahmad,UNEP, Nairobi,Kenya
Peter Bartelmus,uNso,NewYork,U.S.A.
RameshChander,WorldBank,Washington,D.C.,U.S.A.
AnthonyFriend,StatisticsCanada,Ottawa,Canada
AnneHrisn, Stat, Ckbrdgt, U.K.
AnneHarrison,consultant Stockbridge,U.K.
Peter Hill,OECD, Paris, France
DavidJames,Institutefor EnvironmentalStudies,Amster-
dm h ehrad
dam,The Netherlands
Lorents Lorentsen,Central Bureau of Statistics, Oslo,
Norway
Ernst Lutz, WorldBank,Washington,D.C.,U.S.A.
Pablo Mandler,Central Bureau of Statistics,Jerusalem,
Israel
AnilMarkandya,UniversityCollege,London,U.K.
Brian Newson,StatisticalOfficeof the European Com-
munities,Luxemburg
GrahamPyatt, Universityof Warwick,Warwick,U.K.
SelangeSenaize,UNEP, Nairobi,Kenya
Carsten Stahmer, Federal StatisticalOffice,Wiesbaden,
FederalRepublicof Germany
JacquesTheys,Ministryof the Environment,Paris, France
J. L. Weber,Ministryof Research,Neuilly,France
 Index
Alfsen,Knut,5
Arens,Peter L., 6
Assets.See Capital
Ayres,R. U., 26, 65
Balancesheet accounts,40, 42, 44; guidelinesin S9A for, 81-82; for
naturalresources,24
Bartelmus,Peter, 6, 85, 86 nnl, 2, 89
Baumol,WilliamJ., 62
Bax,J., 37
Benerit-costanalysis,56; as methodto achievemonetsryvalue,62-63
Blades,Derek,2, 59
Bounded-knowledge synthesis,56-57
Bower,BlairT.,77 nl
Bye,Torstein,5
Canada:input-outputmodeluse in, 26; resourceaccc,untingin, 5
Capital:aggregationof physicaldata for environmental, 63; benefitsand
disbenefitsassociatedwith environmental,66, 67; consumptionof
fLxed,20; consumption of natural,6; countryapproachesto assessing
naturalresource,4547; depletionfactoras invesimentand, 13-14,
17; deteriorationof environmental,60; distinctionbetweenincome
and, 11, 17; environmental enhancementas format:ion of,22-23, 33;
environmentas formof social,66; gain and loss,68-69; in national
accounts,79; physicaldepredationof, 68; quantifyingenvironmental
stocksof,44; relationto income(Fisher)of, 66; seirviceflowscreated
by environmental,63, 66; substitutability of, 62; use of salesof, 13.
See also Capital,man-made;Depletion;Depreciation
Capital,man-made:accountingtreatmentbySNAfor,3; consumption of,
6, 20-22; depletion of, 9; expendituresfor, 20; measurementof
consumptionof,22
Centralaccount(stockaccount).See Environmentalaccounting;Natural
resourceaccounting
cEs (Conferenceof EuropeanStatisticians), 54, 80
Christiansen,Gregory,78 n12
Churchman,C. West,57-58 n5
Classification
of outlaysof industriesby purpose(coIIw), 83
Classification
of the functionsof government(coFo(,),UnitedNations:
data on pollutionmonitoringand abatement,28
cowp(Classificationof outdaysof industriesbypurpose),83
Conferenceof EuropeanStatisticians(CEs),54, 80
Consumption: of environmental capitalservices,63; of fixed,man-made,
97
and natural capital, 6, 20-22; sustainablelevel of, 11. See also
Depletion;Depreciation
Corden,W.M., 17 nl
Cost(s):best estimatesof,29; of defensiveexpenditures,33-38; experi-
ments to measureenvironmental,37; of extraction,17; need for
estimatesof environmentaland nonmarket,60; of pollutionabate-
ment,27-29, 31; shadow,60, 64 n3
CMted'lvoire:establishmentof soilaccountin, 50
Cremeans,J. E., 28-29
Daly,HermanE., 2, 3, 6, 20, 84, 85, 90
Data,environmental: internationalorganizationsas coordinatorsof,54-
55; as link to economicinformation,65; as sourcesfor developing
countriesof,47-49; as sourcesfor pollutionabatement,monitoring,
and controlcosts,28-29. See alsoStatistics,environmental
DataResources,Inc.,72
Defensiveexpenditures.- categoriesof, 9; for environmentalprotection,
20-22, 33-38; existingand proposedaccountingtreatmentfor,1-3.
See also Costs;Pollutionabatement;Sustainability
de Groot,R. S., 37
de Jouvenel,Bertrand,42
Denison,EdwardF., 72
Depletion:accountingin industrializedcountriesfor, 10; factoras cap-
ital investment,13-14, 17; failureto assess naturalresource and
environmentalcapitaland, 60; of renewableand nonrenewablenatu-
ral resources,3, 9. See also Capital;Usercost approach
Depreciation:accountingfor, 3-4, 12-13; additionsin national ac-
countsto ordinary,73; of capital,68; conceptin nonmarketaccount-
ing for,61; of nonrenewableresource,37
Deterioration,physical:as distinctfromeconomicdepreciation,60
Developingcountries:environmentalaccountingfor, 45-48; environ-
mentalinformationin, 4749; nationalaccountingmethodsin, 10;
nonremunerative sector of, 47; problemsof implementingenviron-
mentalaccountingin, 47-50, 60-63; proposalsto implementenvi-
ronmentalaccountingin, 49-52, 63-64; resourceaccountingfor,6;
SNAaccountingtreatmentof natural resource revenue, 10; use of
economicdepreciationin nonmarketaccountingin, 61; valuation
methodsin, 63
Disasterexpenditure:accountingtreatmentfor,23
Disbenefits:valuationof, 66, 67
Discountrate, 14, 16. See alsoInterestrates
Donovan,D. G.,37
98 Index

Dravie,ChristopheA., 50 IARIW-ECA (InternationalAssociation for Research on Income and

DutchDisease,10-11, 17 nl Wealth-Economic Commission for Africa),48
ICORS(incrementalcapital-output ratios),12
Economicand SocialCommission for Asia and the Pacific(ESCAP), 48 Income:fromcapital,66; creatingpermanentstreamof, 13-14; defini-
EconomicCouncilof Japan,37 tion for sustainable,24; from depletablemineralresources,12-13;
Economicdevelopment:preferencefor sustainable,55-56; standards distinctionbetweencapitaland, 11, 17; Hicks'sconceptof, 2, 8, 9,
for sustainable,35-37 11, 24, 90; Marshall'sdefinitionof, 67-68; sustainability
of, 8; true,
El Serafy,Salah,2, 4, 6, 18 n6, 20, 85, 89, 90 2, 4, 11, 13-14; use of conceptsof net and gross,13
Environment:benefitsand disbenefitsof servicesgeneratedby assets Income,national.See Grossdomesticproduct (GDP); Nationalincome
of,66-67; conceptualand measurementproblemsfor darmageto, 80; (NI)
effectof changesin qualityof, 72; estimateof effectsof policieson, Incomeaccounting,1
40; as form of capital, 23, 66; functionsof, 38 nl; methods to Incrementalcapital-outputratios(ICORS), 12
measurecomponentsof, 65; programsto reversedegradationof,22; Industrialaccounts,73
standardsettingfor functionsof,35-36; statistics,19, 80 Industrializedcountries:plannedenvironmental accountingin, 45
Environmentalaccounting:actions to start implementation of, 51-52; Information,environmental: bounded-knowledge synthesisas conceptto
countryplans for,5, 37, 45; for developingcountries,45-48; devel- combine,56-57; in developingcountries,47-49; and preference,34;
opmentof internationalstructurefor,52; in France,40-53; andlink for socialdecisionmaking, 55; systemin Francefor,40-42. See also
to SNA, 80; physicalunits in, 55; problemsin countryimplementation Preferences;Valuation
of,47-50, 60-63; proposalsfor developingcountriesto implement, Input-outputanalysis:environmentalservicesand damageas, 84; MEB
49-52; use for policymaking of, 49. See also Patrimonyaccounting; as extensionof,81; pollutionassessmentusing,26-28
Naturalresourceaccounting Institut Nationalde la Statistiqueet des Etudes gconomiques(INSEE),
Environmentalcapital.See Capital 81, 83
Environmentaldamage:accountingin modifiednationalaccountsfor, Interestrates: as expressionof value,55-56
73-76; correctvalue of actual,84; disasterexpenditureaccounting InternationalAssociationfor Researchon Incomeand Wealth-Eco-
for,23 nomicCommission for Africa(IARnW-ECA),48
Environmentalprotection.See Defensiveexpenditures InternationalUnionfor Conservationof Nature and NaturalResources
Environmentalquality:measurementin theoretical,modified,and con- (IucN), 79
ventionalnationalaccounts,76 Investment:use of grossand net conceptsin accountingfor,67-68. See
Environmentalstock.See Capital also Capital
ESCAP (Economic and SocialCommission for Asiaand the Pacific),48 IUCN (InternationalUnion for Conservationof Natureand NaturalRe-
Ewers,H.J., 38 n3 sources),79
Extraction,mineral:assessmentof ratesof, 4, 16; depletionaccounting
for, 12-14, 16 Japan:resourceaccountingin, 5
Jazsi,George,29
FDES (Frameworkfor the Developmentof EnvironmentStatistics),19, Johnson,HarryG.,66
80
Feyerabend,Paul, 57-58 n5 Kapp,K. W.,34, 38 n3, 54
Fisher,Irving,66 Keynes,John M., 7 nl, 17n2
Frameworkfor EnvironmentalSatellite Accounting:proposed hand- Kneese,A. V.,26, 31, 38 n3, 65, 77 nl
book for,90 Kuznets,Simon,32, 33, 34
Frameworkfor the Developmentof EnvironmentStatistics(IDES),19,
80 Landefeld,J. S., 60
France:environmentalsatelliteaccounting,83; resourceaccountingin, Leipert,Christian,2, 9, 33, 34, 83
5, 40-53 Leontief,Wassily,65
Lind,Robert C., 61
Grainger,Alan,37 Little,1.M. D.,60, 64 n3
Gross domesticproduct (GDP): classesof expendituresusecl for, 33; Lorentsen,Lorents,5
conceptsfor environmentalprotectionand use of,84-85; correction
for environmentalmeasurementof, 32-34; definitionand shortcom- M.60standards,90
ings of,1-2; macroeconomic analysisand, 13, 23; willingness-to-pay Macroeconomic analysis:use of grossand net productin, 13
methodto correct, 34-35. See also Nationalaccounts;Nationalin- Magrath,William,6
come(NO) Marglin,StephenA., 64 n6
Grossnationalproduct(GNP), 1, 8; effectof environmentalfactors on, Marketmechanism:naturalresourcecapitaland, 61
71-72; theoretical modificationsto, 75-76; in U.S. national ac- Marshall,Alfred,12, 67-68
counts,70, 71 Materials-energy balance(MEB): compatibility
withSNA,80-81
Measurement:correctionof GDP,32-34; experimentsfor environmental
Handbook:proposedfor integrationof input-outputstatistics,81; pro- cost,37; of marketvalueof abatement,30-31; of preference,34-35
posedfor satelliteaccounting,90 MEB. See Materials-energy balance
Harrison,Anne,2, 3, 6 Mirrlees,JamesA.,60, 64 n3
Hicks,John, 2, 8, 9, 11-13, 17 n2, 24, 79. See alsoIncome Mishan,E. J., 34
Hines,J. M.,60 Monetaryaccounting:in France,40
Hirsch,Fred,77 n9 Myrdal,Gunnar,49
Hotelling,Harold,12, 18 n4
HotellingRule,16 Nationalaccountingconcepts,70
Hueting,Roefie,3, 34, 37, 38 nl, 83 Nationalaccounts:correctionfor defensiveexpendituresin, 32-34, 37-
HufschmidtM.,57 nl 38; input-outputsystemsand, 81; measuringpollutionin, 26-31;
 Inidex 99

modificationof structurefor,69-76; naturalproductionaccountfor, Repetto,Robert,6, 63, 78 nI3, 81, 84

84; problemsand proposalsfor expanding,63-64; proposedtreat-
ment of depletionand extractionin, 12-16; role of householdpro-
ductivityin, 71; theoryto linkeconomicand environmentaldata in,
65; in UnitedStates,70
Nationalincome and product account: comparisonof modifiedand
conventional,73-76; of theoreticaleconomy,70-73, 76
Nationalincome(NI): in U.S.nationalaccounts,16, 70-72
Naturalcapital.See Capital
Naturalproductionaccount:as proposalfor nationalaccounts,84
Naturalresourceaccounting:developmentof central accountfor, 42,
44, 51; treatmentof capitalby SNA in, 3
Naturalresourceeconomies.See Developing countries
Natural resources:as assets with permanentincomest-eam, 13-14;
effectof treatmentas capitalassets of, 21; exploitableand perma-
nent, 20-22; in French patrimonialaccounting,44; importancein
developing countriesof,13, 46-47; measurementof capitalconsump-
tionof, 22-23; separationof capitaland incomefor depletable,17
Natural
See resource
Capital stock.
Naturalresourcestock. SeeCapital
Net domesticproduct (NDP):newconceptsfor environmentalprotection
and, 84-85 and, 84-85
benefit: in modifiedconventionalnatiDnalaccount,
Net environmental
73
Netherlands:proposalfor sustainableeconomicdevelopmentin, 37
Net nationalproduct (NNP): as sustainablesocialnet nationalproduct
(ssNNP),8; in U.S.nationalaccounts,70, 71
Net product:in macroeconomic analysis,13,23
NI. See Nationalincome
NNP. See Net nationalproduct
Nonmarketaccounting,1,55-56, 59-61
Nordhaus,William,79
Norgaard,RichardB., 5, 6
Norway:resourceaccountingin, 5, 81
Odum,E. P., 34
OECD (Organisationfor EconomicCo-operationand Development),28
Olson,Mancur,76
Organisationfor EconomicCo-operationand Development(OECD), 28
PACE (pollutionabatementand controlexpenditures),
28-29
Page,Talbot,61
Patrimony:descriptionof natural,42-43
Patrimonyaccounting:as French approachto resourceaccounting,5
40-53; as part of environmentalinformationsystem,40
Pearce,D.W.,28, 31
Peskin,HenryM.,2, 6, 30, 47, 57 nl, 63, 65, 78 n14, 84, 85
Peskin,Janice,30, 78 n14
Physicalaccounting,80-82. See also Naturalresources; Naturalre-
sourceaccounting
Physicaldepreciation,68
POLLEN model,40-41
Pollution:baselinefor measurementof,28; benefitsof abatementof, 2,
28-31, 66; disbenefitsassociatedwith,66-67; measurementof,2-3,
26-31; output,effects,and costsof,26-31
Pollutionabatementand controlexpenditures(PACE), 28-29
Popper,Karl,57-58 n5
Preferences:for sustainableeconomicdevelopment.35, 37-38, 55-56;
use of surveysto determine,34-35
Preventiveexpenditure.See Defensiveexpenditures
Price:determinationof naturalresource,12;shadow,32, 35, 60, 64 n3
Quasi-rent,12
Rehabilitation
expenditure:accountingtreatmentfor,22-23
Rent,4, 10, 12
Resourceaccounting:developmentof,4-5, 6;for flowsof materialsand
naturalresources,80-81. See also Environmentalaccounting;Natu-
ral resourceaccounting;Patrimonyaccounting
Resourcestocks,44
Ricardo,David:andconceptof rent, 18 n5
Robertson,Dennis,18 n7
Ross,M. S., 37
Ruggles,Nancy,77 n7
Ruggles,Richard,77 n7
Samuelson,Paul, 79
Satelson, al,79
Satelliteaccounts:alternativestatisticalsystemsas, 19, 24; in France,
40; informationon physicalunits in, 33; proposedenvironmental, 82,
89. Seeaso Frameworkfor EnvironmentalSatelliteAccounting
Saunders,C. T.,27
Schulz,Werner,38 n3
Seneca,J.d1.,30
~~Services,
environmental: benefits and disbenefits of, 66-67
Smith,Adam:distinctionbetweencapitaland incomeof, 11
SNA.See Systemof NationalAccounts,UnitedNations
Socialcost. See Capital;
~~~~~~~~~~~~~~Social
restoration Costs;Disbenefits;Pollution
cost, 83
Solow,Robert,12, 18 n4
SPIREmodel, 1
SPIREmodel,40-41
SSDS (Systemof Socialand Demographic Statistics),80
Statistics,environmental:methodsto improve,19; physicalstatistics
developmentfor, 80
StatistiskSentralbyra,81
StimulationProgrammefor Researchin HumidTropicalForest Lands
(Tropenbos), 37
Substitutability,
62
Sustainability:of consumption,11, 35; of development, 33, 37-38, 55-
56, 61; factorsof, 24-25; of incomeand output, 8, 11, 23-24. See
also Economicdevelopment;Income;Net national product (NNP);
Preference
Sustainablesocialnet nationalproduct(SsNNP),8
Systemof NationalAccounts(SNA),UnitedNations,1-2; balancesheets
and reconciliationaccountsin, 81-82; developmentof, 55; environ-
mentaspectscoveredby, 81-84: Frameworkfor Environmental Sat-
elliteAccountingand, 86; functionalclassifications for,83; inadequa-
cies and proposedrevisionsof, 19, 24-25, 88-89, 90; omissionof
environmentalaccountingin, 5; proposalto combineeconomicand
environmental informationin, 80; treatmentof man-madeandnatural
resourcesby,3, 10
Systemof Socialand DemographicStatistics(SSDS), 80
Taussig,M. K.,30
Theys,Jacques,5, 33, 81
Tinbergen,Jan, 34
Tobin,James,79
Ulrich,Werner,57-58 n5
UNEP (UnitedNationsEnvironmentProgramme), 49, 80
UnitedKingdomAtomicEnergyAuthority,27
UnitedNations,19, 48, 79, 80, 81, 82, 83, 88
UnitedNationsEnvironmentProgramme(UNEP), 49, 80
UnitedNationsStatisticalCommission, 80-81
UnitedNationsStatisticalOffice(uNso),80
UnitedStates:nationalaccountsin, 70; proposedadditionsto national
accountsof, 73-77; resourceaccountingin, 5; valuationof environ-
mentalfunctionsin, 63
U.S.Departmentof Commerce,77 n7; Bureau of EconomicAnalysis
data for pollutioncostsof,28, 29, 73
UNSO (UnitedNationsStatisticalOffice),80
User cost approach,3-4, 7 nl, 14, 16, 17; in irreplaceablenatural
resource,12. See also Depletionfactor;Nationalaccounts
100 Index

Valuation: and concept of willingness to pay, 62-63; concepts and
methods for environmental, 67-69; of consumption of natural capital,
22; of depreciation of nonrenewable resources, 37; of environmental
damage, 80; of nonmarket goods and services, 30, 55-56, 60; of
pollution abatement benefits, 30-31; as a reflection of social objec-
tives, 63; use of marginal, 67. See also Disbenefits; Nonmarket ac-
counting; Willingnessto pay
Valueaggregation, 55-56, 57. See also Environmental accounting; Non-
market accounting
Value depreciation: and income, 68-69
van Tongeren, Jan, 85, 89
Victor, Peter A., 26
Ward, Michael,4, 85
Welfare,economic: concept and measurement of, 1, 79-80: definition
of, 55; loss by certain groups of, 38; maximization of, 66; as under-
lying concept for environmental accounting, 65-66; using national
income to indicate, 32
Well-being:ability of national accounts to reflect, 70-72
Willingness to pay: to avoid disbenefits, 67; to correct national income
accounting, 34-35; to obtain valuation of nonmarket services, 62-
63. See also Information; Preferences; Valuation
World Bank, 49; recommendations for accounting for defensive expen-
ditures and depletionldegradation, 88; recommendations of Group
Meeting on environmental accounting, 86; and UNEP workshops, 80
World Commission on Environment and Development,79. 85-86

The complete backlist of World Bank publications is shown in the annual Index of
Publications, which is of value principally to libraries and institutional purchasers. The
latest edition is available free of charge from Publications Sales Unit, The World Bank,
1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World
Bank, 66, avenue d'Iena, 75116 Paris, France.
I
 UTHE WORLDBANK

That the pressures for development may cause long-term damage to the environment and the
natural resource base has become of increasing concern to environmentalists and economists
alike. A great deal of work is now being done to clarify the linkages between development and
the environment. Such work should make it possible to integrate concerns about environmental
and natural resource management more effectively into the process of economic analysis and
decisionmaking.
An important problem has been that current calculations of gross domestic product (GDP), the
most commonly used indicator of national economic performance, ignore the degradation of the
natural resource base and view the sale of nonrenewable resources entirely as income. The papers
in this volume were selected from a series of international conferences jointly organized by the
United Nations Enviromnent Programme (uNEP)and the World Bank during the past six years.
The authors are either economists, who argue that present methods for calculating national
income fail to reflect the consumption of natural capital and that present practices inflate income
from defensive environmental expenditures, or they are income statisticians, who offer ways to
adjust for shortcomings in systems of national accounting.
The combination of economic and statistical perspectives on enviromnental and resource
accounting provided in this volume is rare. Throughout, the environment is treated as natural
capital that needs to be accounted for and maintained. The authors both deal with the conceptual
side of accounting and propose a variety of methods to help produce better measurements of
income, which, in turn, could better guide economic policies in general and also encourage a
more sustainable use of natural resources.
Yusuf J. Ahmad is senior adviser to the executive director of the United Nations Environment
Programme, Salah El Serafy is economic adviser on the Economic Advisory Staff in the Office of
the Senior Vice President-Operations at the World Bank, and Ernst Lutz is a senior economist in
the Environment Department, also at the W`orldBank.

OTHER WORLD BANK PUBLICATIONS ON THE ENVIRONMENT

Environment and Development: Implementing the World Bank's New Policies

Development Committee of the World Bank and International Monetary Fund

Environmental Management and Economic Development

Gunter Schramm and Jeremyvarford, editors

Forest Industries Sector: An Operational Strategy for Developing Countries

Andrew J. Ewing and Raymond Chalk

Government Policies and Deforestation in Brazil's Amazon Region

Dennis J. Mahar

Techniques for Assessing Industrial Hazards: A Allanual

Technica, Inc.

Wastewater MVanagementfor Coastal Cities: The Ocean Disposal Option

Charles Gunnerson, editor

Wildlands: Their Protection and Management in Economic Development

George Ledec and Robert Goodland

ISBN 0-8213-1224-3