w '-I " 1' 'I I I _~~~~~~~~~~~~~~~~~ j 4,., TECONJJICSTiANDE Committee of lIaternafiona1 Developrnent Institutions on the Environment (CIDIE) !Ij~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Environmental Economics and Natural Resource Management i Developing Countries Edited by Mohan Munasinghe Compiled by Adelaida Schwab Committee of Intefational Development Institutions on the Environment (CIDIE) Disibuted for CIDIE by The World Bank Washington, D.C. This publication may be reproduced in whole or in part and in any forn for educational or nonprofit uses, without special permion finm the copyrg holder, provided achwwW of the soUce is made The Committe of Intenahl Dcvlpnlt nstutions on the Envronmen (ClDIE) wold appreciat receivig a copy of any publication that uses this publication as a souret The copy may be sent to ffth Enironment Departent, The World Bank, 1818 H Steet N.W, Washigton DC 20433, USA- No use of this publicaion may be made fior resale or other comiacial p witat the prior vwrtm permission of the b copyrihtholders. The designatios of geographical entities in this book and the presentation of the mateial, do not imply the expression of any opinion wasoever on the part of CIDIE or its member institions concering the legal sttus of any country, tertory, or area, or of its authrities, or concerning the delimion of its fixoiers or boundaie- The inte ions and conclusions in this report are dtose of the authors and do not necesry repesent te views of CIDIE or its member insiutiocs.- CopyrightO 1993 The Interational Bank fir Reconstruction and DevelopmentTe World Bank 1818 H Street, N.W. Washington, DC. 20433 USA ISBN 0-8213-2670-8 Committee of International Development Institutions on the Environment (CIDIE) CIDIE was established in 1980, and can- commission of the EuropeanCmunte sists of 17 multinational organizations active in. (CEC) the field of ecomomic development Thbe mem- *~f~Rcntuto e berhip of each such insfitituion is oomposed ofept EBD governments rathcr tha private. or nongovem- RD).IB mental oanzaftions. Erpa netetBn EB CIDIE. serves -as an inter-instxittional Food anad Agritmcltr Organzation of the mechanism which promotes dilgeamong UN (FAO) members, helps them pool knowledge and in- *~Bn 1B foriation, faicilitates discussion of comimon en- viroinentl isses ad idetifiationof *International Fund for Agricultural Devel- cooperative stratgies to address these issues, ~(FD and im-proves com ctoswrith other inter- *Nordic Investnent Bank (NIB) ested agencies;and organIZatIons. Organization of Amnerican States (OAS) The current membership includes: *uie ain eeomn rgam *African Development Bank (ADB) (UNDP) *Arab Bank for Ecornoic Deve3opment in *United Nations Environment Prowramme Africa (BAD]EA) (TJNEP) *Aian Developmenit Bank (AsDB) *World Bank Group (WiSj-) 0 Canibbean Development Bank (CDB) *World Food P?rogramnme (WFP) *Central American -Bank for Economic In- tegralion (CAB3EI) ----------------------------- ----------- SUO woUO) ploqsnoH JoqjIAJnS ceuuopsano :yw s,ouuV ZI- - SU.SSIa puile Olopoqliow aipjo uaowsossV Oil0 --------------------------- ------ - ..pni,s qIoJs6p SoIP 901--- - £pnus ojouonduosocj Lot ---- - - ---- - --------------------------o poq-a- - - yaysIpqXIyIf;N opvpyZ p1wi tausof umm-0q3F^) 'Xuw7qmq jJ .mqufXy £01 ... 4pniS asg_a V :uaqqpurj aql ui walisS a2iUJaMS 3lqnd ejo uopLsanOf aqjLV £6 - ---- - --------- suoxsooa n!w.IoA u! saonueuiauldmuwo3 2aiSarS nis :y%DouuV . hS . ...!.Iod olqnd pu arnlpej uotemoHV LU -------------------s--o----p u S slpf3J SuiipoS 98 -D!3JV uetexqE-qflsjo S0J011303 jepaOs amuoS -- -- ----------- --- spool nnpod se umpp!4 pue uopepeflap IuiuoDum AUB £8 ES- ~ ~ ~~ ~ ~ ~ ~ ~~ ~ ~ ~ ~ -~ - .. -... -.. -....... spoof 0oummsu! pue uoidumsuom se uaIpliUj 08 --------------- uoneonpa olmiJ pue jiuoa qf!l 6 .-- -.- -.- -.- - -.- - - - - -.&.gumosini Iewoos snraA ploqosnoq :sanijeuwi m noneindod LL ---- -- --- - -- ------------------------------------------- -ley6ol > 9L -------------------------- --------------------- wajlqold au -------- luiU s,I wu uuOwILua Uol 41pOOj puu I4IijqJq auwoul .-d'&vsa ZyXcfU £L ----'--'-'-''.'''''''' - ----'' ' -- - duqsnau:puj aA!ipnpoudwaj - c Se PloqasnUoH aLjL :ACUjpJa3 pUU P safI nfosagJ 'MAOJa a -c --*--- --------*--------------------- sa.PUassEwJ UOntEI eA tmuauiuousPwJu3O sazimnuum S :&yZ auuy Lt -------------------------------------------O---:------I-------------------PSm Ud IS Su'sf I :VnnMUUv ,,,.,,,,,, ,.......................... -..............-......................... -----SUOIsflfloJ wc ---------- ---- -- ----------- *--------------- - - - spedux IlIlUmaw iI0 JO lOnIeA sfl)pnis m5B3 Z ~ ~- - ---~~~--2- guI1UmOsIOp O -UO ioj S!(IeS U iuaq-isco jnuau'uoiptu= ST91 ~~ ------------------- ------*--~~-~~-~ ~-- - ------ l;U2tUUCll,LW a)i pue SD)nuOUoCX uguw i 2113I7 MIJ3i punP aqInwunjstyunliqopy Li .... 2DInIEnlUo!s3GF lu amdojpAaq U! u olenicA pug S31DmoU033 IUUaIUOmIIAU3 7Z Zl -------------------I------- --U--------------- ..........l..OUo0 OT~~~~~~ ~~ --- .-,-.,,.,-.,,, ,,,,-. ................... ....................................... ...... ... Ss e el i a l o 01 s!s{~~~~~~~~~~~~~~~~~riEue [IWUW0= 31Iu :gX-~~~~~ ~~ ~~~ ~ ' * -* ----* -- *--.. -.--.-.---.sisicleuce IeCuowvsuj L----------- -- '-- -- ------ -----I------- ------ smidweIM MCDrUI-4 * ~~~- .... ....................... -.--------------------------- --- -- ---.------------------ - t S ~ ~~~ ' - -- -~-- -~~----~~- ------- --- --- sis----ue -e---- £ - --- ---O------------------- -----------O------- .. .u LZ- u"!sp pofoid joj suonmp!isuo 1elUwu vzqV ..3ISSJJJJ -- -'''''--''~ spfJaSoi a Xuajpmdopwaaa aqnc uusnsu oj A£ojopoqpjg IsnpuddV -T AX .............................. sJolnqp)juoj 31(3 noqV ja] --+------------------------------- ----------------------- ---------- - Pnuu )AOMQlRo1 AIX--- sXua m upalAsomyz - Ixuo.oo ju 1X '''''''''--'''''''''''''''-''-''''''''' JOM3npOJ n --**ve--@-------<---SJU@$UOJ---z------- poeao ~vi 5. Environmental Economics and Natural Resource Management The African Development Bank Experience . .................................... 119 AndlMarkady and CrlosMnoz - Enronmental policy paper ................... 120 Integrating economic and environmental concerns ................................................... 123 Analysis of projects. recent experience .............................................................. 129 Conclusions on the review of projects 138 -Environmental issues in sectoral bank activities . 139 Condusions and recommendations .............. ...................................................... 139 6. The Rural Development and Environmental Protection Project in the Day Forest inl Djibouti: A Case Study ..... 143 Nessim J Ahmad The project ............4....4... ............................... .......-.: Economic analysis14 ERDnorlc nLdQis........... ..... ... ........................... ................. ..... ..... :......14B........ -4 Some conclusions . ------------ 166 7. Incorporating Environmental Cosbs into Power Development Planning: A Case Study of Sri Lanka ......_ 169 Peter Meier andMohMa Munosingke Backgrwoud ...... .----.--------...------...--.---170 Envionmetalissues Of powe sector developmntin Sri L anka17 ldrrvlnon ~~~~~~~m ert a slsoFsrretrdvi 1SnLrfb . ............................................. 176 Ouline of the proposed metbodology ............................................................... 180 Evaluating options . ............................................................ ....... . 185 Assumptions for the case study .1.9.2..................................... . 192 The analytical Mode -l.------.--.--.......-----------.. - .194 Technology and policy scenarios.----------------------------------199 Resuaos ----........ult.s ......................l.s...... -............. 19 9 Resis ~~~~~~~~~~~~~~~. ......................... ........................ ...... . ........ ...20 Candidate list and policy implications .......................................................... ..... 219 Towards a robust energy-environmental developent .......................................s . 225. S National Economic Cost of Soil Erosion in Zimbabwe ........................... 229 David Morse and Resnma Saigal mpact of Soil erosion on productivity -------------------------------.-.-.......----.-.....---.-- 230 Methodology of tle Study - -2.2.......................................................... 232 Main findings ofthe study .................................. ................... ........... ...... .... 233 Physical and country context resuts ............................................... ..... 236 Limitations of the stuly ------------------------------------------- 236 9. Environmental Management An Economic Proposal for Uruguay .242 Roberto Alfiedo Recalde Objectives of the proposal .............................................................................. 243 Problems identified ............ ..................................................................... 243 Analysis of causes .245 Proposed economic mneasures ........................................... 247 10. Integrated Environmental and Economic Accounting: A Case Study for Mexico ... 242 Jan van Tongeren, Stefan Schweinfestg Erns Lutz, Maria Gomez Lunea Francisco Guidlen Martin Geneal orientation of the pilot project ....................... 258 Environmental concerns addressed and mesured ........................................ 260 Accounting for envirnmUntal impacts on economic activities .............................263 Limitations of the study and follow-up .............................. .. 289 YE Il.issues and Options in linpleumenting the Montreal Protocol in IDeveloping Countries ............ .. .. .... . 291 Mo/han Munasinghe and Kenneth Kin - Balacing gw and environment nal dS. ....................................................... 292 sses in te plenatiooffte multiaralfim ............................................... 293 Ias and optons for nonal ozone polcies n deelopig countries . ....... 303 Concusion ... ........ 307 Appeudix List of Workshop Participants ....................... 309 Foreword The decade of the 1980s has witnessed a since its inception in l9SO, the Conmnittee fimancralchang in the, way governments of International Development Institutions on and development agencies think about -ihe envi- the Environment (CIDIE, now up to 17 mem- roniment and de-velopmentu The two are no bets has played a key role in coordinating longer regarded as mutually excusive- It jis work and fciiRtating information exhneon now recognized that a healthy environment is the flidl range of sustaiable developxmet is- essentia to sustainable development and a sues. This volume provides ample evidence of healthy ecoonomy. Moreover-, economists and prores made by -the gopin the area of envi- planners are beginning to recognize, that eco- ronuenatal economics-which is a vital preatq-, nomic develomaent which erodes natural capi- nisit to help incorporat niomna tal is ofikm not succesfuiL In tlct, development cnerns into developrmen esomkng. I strategies and programs which do not take ade- amu sure that the results of the workshop pres- quate account of the state of critical resour- ented here are a significant step forward in the, ces-forests, soils grasslands, freshwater, search for sustainable development options co2-stal areas and fisheries-ay degrade the and, therefbre, wfi be most: helpful to fth work resource base upon which funure growth is of aUl ClIDEE member instituirons, as well as depedent. othrs in the development community. Mobamed 7? EI-Asbry chief EvrnetlAdviser to the, Presdent and Director of Environment The World Bank Introduction -At the CIDIE Armual Mceting of April 29- methodology for sustainable developmcnt proj- May 1, 1991 in Washington, D.C., a statement o cts. Background information is disussed, as was issued on behalf of the heads of member or- well as the advantages and drawbacks of differ- ganizations which endorsed the principles of CI- ent kinds of analyses (technical, social, finan- DIE, called for a global partnership involving ial institutional, env al, and greater collaboration among CIDIE members economic)- and other o ons and urgcd CIDEmenm- Munasinghe and Lutz (World Bank) de- bers to promote an active collaborative scribe the role of cnvimental economics in prograL development decisionmaking. The paper briefly reviews the analytical background, including the CIDIE Confereuce on concepts of economic valuation of envinnen- Environmental Economics tal assets, as well as the practical teeboiques of -As a respose to this communique, CIDIE - deteining such values. A number of ca stu- launched a series of cperative activities, te dies that illustrate the application of various val- fit of which was the Conference on Environ- - teChliiCs in developing couties are mental Economics and Natural Resoure Man- . 'bed, and diecions for fbture work are agenent in Developing Countries, organized by smnzL the-World Bank in January 1992 The contents The special invited pap by Dasgua of this volume are based on thfe conference de- jicige Unirsity) examines the linge between family size and environmental degrada- - - slibeaions' (see Appand& fo;r fili list of imlszeadcoa paratiopants(o toon (for example, vanishing sources of water, Ths meeting proved to be most s, and reducing sources of fodder and household - not only as ameans of exchanging the latest in- . fuel) for poor rural households. There are differ- formation on environmental economics among ant motivations for parents to want many clii- members, but also stimulating ideas for new dre among the mo mon arge tiatives In view of the high pnronrty accorded to that chldren are de in themselves to cay the subject matter, and the need to emi on the family lineage and that they serve as se- the material as rapidly and widely as possible, cuity for old age. In poor counies, a dtird mo- the World Bank was requested to put together tivation may be that children are usefid as meonc-eingassets. They are needed on a this monograph on behalf of CIDIE. A second d m g b r sehol to rve Jf t - follow-up meeiing focused more on daily basis for dzec household to surv:ive. Ifthis follow-up metngi a fuse mges was third motivation is important, a different set of held in February 1993 in Washington, DXC government policies may be needed in order to manage population growth Outline of volume -The next four papers describe specific ap- The first three papers in this volume plications of economic valuation tcchniques in a - The fixs tbrec papers in this volume arc ailrfret broad in scopt covering topics ranging from en- vDarling, o ifne, and Nipoitsclcct vironmenLal econcnics, valuation and scope for Am B p t . - .... , . American Devc1opment Bank-) prescnt a case application (as a step towards sustainable devel- . - X ~~~~~~apmnd) to sovty oul?adyevrn of a public sewage system on a Caribbean 0pmenTt), to povCrty, popu1tion, island, with the objective of evaluating the eco- lt linkages . . - .nomnic tasibility of the constuction of such a The paper by Abaza (Umted Nations EnvI- t T moment Pro e e The mmtidolhy used i aanti - ~~ronmesm Prormme) examines thre appraisal-. xii valuation, with risk analysis (through the cal- fonncr southern Rhodesia in the 1950s and, culation of ranges for all -benefits). It is then 1960s. The impact of erosion is quantified, and possible to determine how sensitive the project the cosl of replacing lost nutrients is. calculatcd. is to different assumptions. and to estimate the The methodology used is subjecL to a number of cxpected costs of making the,%Tong decision. qualifications and assumptions, and the papcr The paper by Marklandya and Muiioz goes on to describe the limitations of the meth- (African Developmenl Bank-) addresses the ex- odology and the options for refning it. perience or the African Devclopment Bank- inr T'he paper. by Recalde. (Organization of envirornental econonmics and natural resource American States) provides an economic- analysis management The role of cost-benefit analysis is of the main national level environmental prob- examined, as well as different techniques for [ems of Uruguay and proposes a set of specific monetary valuation. Thereafter, they review the economic measures to deal with them Previous- environmental issues in a number of African ly, environmental resources in Uruguay had Development: Bank- projects and conclude by been handEed using a command and control ap- discussing means to expand the analysis of en- proach. The paper discusses an economic strate- vironmental aspects in project appraisals. gy for environmental mana t and proposes Alunad (Intenational Fund for Agricultural its use as a complement to existing methods. Development) carries out an economic valuation The causes of the problems are identified, and of environmental changes in the context of mar- economic measures. are proposed to incorporate ginal areas, with a case study of a project in the . evironmental objectives in the 'country's coo- day forest of Djiboutih Costs and benefits are nomic policy-- quantified and valued,: and riskl and uncertainty -The van Tongre, Schwveinfest, Lutz- are discussed. as well as. the fiscal impact and Luna, and Martin (UN Statistical Office and income distribution effects of the project The. World Bank;) paper contains a case study of i- paper concludes with a discussion of the impor- tegrated eironmental and economic account- tance of aended tine horizons, the applicabil- ing for Mexico. The national accounts are ity of -valuation- methods, fiscal obstacles to loan expanded to. include productive asset balances financing of environmentally focused projects,. Oil depletion; degradation of environmental as- equity concerns, and constraints on data avail- sets, deforestation, and land use are integrated ability for projects in marginal areas. into the national accounts. Valuation ap- Meier and Munasinghe (World Bank) pres- proaches are discussed, and the environmental ent a case study descnbing attempts -to incorpo- variables are then expressed in monetary terms. rate nvironmental considerations into energy The depletion and degradation caused by differ- decisionmaling in Sri Lanka. Enviromnental eMt sectors are identified. impacts of both thermal and hydro power proj- The paper by Munasinghe and King ex- ects are valued and input ito a power ystem amines issues that arise in the cae of a global planning model. Environmental impacts that environmental problem-ozone layer depletion cannot be valued are physically quanfified and A framework for efficienly eliminating ozone used in. a multicriteria analysis that trades off depleting substances is discussed Practical con- environmental issues against other concems straints that impede implementation of the Mon- (lik-e economic costs). The results provide a ba- treal Protocol are descnrbd- and methods of sis for improving energy policymaking in Sri overcoming them are presented. Lanka. The last four papers in the book go beyond Summary of principal issues and local project level to deal with enviromental findings at the conference economic issues that are sectorwide, national, or T p a t C Cnenc .... .. . T~~~~~~~~'he participants at the CWIE Conference global in scopae on Environmcntal Economics agreed that there Norse and. Saigal (Food and Agriculture Or- - - is a growing awareness of the need to incorpo- ganization) describe a method developed to as- r e c -- - . . . . 9 ~~~~~rate environmental conccrns mto projects. The sess the cost of SOIl. crosion In Zfinbabwe- The - - --. vital role of environmental economics m the de- methodology is applied to data collected in the i p h b accepted- -cisiormiaing process has been acceptd by Introduction -ij CIDIE member institutions. At the same time, Environmental valuation techniques were the majority of participants felt the need to work- considered a useful cxtension of standard cost- toward a consensus among CIDIE member in- benefit analysis (CBA), despite the dearth of ap- stitutions (and oLher relevant international in- plications in the developing world, and prob- stitutions) on the appropriatc application of lems of identifying enviromnental changcs economic instruments for environmental man- (particularly in the case of complex biological agement and protection. More specifically, systems). based on practical applications and the experi- It was felt that the linkages between envi- ence gained, economic methodologies best ronnental impact assessment, traditional CBA, suited to the valuation of environmental costs and environmental economics needed to be more and benefits in developing countries should be clearly defined. The difficulties of applying di- identified, and guidelines devised that will es- rect valuation methods (contingency valuation tablish a standardized methodological approach method, travel cost method) in developing coun- for the implementation of thesc techniques. A tries were discussed. However, participants were system of data gathering, updating, and analysis virtually unanimous in their belief that environ- should be introduced and established to facili- mental valuation tchniques should be immedi- tate exchange of information and ideas among ately incorporated into project analysis, while relevant institutions. recognizing the need for further development Developme:t agencies should ensur hdat and refinement of these techniques. It was sug- staff members are trained in environmental. gested that the multicriteria approach could be assessment techniques and applications of envi- adopted in cases .where environmental impacts ronmental economics in project analysis. Staff were hard to quantily members should be encouraged to utilize these Participants recognized the need for a techniques when sufficient data is availablc, fiamework for determining and prioritizmg in- Participats stressed the importance of fornation needs for developing countries in- identifying and evaluating the sectoral, econo- cluding the identification of the type of data mywide, and global impacts of development required, as well as the system of gathering, projects Projects should be examined within the storage, analysis, harmonization, and coordina- context of an integrated framework for sustain- tion of data used by different ministries and able development, in order to ensure adcquate government bodies. The framework should also rcognition and fimding, and to achieve quanti- help idwntify the type of technology to be used, tatively specified-environmental targets in a giv- as well as expertise and insttutional mecha- en sectoral and geographical area. Several 'nisms requied to collect and analyze data. participants called for more research into the in- Participants stesed the importance of in- tegration of environmental economics at the see- formation dissemination, both within CIDIE toral level. Furthermore, in order to provide member insfitutions and other international consistent guidance on structural adjustment po- agncies, and through public awareness cam- licies and macroeconomic policy reform, the en- paigns and technical assistane. As an inmedi- virorunntal impacts of such economywide ate follow-up, participants agreed to policies should be emined systematically exchange information on envi- Participants felt hat the causal link betwee ronmental economics,- especialy case studies of poverty and the environment merited further et- applications in developing countries. The World amination, and that the distributive aspects of a Bank was requested to: (a) prepare a mono- project were a key factor in the decisionmalcing graph containing selected papers presented at process. The importance of relating the achieve- the conference, as well as the principal findings; ment of environmental objectives to social and and (b) organize another conference on environ- economic objectives was highlighted. It- was mental economics in 1993, with the focus more suggested that a specific workshop to further on the link-s between the environment and poli- explore this issue would be beneficiaL- cies at the sectoral and macreconomic leveL Mohan Munasinghe July 1993 Acknowledgments The assistace' provided by Noremn Beg, Specia thanks are owed also to the confer- Annika Persson, and Luz Rivera, as well as ezie participants and pape authors for their Mazy Louise Hollowell, Rebecca Kaiy, -Timo- valuable contributions and cooperation. thy Kary, and Mahissa Ngo Van flue of Alpha- Qmcga Services., hic., in the preparation of this volume is gratefully acknowlecdged. About the Contributors ussein X Abaza is Assistant Policy Advisor Pet Meier is Consulant to the World Bank, to the Executve Director, United Na- Depar t, and Chief tiosns ironent P-grmn-' Economist, IDEA, Inc- -Jessim J. Abmad is a Resource Econamist in Mohab Munasinghe is Chit Environmental. the Technical Advisory Division, of Policy and Rsech Division, The the Intenational Fund for Agriukwal Word Bank- Developmet (IFAD). Carlos Muct, Director, E m Eco- Arthur ET Darling is Senior R Beam- nonics Dqeurent, Ministry of Social mist, Prect Advisors Office, Inter- Development, Mexico City, Mexico. American Development Bank Mario Nikditscbek is Economist, Sanitation Nartha Dasgua is Professor of Economics, and Urban Development Divisi, Cambridge University, and Chaiman, Inter-Amarican Development Bank- Beier Iernational Institute of Eco- David Norse is Senior Research Fellow, Envi- logical Economics, Stokolm ro-nementa! Change Unit, Uniesity of Christian Gomez is Advisor, Office of the Ex- Oxford, and Research Associae, OD, ecutive Vice-President, Inter-Amercan London He was fomy Senior Poi- Development Bank. cy and Pling Coordatm, Agricul- Keneth King is Senior E i Sp- ture Deparnnent, FAO. calist in the Envronment Dearent Roberto Alfredo RecIde is a Developmat of the WoddBak: Ecnmsin ihe Depanmen of Re- Maria Gomez Lisa is Director of te Naion- gion Development and Enviomnt, al Account and Economic Statics The Organiation of American States. Depatment of the Nstu Nconal Reshma Saigl is Conutant, Agricultu De- deE ca Geogafica y nrna- parti, FAO. ca (IEGI) in Mexico. Stefan Schweinfest is Statstcian in te Na- Ernst Lulz is Senior Economist in the Emn- tional Accounts Section of UNSO. ronmena Policy and Research Divi- Jam Tongeren is ChiefoftheNatial Ac- sian of heWold Bamk. couts and Classiions Branch in AMi Markadya, Harvard Institlte for In- the Uitd Nation Statistcal:Offine terna Devlopment, Harvard Uni- (UNSO). versity, Cambridge, Mass. USA. Francisco Gilkln Marti is Sub-Diector of the National Accounts and Economic Statistics Deparm t Of lNEGL 1. - Appraisal Methodology for Sustainable Development Projects Hussein Abaza There has been increased awareness of the need to incorporate environmental con- siderations in -the appraisal methodologies ofprojects This paper reviws tke various analy ses, including informatiog. technical, economics, fiancial, socid4 istilutonal and environmental analyses, which have to be made at the various stages of the project cycle in order to ensure tha projects implemented are environmental7y sound and sustainable- It highlights elements of environmental concerns which need to be consideredfor possible in- tegration in the project cycle adopted by internaional development assistance organizations. and governmental and non-governmental institutions in order to ensure the development and implementation o,f environmentally sound projecs. 2 Different terms have been used to describe tance arganizations and by governmental and the various stages of the project cycle. Accord- nongovernmental institutions. ing to the World Bank (Baum and Tolbert It is intcnded that this documcnt be n- 1987), the project cycle falls into the following vised and refined on a continuous basis as main categories: more experience and knowledge in the analy- Identication During this phase the proj- sis of development and execution of projects ect ideas which represent a high priority a use of thre country's ursoes to achieve - - The emphasis given to the vanous stages - development need;s are identified mhese of the project cycle and to the various analyses project ideas should meet a preliminay test will depend to a very large extent on the na- of .basibility. This involves ensuring that tae and type of project m question and the of feasiblit hl v lvc cs .n Thatmole eg - - . - zw~~~~~~~~~peailing condifians and. stat3 of develop- institutional and technical solutions, withln casts comenurtewihheement in which the project is being implem- costs commcnsurate with thie expected- ented, as well as to the mandate of the benefits, will be sought and appropriate it iitmatianal development instituton support- policies adoptedL m supor big thie project * Preparation- Once a project passes the identification phase, a refinement of the General Considerations For project in all its dimesions-technikal, Project Design - conomic, nionn fi=nanL socia, .CO .OIIIIC, environmental, financil, social, Each project should be designed using a institutional-has to be mau&le. framework to ensue a logical relationship and * Appraisal This phase involves the asses- hierrchy between th projects activities, out- sment of the overall soundness of the proj- puts, and inputs. Project design should clearly a and justification for its implementation. identify the following: Implenenwtaon This stage covres the ac- * Problem(s) or issues) the project intends tMal development of the project until it be- to address comes operational, including monitoring Long and shortm objectves of the and supervision of project activities. Loject Evahatior An ex post evaluaon ofte the Targ ps0rmproje cire pxoject is intended to determine the extent to which the project's objectives have been Acves eto achievetie objectves realized and to draw lessons from the epe- * Fmancial and human resources required to -nence gained in dte implementaion of the achieve the set objectives project fbr use in fiture projects Ma. gment responsibility and strutarem Throughout the various stages of the proj- - p. mat - : * linplementaton~~~~~modltes ect cycle several analyses have to bre made- The project should be designed to es including infomaton, teical socal, fina- - . . . - . - . . ~~~~~~replicability widiin the boulndaries ofthe coun- cial, institutional, economic, and enviromen- :aL This document attepts to 1ig1il tbese --etry or egion in which it is being implemented analyses and the ideratinsDbtand in countries with similar socioeconomic analyses*and the considerations to be taken anci environmental conditions. Repliation into account in order to ensure hat projects implemented are environmentally sound and could be in one of the folowing fonns: sustaiable. It provides elements of environ- thnology, admini ve setp ( mental conerns that need tobe considered fior ment), processes and styles, or expenence. their possible imegraton ino the project cycle The projec should be designed to allow adopted by international development assis- -fleibilitY in the modification of its acivities to anain its set objectives and in the modification L Te -author is most graeful to his colleagues at the United Nations Enviromental Programme UNE) -whohave provided comments and contibutions at the vaious stages of the prepartion of tis paper hav p-ie comet an.ot ofti 1. Appraisal Met hdologyforSustaiuable Development Pro jets 3 of those objectives to reflect the changes in tpes of inflrnation listed below will be more govermnent policies and in the priorities of the relevant depending on the nature of partcular government and the local population (Harrison projects. 1987)- Flexibility is also recquired to allow for The analysis of background information is delays in the implementation of project activi- -ifntded, among other things, to: ties that could be attributed to one or a number * Providc a preliminary understanding of the of the following; (a) withdrawal of one or more prorities given by the goveniment to envi- of the project collaborators; (b) delay in the de- r entl concems. livery of imported components of equipment * E and material required for the project; (c) delay in the delivery of local inputs by the govem - pItvous activities, and ongomg and mat implementing agency; shortage of o-planned activities being undertaken in the ment implementing agenC-r, (d) shortag of lo-- cal -material and equiment and spare ps country concered, whether by the gover- calmaeral ndcqipmntan spre prt ment, nongove nmentnal orgmlzadons needed for the implementation of project activi- ties; (e) change in government leadership or (NOOs),UniedNations (UN) agencies, or strucrs; and Iocal commotions and donorcountes. disubances. * Show the relation of prposed projects to The project should also be flexible enugh the broad national development, its replica- to alow for adjustments and changes in its de- bility (similuar cras, simiar situations). sign to meet changing rquirements or flaws in Projects should provide experience that can its design The project design should provide beused in abroadercontext. for suspension if it becomes apparent that un- * Ensure complementarity of activities and fomseen negative impacts and envirom avoid dupliation. damage are resulting from project * Enable the preliminary identificaton ofpo- implementation- tential project beneficiaries and A mechanism should be built into the proj- collaborators. ect for the recoding of the experience gained * d durmg the process of implemention and re- sults and outputs obtained. Provisions for the OlIaioIS periodic review and evaluation of project acti- Informaton gathered should include t vities involving all project collaborators should - - n be conducted at least once a year or at shorter *Data on the country in question, including intervals as deemed necessary. The findings of demographic stucture, economic situaton the project reviews will be reflected in adjust- of the country, national development plan, merts and modifications in project design. The debt stucture, and political system project should also include provisions for ex adopted. post evaluation in order to learn more about its * General environmt suation of the impact and the extent of its susinbility (UN country, ginug climatic conditions, to- Joint Inspection Unit 1989). pography, the natal resoure base, the . -dependence of development on this base, Information altlYsEss- its degrdation and depledon (for example, Preliminar backgronue informatin .land degradation and soil erosion, defores- rreliminary biftroun. information tation), major environmental and resource Relevant information on the country in constaints on development, environment- which the project will be implemented should related activities being implemented, and be gathered, to assist in the identification as existing govemment institutional strc well as the design of the project Exing re(s) dealing with the environment sources (such as national environmental pro * e pe a a uea files an oselated projects ard acvatintes und les and- conservation strategis) should be or being implemented by international de- consulted to minimize the amount of time and velopment assistance organizations, UN effort devoted to gathering infornation. Certain agencies, and govermental and .~~~~~~~~~~~~gnis an.Dennna n 1. A pEra Mt hod forSn Del pwet Projets .S * Priority given by thc govcnment to the available for use mprojects implemented in do- various rgions. velopingcountrie Extent of security and social stability of For iany developmg counties, an ap dhe different mgions - - - propriate technology would bc one that makes th-ftr s full use of labor while minimizng the use of Timing capitla and advanced skills. In the selection of Consideration of the duration and the most the right technology, therefor inappropriate appropriate time to start a project bas not al- govemment policies such as overpricng of la- ways been given enough attention. Determining bor, underpricing of capital though subsidied the right time to start a project, given the avail- s - tes, and undervaluing of exchange ability of the roquired financial resources, de- r should be taken into acwount otherwise pends on the demand for a projeces ourpns, the project might lead to the adoption of ec- the availability of the right technology, the ex- nomkay inefficient technologies (Baum and isting stare of the enviromental and n Tolbert 1987). Better still would be the pursi sources, and the extent of the impaOs that a of ma omic policies conducive to te project is likely to have on them It is some- adoption of a appropriate technology.- times too late to implement a specfic project The project should allow for the proper that aims at recffying a ceain envmronmenl m - n of equipment and for onsite tra- issue- At other times, such a project may need ing for fteir use and maintenane it is there- to be redesigned for a different lanch dat fre oflen encouraged to use local equipmcnt (Baum and Tolbert 1987). with which the local popation will be more The project shoud also be realistic in the familiar, and local skills that would be avail- planning and schdung of activite. A reason- able to provide training for personneL. able take-off or lead period at the beginming of A major objective to be puraud in the a project should be provided for, during which selection of technology is that of overning all the essential elements would be assembled shortages of human and institutional resources. before project activies move into full swing In addition, governments sbould particularly (UN Joint lnspection Unit 1989) - focus on eahancing public awareness ofthe use of apprpriap technologies, encouuging indus- Technoiogy . . . try and the prvate sector to devote sufficient Apprpriate technologies should be used in resources to tcnological research for the de the implementation of project activities. Such velopment and inDduction of sound technolo- technologies normally would be descried as gies (Baum and Tolbert 197 suiting local conditions and the available skills- and taining and being within the fancibl Soci an*s means of the population. The use of traditional Social analysis of the project shod begin technologies and material would reduce the de- at the outset of project idntfication and con- pendence of.a project on outside technologies tiue though the vmaious stages of the project that are usualy expensive and require hard cycle- Projects that ignore the traditions, val- curencies in short supply in:developing coun- ues, and social organiztion of the inended ties; destruc traditional technologies, how- benefciaries have little chmnce of success ever, should be replaced by -envirometlly (Baum and Tolbert 1987). sound altmnatives. Idealy a combination of - prcpa- both modern and- tradiona tcologies Publi partiipation should be used, wL-ch leads tO updating of the The various levels of authority in the gov- former by the latter. ernnment-the cental goverment authrity, Only fully tested new technologies that are and provic and local authorities-es well as le to the local popuation and that are r e preseNaIVes of the local populaons, in par- likely to function in the project area should be ticular, the projec beneficie, should be in- introduced. Moreover, the tansfer of environ- volved in all phlses of project formulation and mentally sound tecmologies should be made. design, and in the various stages of the project 1. Appraisal Mfethodoog for SainaNe Development Projects S * Priority given by the govcnmnent to the available for use in project implemented in de- various regions. veloping countries. Extent of security and.social stability of For many developing countries, an ap the different regions. propriate technology would be one that makes full use of labor while minimizing the use of Tlming capital and advanced skills. In the selection of Consideration of the duration and the most the right technology, therefore, inappropriate appropriate time to start a project has.not al- govrnmment policies such as overpricing of la- ways been given enough attention. Determining bor, underpricing of capiLtal through subsidized the right time to start a project, given the avail- hiterest mtes, and undervaluing of exchange ability of the required financial resources, de- mates should be taken into account; otherwise pends on the demand for a project's outputs, the project might lead to the adoption of eco- the availability of the right technology, the ex- noTmcally inefficient technologies (Baurm and isting stae ofthe ernvironmental and natur Tolbert 1987. Better still would be the pursuit sources, and the extent of the impacts that a of macreconomic policies conducive to the project is likely to have on them. It is some- adoption of a appropriate technology. times too late to implement a specific project The project should allow for the proper that aims at rectifying a cermain enviro nml maitm ance of equipment and for onsite tin- is:c At other times, such a project may need mig for their use and maintenance. It is there- to be redesigned for a 'different launch datE fore often encouraged to use local equipment (Baum and Tolbert 1984 . with which the local populMion will be more The project should also be realistic in the familiar, and local skills that would be avail- planning and scheduling of acfivities. A reason- able to provide taining for personneL able take-off or lead period at the beginning of A major objective to be pursued in the a project should be provided for, during which selection of technology is that of overcoming all the essential elements would be assembled shortages of human and institutional resources. before project actvities move into full swing In addition, government should parficularly (UN Joint Inspection Unit 1989)- fbcus on enhancing public aware of the use Techno-ogy of appropriate techbnologies, encouraging indus- TechnoIogy -try and the private sector to devote sufficient Appropriate technologies should be used in resources to technological research for the de- the implementtion of project activities. Such velopment and introduction of sound technolo- technologies normally would be described as gies (Baum and Tolbert 1987). suiting local conditions and the available skills and traning, and being within the financil Social anal means of the populatioa The use of traditional Social analysis of the project should begin technologies and material would reduce the de- at the outset of project identification and con- pendence of.a project on outside technologies tinue through the various stages of the project that are usually expensive and require hard cycle. Projects that ignore the adtions, val- currencies in short supply in developing coun- ues, and social oranzation of the intended tries; destructive tradiional technologies, how- beneficiaries have little chance of success ever, should be replaced by environmentally (Baum and Tolbert 1987). sound alternatives. Ideally a combination of both modem and traditional technologies should be used, wh-ch leads to updating of the The various levels of auhority in the gov- former by the latter. enmnent-the central govemment authority, Only fully tested new technologies fhat are and provincial and local authorities-as weli as acceptable to the local population and that are representatives of the local populations, in par- likely to fimction in the project area should be ticular, the project beneficiaries, should be in- introduced. Moreover, the transfer of environ- volved in all phases of project formulation and mentally sound technologies should be made design, and in the various stages of the project cyce. This involvemnct would ensure that the land tenure system, water, forcst, and needs and problems of the population con- pastu resources rights cemed are addressed and that their support and. * Gender roles in the local society, particu- participation in project activities are secured. It My with rcspect to rcsource management. should be bome in mind, howevcr, that there wr wil be ustances when comm-uty .rn * Basic mffastructure available in the tions will pursue narrow interests that would area-for example, rads, schools, hospi- be of much value in this process I tals or clinics, and cooperative societies, not be of much valuse in this process (13amm andTot bert 1987). - and the extent to which the population is and-Tolbr 1987). - .satisfied with available ficilities and ser- The prject design should ensure an m-iv creased role for the local population in project management and a gradual takeover once ex- Sociocultum-l and demographic character- ternal project collaborators pull out Through istics of the project population including this process it is also anticipated that the capa- ethnic, t-bal, and class composition bilities of the grassroots population in organiz- * Environmental and development problems ing themselves for self-reliance will be - cing the population and how they foresee improved- the solution to them Partiipation of the local population not - - * ~~~~~~~~~~Epidemiological and demographic- fatZrs only minimizes opposition, mobilizes support, that are kely to affct thc potential par- and increases the impact and susmambility of - ~~~~~~~~~~~~~~~~ticipation ofthe population projects, but it is also a means to gather rele- t infonnasion fr ploect dsign alcin *-.Effective local-groups in the area, their or- vant information for. project design- Takidnggn*oaseu,adtysofcivis into consideration the social aspects in project in al tey are ind. desig,n is needed -iot only to adapt project de- . Whihtheyare invol sign to the tadiLons, and social and culturl te projc wshold beousedainfesOmu- norms of local populations, but also to orga- laig the proectfaewo modit o nize and mobilize people to foster the process implemetaion, and a segy to secure the of social cage (Baum and Tolbert 1987). oment of the project populai Contact with the concerned population tEroughout the project cycle- It will also assist should be organized by the local authorities in the selecton of technologies to be adopted through the central govermment or government in the prject implemention (icluding new adimnovative technologies). implementing instiion(s). The various seg- * ~~~The project should include a public awarc ments of the local population should be con- T p s i taed, to ensure that the interests of the ness component aimed at sensiizing the pOP- - diffierent social, Economic, and religious groups . ulation to the positive effects of applying the environmentally sound approaches and are accounted for.. Communication with the lo- tcnlge,a ela h ermna fet cal population in their local language should be technologics, as well as the deLneetal effects made possible. - on the environment, the populaon, and the a economy of using inappropnate technologies - - ~Contact with the locat popuon is m- tended to establish the following: Project beneficiaries A Main economic activities ofthe population In any project, the beneficiaries must be * Mjorecological zonesintheprect e clearly identified. While the cost of environ- * Size of the population in the project area mental management projects, such as the set- ting up of a national wildlife conservation area, and project beneficiares or tret groups may fall on the local community, benefits * Stucture ofthe local community and exist- might accrue more to fiture generations and to mg leadership and extent of its influence on the mterational community at large. Mecha- the population .isms for defraying the costs on the local com- * Local systmsn goveming the access to and munily must therefore be worked out control of naural resources-for example, The project should aim, where possible, to beneft the poorest segments of the population L. Appr~ Mcd hedology for Snuoiuable Development Projects 7 (rimberlake 1985). Aside from the issue of in- a project should determine the ideal recovery tragenerational equity, there is also an environ- policy. This is normally a policy that secures mentl argument for targeting the poorest of maximum financial benefits to the economy the poor. The poorest segments of the popula- from a project, uking into account its impact tion are often trapped in a vicious cycle of pov- on the distnrbution of income and on the gov- erty and environmental degradation. Helping emient fiscal policy (Baum and Tolbert them to break this cycle and lengthen their time 1987). horizon beyond a day-to-day existence may The financial profitability of a project can significantly ease the pressure on the resource be measured by the internal financial nre of re- base. r. This is the discount rate that equlizes the In identifying projea beneficiaries, gender present value of the steams of financial costs rmles have often been neglected. Women's ac- and benefits over the life of a project The dif- cess to local resoutces (both.natual and other- ference between the fincial and the economic wise) is frequently markedly different from mate of retur is that the latter adjusts various men's. This information must be incorporated costs and prices to eliminate the impact of into project design from the beginningto ensure trsfer payments such as subsidies, duties, tat the effects of projects, particularly on re- taxes, and other distortions in the valuation of source use, are those intended. The role of traded and nontraded goods by using shadow women, as distinct frm that of men, needs to prices that reflect their true vale to the econo- be clearly idenified n project implementation. my The main difference between the financial. The project should be designed to lead to and economic analysis is that while the fonner an equitable distribution of benefits among and deals with the costs and benefits from the point wihin target groups and not to exacerbate any of view of the individual enteprise, the latter nequities among local popuations. does so from the point of view of the economy Training , , -- --.as a whole (Baum and Tolbert 1987). The following financial consideraions Since no project, progm, or plan can be should be taken into account in project desigr successflully implemented without sklled and - aTe ject shouw include the -traned manpower, he project should provide - contribution of the local population- This -fior ~an aDprp---t ewIvromnent tlaminig com- forWan ppropriateM=ctraining cor- could be in the form of land to be provided ponent The training of a locally available, fortheprojectsiteoractivities,oflaborto silled worlcforce will -help ensure project s- udrtae some ofte pr activities, or - - un~~~~~dcrae some of the pmject aedvkies, or tainability and contribute to capacity buildin&o m - of. material and equipment This contribtlo The project should be designed to allow for tion wi ensur that the proect reflects e the training of the local population in m - and needs of ocal pop ions and duced techniques whether these relate to the win secure their commitment to the project implementation of project activiies, to the - The project should clearly demonstat that management and monitoring of project ivi- . it will resulk in finacial benefits to 1he tes and their evaluaton, or to the operion population. For certain types of projects, it - and mainenance of o ational systems es- . m;y also be necessaiy to consider the abil- tablished through projects Where possible, this y ofthe local populaion t partcipae i, tmining can be uftgmted with technical train- and contute to, the project aum and imn Training provided should be paut of or Tol 1987) inkled to, training programs in the country in question. * The project should be designed to ensure -tat its activities and achievements will F inancial analysis continue after financial support from do- An intgl part of every project is setting . n ad poKn agecies ceases. a fincial plan that ensures the availability of Funds required to cover the continued nm- ning and operating cost of a project should fimds for completing as well as operting and. -: maintaii a project. The financial analysis of be made available whether through income generated by project activities or through 8 community fiud-raising initiatives. Project period of five to ten years, careful consider- design should clearly identify who will ation and thought should be given in project cover thesc costs. design to all possible collaborators and benefi- * Project costs should be kept to a minimum, ciaries. These include: (a) the ministry of en- s s s r L ~~~~~~vironment or grovcrrunent agency responsible particularly those related to fbrgone bne-t gen fits. This str y wfor the environment; (b) the ministry of plan- financin of the projcitas twesl ' ning or finance; (c) other relevant ministries or financing of the projcct as well as ItS - -..-. -. - .................. -govemment intitutions expected to be in- adaptability and replicability (Harrison volved in a particular project; (d) donor coun- try or agency providing funding for te project * The use of imported components in proj- i question (e) United Nations Developmen ects should be avoided or kept to mini- Programme (UNDP) office in the counuty; ({) mum-whethcr to financial resources, relevant UN agencies represented in the coun- labor, management, skills and expenence; try, (g) relevant NGOs. or equipment and material (Timberlake. Where possible, a preliminary field visit 1985). should be undertaken to expand on conacts with possible collaborators and beneficziais - Institutional analysis lOTis visit isiended Eo complement theinitial The institutional dimension in project backgrund infonmation (see Information analysis has until recently been receiving less analysis section) by providing more detailed atention from intemational development as- informationtrgh diret contacts and prima- sismce institutions than the technicaL eco- ry sources. The field visit will mvolve con- nmi, and financial ones. It has now become tacts with all die agencies and institutions evident that the availabhity of the necessary listed m the preceding paragraph. The visit is intitutional structures and arrangements is a . itendedto asermaiidenfifyd primary means of ensuring the efficient imple- * Main govemment implemeing agency for -mzentafin of stambedevelopment projects-- -- -- the project and various collaborating mm- 'Bmim and Tolbert 1987) -- istries or departments Indequate institutional arrangements in * - - ' tE @ ~~~~~~~Envimo}menltal goods anldservices pmvided developing countries are characterized by a se- a - - - - > . oor~~~~~~~~~~~b the project area and the lfcely impact of rious shortage of skilled manpower, and expe- the project ae a e l rienced staff, a lack of trainng and staff theprojectonthem incentives, overloaded facilities, low salaries, Major economic and enviromnmental prob- and counterproductive government policies lems ficing the country as identified by the and legislation (Baum and Tolbert 1987). govermment Govemment policies have their repercus- - - p a -oh - - * (i~~~~~~~~~~~overnmentpriorities and policies sions on institutional anrangements necessary Region or project .area for the imzniementation of sustainable develop- ment projects. These policies include ex- * Main sources of fimding forthe project change rates, trade policies, pricing policies, * Goverment cou rt contrb- taxes, and export duties. Such policies affect utiowhether in cash or in kind the economic performance of projects and the use of envionmental goods and serviLces mi- vo-vecL Moreover, inadequate policies . d- expected financial contribution of each volved. Moreover, inadequate poHTcies can di-. vaet atention fr,m the development functions ° Starting date of the project and its esti- of government institutions and undernine in- mated duration centives, morale, and management capacities * Main projects or activities which the proj- of the institutions themselves (Baum and Tol- ect activities could either be linked with or bert 1987). complement Because the formulation of field projects * Rol f UNDP(Offc fProetServices) requrs the design of complex, mterrelated i t of the project activities that could probably extend over a -1. Appraisal MerodologyforSustainable DevelopmentProjeta 9 * Role ofNGOs reduction of -CO', stabilization of climate This preliminary field visit should also change, coastal area development, and new identify the people that are likely to be involved patterns in - land use and settlemens in, or affected by, the project, as well as their development). social and community organizations or other bodies that may be important in ensuring Projectmanagementand public participation in plamning and adminstraton implementationL -Institutional arrangements rquired for The ministry of planning or finance should successful project implementation-and the be involved in the early stages of the project role of each collaborating -agency or bo development in order to ensure that the project dy-should be clearly defined, with the timing is a government priority, that the project is well of the delivery of specific outputs identified. itegrated in the coimtry's -national plan; that Srucures and capabilities of existing local in- project activities are coordinated with other on- stitutions should be assessedL The upgrading of going or planned activities; and that fncial existing structures, skills, and adopted manage- support can be allocated for the project, wheth- ment and organzational systems by local n- er through govermment sources or through bi- stitutions should be enhanced through their lateral and multilaterml . channels. The participation in projects. involvement of the ministry of planing or fi- To reduce the negative impact of the rapid nance will also acquaint officials of the country trnover of counterpart staff the project should with the-process of developing environmentally be designed to be part of the existing gover- sound projects for futue applications. ment administrative setup (cena, regional, or *The natu and type of project determines local) rather than openate through ad hoc proj- the level of istitutional structures and arrange- ect management, which generally relies on ex- ments required for its implementation. Inst- patriates (UN Joint Inspection Unit 1989). tional arrangements fir impleme Also, to reduce the adminsative burden of the- infrastructure projects are familiar and can be recipient governent, d alition of projF specified closely with the required technology ect support and monitoring functions from the identified. In the case of enviromnental proj- cental to the regional or district level should ects, these institutional arrangements are not be encouraged (UN Joint Inspection Unit always likely to be available. 1989). Establshing and strengthening local in- Institutional arrangements in project design stiutions and capabilities for sustinable devel- should include the costs and benefits of con- opment involves a lengthy process of servation and rehabilittion of the manmade,as experiment and adaptation. It is a time constnn- well as the natal, capital stock ILstitutional ing process since it involves economic and anrangements should thus be put in place to en- technical as well as social and cultural con- sure the self-financing of projects straints. Fostering this process has been ham- pered by the shortage of required budgetary Indigenous capabilities resources (Bam and Tolbert 1987). Projects should be supportive of the cre- Attempts should be made in project for- ation and strengthening of indigenous institu- mulation to address the issues of institutional tions dealing with the assessment and strengthening and capacity building. For proj- development of environmentally sound models ects with specific enviromental components, andtechnologies; consideration should be given to the integration The participation of local institutions and of enviromnental considerations into entities such as farmers' associations, coopem- development plans, the introduction of ap- tive societies, nongovernmental oraito propriate enviromnental legislation and modes and women's groups should be encouraged. of their application, and demonstration-type They are usually the most organized local enti- projects (for example, development and testing ties, possessing both the minimum organEiza- of new technology for energy conservation, tional structures required to undertake specified 10 set of activities and the kmowledge of existing enhanced economic and social benefits or re- conditions as well as needs and rquirements of duced economic and social costs when com- the population, pared with the risk of irreversible damage or Integration of project activities the cost of remedial action that could result. Inltegration of project activties - angenmnlcodeaisiit Takidng entvirnmental considerations into The project should be filly integrated into account in the various phases of the project the country's national development plans and cycle must not be seen as placing undue con- well coordinated with other related activities strints on the development options of a coun- being implemented by othergovermentminis- try. If a project is to be sospended on tfies and departments, UN oraniza donor enviromental grounds, altenative options that agencies, and NGOs. This coordination could are environmentally sound must be provided to be achieved through the establishment of a meet the country's development needs. More- coordnating committee in which the various over, implications of enviromental impacts as- government departments or ministries, as well sessed from the global standpoint cannot be as outside agencies concerned, are represented. isensitively trnlated into specific action in Project fbrmulation may also entail an developing countries in the absence of concrete assessment of the' policies and regulations aleatives that would enable the poor coun- adopted by a country and the extent to which tries to relate the short-tenn well-being of their such polcies would need to be adjusted to tace populations to their long-tern well-being and enRvironental considerations into account in to that of the world. the development process. For most projects, particudy those in- volving lare puc investment in areas such Environmental analysis as a , sQ envirental impact The issue addressed here is how environ- assessment (ETA) should be carned out and mental damge can be avoided or reduced to linked with the cost-benet analysis ensure that development initiatives and their (CBA-see Economic analysis section). The benefits are sustinable. The objective of envi- objective of ETA is to ensure tha environmen- onmental managemet should be to achieve tal aspects are addressed and potential prob- the greatest present benefit possible from the lems are foreseen at the appropriate stage of use of natual resources without reducing their' pmject design. The ETA should be envisaged as potential to meet-fiture needs and the carring an integal part of the plnning process, initi- capacity of the environment (Baum and Tolbert ated at the project level from the start 1987)- Various guides on ELA are available; the The taditional approach to environmental main steps are as follows (Ahmad and Sammy management of undertaking projects with pri- 1987): marily enviro objectives or ensuring * Prelimary acrivities include selecting a that components of other projects include ele- coo r for e ETA and collecting ments to mitigate negative enviinmental im- -background information.' This action pacts, though important, is inadequate. A more should be undertlaen as soon as a project comprehensive approach where environmental been identified. and natual resource management is integrated into economic an socia policy should supple- dno vle on intomeconomic and secial prolicy should h sue-analysis of the impacts of project activities ment the project-by-project .approachwihavetodndigtosimcs (Schramnm and Warford 1989). with a view to identifring those impacts Taking environmental considerations into are worthy of a detailed stud. account in development planning does not im- * Baseline study entails the collecdon of de- ply that the pace of socioeconomic progress tailed information and data on the condi- wlll be slowed. According to the World Bank tion of the project area prior to its (Bum and Tolbert 1987), in almost every se- implemenon. tor ofthe economy, projects have demonstated * Impact evaluation should be done in quan- that incorporating environmental safeurds titative terms, where possible, and should L. Appraisal MekhodoloM for Sustainabfe Development Projecs 11 include the working out of potential mid- * Be designed to meet che needs of the gation measures. Impact evaluation cannot present generations without compromising proceed until project alternatives have been the ability of future generations to meet defined but it should be completed early their own.needs. Projects should therefore enough to permit decisions to be made in a emphasize intergenerational equity. Sus- timely flshion. tainable development projects should en- : ' Assessm2ent -involves combining environ- sure that they do not result in a net capital mental losses and gains with economic loss for futuregenerations. (Whereas some costs and benefits to produce a complete economists argue that sustaiable devel account of each project alternative. Cost- meet can be attained by ensurng at capi- benefit analysis should include environ- t Wealt whether man-made or mental impacts where these can be eva- environmenea does not decline over time, ated in monetary {see Economic -others emphasize that a constant or in- analysis section).. creasing stock of environmental capital Doc- mentation is to describe - should be transferred to fuitre generations eDoumenzalion is prepared tO dese8be the work done in the Elk. A working docu- Pearce I989). - menlt iS prepared tO- provide clearly stated * Ensure that their activities are compatible - and argued recommendations for immedi- with sislainable enronmenal manage ate action. The document should contain a melt and lead to development witou list of project options witi comments on & oii. the environmental and economic impacts of Involve activities which at least leave the cach. level of environmental quality intact, or * Decisionnzalng begIns when the working - which compensate for its loss in valu document reaches the decisionmaker, who terms. re project should lead to the maxi- will either accept one of the project alterna- miion of net benefits of economic devel- t-dvesrequest frther study, or reject the - opment provided that the value of services proposed action altogeher -and quality of basic natual resources are - Pos- audi.s are made to determine how being maintained (Barbier, Burgess, and - - w Post ausare made to determine how .-Pax18) Pearce 1989). close to reality the EIA predictions were. - Due to the negative impactsat many 3 Not disrupt ecological systems through previous development initiatives have had on over-exploiton of natural resources. the environment safe minimum standards - Project activities should be designed to use with which projects have to comply have been natuml' resources m a rational way without increasingly introduced into project desi. hampering or denyig fture genetions addition, safeguards to prevent, m sinize, or the use of such resources (Timberlake compensate for the negative impacts of pMj- 1984 Activiies leading to an irrevesble ects on the envimrnment are being incarpo- loss of natral capital such as complex rated into project design from the start (Baum life-support systems, biological diversity, and Tolbert 1987). aesthetic functions, and climatic conditions Pursuing sustainable development projects are not compatible with sustainable devel- should be regarded as an essential process opment and therefore should be avoided. thmugh which environmental considerations * Where applicable, lead to the conservation are intgted with economic planning and con- of ntural babitats, forests, and species; re- sequently with the national development plans verse deforestation; introduce sustinable of countries. The following are major cnteria rwatr management schemes; reduce acidifi- to be taken into account in the design and for- cation and hazarous wastes; develop and mulation of environmentally sound field introduce environmentally safe industrial projects. processes; slow and eventually stabilize Environmentally sound projects should: rapid population growth; establish pattemrs - . - .... :~~~ 12 of consunption that respond to local needs from the society's point of view through the and contribute to sustainable development. application of cost-benefit analysis. CBA seeks * Incorporate the costs of rectif ying any ex- to determine not only if a project can be ex- - . . pected negative ordetrimental environmen- . pected to provide a satisfactory return to the - . tal effects. This is referred to as.the - -economy, but also if there is an alternative way anticipatory approach and is preferable to of achieving a project's objectives that would - - the reactive one, which involves waiting offer a higher return. This is done by compar- - until problems occur before taking any me- - ing the costs and benefits of projects .through- medial action (Pearce 1989). out their lifetimes (Baum and Tolbert 1987). - - . - . - - . - ~~~~~~~Bccause- market prices of project inputs. - Bc designed to introduce or increase the. a B m p o undersandin ofo and outputs are distorted by taxes, subsidies, understanding of the local population and quta an ote.oenetitretos .. - . . ... . - ~quotas and other govemnment irterventions, their participation in activities leading to these p.ces may not reflect to two costs and the conservation and protection of the envi- -ronent, particularly those projects relat- benefits of a project to society. Shadow pnces ing to efficient e-nergy prodwtion and are therefore used to reflect the value of inputs and outputs. Shadow prices reflect the true consumnption. . . - willingness to pay by individuals to have ac- ° Restrict the use of chloroflourocarbons cess to certain goods.and services whether (CFCs), pesticides, chemical fertilizers these were marketed or not It is the pnce with negative environmental effects, and which would prevail if all resources in the material or components producing carbon economy were optimlly allocated (Barbier, emissions; and introduce and encourage the Burgess, and Pearce 1989). Social CBA ex- use of enviromnentally benign substitutes. pands on. this procedure by giving -special - Be supportive of and consistent with an ac- weighting (in a systematic manner) to benefits ceptable environmental strategy or pro- and costs of a project that help promote the at- gram adopted in the country as well as tainment of broader social goals, such as in- with global and regional environmental come redistribution (Baum and Tolbert 1987). conventions and protocols (Montreal Pro-. To measure a project's profitability, three tocol on Substances that Deplete the Ozone methods are used- the net present value, the in- Layer, Vienna Convention for the Protec- temal rate of retun, and the benefit-cost ratio. -ion of Ozone Layer, Basel Convention on The net present value of a project is the value the Control of Transboundary Movements of the benefits after deducting the costs both of Hazardous Wastes and their Disposal, discounted at the opportunity cost of capitaL Convention on Biological Diversit-, and For a project to be economically viable the Convention on Climate Change)- present value of the net benefits must be zero. or positive,. and hi.-r or at least as high as Economic analysis. that of mutually exclusive project altematives. Sustainable development ppmjects should The intemal rate of return is the discount rate be considered within the context of projects which results in a zero net present value for a pmmoting development rather than growth in projeeL A project can be justified if the dis- its strict sense. Development includes all fac- count rate.equals or exceeds the opportmity tors that lead to increases in the well-being of cost of capitaL This method should however be the population concemed, with economic avoided in comparing mutually exclusive proj- growth being one of the important sources of ect altematives. The benefit-cost atio is a var- increases in welfare (Barbier, Burgess, and ant of the net present value in which both Pearce 1989). . benefits and costs are defined in terms of their present value. If the net present value is posi- Cost-benefit analysis f - .- tive or zero the benefit-cost ratio will exceed or Economic analysis of prjects usually equal unity (Baum and Tolbert 1987). takes place during the appraisal stage and at- Cos3ts to a project normally fall into three tempts to measure the profitability of a project categories, those being direct cost, indirect 1. Appraisal Methodologfor SustainalNe Development Proj es- 13 costs, and opportunity costs. While indirect assessment allows many environmental prob- costs refer to external costs caused by a pmj- lems to be measured and compared in common ect, the opportunity cost of a resource used in a terms and different risk reduction options to be certain way refers to the value that would have evaluated from a common basis. becn derived from its best possible alternative Despite uncertainties and controversies as- use. Opportunity cost plays a central role in sociated with the evaluation of environmental cost-benefit analysis. It can be applied in hazards, comparative risk assessment serves as market-oriented, centrally planmed or mixed a tool for indicating the most promising road to economies (Baum and Tolbert 1987). follow, for the targeting of limited resources Cost-benefit analysis and environmental and for mobilizing and deploying expertise in an efficient and rational manner. Establishing considerations environmental priorities on the basis of rclative The application of methodologies and risks is f*aught with contentiousness and diffi- evaluation criteria used in cost-benefit analysis culty. It requires a tolerance of uncertainty, a to account for environmental considerations willingness to deal with error and to learn from are, however, not straight forward, since the mistakes, and the capacity to adapt to new in- environmental costs and benefits often occur formation and changing circumstances. over a long period of time and are difficult to The rankdng of risks is, however, hampered predict and quantify. Nevertheless, the extet by the gaps in key data sets and the unertainty to which environmental costs and benefits are associated with existing data. Moreover, meth- incorporated into CBA, with values assigned to odologies used to estimate the costs and bene- environmental services and damages, will serve fits of risk reduction are also inadequate and to establish the fkLc that environmental services therefbre hinder the assessment and compari- are not ftee and damages accruing from a proj- son of risk. An additional difficulty is that val- ect should be accounted for (Pearce 1989). ue judgements must always be made in For enviromnental costs and benefits to be comparing and rning environmenal risks. incorporated ino CBA, the environmental im- Further efforts are therefore needed in the area pacts of a project must first be identified and of risk assessment, particularly to bridge the quantified. Since natral systems are very com- gap that exists between the public's perception picx and their reaction to exogenous impact of environmental risks and what is considered are charcterized by uncertainty, a comprehen- most serious by the technical professionals sive environmental impact assessment (EIA) charged with reducing such risks. should be carried out (see Environmental analysis section). Apart from its own intinsic Vulnation of environmentl goods and value, an EIA serves to systematically identify. and quantify the impacts of a project on the Once the environmental impacts and risks natural environment of a project have been identified, the next task is to value these impacts and risks (Dixon et aL 1988). Man-made and human capital may be In conjunction with environmental impact valued relatively easily by observing existing assessment and cost-benefit analysis, a third market systems. However, valuing enviromnen- tool, environmental risk assessment, should tal assets and their fimctions, such as the ozone also be used where appropriate to integrate en- layer, fimctions of tropical forests, wetlands, vironmental and development considerations. and the like, is difficult since the market system Risk assessment is the process by which the is unable to reflect the full contribution of thiese fonn, dimension, and characteristcs of the assets and fimctions to economic activity and risks ta environmental problems pose for hu- human welfam Because there exist market man health, ecosystems, the economic system failures, such as the presence of monopoly and or the quality of hum life are estimated. Risk external costs and benefits, market mechanisms management is the process by which these risks will not automatically secure an opfimal alloca- are reduced. The concept of environmental risk tion of resources. For example, the market L. AgPaai.u 16dwodoftfor SwainaMe DeweeI,mcffw Pr~ejeI 1 project to te ovall benef or cot might be at the national, regional, or sectoral levl Ex- margirmal (Barbler, Burgess, and Pearce 1989). tensive effors are underway t0 develop and A progrm may entail a significant environ- pmmote these techniques. Var methodolo- meall cost or benefit to society as a whole, gies have been proposed for rescurce account- while individual project apprisals may not ac- ing that can- filfil - one or mom fimctions count entirely obr this benefit or cost incl ding, among oters, the provision of an in- Another serious problem in the application foron system on of cost-benfit analysis anrses fiom the use of a linkages and the calculation of revised national discount factor to arrive at the net pesent val- . mcome figars. Such accounts cold be used to ue of fite costs and benefits. The analysis is reflec the eflects of a project or a progm on innded to relate filuare costs and bmneft to a countys corrected n_ional incm- purt- their present value. Many of the importat larly where a project, or set of projects, signifi- benefits of environnical _agement a e, candy affects the exploitation of teb couny's bw:eve, to fiua generdons and become dis- natural as is often the case in large- cemible only afte the time at which sandar-d scae development projects. Naual resomurce cost-benefit analysis discounts frture benefits accounts have not yet been established by any to zeo, this discouaging invesments inl proj- developing country government and by only a am specificaRy designed to provide nfw developed countries It is hoped, howeve, environmental or conservation benefits far into that as prescriptios emerge fom pilot studies the fitue (Baum and Tolbert 1987) ' and reevat exeine in deloped countries on the most useful and cog-effective method- Suestions for ensuring better f or NRA, more will adoptthis -aiccounUng for the environment h - tool as. an integrl part of the decisnmakin poject appraial pmces Nevet ,less, fiurher reseach and de- No conss exists on how CBA should velopment of analia fiamewors to iate- be modified or sWpplemoted to properly micor- grae environent consideraio into pome environmenal i ns The sug- nacroeconomic policy anaybsis and .frml- gesfios listed below are only prelimiay tion is.required snce many aspects of macroe- a--empts at putting into operation able comc policy (for example, fsa polky) cmn development in economic anflysis, and they have significant effect on the use of envirn- have not been broug into widesptead -use mental and natural resoures (Muzondo, MG- -The proponents of suh mOdificons usually randaandBovenberg-1990). place them wihi the contex of a whole pack- Corcting madket prices to reflect the full age of reforms including- costs aind benefdis of using teb envio inm in- Austing the national accounts system to vexemagin ---reflecteber elation Of ~ in e xn ties. Once this is done, the users: of --nment to cbanges in the economy m-- nenal goods and serices would t - (envronmn andi natumal - n sbe forcedto tale into accoUntthe hfll costs and acut) bemnfts of usmg the enviroment in tei ded- * Austing or coecting market prices (for smnmang process example with the use of taxes) to reflect - uof tDIn tg om t ofaconventional discount rate, someobevr the full costs and benefits of using the ad p ones have usi lwer environ.ent (than usua) or even negative disont rates - Providing for the economic value of errvi- that place a higher value on fute benefits and romnental impacts in pmject appraisals weigh fitur costs heavily. However, for sever- (Babier, Burgess, and Pearce 1989)-- al reasons-because money has a time value, Envirownental and natul resourue ac- became a low discount rat could admit proj- counting has remntly been rcoopized as one ects with a low return, and beocase use of dif- instrment that can b: used to integrate anvi- f discout raes for eiromntal and rmnmental considertions irto economic policy other projects- could distort- inestment 1. Appraisa Meoddologyfor Sustinable Development Projects 15 project to the overall benefit or cost might be at the national, regional, or sectoral level. Ex- marginal (3arbier, Burgess, and Pearce 1989). tensive efforts are underway to develop and A program may entail a significant environ- promote these techmiques. Various methodolo- mental cost or benefit to society as a whole, gies have been proposed for rescurce account- while individual project appraisals may not ac- ing that can flulfill one or more fimctions count entirely for this benefit or cost including, among others, the provision of an in- Another serious problem in the application fbrmation systm on environment-enomy of cost-benefit analysis arises from the use of a inkages and the caculation of revised national discount factor to amrve at the net present val- income figures Such accoun could be used to uc of fire costs and benefits. The analysis is reflectthe effects of a project ora program on inatnded to relate future costs and benefits to a country's corrected national income, particu- their present value. Many of the important larly where a project or set of projects, signifi- benefits of enviromniental management accru, candy affects the exploitation of the counr's however, to fuire generations and become dis- natural resources, as is often the case in large- cemible only after the tine at which standard scale development projects. Natal resource cost-benefit analysis discounts future benefits accounts have not yet been establised by any to zero, thus discouraging investment in proj- developing county government and by only a ects specifically designed to provide substantial few de-veloped counties It is hoped, however, environmental or convation benefits far into that as prescriptions emerge from pilot studies the fitue (Baum and Tolbert 1987). and relevant experience in developed coutries Suggestions for ensuring better on the most useful and cost-effective method- accounstiong for treenvirornment ologies fbrNRA, more countries will adoptthis tool as. an integral part of the decionnakng project appraisal process. Nevertheless, further research and de- No consensus exists on how CBA should velopment of analytical flameworks to ute- be modified or supplemented to properly incor- gmte environmental consideations into porte environmental considertons. The sug- macroeconomic policy analysis and formula- gestions listed below:are only preliminry tion is required since many aspects of macroe- altempts at putting into operation susUinable conomic policy (for example, fiscal policy) can development in economic analysis, and they have significant effects on the use of environ- have not been brught into widespread use. mental and nturnal resoures (Muzondo, Ml- The proponents of such modifications usually randa, and Bovenbeg 1990). place them witiin the contxt of a whole pack- Conecting market prices to reflect the $x11 age of reforms including: costs and benefits of using the environment in- * Adjusting the national accountsstem to volves ig envir enal exte - reflectthe relation of changes inthe envi- tes. Once this is done, the users of romnent to changes in the economy envrronmental goods and services would then (environmental and natual resource be foced to take into account the fu costs and ac counting) : - - - u - benefits of using the environment in their deci- * Adjusting or correcting market pric (for siomakingprocess wDue to the problems arising fom the use example with the use of taxes) to reflect :o cneistdsom ae oebevx the full costs and benefits of using the and pctitoners have advocated using lower enviromnent envronment {(than usual) or even negative discount rates * Providing for the economic value of envi- that place a higher value on fiture benefits and romrnental impacts in project appraisals vweigh fitumre costs heavily. However, for sever- (Barbier, Burgess, and Pearce 1989)- al reasons-because money has a time value, Enviromnenal and naurl resource ac- because a low discount rate could admit proj- counting bas recently been recognized as one ects with a low reurn, and beca use of dif- iinstument that can Nu used to intate on- ferent discount rates for environmental and ronmental considerations into economic policy otiher projects could distort investment - 2 Environmental Economics and Valuation in Development Decisionmaking Mob. Mae and -En Lutz One essental step towards acbievig=sutaiable dewelopnmen is th eoomca ef- ficient manageentd of i(itual resowce& Thipaper e.Wfain the key role ofeninmtl ecoomcsinfailtain th mrecieciv icoporation ofwnironmenta concern into de- velopent decisinraidng. It also reviews concepts and kdecimiwusb vahuton of environ- mental &Wnac& dial enable such enviromental cmideraftom to be explc* comi&derdin die conventonl cost-benft~t calculus used inz economc*deWionmaling. Key related aspects incldin7g envronmental impacts Of econmaywide2polides, diwscon rate ies,is anAd multi- criteia ana!yiis are Prevewed The prx=s ofr internalizin environena extermalites may be faciliated by mnaking even rozgh qualiatve assmewnts earlyo in the project eul- .ion cy The a g qiuc an approac include: (a) the .earl ecluion of op.on that are not sound envirnmenally (Z) jmoe- efifectiv in-depth consderatio of thoe abLter- native that are preferable from the envronenal vieumpoint and (c) better opportntes for redes"ggngprcts andpolies lo achieve sustainbl developmenot gf-- A manber of developing cowtiy case sties, which cover a wide rag ofpracal- valicon methods, f e reviewed We may canclude genemr that frther applicaton - t pr=acavblems in deveopirg coImtres is reqired (rt thavn futer thoretical dvel- opm- ), of te eiromnta aluation concepts nd tciques presented in the pp Such case stdy work can be most effecie when car7ied out asprt of prject preparation 2 Environmental Economics and Valuation in Development Decisionmaking Mohan and Ernst Lut One essential step towards achieving sustanable development is the economically ef- ficient nmangemnent of natural resources. This paper explains the key role of envionmenta economics in faciflating the mnore effective.incorpration of environmntal concerns itwo de- -velopment decisionmaJdng. It also reviews concepts and technques for valuation of environ- mental imipacts that enable such environmental considerations to be explcitly considered in the conventional cost-benefit calculus used in economic decisionmalcng, Key related aspects Jincluding environmental impacts of economywide policzes discount rate issues, and multi- criteria analysis are reviewed The pnicss of internalizing environmenta exernalities may be facilitated by making 'even rough qaiaveassessments earl -on in the project evalu- ation cycle. The advantages of such an p include: (a) the eary excusion of options tha are not sound environmentally; (b) more effective in-depth consirderation of tose ater natives that are Preferable from the environental iewpoint; and (c) better opporntuitiesfor redesigningprojects andpolicies to achieve sustainabl development goals. *A number of developing country case studies which cover a wide range of practical valuation methods, are reviewed We may conclude generalfr that further application to practical problems in developing countries is required (rather than further theoretical devel- opment), of the environmental valuation concepts and techniques presented in the paper. Such case study work can be most effective when carried out as part ofproject preparatiom- 1R In recent times, the environmen has able life support fimctions (for exarnplek the emerged as a mTajor worldwide concern. PoUu- stratospheric ozone layer that filters out harm- ton in particular is perceived as a senous fiu ultraviolet rays). thrat in the indusialized countries, where the This paper reviews how envi al quality of life bad hitherto been measured economics facilitates the efficient use of natu- mainly in trmns of growth in matal output- ral resources (both mineral and biological), as Meanwbiue, natural resource degradation is be- well as manmade capital and human resour- commg a serious impedinent to ewnomic de- ces-an objective which is a vital prerequisite velopment and the alleviation of poverty in the for susainable development Special atteton developing world. is paid to the key role of environmental eco- Mandkid's relationship with the environ- noinics in helping value environmental and met has gone through several stages, starting natural rsources more precisely and in inter- with primitive dine in that human beings lived nalizing the costs and benefits of using such re- in a state of symbiosis with natur, followed by sources into the conventional calculus of a period of increasing mastery over nature up economic decisionmakig. More generally, the to the industrial age, calminating in the rapid identification of sustainable development op- materia-intensive growdL pattem of the twen- tions requires: tieth centuy that adversely affcted natural re- * Good u ding of the physical, bio- - sources in many ways. The initial reaction to logical and socil of human such environmental damage was a reactive ap- activities proach chdatized by increased clean-up ac- tivities. Inl recet deades mankind's attitude Of damage to the envionent that improve: towards the environment has evolved to encom- pass the more proactive design of projects and te design of policies and projects and lead policies that help anticipate and avoid envion- to environentaLly sound sn metal dgadation. The world is currently ex- deion ploring the concept of susainable * Development of policy tools and strh- development-an approach that will permit ewing of buman resources and institutions contimning improemnents in the present quai to iiplement viable strategies and manage of life and will lower the intensity of resource natal resources on a sustainable basis use, thereby leaving behind for future geLerki economic.and-th dons an undiminished or even enhanced stock Ling economics and the of nahral resources and other asse&.- environment Eavironmental assets that we seek to pro- tect provide three main types of services to hu- The role of environmental economics man society, and the consequences of their Environmental economics plays a key role degradation must be incorpor' into the deci- in identfiing options for efficient natiil re- sionmaking process- First, the environment source management that facilitate sustinable provides esstial raw materials and inputs that development It is an essential bridge between support human activities. Second, it serves as a the traditional techniques of decisionakingg- sink to absorb and recycle (normially at little or and the emerging more eronmentally sensi- no cost to society) the waste products of eco- tive approach Environmental economics helps activity. Fimally, it provides irreplace- us inco ate ecological concerns into the @The authors are grateful to Edward Barbier, Robin Bates, Noremn Beg, Jan Bojo, Wilfrido Cruz, John Dixon, John English, Gunnar Kohlin, Karl-Goran Maler, David Peare Adelaida Schwab, Jeremy War- ford, and Dale Whittington for comunents and contnbutions at various stages in the prepaation of this paper. This work was supported in partby a grantfrom the Government of Norway.) This braa dfion is based on OurCommn Fare (World Cnmi on Enwfromnand Develp 1937) Forarecnt review of aa_iive definitions oftsainble depmt, see Pcry 1992. 2. Ervironnmetal Economics and Valuation in Dadopmnem Decisionanain- 19 Environmental Analytical Socioeconomic System Tools and Methods Structure Global Irnter- .~~~~~~~~~~~~~~~~~~~C . U . WA 4 -anationa -l National Ns 0 . ~ ~~~~~~~~~~~~~~~~ . 0 a-~~~~W cn 0c Natural ~~~~~~~0 a- C 4~~~ CU C C~~~C zonveniional fionvvork of _Dsciyas E bi ae oihceiomla 3E W -. C 4aŽ .~~.. C-~.U C COOc Rgoa Water 2-E. - -ieegiona > S Wa > 0C Ae global and inns- level consists o acrs Ah d=' o str ofh C) and Air -Proiect Figure 2-1: Ineorporatiag Envronmental Concen into Deciuionnaking conventiona framework of human society, as a In each case, a hoistc nvien imental .Shown in Figure 2-. analysis would seek to study a physical or eco- The right-hand side of the diagram inch- logical system, in its entirety. Complications cates the hierarclhcal natre of modem socty arisebecause such natural systems tend to cut he global and tnational level consists of across the d o g stnructue of human sovereign nation states. In the next Level are in- societ - For exampvle, a -forest ecosystemn (hlk dividuaI countries, each having a muluisectoral the Amazon) could span severan countries, and macroeconomic structure. Various economic, also intaect with many diferent economic see- sectors (such assenergy, industry, agriculture, tors withnm each contiry. tralsort, etc) exist within each country. Fi- The causes of enviromnental deadation naly, each sector consists of difepent subsec- arise from human activity (ignoring natural di- tors, projects and local schemes sasters and otihr events of nonhuman oigimn), Unforstuntely, the analysis of the environ- and therefore, we begin on the right side of Fig- ment cannot be caried out readily using the ure 2-L- The physical (including biological and above socioeconormic structuring. As shown on social) effects of socioeconomnic decisions on the left'side of Figure 2-1, one convenient the environment must then be traced through.to breakdown recognizes enviromnmental issues -the left side. T'he techniques of environmental. that are (I) global and tasational (for exam- assesmen (EA) have been developed to facili- pie, climate change, ozone layer depLetion); (2) tate this difficult analysisTor example, defor- natural habitat (for example, forests and other estaton of 'a primary moist tropical forest may ecosystems); (3) land -(for example, agricul- be caused by hydroelectric damn (energy sector tiural zone) (4) water resources (for example, policy), roads (trnsport sector policy), slash river basin, aquifer, watershed); and (5) urban- and burn fanning (agricukure secor policy) industrial (for example, metropolitan area, ming of mieral (industral sector policy), 20 land clearing encouraged by land-tax incentives Recent developments (fiscal policy), and so on. Disetangling and Altog the consolidation of environme- pri&ozing these multiple causes (n side) tal economic theory and its application to em- and hicir impacts (left side) will involve a comr- pirical isses sprad throughout the developed piex environmenta assessnent exercise. world in the 1970s, the incorporation of envi- - Menwhile, the usual decisioonaking P~O ronmental issues into development planing is cess on the right side of Figure 2-1 relies on a relatively recent phenomenon. A review of techno-cngincering, financial and economic the liteure in the field reveals the paucity of analyses of projects and policies. In particular, witin touching upon the environmner. Aca- w re note tint conventional economiic analysis demic developmcnt economics barely acknowl- has been well developed over the past- several e he field until recenty (Dasgupta and decades, and uses techniques including project -Male 1990). Much of the literature on the en- evaluationkcost-benefit analysis (CBA), secto- vironme economics of developing counties rallregional studies, inultisectoral fl = ~has nemrged since the mid-1980s. This is large- nomic analysis, ande inenational ecoaomic ly as a result of the inrasing emphasis being analysis (finance,trad, etc.) a:t the VioUs phced on envirom issues by major aid- hierarchic levels. givig ittions Figure 2-1 also shows how environmental he Bndtland Report rized the role eoonomics piays its bndgmg role, by mapping of economics in s e devep- the EA results onto the framework of conve- i assesg a vioe nfs tional economic analysis. Once again, a Varyt degradation in developing countries, and in de ofi enviromnenta economic techniques mclud- sigi relevant incentives to limi such degra- ing economic valuation of environental im- dation (World Comission on E mt pacts (at the local/project level), integratedlopmcnt 198. Thereceworldwide resour-ce gement (at the sector/regional concern regarding the enviroeinnet has level), environmental macroeconomic analysis s wgthened the emphasis placed on environ- and e ac ting (at the economy- - susr ity as a impor crierion wide, multisector level), and global/ransna- fo, sound natural resource man tional vi l economic analysis (at the A of this concer is rflected i the intemational level), facilitae this pr s of in-a e d iopaid to rgenerational eq-: corporating viromnental issues into tradition- nty and the role of discount rates i economc al decisiomnaking The analytical technques - described above overlap insidemi,13t, and - In recmt years most govemrnents have ad- therefore this conceptual categorizafion should - tir policy objectives to include the not be interpreted too rigidly. p theirne of clrudes Oncethe oreoingstes ar copletd, proper management of natural resources aOnce the mueog steps are oomplted, alongside economic growhi and income proje-fts and policies MUA- be redesigned to re- itbonpvr altsn he projens pohcies flXkSi&lecl ~~distribution/poverty alleviation. These duce their nimpacs and to shift objectives may be so compleme y the development process towards a more sus- but it Is accepted that offs are rired tamable path. Clearly, the formulation and n-. at least in the short term. Given the exsting plementation of such policies is itself a difficult scates of financial and human resources in task. In the deforestation example described devloping countries and the emerging pressure earlier, the decisionniakers who wLsh to protect on environmental resources, it is particularly this single ecosystem are likely to face pob- important to use these limited resources in a lems in coordinating policies in a large number w y that can be sus d and that will have of disparate and (usually) non-cooperatmg thew maximumbenefit in ams ofthe country's ministries and line istiLtutions (such as energyi, objectives. Sound economic analysis of proj-- transport, agriculture, industry, finance, and ectsandpoliciesisanimportantmeansofmak- forestry). ing the allocation prmcess more efficient, and thereby more sustainable. However, economic 2. Environmental Economics and YVauation in Deedopment Decisionmaldng 21 efficiency does not necessarily guarantee accepted by the World Bank's Board of Execu- sustaiability. tive Directors and the borrowing government The growing attention to environmental is- The borrower is responsible for implementing sues in the work of the World Bank over the the project according to conditions mutually past decades culminated in November 1989 in agreed on with the World Bank. Supervision of the issuing of the mandatory environmental im- the implementation process is carried out by pact assessment required for all World. Bank the World Bank through periodic field visits projects? Thus enviromnental analysis has and progress reports from the borrower. Evalu- been elevated to the same level of importance ation is the final stage of the project cycle, fbi- as the three traditional aspects of project evalu- lowing disbursement of the loan. Project ation: financial, economic, and technical analy- performance audits are carried out by an. inde- ses The -valuation of environmental inpacts. pendent World Bank department, and- where takes on added urgency in this context, since it appropriate; involve review of previous project would permit environmental concerns to be in- documents and field visits. This analvsis vields corpomated efifctively into the nonnal decision- valuable experience that helps improve the making process in all World Bank operations. work at all stages of the project cycle. Conventional project evaluation Economic analysis and cost-benefit The succsfil completion of a develop- cnteria malt project usually involves sevi2al wEll- Cost-benefit analysis (CBA) is the key ele-. defined steps The systematic approach used by - - ~~~~~~~~~ment in the appraisal stag of the pect ycle. th- e World Bank in a typical project cycle in- CBA seeks to assess project costs and benefits cludes identification, preparation, appraisal, negoiatins ad fnancng, mplmenttionand using a common yardstick. Benefits are defined goiaions and financ implementation and relative to their effects on the improvements in - - - - supwnision. and evaluation - - human w ell-being. Costs are defined m terms Th2e projecycle - -of their opporunitycosts which is the benefit foregone by not using these resources in the Project identification involves preliminary best available alternative application. selection (by the borrowing coutry and the In addition to this economic test. as pre- World Bank) of potential projects that appear viously mentioned, a number of other aspects to be fasible and conform to national and sec- (including technical environentaL istitution- toral development goals fn the preParation al, and financial criteria) also need to be con- phase wiich may last one year or more the sidered in project appraisal The economic borrowver studies the engineering technical, en- analysis of projects difflrs from financial vironmental, instintuionaL economic and finan- analysis. The latter focuses on the money prof- cial aspects of a proposed project Project its accruing to the project enterprise or firnn appraisal consists of a comprehensive and sys- based on market or financial costs. The eco- - tematic review of all aspects ofthe project, cu- nomic analvsis of a project, on the other hand, minating in an appraisal report that discusses measures the effect on the efficiency objectives the national and sectoral stategies as vell as in relation to the whole economy. Rather than the engineering/echnical, envionental, in- fmancial prices, shadow prices are used that stitutionaL economic, and financial issues. reflect opportuniy cost including valuation of The appraisal report is the basis for justi- externalities wherever practical (as describcd ijring the investment, as well as the negotiations below). Some of the criteria commonly used in at which the borrower and financier (the World the cost-benefit test of a project are described Bank) discuss the measures required to ensure next, with the emphasis on economic rather the success of the project, and the conditions than financial evaluation. for fimding. The resulting agreements are in- The most basic criterion for accepting a cluded in loan agreements which together with project compares costs and benefits to ensure the appraisal report are considerd and that the net present value (NPV) of benefits is 2 For detais, sceWod Bnk Opeatina Directive 4.01. 22= positive: Another fiequently used criterion is the benefit-cost ratio (BCR): T NPV = E (B,-C)(l + ry T T t. - BCR = [E B/(lAri] [S: C/(l-I+r)J. Where Bt and Ct are the-benefits and costs t=O t=: in year t, r is the discount rate, and If BCR > 1, then NPV > O, and the projea is T is the time horizon. acceptable. Both benefits and costs arc defined as the Each of these crteria has its strengths and difference between %hiat would occur with and weaknsss, but NPV is probably the most use- without the project being implemented. In eco- fuiL The NPV test may be used to derive the nomic testing, as described later, B, C, and r least-cost nrle. In certain cases, the benefits of are defined in economic tenns and appropriate- two altenative projects may be equal (that is, ly shadow pnced using efficiency border prices they both serve the same need or demand). (see the foflowing discussion on discount rate). Then the comparis of altratives is simpli- However, for the financial analysis of projects, fied Thus: B, C and r may be defined in financial terms. T If projects are to be compared or ranked, NPVW-NPY V= E [Ca, - C(J/(l+r)t, the one with the highest (and positive) NPV t=O would be the prrrd one, that is, if NPVJ > NPV1 (where NPV; = net present value for since the benefit stram canel out Therefore, project i, then project .1 is preferred to project if II, provided also that the scale of the alterna- tives is roughly the samne More accurately, the T T scale and scope of each of the projects under x q,j(i-)y> x review must be alterd so that at the margin, -=O tO the last increment of investment yields net benefits that are equa (and greater t1han zero) tis implies that NPV1 > NPV0. for all the projects. Complexities may arise in In other words the project which has the the analysis of interdependent projects. lower present value of costs is preferred. This The internal rate of return (IRR) is also is called the least-cost alternative (when bene- used as a project criterion. It may be defined fits are equal). However, ev after selecting by: te lea-cost alternative, it would still be nec- essary to ensure that tins project would provide T apositiveNPV. S(B -C,)/(+IRR)= . t=0 Shadowpncing Thus, the IRR is the discount rate which re- the i world of perect competi- duces the NPV to zero. The project is accept- ton, th interaction of atomistic profit- able if IRR > ARI (Accounting Rate of producers and utility-maximizing Interest), which in most normal cases implies c g r - . . . . . ~~~~~~~~consumers gives rise to a sitation that is NPV > 0 (ignoring cases in which multiple 3 In tiS state, pri roots could occur, which may happen if the annual net benefit stream changes sign several , the true marginal social costs, scarce re- -- * * Xr w ~~~~sources are efficienltly allocated and, for a- times). Problems of interpretation occur if al- . a e a a r temative projects have widely differing life- given income distribution, no one person can be - - . - . . ,- tti~~~~~mae bettr off winthout niakng smomone else times, so that the discount rate plays a critical role. - worse off(Blator 1957). 3 Furthcr details conlcing the summaridmal in this secion may be found in Dasgupta, Marglin, and Scn 1972: --Littleand Mirlecs 1974; Muasinghe 1990b; Ray 1984; and Squire and Van derTak 1975. 0~~~~Lti an Mi-:s17,Mnaigc19b 8 2 Enviwonmental Economics and Vaiwadan in Deudopmen* Decisionmxahing 23 However, conditions are likely to be far Two basic types of shadow prices exist. from ideal in the real world. Distortions due to These involve whether or not society S indif- monopoly practices, exa economies and ferent to income distribution considerations. To diseconomies (such as evironmental impacts iEustae this point, consider the simple nation- which are not internalized in the private mar- al goal of maximizing the present. value of ag- ket), interventions in the market process gregt consumpton over a given time horizon. through taxes, import duties and subsidies, all If the consumption of different individuals is result in market (or.financial) prices for goods added directly regardless of their income levels, and services which may diverge subsntially then the shadow prices derived firom such a from their shadow prices or true economic val- model are termed efficiency prices because ues- Furthermore, the reliance on strict efficien- they reflect the pur efficiency Of resource al- cy criteria for detrmining economic welfare location. Alternatively, when inmcasing the implies the passive acceptance of the existing consumption of the lower income groups be- (skewcd) income distribution-this may be so- comes an important objective, this consider- cially and politicatty unacceptable, especially if ation is given a greater weight in evaluating there are large income dispari5vs. Such consid-. aggregate consmrnption. Then, the resultant eraiions necessitate the use of appropriate shadow prc are called social prices. shadow prices (instead ofmarket prices) of The goal of shadow pricing is, therefore, project inputs and outputs, to determine the op- eiher efficiency or socially oriented. In brie£. tmal investment decisions and policies, espe- efficiency shadow prices try to establish the ac- cially in the developing countnes where market tal economic values of inputs and outputs, distortions are more prevalent than in the in- while socially oriented shadow prices take ac- dustialized countries. count of the fact that the income disibution Consider a general equilbrium model of bewee different socital groups or regions the economy in which the national goal is em- may be distorted in tms of overall national bodied in an acceptable objective function such objectives This may call for secial adjust- as aggregate consumption. This consumption is mens-vig greaer weight to benefits and to be maxmz subject to constaints that costs accing to the poor relative to the rich, might include lmits on resource availabities, for exampler In practice, such formal weighting distortions in the economy, and so on. Then, schemes are seldom used in project evalu- the shadow price of a given scarce economic ation-instead, income distributional and other resource represents the change in value of the social issues are addressed thrugh direct tar- objective function, caused by a marginal geting of beneficiaries and similar ad hoc ap- change in the availability of that resource. In proaches. In our analysis, ve wfll place the more specific conte of a mathatical prinary emphasis on efficiencry shadow programming macroeconomic model, the opt- pncing.- mal values of the dual variables (that cone-L- Nonpriced inputs and outputs must be spond to the binding resource availability shadow-priced to reflect their economic oppor- constraints in the primal problem) have dinen- tunity costs (see Annex 2A for details of border sions of price, and could be interpreted as shadow prices and conversion factors). Major shadow prices (Luenbeger 1973 and Sassone categories of such nonpriced iputs and out- 1977). While the general equilibrium apprach puts are public goods and externalities is concepually.important, it is too cumbersome (especially enviromental impacts). Public. and data-intensive to use in most cases. In goods are defined as those goDds and services practice,. partial equilibrium techniques that that are free to all without payments once they evaluate the impact of the change in the avail- have been made available, such as police ability of a given resource on a few key areas, protection and tansportaion and navigation ratder than throughout the ecoomny, may be facilities. Externalities are defined as beneficial used. (See the following sections on environ- or adverse effects imnposed on others for which mental accounting and environmental impacts the orginator of these effects cannot charge or of economywide policies.) be cbarged (as the case may be). 24 Unfornately, many extenalities are not 1974). Essentially, this unit is the same as free- only difficult to measur in physical temns but ly disposable firegn exchange available to the even more difficult to convert into monetary govement, but expressed in terms of units of equivalents (that is, to measure the "willingness local currency converted at the official ex- to pa/" of the parties affected by the exerali- change mrt. Anne 2A conins a discussion of ties). Quite often therefore, the approach take this particular yardstick of value. The- border- is to inpose regulations and stmndards, ex- priced merair is particularly relevant for the pressed in physical measurements only, that try foreign exchange-scarce- developing ounies. to eliminate the perceived enal damages. It represents the set of opporunities available However, this approach may not be effective, to a country to purchase goods and services on because no attempt is made to compare the the intemational market. .costs of compliance with the real benefits pro- - vided (that is, damages avoided). The basic Natioal icome accounts and concepts and echniques for economic valUa- macroeconomic performance lion of enviromnental impacts are di-sssd - n order to accutely recognize and in- later in the chapter. dlude enviomatal concers hi economic . -merai-e analyses, standard income acounting teh- nmques must be re-exan Performance is To derive a consistent set of economicn currdy mesured by the growth in gross do- shadow prices for goods and servics, a com- mestic product (GDP), and policy reforms are man yardstick or mnixaire to measure value is justified routinely on the basis of their short-, necessay. The choice of the umraire, like the medium-, or long-term contr on to such choice of a currecy unit, should not influence growth. While GDP measures market activity the economic criteria fir decisionmaking pro- reasoably well, it has been crticized for its vided the saune consistent fimeworkx a as- neglect of nomarket value added- More IMpor- sumptions are used in the analysis tandy, since GDP does not consider depreci- The same nominal unit of currency may atiol of manmade capital and -also leaves out have a difitrent value depending on the eco- the degradation ofVnaunl capital," it is an in- nomic crcumstancs in which it is used. For - rate mau of tre, aile ince example, a rupe-worth of a certain good par" In terms of the envinment, there are three chased in a duty free shop is likely to be more specific shor s in the current national han te physical quantity of the same good ob- acounting frameworlc taimd for one rupee from a retail store, after 4. Natural and environmental resources are import duties and taxes have been- levied, not included in balance sheets; tbere, Therefore, it is possible to distinguish intuitive- national account represen limited indica- ly between the border-priced rupee, which is tors of national wel-being, since tey are a used in intemational markets free of import tar- poor, or even "evese;, measure of iffs, and a domestic-priced rupee, which is used cbanges in environetl and resource in the domestic market subject to various dis- conditions. tortions. A more sophisticated example of the s. comendonal national a s fail to re value differences of a currency unit m vanous cord lhe depreciation of natural capital, uses arises in countries where investment for such as a nation's stock of water, soil, air, fr econoirc gmww is consideTed inae- -eal ,adwlens futur econmic VOW~~t W ~ n~e nonreewable resources, and wifldyno quate. In these instances, a rupee-worth of say- : which are essential for e ns- ings that could be invested to inreethe level tence. Dasgupta and Maler (1991) make of futre consumption, may be considered un- more valuable than a rupee devod to current derpric in the maret-the lower the consumption, - - ---value added, tbe lr is the extent of un- A most apprpriate numeraire in many n- - derpricing of the final product It foLows stances is a unit of uncommitted public income t es that export primary prod- at border shadow prices (Litde and MriATces ins do so by subsidizig them, usual ..~~~~~~~~~~~~~~~~ut dos:ysbdzqgtea say 2. Environmental Economics and Valhution in Dcevopmt Decusionmaking 25. with disproportionately large adverse im- SNA and corresponding performance indica- pacts on the poorest members of society tors would encourage policyrnakers to reassess (who are less able to protect themsel- the macroeconomic sitution in light of envi- ves)-the small cultivator, the forest ronmental concerns and to trace the links be- dweller, the landless peasant, and so on. tween economywide policies and naural Currently, there are no estimates of such resource management (Muzondo et al. 1990). hidden costs or "subsidies." If there were, Based on ongoing work since 1983, the the GDP of many countries could well be World Bank- has actively encouraged the con- significandy lower. In addition, natural re- sideration of environmental issues during the source depletion raises intergeneradonal ongoing revision of the SNA by the United Na- equity issues to the extent that the produc- tions. As an inkerim measure, it proposed that a tive assets available to fiture generations satellite system for irte d environmental are unfairly diminished (see the discussion and economnic accoDuming (SEAA) be creaed. on discount rate in the next section). This approach represents a compromise that 6. Cleanup costs (for example, expendires does not cbange the core of the SNA, its pro- incurred to restore envirnental assets) duction boundary, and the relevant tie ser- are often included- in national income, ics -but at the same tine, the satellite accounts - while environmental- damag,es are not oon- will encourage the collection and compilation sidered. For private firms, defensive evi- of relevant information on natrl resources rnmental expenditures (that is, measums and the enviromnment The satellite accounts are to reduce or avoid environmenl damage) an interim step that will permit filrther progress are netted out of final value added. In con- towards th*e comnputation of an EDP and an trast, such cleanup costs are considered as EDL The World Bank, r wih the UN -p"roductwve contntutionxs to national out- Statistical Office (UNSO), has recnly com- put if they are incurred by the public sec- pleted case studies in Mexico (van Tongeren et tor or by households. The calculation of ah 1991) and PapuaNew mus et GDP is distorted in two way sir- aL 1992), to determine how such accoun can able outputs (for example, pollution) are be P - P&-e: overlooked while beneficial ernment-- The "UNSO Framework" (Brtlmus, related inputs related to environienal Stabmer, and van Tongeren 1989) was used as Deeds are often implicly valued at 2).4 a basis for the above studies. This fimnework The deficiencies in the accounting tech- is a Sysii for environmentally-ad eco niques employed at present point to the need nomic accounts which represents one approach for a system of national accounts (SNA) which to deriving EDP and ENI. Its objective is to in- permits the computation of an environmentally- e vomenal daza sets with exsting ajusted net domestic product (EDP) and an national accounts infornaion, wbile maintain- environmentally-adjusted net income (EDm). ing SNA concepts and principles insofar as Such measures would help to better capture en- possible Environmental costs, benefits, and vironmental services, account for the depreci- natural resource assets, as well as expenditures ation of both manmade and natnual capital, for envionment protection, are presented in exclude relevant categories of defensive envi- flow accounts and balance sheets in a consis- ronmental expenditures, and/or estimate dam- t manner, while mainaining the accounting ages as a result of econonic activities. To the idetities ofthe SNA. extet that national -level decisionmakers and On the basis of the UNSO famework a macroeconomic planners (typically, in a Minis- draf Handbook for Enviromental Accounig try of Finance or National Planbing)relyon the -has been developed (UNSO 1990) which out- c-oonventional SNA to formulate economic poli- lines in detail the possibilities for computing an : cies, a supplementy environmentally-adjusted EDP. Until the EDP concept becomes more 4Dasgpta and Maler (1991) suggest that in Od to avoid doublecounting, exditures that enhance resouce bases, such as forests, should not be included in national income computations, as they are already reflected in the value of chamges in the resource stocks. 26 widely accepted and used, decisionmakers- and Real currency devaluations have the effect policy analysts should keep in mind the limita- of increasing international competitiveness, and tions of current national accounts infbnmation. raising production of intemationally tradable Ultimately, the most desirable policy reforms goods (for example, forestry and agricultural are the ones- that would focus on increasing products). If the agricultural response occurs EDP rather than GDP, since EDP more accu- through crop substitution, environmental im- rately measures "sustainable" income (see also pact would depend on whether the higher Peskin 1990, and Repetto et al. 1989, for re- prict4 crop had environmental benefits (for ex- lated research). ample, tea, cocoa, and rubber) or was environ- Economywide policies and the mentally damaging. (for example, tobacco, environment sugarcane, and corn). Environmental impacts would also depend on whether increased pro- Economywide policies (both macroeco- duction led to timing on new land (which nomic and sectoral) play a significant role in could result in increased deforestation) or re- the rate of depletion of natural resources and sulted in more efficient use of existng fnn-- the level of environmental degradation. Fiscal land. By increasing the competitiveness of and monetary policies, structural adjustment world exports, it has bee demonstrated by Ca- progrms, and stabilization measures all have pistrano and Kiker (1990) that the opporunity an effect on the natural resource base. Unfortu- cost of keeping timber unharvested increases as nately, interactions between .th economy and a resulL This could lead to forest depletion hat the environment are complex and our under- significantly exceeds natual regenerative standing of them lmited. Idealy one would capacity. wish to trace the effects of economywide policy Overvaluation of thebexchange rate results reforms (both macrocconomic and sectoral) in negative terms of trade, encouraging the pro- through the economic and ecological systems. duction of subsistence crops at the expense of Time and data limitations generally preclude s interatonally tradable products. If these sub- the use of such comprehensive approaches in sistence crops are environmentally harmful, developing countries. Practical policy analysis then cunrency overvaluation leads to environ- is usually.limited to a more "partial equilib- mental degadation. In addition, decreased rium" approach that involves an effort to trace competitiveness of products and lower fann- the most important irnpacts of specific econo- gate prices push small cultivators onto margin- mywide policies, at least qualitatively, and al lands in an aempt to absorb. the effects of wherever possible, quantitatively. the price changes. In a recet review of the links between Macroeconomkcpolicies Macroeconomic - ilcies growth, trade policy, and the environent In the 1980s, debtor countries adopted (Lopez 1991), the author argues that the output emergency stabilization programs that necessi- froin a natural resource such as a forest or tated currency devaluations, controls on capi- fishery (where production depends critically tal, and interest rate increases. When income on the stock) also will be affected by other fac- levels dropped, tax revenues decreased accord- tors (for example, property rights). Thus, if ingly As unemployment inased, govern- trade policy increased- the value of output (for ments fell back upon expansionary financing example, imber or fish exports), then the de- policies, which led to increases in consumer gree of ownership would influence how pro- prices. The effect of such policies on the poor- duction and resource stocis were managed. est. population groups often drove them onto Reactions might range from more investmt in marginal lands, resulfing in soil erosion and de- and maintenance of assets (if environmental sertification. Fuel price incraLses and lowered costs were internalized by owner-users) to rap- incomes combined to cause deforestation and id depletion (when the users had no stake in the reductions in soil fertility, as the poor were resource stock). forced to use fuelwood and animal dung for heating, lighting, and cooking. 2 Environmenral Economics and Valuation in Dewveoment Decisionmnkiag 27 Structiral adjustment intensive hcavy industries sector that has geCI- erally benefited from protective industrial and The conditionality clause embodied in trade olicies. structural adjustmcnt loans (SALs) implem- ented by the World Bank covers many of the Prublic inives:menVlexpendituire reviews elemcents-discussed above: increases in agricul- euralemens discussedcy above:aintiases ind agre - Reductions in public expenditure are an in- at of agpricsturrl subsidics (Sebastian and tegral part of many SALs, and usually emerge Atlofiagricus ltura subsidites (Sebastian and of from recommendations on spending prioritics Alicbusan 1989). Export taxes arc often one of maeiPulcIesenExndtrR- the policies included. Such taxes, which are a means of generating public revenues, result in vieVs (PIIERs). The main purpose of PlIERs is lower agricultural prices. Again, environmental to provide recommendations to governments on impacts depend on die nature of the crops. the size and composition of their spending pro- SALs often call for re&uctins in energy grams and on ways to strengthen local institu- subsidies, to decrease pollution and encourage tions in ways that enhance county capabilities energy conservation, However, such policies to design and implement such programs. They energy onservcon.uHowever,. such policies have also bcen used to carry out basic sector may force poor consumers -to substitute fuel- wokadtienfypjcsaprraefr woodand ninA dug fo liqid uels eheeby work and to identify projects appropriate for- wood and animal dung for liquid fuels, thereby World Bank supr. PlIERs can foran the increasing deforestaion and decreasing soil Wol aksupport. lEscmPr h increasing deforestation and decreasing soil heart of the World Bank's country dialogue, as SALs usually entail the adoption of poli- ere decisions by t coreplanning and finance agencies are central to the key objec- cies designed to (a) promote trade liberalization ofsrcuaadstn,pveyalvi through removal of barriers to external trade, and encouragement of exports; and (b) impl- ation, and sound management of natural resources. ment institutional reform through privatization, Pubc i p n m -on encoument f forign ilrestera, etterPublic investment programs in most coun- - eancouragandement of foregnd investment,bette may not have given adequate weight to en- trainng and marketing, and reform of landena obctvs scmprd o ownership. - : -- ~~~virornmental objcctives as compared--to --- ownership. The benefits and costs of such projects are efficiency and poverty alleviation The poten highly country-specific. There is some concern tial clearly exists for investment reviews to ap- that encouraging foreign investment and privat- propriately elevate environmental concerns and ization might lead to the growth of "polution thereby help to avoid makng iomestnents tht havens," given the weakness of environental ave seous lontrm environenal cse- regulations in most developing countries. Trade quences. The expenditure reviews are perhaps less crucial; nevertheless, these could be used liberalization also could encourage the growth esure, fortele, that conservation- of energy-intensive and/or highly poliuting in- ore, ac n extension pogsetva dustry. However, pollution caused by industri- &ir areocuent ernent pendtues alizaun cold beoffEet bs affoestatofair share of current govermment expenditures.- alization could be offset by afforestation (although this does not necessarily compensate Sectoralpolicies residents of polluted areas), and limited by ap- propriate taxation policies that encouraged the Pratical examples of the efs sectral use of poilution abatement technologies. In a policies have on the environment have been de- recent review of the Lafin Amefican experi- scribed in several studies.' Some countries sub- ence, Birdsall and Wheeler (1991) conclude sidize urban consumers by placing pric - that there is no evidence to show that open ceilhlgs on food. In such cases, the environ- economies are more prone to pollution. They mental consequences will be the same as fot argue that the inflow of foreign technology and Cw- cy on/SllntioD, as both result in low- capital would tend to bing in better pollution d incentives to increase production of in- standards. At the same time, it is the pollution- temnationally tradable crops. 5 See for example: Bsiwanger 1989, Mahar 1989, RepetD 1988, Sebastian and Alicbusan 1989, and Lutz and Young 1990. -28 Lutz and Young (1990) traced.the effects achtally being done in .practice, the question of agricultural policies on the natural resource arises whether it is realistic to expect a more base. They found that where the removal of a thorough treatment of envirornental issues. fertilizer or pesticide subsidv is being consid- Our view is that even in situations where ana- ered in an adjustment program, government ex- Iysts have limited resources at their disposal, penditurcs will decrcase. famners use.of these natural rcsource and environmental issues- may inputs waill decrease, and adverse environmen- be critical ones that can rake a difference be- tal side efibets will tend to diminish.as well. tween success and failure of a project or poll- In the case of Brazil. Binswangier (1989) Cy. Therefore,. depending on the nature of the showed that general tax policies, special tax in- project or policy reform package, a significant centives, the rules of land allocation, and the share of available resources should be devoted agricultural credit sysem all accelerate defor- to enviromnentally oriented economic analysis, estation in the Amazon. These policies also in- preferably early in the project cycic. Even in crease the size of landholdings and reduce the cases wvere the physical impact on the envi- land available to the poor. ronment cannot be easily valued in economic Mahar (1989) traced many of today's terms, techniques of multiobjective analysis problems in the Amazon in Brazil to the deci- may be used to improve the quality of the in- sion in the mid-1960s to provide overland ac- vestnent decision. cess to Amazonia. This decision was made The following section contains a summary before enough was known about the region's of -best practice" that can be pursued to better natural resource base and how to develop it in intgate natual resource and environmiental a sustainable manner. The inial error was issues into economic analyses of projects and c.41pounded by subsequent decisions to pro- policies. Four key aspects are discussed: (a) vide generous incentives to investors willing to determining the physical impacts and valuing undertake environmentally questionable live- the. impacts in monetary terms; (b) mult- stoc:k projects, and more recently, smelting objective decisiomnakin, (c) discounting; and projects in the greater Carajas areas. Official (d) risk and uncertainty. The main emphasis is settlement projects have also contributed to de- given to methods and approaches for valuing forestation although it would be wrong to environmental effects. place all of the blame on the settlers. Pushed bv poverty and skewed land distnrbution in their Valuation of environmentl costs and regions of origin, the settlers have merely re- benefit sponded to incentves created by the govern- The first step in conducti environmen ment in the form of access roads, title to public ly sound-economic analyses is to detemine the lands, various public services, and. in the case environmental and natal resource impacts of of the Transamazon scheme, subsistence the project or policies in question. These physi- allowances. cal impacts (broadly defined to include also biological and social effects) are determined by Environmental cost-benefit comparing the "with project" and the "without analysis for economic project" scenarios. For determining such im- decisionmaking pacts, the economist will have to rely on the ex- Recently, Little and Mirrlees (1990) note peruse of engineers, ecologists, agronomists, social scientists, and other experts. An impor- that from the iid-1970s to 1990, there had tant issue, outside the scope of this paper, is been a rise and decline of project appraisal in i a cp a - - . , - - ~~~~~~~~that such physical mipacts are complex and of- the World Bank and elsewhere", and they con- - t undestood cluded that currty the incentives were inade- T s s i c - -- . . - ~~~~~~~~~~~~The second step in considering envlron- - - quate for. project analysts to undertake. mental effects involves valuing the physical -im- thorough, in-depth evaluations of projects. p a r. A - - ~~~~~~pacts and rd- honsips. An environmenta Given the existing discrepancy between impact can t in a meable change in what ought to or could be done and what is pno<on an&'or change i enviromental ;~~~~~~~~~~~~~~~~~~pouto an/o chng -i enioeta-X--t--.-- -.. 2. Environmenodj Economics andl Valuation in DvELpment Decisionmalng 29 Total Economic Value Use Values Non-Use Values Direct Use Indirect Use Option Beque Eistence Values Values V Values: - : :.- 1~~... 1 I I1 , u - - . 1- ' ' - - - Output that Functioned Future direct Vaiue of Value from can be beneft and Indirect leaWing use and knowledge of consumed us values non-use values continued directly for offspring existence, based on. eog.. moral | . . I , cornvieion * Food E Ecological * Biodiversity - Habitats * Habitats B i3omnass functions * Conserved * Irreversible * Endangered - Recreation * Flood control habitats changes species - Health * Storm protecion Decreasing 7angibiFtt of value to individuals Figure 2-2: Categories of Economic Values Attnrbuted to Environmental Assets quality A number of concepts of value, and TEV = UV + NUV or practical valuation tebniquies have been. devel- TEV = [DUV + IUV + OV] + [EV + BV] oped to tace tde wdfar inpacts of these changes. Figure 2-2 shows this dia on of TEV in schematic forn Below each valuation con- - ~Basc concepts of economic value Basic c cept a short description of its meaning and a ConceptaHlly, ite total economic value few typical examples of the environmental re- (1EV) of a e consistsofits a) useval- sources underlying the perceived value, ar uc (UV) and (b) nonusc value (NUV). Use val- providedL Option values, bequest values, and ues may be broken down firther into the direc existence values are shaded to caution the ana- use value (DTV), the indirect use value QV, lyst concemning some of the ambiiguities asso- and the opon value (OV) (potential use val- ciaed with defing these concepts. As shown ue). One needs to be careful n to double in the examples, they can spring from similar count both the value of indirect supporting or identical resources, and their estimation fumctions and the value of the reultig direct could be interlinked as welt However, theser. use.6 The categones of nonuse value are exis- concepts of value are generally quite distinct tence value (EV) and bequest value (BV)? Option value is based on how much ind u Therefore, we may write: are wiling to pay today for the option of £ For a discssion and example ofthis, see Aylward and Barbir I 992. 7 The vaious teCms in ft equation for TEV may be grouped in somewhat diffrent ways, fer convenience. See, for ac-. 4ampe, Walsh, Loomis, and Gillman 1984. In order to measure wiliess to pay for wildeness protion, they soug to saate (a fizture-oriented) preservation value from recreatiaoal use value (in curmnt use). Accordingy, these u- fiers deed preservation value (PV) as option value plus existence value plus bequest value, that is, PV ='[OV + EV + BV]. .30 - - Table 2-1: Taxonomy of Relevant Valuation Tecbniques Conventional market Implicit market Constructed market Based on actual Change of productivity' Travel cost Artificial market behavior Loss of earnings Wage differences Defensive expenditure Property values Based on potential Replacement cost Contingent valuation behavior Shadow project preServing the asset for fuiture (personal) direct foregoing comiments apply to change of- and indirect use- Bequest value, while exciud- productivity. ing individuals' Own use values, is the value Often, the result of the imnpact cannot be that people derive from 'knowing that others directY related to a nmaket activity. in some of (perhaps their own offspring, Will be abile to these cases, the WiT could be estimated at benefit from the resource in the future. Finally, conventional market value by using.a. closely exiseence value is the percived-value, of the en- related proxy. Care should be eecsdo h vironmcntal asset unrelated either to current or following points: (a) the relevant attributes afZ optional use, that is, simply becuse it exists.81 fected by the eniometlipact might, in A variety of valuation techniques; may be the case of -the proxy measure, be mixed with used- to quantify the above concepts of value- other attributes, thereby.aflecting the value of Thbe basic concept of economi-c valuation un- pmroxy and (b) if the poyattributes are identi- derlying all thezc techniques is the wilingness cal to the ones lost by the impact, the the vat- to pay (WTP) of individuals ibr an environ- ue given by the proxy is only a lower bound for mental service, or resource., i.e., the are under the. true WTP. This approach applies tothe fol- 'the cmestdor Hlcksiant delmand curve.' lowing techniques: loss of earings, defbnsive, As shown in Table 2-1, valuation methods can expenditure, replacement cost, and. shadow be categorized, on the one hand, according to project. which type of market they rely on, and on the In certain cases the WTP can be estimated other hand, by considering how they make use through derivation of a demand function for the of actual or potential behavior, environmental. asset throgh analysis of actua Next, we group the valuation techniques in behavior. Examples of this approach (also Table 2-1 according to analytical method and called surrogate market techniques) in clude provide sOme geneal comments, before dis- travel- cost, wage difirential, and property cussing each -technique in greater dea-il valuiation. Under specific conditions, such a's when The WTP can also be elicited through a the environmental imnpact leads to a marginal controlled experiment or direct interviews, us- change -in the supply of a good or service that ing the artificial market and contingent is bought on a competitive market; the WIT valuation. can be estimnated directy in terms of changes Dietfecsvldononntna valued at prevailingz market prices. If the mar-efet vle o ovetoa ket is not fully competitve~, then the market * zrkt valuation will be a partia measure-, and shad- The prmnary fratur of th e methods consid- ow price corrections may need to be made. The ered in th,is section is. that they are &diety based on miarket prices or productivity. This is 'See for exmple Randall and Stol 1983. For an up4o-dawte epsition, s,ee Braden and Kols,tad 1991. 2. Environmental Economics and Vauation in Developmen DecisioninakiM v 31 possible where a change in environmental involved (also called the hunan capital ap- quality afftcts actual production or productive proach) (sec Frcdriksson and Persson 1989). capability. Actual defensive or preventive expendi- Change in prodctuivity. Development tures. Individuals, firms, and govermments projects can affect production and productivity undertake a variety of "defensive expenditures" positively or negatively. The incremental out- in order to avoid or reduce unwanted environ- put can be valued by using standard economic mental effcts. Environmental damages are of- prices. There are examples of this in the foL- ten difficult to assess, but infbnmation on lowing case studies. In the study on soil con- defi:nsive expenditures may be available or can servation in Lesotho. the increased production be obtained at lesser cost than direct valuations from conserved land is estimated. In the valua- of the environmental good in question. Such tion of I hectare of Peruvian rainforest, the actual expenditures indicate that indviduals, values of difflerent production schemes are firms, or govemments judge the resultant bene- compared Oer examples include impacts on. fits to be greater than the oosts. The de£fnsive tropical wetlands (Barbier et al. 1991) and the expenditures can then be interpreted as a mini- effects of sedimentation on coral diversity and mum valuation of benefits.' However, caution ultimately on fish production (Hodgson and is advisable with this approach, especially in Dixon 1988). cases where defensive expendit=s are arbi- Loss of earnigs Changes in environmen- trarily mandated by governments, with little or tal quality can have significant eflects on hu- no consideration to market fbrces or free man health Ideally, the monetary value of choices by informed economic agents. health impacts should be determined by the P p wiingness to pay, of individuals, for improved P mariceto health In practice, one may have to resort to "second best" techniques such as using fore- Replacement cost Under this approach, gone earmngs m cases of pr a deat, the costs that would have to be incurd in or- sickss, or absenteeism (and increased medi- der to replace a damaged asset are estimatec cal expenditures, which can be considered a - T esthax is not a measure of benefit of type of replacement cost). This approach may avoiding the damage in the first place, since the be relevant, for cxample, when considenng damge cots may be higher or lower than the road and industr plant safety,- and projects replcment cost However, it is an appropriate that affct air pollution in major cities of devel- technique if there is some compelling reason as opmg cuntries. to why the damage should be restored, or cer- The "vahie-of-life approach is often ques- . ny tbat this will occur tioned on ethical grounds It is argued that it T plcment cost appc has been dehumanizes life, which is considered to have applied to protecting groundwater resources m infinite value. In practice, however, socieLy fm- the Philippines, by determining the cost of de- plicily places finite values on human lif and veloping alternative water sources health in policy. and project decisions that af- (Munasinghe 1990c). A second type of appli- fect enviromnental quality, workers' safety, or cation involves estimating erosion prevention health. If this was not so, we would be justified benefis by calculating the value of fertilizer in spending all of GDP on health needed to replace the nutrients lost through soil improvement. erosioa The metiod is only relevant if, in the In the case of an increase or a reduction in absence of erosion control measres, the fertil- the probability of numbers of deaths, an am izer would actoally be applied. Another exam- proximate estimate of value is the loss in esti- ple woud be the cost of an artificial fish mated fitre eamrnings of the individuals "' As mentioned in the section on national income accuts, wrk is ongoing to iden* defensive exeditures Such cx- penditnes by firms are teated in the current stem of national accounts as inumediate costs and am atrefore not part of value added or final output Defensive expenditures by hoaseholds and governments, on the other hand, are tated as final ecpenditur and included in GDP. Prst reseh seeks to addrKshis and othcr issues and inconsi- tencies in the SNA (see Lutz and Mniasinghe 1991). 32 nursery to estimate the value of wetlands that lower than that for more distant users. Analysis might be impaired by a project. The same tech- of the questionnaires enables a demand curve nique of estimating potential ex-post mitigation to be constucted (based on the willingness to expenditures represented by the increased costs pay for entry to the site, costs of geting to the of health care, is used in the tobacco case site, and foregone earnings or opportunity cost study. of time spent) and an associated consumers' Shadow project. When evaluating projects surplus to be dermined. This surplus repre- that have negative environmental impacts, this sents an estunate of the value of the environ- approach involves die design and costing of mental good in question. one or more -shadow projects" that provide for Two of the case studies summarized below substitute environmenta services to compen- use the travel cost method- In one study, the sate for the loss of enviomental assets under travel cost for domestic trips to a forest resr the ongoing projects. This approach is essen- in Costa Rica is used. In the other, a study on tally an instittional judgment of the replace- the value of elephants in Kenya, the travel cost ment cost, and is increasingly being mentioned of tourists from Europe and North America is as a possible way of operationalizing the con- used as one smateof consumer surplus-. cept of sustainability at the project level It as- Properry value. Also referred to as a sumes that maintaining envim capital "hedonic price" technique, the property value intact is a constaint. Its application, therefore, method is based on the general land value ap- could be most relevant when acnticalj emniron- proacht The objectv is to dermine the im- mental asscts are at risk plicit prices of certain chacst of . . - . - ~~~~~~~properties- In the -mrnnnlarea, foir m- Valuation using implicit (or surrogate) stanc, the aim of the method is to place a val- ue on the benefits of environental quality The methods and techniques described in improvements, or to estimate the costs of a de- this section use market infor on indirecly.. terioiation (for example, the efiects of air The appraches discussed here inmlude travel pollution in certain areas)- cost, property value, wa d a and The property value apprach has been used marketed goods as surngates for nonmarkcted to analyze the effects of air pollution in certain goods. Each technique has its particular advan- areas. Where pollution is localzed, the method tgs and disadvanages as well as its specific compares prices of houses in affected areas quiemen for data and resources The task wih houses of equal size and similr neighbor- of the analyst is to determine which of the tech- hood characterstics elsewhere in the same met- niques might be applicable to a particular rOpolhtm area The approach is based on the situation assumption of a competive real estte market, Travel cost. Most oft conned with and its demands on infonnation and statisticalI recreational analysis in industial countries, the analysis are significant Its applicability to de- travel cost method measures the benefits pro- veloping countries is, therefore, limited. duced by recration sites (parks, lakes, forests, Wage 4fferenflal. This metlhod is based on wilderness). A related method can also be used the that in a competitive market the do- to value "travel time" in projects dealing -with mand for labor equals the value of the marginal fuelwood and water collection.' product and that the supply of labor varies with In this method, the area surrounding a site wodring and living conditions in an area. A is divided into concentric zones of incrsing higher wage is therefore- necessary to attract distance. A survey of users, conducted at the workers to locate in polluted areas or to under- site, determines the zone of origin, visitation take more risky occupations. Again, as in the rates, travel costs, and various socio-econonic case of property value, the wage differential characteristics. Users close to the site would be can ody be used if the labor markt is very expected to make more use of it, because its compettive (see earlier discussion on shadow implicit price, as measured by travel costs, is pricing). Other considerations are that ths "For a recent example, see Halny 1989. 2 ERvironmnen olEconomics and Vation in Deveopment DecisioNsaking 33 method relies on private valuation of health done i industrial countries they found that the risks, not necessarily social ones. In this con- overlap of estimates is complete, if accuracy is text, the level of infonnation concening oc- expressed as plus or minus 60 percent of the cu-ational hazards must be high in order for estimates computed. This result provides'some private individuals to make- meanngfil trade- rassurnce that a rigorously applied contin- offs between health risks and remuneration Fi- gent valuation methcod, while not beig very nally, the efl:cts of all factors other than pe, neverdheless can produce valuations environment (for example, skill level, job re- that are of the right order of magnitude and sponsibil, and so ftrth) might i ;nfluence that may be sufficient to rule out certain alter- wages must be elimnatd, to isolate the im- native projects or favor ohers. pacts of environment. The contingLzt valuation metiod has cer- Marketed goods as proxies for nonmar- tain sho in, cluding problems of de- keted goods. In siuations where ai I signig, implementing, and intpreting goods have close substitutes that are marketed, quiomuaires' While its applcability may be the value of the envi t g-d in question limited, there is now considerable experience in can be approximated by the observed market applying this survy-based approach in devel- price of its substitutes. For example, the value oping countres, for example, to evaluate the of a nonmarketed fish variety can be valud at quality . of supply of potable water and the price of the most similar fish being sold in electricity services.13 local mars (Barbier et aL 1991). In certain ontingent valua- tion may be the only available tchnique ftr es- Valuation using consct ed markets imaing benefits. It can be and has bee Contingent valuation. In the absence of applied to common propty resoures, ameity people's preferences as revealed in mad th, e resources with scc, ecologial or other car- coninet valuation method tries to obtain in- actenstics, and to other situations where mar- foiration an consumers' pere. by posing ket information is not avaiable (Mitchell and direc questions about wlliness to pay. It na- Carso 1989) Caution should be exrised in sically asks people what tey are willig to pay seeking to pursue some of the more abstract for a benefit, or what they are willig to accept bhifits of env ment assets such as exs- by way of c on to tolerate a cost (or tence value (ofan asset ht may never be used. both). Ibis process of asking may be itber but pmmises psychic satisfaction merely be tough a direct quoaiesurvy, or by ex- causeitexists)'4. peinental tecniques in which subjects re- Two examples of contingent valuation sur- spond to various stimuli in laboratoiy" veys are given in the case studies The use of conditions What is sought are pesonal vai- fte conint valuton method for estimating tions of the rsondet for ceases or e- alue of viewng elephans m Kenya shows creases in fte quanty of some good, that it is possible to achieve an understanding contigent upon a hypothetical market Will- of the order of magnitudde of he benefits with bigness to pay is constrained by the mtcome only small means. The study described below levd of the respondent, wheres wigns r services in accept payment for a loss is not contaed sother Haiti ts the methodology for diff&r- Estim of willingness to accept tend to be eat biases, provmg its reliability- significantly higher than willirgnessoi-pay Artificial market- Such markets could be esimates- constructed for expeimental purposes, to de Pearce and Makandya (19B9) compared t-nme consumer willingness to pay for a good the tiaent luatli methd with otbe, or sercice For example, a home water purifi- more market-based methods. [n seven stdies cation kdt might be marketed at variao price - L, S-= for exmple seveal papas appeaig mi The Ener Jonal, D _ener 198A. "For examples of Congentaluation sies, se the case studies on Haiti and eya in tbis pape, and in Mu ge 1990a. See forexample'adall and Stoll 1983. 34 levels or access to a game reserve might be of- - such a simple choice between solutions B and fered on the basis of different admission fees, C since the forner is better than the latter with thereby facilitating the estimation of the value respect to objective Z, but worse: with respect placed by individuals on ster purity or on the to Z,. In general, more points (or solutions) use of a recreational facility, respectively, such as B and C may be identified to define the set of all non-dominated feasible solution Multi-objective decision making Multi-objective decisionmaking points that form a 'non-inferior" curve (or The meffiods described. above are used to curve of best options). estimate costs and benefits of a given project in' For an unconstrined problem, fiurther monetary terms. When projects/policies and ranking of alternaivescannot be conducted their impacts are to be embedded in a system of without the introduction of value judgements.- broader (national) objectives,. some of which Specific information has to be elicited from the cannot be easily quantified in monetary terms, decisionmaker to deterrnine the most preferred multi-objective decisionmaking offers an alter- solution. In its most complete fom such in- native approach. formation may be summarzed by a family of Desirable objectives need to be specified. equi-preference curves that indicate the way in These often exhibit a hierarchical structure, which the decisionmaler trries off one objec- The highest level represents the broad overall tive against the other, as illustated in Figure objectives (for example, improving the quality 2-3. The preferred alterative is that which re-. of life), often vaguely stated and, hence, not sults in the greatest utility-which occurs (for very operational. Some of these, however, can contiruous decision variables as showm here) at be broken do wn into more operational lower the point of tngency D of the highest equii- leve objectives (for example, increase income) preference curve, with the Pareto optimal so that the extent to which the latter are met curve. Ln this case, point E (on an even higher may be practically assessed. Somctimes only equi-preference cuve) is not attainable. proxies are available (for example, if the objec- Several multicriteria methods have been tive is to enhance recreation opporunities, the developedL!S Which practical method in par- attbute number of recreation days can be ticular is suitable to detemine the best alterna- used). Although value judgements may be re- tive available depends upon the nature of the quired in choosing the proper attnrbute decision situation. For instance, ineractive in- (especially if proxies are involved) in contrast volvement of the decisionmaker - has proved to the single-criteron methodologies used in useful in the case of problems characterized by economic cost-benefit analysis, measument a large number of decision vanables and com- does not have to be in monetay terms. More plex causal inrlationships. Some objectives explicit recognition is- given to the fact liat a can be dealt with through direct op on, variety of concerns may be associted with while others require the satifaction of a certain planing decisions standard (for example, level of biological oxy- An intuitive understanding of the fimda- gen demand (BOD) not below 5 miflgrams per mentals of multi-objective decisionmag can le). be provided by a two-dimensional graphical ex- The major accomplishent of multi- position such as in Figure 2-3. Assume ti a objective decision modls is that they allow for project has two non-conmensurable and con- more accurate rep on of decision prob- flictng objectives, Z, and 12. Assume firther . lems in the sense that several objectives can be that alternative projects or solutions to the. accounted for. However, a key question con- problem (A, B, and C) have been identified. cerns whose preferences are to be cnsidered. Clearly, point B is superior to (or dominates) A The model only aids a single decisiosmaker (or in terms of both Z, and Z.. Thus, alterative A a homogeneous group). Various interested may be discarded- However, we cannot make groups will often assign difcrent priorities to t3 For an inloductory overview relevnmt to nahtral resour analysis, see Munsinghe 1992. An exa:nsive survey includ- ing refernces to about 150 applications has been done by.Romero and Rlbman 1987. A shorter but more recent suvey is by Petry 1990. X En'ironmentad Economics and Vauation in Devdopmeut _ecid-onmalda 35 the respective objectves, and normally it may Often, the rate of time preference and the not be possible to determine a single best solu- rate of capital productivity are not equal, be- tion via the multi-objective model. Also, the cause of imperfect financial marke and gov- -mathematical framework imposes constrints ermnent distortions intouced by taxation -upon the ability to effectively represent the Also, individual decisions difflr from social de- planning problem. Nonlinear, tchastic, and cisions in that individuals are relatively short- dynamic formulations can assist in better defin- lived, whereas societies persist for much longer ing the problem but impose casts in tenrs of penods. Thus one strong reason for individual complexity in fornulation and solving the mod- preerence for the present-the cetinty of el (Cocklin 1989). death coupled with the uncertainty of when it Nevertheless, in constructing the model the vwMi occur-is absent from the viewpoint of so- analyst communicates infbnnation about the ciety. Tbe conunumity has reason to discount nature of the problem. He specifies why factors the fixture less than individuals. are important and how they interact Liebman The rate of capital productivity is often (1976) observes that "modeling is thinidng very high in developing countries, because of made public" and considers the transfer of the scaity of capital. In the poorer cotries, knowledge as perhaps the most important con- the rate of time preference also is elevated in tribution of modelling. With respect to the se- many cases, because ofthe urg of satisfy- cond point of criticism (that is, diverse ing immediate food needs rather tan ensurng preferences), Liebman suggests that there is long-tm food security (Peare and Turner value to be gained in constructing models from 1990). diflinng perspectives and comparing the Higher discount rates nay discriminate results. against fie generations. This is because projects with social costs occurring in the long Discount rate - - term and net soci benefits occurrmg in fte Discounting is the process by which costs near term, will be favored by higher discount and benefits that occur in different time periods rates. Projects with benefits accing in the may be comparedL The discount rae to be used long run will be less likly to be undertaken un- has been a general problem in cost-benefit der high discount rates. It is therefore a logical analysis", but it is particularly important with conldusion that fixe generations will suffer regard to environental issues, since at least from market discount rates mined by high some of the assoiat costs and beft are rates of acrent geneation time preference long-term in nature. and/or productivity of capital. In standard analysis, past coss and bene- Based on the firegoing, it is oftn argued fits are treated as "suniC and are ignored in de- that discount rates should be lwered to reflect cisions about the present and fixure. Future environmental concerns and issues of tegen- cost and benfts are -discounted to their eratonal equity. However, this would have the equivalent present value and thencompared. In drawback that not only would ecologically theory the interest rate in a perfiet mrnket re- sound activities pass the cost-bnfit test more flects both the subjective rate of time prefer- fiequently, but also a lag number ofprojects ence (of privatc idividuals) and the rate of would genrally pass the test and the resulting productivity of capital These rat are equated increase in investment would lead to additiona at the margin by the market, so that the rate at environmental stress. Norgaard (1991) argues which individuals are.wlling to trade prcsent that low,ring discount ra can in fat worsen for fiture values is just equal at the margin to -environmental degradation-lowering the cost the rate at which thy are able to tranisform of capital and theby lowering the cost of pro- present goods (in the form of foregone con- duction such that more is consumed in the sumption), into future goods (through capital near-term rlative to the case where discount investment). rtes we higher. "For more ddails, see Dasgupta 1972, Hrberr 1976. Uile and Mlics 1974, Mglin 1963, and Sen 1967. 36 Manyv nvironmentalists believe that a zero incorporates these increasing benefits of discount rate should be employed to protect fiu- preservation by including preservation benefits ture generations. However, cmploying a zero foregone w:ithin project costs. The benefits arc discount rate is inequitable, since it would im-. shown to increase through time by the use of a ply a policy of total current sacrifice, which rate of annual growth. While this approach bas runs counter to the proposed aim of eliminating the same effect on the overall CBA as lowering discrimination between time periods-especial- discount. rates, it avoids the problem of dis- -v whvn the present contained widespread pov- torted resource allocations caused by arbitrari- crtv (Pearce 1991). ly manipulating discount rates. Norgaard makes the case that manipulating In summary, the following conclusions discount rates to reflect sustainability concerns -may be reached, witiin the context of environ- results in an inefficient use of capital. Instead mental cost-benefit analysis: (a) the standard he suggests direct income transfers to compen- opportamity cost of capital (for example, 6-12 sate for environmental degradation. He utilizes percent) should be used as for NPV calcula- a general equilibrium model to demonstae ons, and as the comparator when the IRR is that income tiansfers to future generations, computed;(,) eIftbrt,s should be made to ensure through the efficient allocation of resources, re- that compensating investents offset capital sults in new levels of savings and investment, a stock degradation within a framcwork of policy shift in the types of investments, and a diffrent and project decisions; and (c) in the case of interest (or discount) rate. The rate of interest projects leading to irreversible daLage, CBA miay increase or decrease, but this is irrelevant, should be adapted to the extent possile, to in- since it merely serves as an equilibratin price. clude a measurement of the foregone. benefits In order to facilitate such intenerational of preservation in the computab t of costs. transfers, one altemative might be to impose a u sustainability constraint, whereby current well- being is mamized without reducing the wel- All projecs and policies entail some cle- fare of future generations below that ofthe cur- ment of risk and uncerainty. Risks are mea- rent generation. In practice, this would entail sured usually by the probabiities at can be monitoring and measuremnent of capital stocks assigned to the likelihood of occurrence of an (man-made, human, and natural) and an over- undesirable event (for example, an industrial arching investment poicy that sought to ensure accident). Uncertinty descrbes a situation that sating investments offet dpreci- where little. is known about future imnpacts. ation of existing assets (Pearce 1991). The aim Therefore, no probabilities can be assigned to would be to ensure that the oveall stock of definite outcomes, or even the outcomes are so capital (broadly defined) is preserved or en- novel that they cannot be antcipated. hanced for fiture generations. Apart from the. Risk can be teted probabiistically on the previously detailed atternpts to include depreci- basis of known or estimated data, and therefore ation of natural resource stock in national in- ind against and treated like any other proj- come acco , little has been accomplished ect cost. However, uncertainty defies actarial in this area- principles because of undefinable outcomes. In the case of projects leading toirrevers- For example, stratospheric ozone layer deple- ible damage (such as. destruction of natural tion was an unklown outcome of the introduc- habitats, and so on), the benefits of preserva- tion of chlorofluorocarbons (CFCs) and could tion may be incorporated into standard cost- not have been costed as a risk when they were benefit methodology using the Krutilla-Fisher first used. Uncertinty is especially imtportant approach (Markandya and Pearce 1988). Bene- in environental issues. As projects grow lag- fits of preservation will grow over time as the er in scale and introduce new substmnces into supply of scarce environental resources de- the environment, the category of uncertaity creases, demand (faeled by population gowth) looms much larger than risk The proper re- increases, and, possibly, existence value in- sponse to risk is to counw it as a cos in ex- creases. The Kntllai-Fisher approach pected value computations. Howcver, the use 2. Enrironmenfal Econamics aFd Valuation in Development Decisionmaking 37 of a single number (or expccted value of risk) facilitates. a more informcd decision. If in- does not indicatc thc degree of variability or the formation growth is contingent upon the devcl- range of values that might be expected. Addi- opment taking place,- which is unlikely in an tionall, it does not allowv for individual percep- environmental context, then QOV is positive tions of risk. The proper response to -w.hen the uncertainty regards the bcnefits of uncertainty is to procced with caution-if the preservation, and negative when the unccrtainty future cannot be perceived clearly, thcn the is about the benefits of the developmcnt)' speed of advance should be tailored to the dis- Environmental policy fonnulation is com- tance over which the clarity of vision is plicated by the presence of numerous forms of acceptable. uncertainty. As an illustration, Bromley (1989) In practice. the way risk and uncertainty identified six different aspects of uncertainty in are included in project appraisal work is the case of air pollution resulting from acid de- through sensitivity analyses, which detennine position. Thev are (I) identification of the how the 1RR is dependent on different van- sources of p; Iicular pollutants; (2) ultimate ables. Using optimistic and pessimistic values desfination of particular emissions; (3) actual for diffirent variables can indicate which vari- phvsical impacts at the point of destinaton (4) ables wiii have the most pronounced effects on human valuation of the realized impacts at the benefits and costs. Although. sensitivity analy- point of destination of the emissions; (5) extent sis need not reflect the probabilitv of occur- to which a particular policy response will have rence of the upper or lower values, it is useful an impact on the abovementioned factors; and for deternining ws%hich variables are most im- (6) actual cost level and incidence of those portant to the success or failure of a project costs that are the result of policy choice. (Dixon et al. 1988). Much work has been done Bromley suggests that the way in which on the subject of risk and uncertainty in project policymakers address these uncertainties de- appraisal (for a recent treaUment, see Anderson pends on their perception of the existing entitle- and Quiggin 1990). ment strucure. The intrests of the future are The issue of uncertainty plays an important only protected by an entitlement structure that role mi environmental valuation and policy for- imposes a duty on current generations to con- mulation. Option values and quasi-option val- sider the rights of futur generations He terms ues are based on the existence of uncertainty, them '"missing markets," noting that "future Option value (OV) is essentally the premium generations are unable to enter bids to protect that consumers are willing to pay to avoid the their interests? In the absence of such a stnUC- risk of not having something available in the ture, demisionnIakers may tend to follow a polh- future (see the study on valuation of elephants cy that ignores costs to future generations, and in Kenya). The sign of option value depends minimizes costs to current generations at the upon the presence of supply andlor demand un- expense of the future. If the entidmet struc- certainty, and on whether the consumer is risk tue is adjusted, the policymaker can then ex- averse or risk loving. amine three policy instruments to ensure that Quasi-option value (QOV) iz the value of future generations are not made worse offt preserving options for future use in the ex- mandated pollution abatement; full compensa- pectation that knowledge will grow over time. tion for future damages (for example, by taxa- If a development takes place that causes tion); and an annuity that will compensate the irreversible environmental damage, the oppor- future for costs imposed in the present. In the tunity to expand knowledge through' scientific free of uncertainty, the first option would ap- study of flora and faina is lost Unceruinty pearto bethe most efficient about the benefits of preservation to be derived Other important sources of uncertinty through filture knowledge. expansion (which is linked with envirornmental issues include uncer- independent of development) leads to a positive tainty about land tenure, which leads to defor- QOV. This suggests. that the development estation and unsustainable agricultural should be postponed until increased knowledge practices, and uncertainty about resource 1' See Pearce and Turner 1990, and Fisher and Hancanann 1987. 38 rights, which can accelerate the rate of deple- one year only, does not take into consideration tion of a nonrenewable resource. Policymakers the impacts on soil conservation. CaLtulations can address these issues by instituting land re- for maize and sorghum show a negative real forms, and by designing appropriate taxation marginal IRR of 21 and 30 percent, policies that return rcnts to public sources rath- respectively. er than to private agents. An economic cost-benefit analysis is per- fonned, distnguishing between productivity Case studies of valuation of impacts due to increased use of commercial in- environmental impacts puts, and impacts due to increased soil con- In is section, selected case studies from servation. Costs are prinanly labor and developing countries arc briefly outlined to il.- material input costs. Primary benefit categories lustrate the implernentation of some ofithe t include increased production of sorghum and niques of environmerntal impact valuation. maize (incremental yield due to conservation Further details of these case studies are pro- reive to future decline in yield due to ero- rided in Annex 211. These illustrations apply sion), fruit, fuielwood, and fodder. Given vari- primarily to the direct and indirect use value ous assumptions regarding the rate of growti categories shown in Figure 2-1. Because devel- of project implementation factors (see case oping country examples describing aepts to stdy in Annex 2B), the results of the base case estimate option, existence, and bequest values are an NPV of -M 7.0 million at a discount are rare, several studies applied to the indus- rate of 10 percent and -M 5.6 million at 1 per- tiaized countries are-al e cent (Table 2B-2 demonstrates that conserva- tion crop benefits are more significant at a Change in productivity method lower discount rate, as they are slow to * ~~~~~~ ~aeializ). Cost-benefit aaysis of lnd improvement The qualitative interpretation is that the in Lesotho project makes a loss that is significant in com- The Farm lImprovement with Soil Con- panson to the resources invested. In terms of servation (FISC) project was initiated in South- the overriding target of the project, to raise ag- em Lesotho in 1985, with the overriding aim of nculuproduction among smallholder frm- ing households, the project cannot be termed raising agricultnral production among small- scesu.Hwvr h eeiso olcn holder faming households through soil con- successf.io However, the benefits of soil ont- servation measums, subsidized inputs, and dervto mprove elds (FS h apo affioxescfion."' In pursuit of this goal the prog desire to improve crop yields (FISC has a pro- ecthasorehabililated In d pursuit of sthisgoal con duction, as opposed to a preservation, onenta- ect has rehabiltated old terrace structures, con- on). Given (a) demand uncertainty arising structe new oes, andadded ther ~from lack of infonination on fiture population,, for conservation. It has promoted the use of hy- from habits aturn ol and ca- brid maize and sor , the planting of fdder fod habits, agricultural f tcnology, and ca- grasses, and tree planting. It has also promoted pacity to Emport food, and (b) supply'iincer- rotational grazing on communal rangeland. In tainty about the possibility of droughts leading this study, conservation is defined as the to soil losses, it may be advisable to ascribe a promotion of optimum use of land macpositive option value to soil conservation ef- dance with its capability so as to assure its frts, to be icorporated ito the stream of maintensance and improvement,costs and benefits. Again, if long-term protec- A financial analysis is carTied out dhat tion of the land base is assigned a higher pnror- ity relative to immediate productivity increases, compares a high input altemative (implying the thre woud be a basis for more substantial use of commercial fertilizer and bybrid seed) subsidization of cover crops such as fodder with the traditional altenative (implying no use of fbrtilizer and locally saved seeds as opposed glasses at the expense oftraditional cropping. of frtflzer nd lcaR savd seds a oppsedAside from the debatable benefits of its to hybrids). The calculation, which is valid for prdcioa opose to cervton produc"ioneas opposed to conservat1on : - " TTis cm study is derivedfrm Bojo 19 1.---- --.--- ; Fwireanmental Economics and Valuatien in Deni:pnet Decihionm ad- 39: -oriented approach, FISC serves as an impor- Loss- of earnings method tant model in terms of its emphasis on con- sultation and community participation and of Economic analysis of a waler supply and its reliance on low-cost, labor-intensive field healfh program in Zimbabwe techniqcuCs. Fredriksson and Persson (1989) evaluate Valuation of an Amazonian rainforest the Manicaland Health, Water, and Sanitation financial appraisals of .rop~al ~ Program in Zimbabwe using social cost bmefit Most finanaal apprals of trpical i- analwysis. The objectives of the program are to. rests have focused exclusively on timber improe living conditions in-the commnal resources and have ignored the market benefits aras of Manicaland through: (a) imp of non-wood products, thus providing a strong exin g and constucting new supplies market incentivc for ctive loggmg and that ensure an acceptable quantity and quality widespreaid fiorest clearing. In an effert to 11- of water for domestic use and that are reliable lustrate the values of nonwood forest products, and accessible to the conrnunity (b) imp the authors present data concerning iventory, saion cditions by constructng lates production, and current market value for all the and consequently preventing flies from bred- commercial tree species occurring mi 1 ice ing and diseases from being transmitted; and of Amazoniani forest. . The˘y arrive at -a - i ha ed -h of Amazman foest. Thy arrie at a (c) providing health education to. improve hy- combined NPV of US$6,820 for a fruit and la gienie prces and motivate behavioral tex production and selective cutting prqject, with logging contibuting just 7 percent of thelr sed as a total. This compares favorably with an esti- nuraire The shadow price of foreign ex- mated NPV of US$3,184 for timber and pulp- chang was to be Z$ 175IUSSI in wood obtained from a 1-hectare planion in t analysis Brazilian Amazonia, and an NPV! of US$2,960 me shadow price of unskilled labor is set for fully stocked cattle pasturtin Brazil. . zeoi the dry season, due to unemployment In order to extrapolate the value offte in Zimbabwet h the peak rr- there is a project, which is based on the value of 1 of labor. As u y exist be- hectare, it would be useful if attempts were cause of possible delays ia the harvest, a SC- made ;to incorporate elasticitis of demand for tivity analysis is made in the evaluation, wher the products. (Tis approach is being followed the shadow wage of unsked labor is set to for a World Bank study of a similar nature in 100 pe, 75 percen, and 50 percen of the Inia.) . - . - - market wage of Z$ 0.46 per hour. The oppor- Another methodological concern is the use b' walue lalaor is set to of returns per hectare as the Unit of companso durig the whole year. (Tis is a little srpris- between different uses. A recent World Bank i given tht children represent a signifint study concludes that an appropriate analysis laborforceimanydevelopigount ) For would compare remuns per productive umit, in- skilled laor, the shadow wage is set at e cluding land, labor and capital valued at their wage. opportunity cost Since land in the Amazon is The social discount rate is assumed to be generally abundant relative to labor, an analy- the rate of return in the private sector, which sis based -on the returns to laibor would better would mean a real social discount of 4.86 per- predict the market outcome than one which fo cent, sice the commercial bank lending rate to cused on the murns to land. This deduction is the industial secor at the end of 1987 was con-ined by behavior in the arket, where ..14.75 Percent and the inflation rate was 9.89 forest extractivism has tended to vanish wher- peent A time horizon of 40 years was es ever labor has bad reasonable alternatives.20 - lhshed, which was the duration tme of bmenfits from the project Materials supplied by the "This ce study is from Pders etaL 1989. :° WorldBank 1992- project wvcrc valued at market prices. Comnu-- percent hcalth improvement, the intemal rate of nity input is valued at the opportunity cost of return was greater than the social discount rate, unskilled labor. Tha constructor is assumed to and the nct prcsent value of the project was be paid the shadow pricc of skilled labor. strongly positive. At a social discount rate of The change in consumer surplus is deter- 7.24 percent, estimated in the alternative ap- mined by calculating the average price for wa- proach, the project is not found profitable even ter collected in the wct season and dry season, if a shadow wage of 100 percent and 100 per- both before and after the project. The price of cent disease reduction is achieved. water in the wet season is calculated using 75 While most of the benefits come from dis- peak period days with no energy savings but ease reduction, it must be pointed out that with a shadow wage greater than zero. In the benefits are probably underestimated through remaining 105 days energy is saved and the the use of a lower bound. This is because the shadow wage equals zero. Using this infonna- value of a saved lift has been underestimated, tion, the chingc in consumer surplus is then through the usc of a lower bound. Secondly, calculated at shadow wages of 50 percent, 75 the paper does not take into account other percent, and 100 percent. benefits such as local industry that would bene- Cost of illness estimates consist of treat- fit from improved watersupply Third, as in- ment costs, costs of lost production, and costs come distribution will probably irnprove as a for exta transportation. Since willingness to result of the project, benefits could have been pay for healthier and longer lives was not esti- given a greater weight The authors conclude mated, the values used significantly underesti- that the project is probably a success from a mate the true values. For treatment costs, the societal point of view as the necessary increase costs for private treatment are used as the op- in estimated lower bound value for a child's portuity cost A life saved would be valued life Cm the case of 40 percent disease reduction very highly by the individual concemed and his and a shadow wage of 50 percent of the market or her family. From societv's vievwpoint, a low- wage) is rather small. er bound or minimum estimate of the gain from a lit saved is derived by the authors using f- Travel cost and contingent valuation ture production and consumption. The methods weighted average of the wage adjusted for both unemployment and the income in naDomestic consumer surplusfrom vists to lands is used as an estimate of a child's fiutu a raii!forest reserve in Costa Rica production, and fuLture consumption is esti- This study measures the value of ecotour- mated as final household expenditure per capi- ism at a tropical rainforest site in Costa Rica ta. The net present value of the dherence using the travel cost method.2' By observing between production and consumption is the net travel behavior, the authors reveal that Costa output gained by saving the child. The same Rican visitors are wiling to pay US$35 per measure is utilized to determine net present household to visit the site. -The study finds that output produced by an adult's lit if he is saved visitation is highly correlated with education today. Sensitivity analyses were conducted for (and therefore probably income), and that social discount rates of 2 percent, 4.86 percat, - households in areas wiff high population densi- 7.24 percent, and 9 percent; disease reductions ties make more trips of 40 percet, 70 percent, and 100 percent; and The paper only considers domestic visits, wet season shadow wages for unskilled labor although foreign, visitors to the site of 50 percent, 75 percent, and 100 percent of outnumbered domestic visitors by four to one the market wage for casual workers on comn- in 1988. Foreign visitation is likely to be worth mercial farms in Manicaland. far more than domestic, as foreign visitors have In the base case, with an estimated social higher travel costs, and a greater value of trav- discount rate of 4.86 percent, shadow wage el time because of higher earnings. Additional- 100 percent of the marlcet wage, and a 100 ly, they provide foreign exchange. T This sudy is derived fiom Tobias and Mendelsohn 1991. 2. Environmental Economics wad Valuation in DeveJepment Dccisiomeaking 41 Nevertheless, if we use the same value of Valueof viewing elephants on safaris in US$35 per visit for all visitors, this would re- Kenya sult in an NPV of US$1,250 per hectarc. This figure is one to two times the magnitude of the The tvel cost method was used to esti- purchase price currently paid by the reserve for mate a demand function for safiaris in Kenya.23 the acquisition of new lands. - The analysis is based on the 80 percent of It is unclcar, however, whether the authors tourists who come to Kenya froin Europe and assess the cost- (in foregone eanings) of time North. America. Surveys of tour operators and spent at site. It would have. been useful to have visitors provided data for the timation of land a clearer definition of their tenn "cost of travel costs, air fhre, and travel time costs. Travel time." In addition, they. use a linear demand time costs wore weighted at 30 percent to re- functioni as opposed to the more popular semi- fleet thefict that vacation time is valued at log functional form,7 as visitation rates from lower than the gross wage rate. A weighted av- many zones were zero. In. a similar study,. erage consumers' surplus of US$725 is esti- Willis and Garrod (1991) made the case for the mated. This gives a total consumer surplus for superiority of the semilog functional form over those on safari iu the range of US$182 million the linear for e Clawson-Knetsch Zonal to US$218 million annually, depending on the Travel Cost Method. They also found that the assumed level of vistation. zonal travel cost method probably overesti- Toidentifythecontrbutionelephantsmake mated the consumer surplus for their sample of to the value of a safari, tourists were asd to travel cost studies. The individual travel cost alocatedtepleasureandenjoymentoftheirtip method comes closer to contingent valuation over various categores of experience. Ele- results. - . phants represented 12.6 percent of total enjoy- As in the Costa Rican study, most travel ment. Th , the estimated economic value ost studies look at single-,ourpose, sinFle des- - of a safari yields a viewing value for elephants tination trips. A more general methodological of US$23 million to U$27 million per year. problem is how to deal with multiple destina- I order to assess consumers' willingness tion trips. In most cases of intemational tour- to pay to i the elephat population at ism to developing countries.the travel cot current levels through increased enforcemet would need to be attributed to many activities actvih,, a survey was designed, the at a number of sites. The problem then be- contingent valuation approach. Attempts were comes to elicit the specific value given to a cer- made to adJus for biases. The average value tain site. An ongoing World Bank study on was US$89 while the median was US$100 environmental valuation of a forestry develop- Thus yields an annual viewing value of US$22 ment and conservation project in Madavscar million to US$27 million and US$25 to US$30 is attempting to do this?2 Their model uses a million, respectively, based on an estinate of household production framework in: which the 250,000 to 300,000 aduLt safaris per year. household examines cost functions for specific -(his is an example of option: valuethe pre- activities within the potental destination coun- mium-consumers are wiUling to pay to avoid the tries to compare the expectd satisfiction and risk of supply u ait.) cost of trips of various bundles of activities. Note that both fmethods produced annual However, this demands much from the empiri- viewing values for elephants of around US$25 cal data. A survey of tourists is being used to million. Although these estimaes are rough, collect travel cost data as well as itinerary a - they are a useful guide to the order of magni- that includes the distribution of time between tude of value. activities for each individual, the costs of pur- suing the activities, and the features of the van- ous activities tiat lead to differences across individuals in their ability to undertake therm. -K Yramnr et al. 1992. 23 Derived rom Brwn and Henry 1989. 42 ! Willingness to payfor walerservices in Contingent valuation methodto Haiti estimate option, exstence, and bequest The contingent valuation metod w ualUes in this study to estimate consumers' willingness Since hardly any developing country exam- to pay for an improved water system in a vil- pies are available in this category, four U.S. lage in southeem Haiti.'. The project was ex- examples are provided below. An ongoing ecuted by CARE. The research team devised World Bank study seeks to determine nonuse tests in an attmpt to correct biascs that could values of tropical forests in Madascar threaten the validity of the survey results, such (Kramer et al. 1992). as strategic bias, stating point bias, and hypo- The following studies all used the contin- thetical bias. The results of te survey, utilizing gent valuation approach to obtain acual mea- an ordered probit model as opposed to a linear sures for option, existence, and/or bequest model, demonstated that willingness to pay for values: The first was quite possibly the original a new water system (whether for a public survey that developed a methodology to at- standpost or for a private comeCion) was tempt to detemine existence valuz. Both the positively correated to income, the cost of ob- first and second studies, utile revelig the taing water from existing sources, and the sgniflcance of existence values, exammine them education of household members. It was ncga- as an adjunct to their main focus, which is rec- tively correlated with the individual's percep- reational use value. The third is perhaps the tion of the quality of water at the traditional first one that imder to examine total pres- source used before the construction of the im- ervation value in d"pth, broken down into its proved water supply systemL The sex of the re- three separate components of option, eistnce, spondent was staistically significant in the and bequest value. The last study is basically model for public standposu but not in the mod- an attempt to measure the ceffcts of infonna- el for private oimectiont tion disclosure (endangered status and physical The mean of WTP bids for public stand- apparance) on existence values for endangred posts represented 1.7 percent of household in- specirs. come, while the mean WTP bid for private - ' . - . ~~~~~Eidstence vah&e of preseWng vrs5fity connections was 2.1 percent of household K nv comes. These bids are significantly lower than The survey atempts to measure annual the 5 percent rule-of-thumb ofen used in mral household willingness to pay (WI?) to pre- water supply plnnming as an estimate of maxi- serve visibility in the Grand Canyon-bet nrum aability to pay" for private connections. WTP if visibility preservation were to be ex- However, the bids are based on the assumption tended to the entire southwestern parklands re- that the public standposts are already in place. gion and WTP to prevent plume blight seen The results of this study show that it is fiom Grand Canyon National Park.?5 For the possible to obtain reasonable, consistent an- purpose of the study, only the major source of swers in a contingent valuation survey con- air pollution in the region, coal-fired power ducted among a very poor, illiterate population, plants, was the focus. Over 600 households in Contingent valuation has the poteni to be- Denver, Los Angeles, Albuquerque, and Chica- come a viable method in developing countries: go participated in the survey. One-third of the both for collecting infomation on individuals' respondents were asked a pure user value ques- willingness to pay for a wide range of public tion: How much would thcy be willmg to pay infriastructre projects, and for enviromnental in higher entrance fees per day for visibility protuion services (such as the treatment of in-. protecion at the Grand Canyon or oer pats? dustra wastewater flowing through residential The other two-thirds of the respondents were areas). asked how much they would be willing to pay in higher electric power bills to preserve 24 Thiscase study is dived f[an Whitingtou e aL 1990. 2s Derived from Schulz et a. 1983. Z Ensiroxmensal Economics and Valuaion in Devalopmewt Decisionmaking 43 visibility in the parklands, a measurement of levels of supply certainty. Existence values are total preservation value (defined by the authors high for grizly bear (US$24 at 5 years, as thre sum of existence plus user value). The US$15.20 at 15 years), but are significantly authors interpreted existence value as the dif- lower for bighom sheep -(US$7.40 and ference bctwecn total preservation. value and $US6.90, respectively). user value. The preservation value. bids are substan- OfOption, existence, and bequestvaluesof tially higher than the user value bids, apparent- wildeness ly sigmifying that existence value is an. The key question posed here is the amount important component of total economic value. of wilderness to be protected in Colorado. A The authors are careful to point out that visita- sample of 218 resident Colorado households tion plans were not an overwhelming factor in participated in a mail survey.27 ReSpondet determining preservation value bids, and that wie asked to report their willingness to pay knowledge acquired tLrough previous visits into a special find to be used exclusively for was also considered relatively unimportant m the purpose of protecting wilderness. This pay- the determination of bids- Morover, preserva- ment vehicle is recognizable to Colorado resi- tion bids did not decline with distnce, which dents, being similar to the state income tax seems to indicate that nonuse value was an im- form's checkoff for nongane wildlife preserva- portant component-in the respondents' bids. tion. Respondents were asked to write down the Option pricemandeistence valueof - naximum amount of money they would be vild.ife - . - willing to pay annually .for protection of cur- rent wilderness, and for hypothetical increases This study measures the option pnce in wilderness depicted on four maps. Once this (option value plus expected consumer surplus) budget allocation was completed, respondents and existence value of grizzly bears and big- were asked to allocate the highest amount re- hom sheep in Wyoming-both of these species ported among four categories of value: recre- being endangered by human activitv in the. ational use, opton, existence, and bequest area26 A mail survey wvas sent out, with ques- demands. Total-preservation benefits were esti- tons being directed towards hunters and non- mated as the- residual after recreation use bene- hunters. Hunters were asked their WTP for a fits have been subtrated from total WITP for "stamp" allowing thern to hunt in new hunting wilderness protection. Preservation values were areas m eitier five or fifteen years for grizzly estimad by developing an appropriate econo- bear or bighom sheep. Respondents were each metric model of Winess to pay by survey confronted with only one time horizoai. The households, and by aggregating values across probability of supply was variable. Nonhunters households in the state. were asked to specifr their WTP for the exis- Results indicate that as the quantity of wil- tence of the animals or for the opportumity to derness increases, annual household preserva- observe thern in the fiture. tion values increase at a decreasing rate, except As expected the overall option price in- for bequest value, which is linear. Option value creased as the probability of supply increasedc had a strong positive associaton with income. Contrary to expectations, no systneatic reta- In-state wilderness users had a much higher op- tionship could be detennined showing that bids tion value than nonusers, indicating that recre- based on certain demand exceeded those based ational use is an important element in the on uncertain demand. Existence values and ob- determination of option value. Existence value server option prices were significant. The mean was positively related to the importance of bids for. observer option prices were in the preservation of natural scenery, ecosystems, range of US$20 for both rizzly bear and big- and genetic strains. Existence value increased hom sheep, regardless of the time elernent. This with frequency of wildemess trips undertaken. is on a par with option bids for hunters at high All income groups valued existence of 24This study is derivcd from Brookshic 19831 - Derived from Walsh etal. 1984. 44 wi!demcss about equally. Intcrcstingly, a wide windfall gain of US$30 amnong preservation rangc of workers (skilled and unskilled) 'would funds for three animal species, given four sce- pay US$1 50 more for cxistence demands than narios containing different levels of information would persons in other occupations. Bequest about physical appearance and endangered value was not influenced by the number of chil- status. dren living at home with rcspondents. This The effects of information disclosure on re- seems to indicate that bequcst value is correctly sponses vas more evident here. Faced with defined as the satisfiction from interpersonal zero information distinguishing species, the transfers of wildemess to indefinite future gen- subjccts' willingness to pay to preserve each erations rather than specifically to the children specics w%as nearly equal. Given infomiation on of the respondent. Retired persons wrce willing physical appearance, they allocated more to the to pay US$6.15 more for bequest demand tlan monkeylike animal as compared with the rab- were other respondents. All income groups val- bitlike or ratlike animal reflecting a strong an- ued bequest demands about equally. thropomorphic tendency. Given information on The authors coiiclude that, even without endangered status, respondents allocated sig- taking into account the preservation estimates nificaniy more funds to the animal that was of nonresidents of the state, adding preserva- endangered but savable as compared with tion value to the consumer surplus of recre- ubiquitous or extremely rare animals. When in- ational value had a substanial effect on the formation wNas provided on both physical ap- benefit value for w-ilderness. pearance and endangered stats, ihe endangered but savable species received the Eirislence valuwe of endanzgered species Existence value of-endangered species highest allocation followsed by the -rare and This study tests the hypothesis that an indi- abundant species. These results suggest that in- vidual's WTP to preserve a particular animal formation about endangered status may be is significantly influenced by information pro- relatively more important to respondents than vided about the animal's physical and behav- information about phy ical characteistics in ioral characteristics and about its endangered formulating preservation bids. status.m Public awareness about endangered In conclusion, it appears that information species and preservation altematives plays an disclosure can influence perceived marginal ef- important part in detmining the rplicability ficiency of investment in a preservation fimd and usefulness of existence valuation results and thereby result in changes of an individual's The experiment was conducted using iso- budget allocation strategy. lated experimental and control groups of paid university-level student subjects in the United States. CVM wms used to measure preservation One essential step towards achieving eco- bids for a humpback whale prcservation fund- nomically efficient management of natural re- The experimental group was- then provided sources and formulating a practical sategy for with more information about the whales sustainable development, is the effective incor- (through the screening of a film), and both poration of environmental concerns into deci- groups were then questioned again. The exper- sionmaking Traditionally, the economic imental group increased their bids by 32 per- analysis of projects and policies (including the cent from their original values, and the control techniques of shadow pricing), has been devel- group increased their bids by 20 percent This oped to help a country make more efficient use may be attributable to the fact that all respon- of scarce resources. 'External efcts," espe- dents had more time to reconsider their bids, cially those arising from adverse environmental and perhaps demonstrates howv preferences are consequences, often have been neglected. leamed through the interview process itself, This rport has revieved concepts and even in the absence of new inforiation. techniques for valuation of environmental im- Finally, all control and experimental sub- pacts that enable such environmental consider- jects were asked to fully allocate a lump-sum ations to be explicitly considered in the u Thi case study is derived frum Samples et al. 1986. 2 Envirenmental &onomics and Vaiwatien in DLe-vopment Decisionmaking 4S conventional cost-benefit calculus used in eco- Also, one can often identify Xt key nomic decisionmalcing- Even rough qualitative envirometal indicators to which the decision assessments early on in the project evaluation is sensitive and focus attention on them. cycle may facilitate the process of internalizing Some modest evidence exists that the val- these environmental externalities. The advan- uation techniques for determining use values tages include early exclusion of environmental- may be applied successfully in appropriate ly unsound alternatives, more effmective in-depth cases. However, examples involving the es- consideration of environmentally preferable al- timation of nonuse values are virually nonexis- ternatives, and opportunities for redesigning tent in the developing world, and rather scarce projects and policies in order to achieve sus- even in the industiaized nations. The use of tainable development goals. multiobjective decsion methods also needs to Certin specific shortomings and difficul- be explored, where valuation is not easible.29 ties associated with the case studies were dis- Nevertbeless, developing countries are at- cussed earlier. More generally, greater tempting increasingly to both improve and application of the enviromnental valuation con- make use of economic techniques to value envi- cepts and techniques to practical problems in a ronent assets. While the academic literature developing country is required rather than fur- usually focuses mainly on the development of ethr theoretical development Such case study the techniques, there are also sector- or topic- work can be most effectively carried out a3s related approaches?0 For pract s, the im- part of project preparation. A major purpose m portant concen is to keep up with and make this endeavor is not to provide fine-tuned num- use of the advances most relevant to their oiw bers but to indicate orders of magitude. Some areas of application. To facilitate this, a range altematives can be ruled out and gross environ- of publications is included in the bibliographv ntal errors avoided in this fitshion. at the end of this paper. 3 A recently completed World Bank case study, (Meier and Munasinghe 1992), involves enegy and environmental issues in Sri Lanka and utilizes this approach 3 See for cxiampic Dixon et al. 1989, Sherman 1990, and Tuner 1988. 2 Environmental Economics and Valaation in Development Decisionmeking 47 Annex 2A: Using Shadow Prices The estimation and use of shadow prices is cost (MSC) of tiis increased supply is the rele- facilitated by dividing economic resources into vant resource cost If the incremental demand adeable and nontradeable items. Tradeables for the nontradeable results in decreased con- and nontradeables are treated differently. The sumption of other domestic or foreign users, values of directdy imported or exported goods the border-priced marginal social benefit and services are already Imown in border . (MSB) of this foregone domestic consumption prices, that is their foreign exchange costs cor- or of reduced export earnings, would be a more verted at the official exchange rate. Locally appropriate measure of social costs. purchased items whose values are known only -The socially optimal level of total con- in terms of domestic market prices, llowevr, sunmtion for the given input (QJ) -would lie at must be converted -to border prices by multiply- the point where the curves of MSC and MSB ing the former prices by appropriate conversion intersect Pnrce and nonprice distortions lead to fctors (CFs). . nonoptinal levels of consumption Q Q(4 char- aclerized by dfices beween MSB and Border (Shadow) Price = Conversion Factor x MSC. More generally, a weighted average of Domestic (Market) Price MSC and MSB should be used if both effects are present,. The MSB would tend to dominate BP = CFxDP in a short-rni, supply constaed siation, the MSC would be more important in the longer For those tradeables with infinite elastici- run, when expansion of output is possible. ties-of world supply for imports, and of The MSC of nontrdeable goods and ser- world denand for exports-the ost,insurance, vices from many seors can be determined and freight (c.ift) border price for imports and through appropriate decomposition. For exam- the fir-on-board (fohb.) border price for ex- pie, suppose one peso-worth of the output of ports may be used (with a suitaH:e adjustment the construction sector (valued in domestic for the marketing margin). If the relevant elas- prices) is broken down successively into corn- ticitics are finite, then the change in import ponents. This would include capital, labor, ma- costs or export revenues, as well as any shifts terials, and so on, which are valued at pesos in other domestic consumption or producfion Cl, C2,... C. in border prices. Since the conver- levels or in incometrnsfers, should be cosid- sion factor of any good is defined as the ratio ered. The free trade assumption is not required of the border price to the domestic price, the to justify the use of border prices since domes- construction conversion factor equals: tic price distortions are adjusted by netting out n all taxes, duties, and subsidies. CCF = £ C; A nontradeable is conventionally defined as i=l a commodity whose domestic supply price lies between the £o.b. export price and c.iL£ import The standard conversion fiactor (SCF) may pnce. Items that arc not traded at the margin be used with nontadeables that are not impor- because of prohibitive trade barriers, such as tant enough to merit individual attention or that bans or rigid quotas, are also included within lack sufficient data. The SCF is equal to the of- this category. If the increased demand for a ficial exchange rate (GER) divided by the more given nontradeablc good or service is met by familiar shadow exchange rate (SER), ap- the expansion of domestic supply or imports, propriatcly defined. Using the SCF to convert the associated border-priced marginal social domestic priced values into border price 48 equivalents is conceptually the inverse of the up form Allowance must also traditional practice of multiplying foreign cur- be made for seasonal activities such as harvest- rency costs by the SER (instead of the OER) to in8, and ovahead costs liketanort e s. convert foreign exchange to the domestic price Based on the foregoing th efficiency shadow equivalent. The standard conversion fictor may wage rme (ESWR) is given b:- be approximated by the ratio of the official ex- change rate to the fie trd exchange rate ESWR= man + cu. (FTER), when the country is moving toward a fier trade regime:- where m ad ar e foregone margi out- put and overhead Costs of labor in domestic SCF = OER = eX + uM prices, and a and c are correpondiconver- FTER eX(1-tJ + nM(1+t,) sion factors to convert thes values into border pnces. where X = foh. value of exports, If we are interested only in effcency pric- M ci. value of imports, ing, then we may stop herem However, if social e = elasticity of domestic supply of pricing is inportant, consider the effect of -exports, hese cbanges on consumption patters. Sup- n = elasicity of domestic demand fir pose a worker receives a wage W. in a newjob imports, and that the income foregone is We, both in do- -- , = averaWe tax rae on e t mestic prices. Note that Wn may not necessan- (negative for sLbsidy), and ly be equal to the maginal product fore gne t = average tax rate on imports. It couldbe assumed, qiite plausibly, that low- Illustrative of inpora tadeable inps income workes consume the entire increase in used in many development projects are capital income (W, - W). Then this increase in con- goods and petroleum-based fixels Some coun- sumpti.will sult in a resource cost to the ties may have other fiels avaiable, such as economy of b(W. - WJ. The increasod con- natual gas or coal deposits. If no clear-cut cx- sumption also provides a benefit given by port market eist for these indigenous energy w(W - W). where w represens the MSB, in resources, then th cannot be teated lie border prices, of incrgdomestic-ced tradcables. If there is no altemative use for private sector consumption by one umit such fiels, an appropriate economic value is Therefire, the MSC of the production or extraction of gas or coal plus a markup for the disAonted value SW s am + cU + (b -w)(W. - Wj of fiut consumption foregone (or "user cost"). If another high value use eists for The symbol b represents the MSC to the these fuels, the opporunity costs of not using economy, resulting from the use of the in- the resources in alternative uses should be con- creased income. For example, if all the new in- sidered as their economic value. come is consumed, then b is the relevant Two important nontadeable primary fac- consupto cnvers ftor or resource cost tor inputs are labor and land, the next subjec Cm units of the numeraire) of makig available for discussion. lThe foregone output of workers to caners one unit worth (im domestic used in the energy sector is the dominant corm- prices) of the marginal basket of n goods that ponent of the shadow wage rate (SWR). Con- they would purchase. In this case sider a typical case of unskilled labor in a labor n surplus country-for exmple, rual workers b = &.CFi employed for dam construction. Complications i=l- arise in esimating the opportunity cost of la- bor, because the original rural income earned where is the proportion or share of the i th may not reflect the marginal product of a gri- good in-the marginal consumption basket, and cultural labor- Furthermore, for every new job CF, is the cor nding caverion fhctor- created, more than one rual worker-may give 2 Environme nal Economics and Valuation in Development Deelsionmane - _ 49 The corresponding MSB of increased con- might be the flooding of virgin jungle because sumption may be decomposed further, w = d/v, of a hydroelectric dani that would involve the where I/v is the value (in units of the numer- loss of valuable timber, or spoilage of a recre- aire) of a one-unit increase in domestic-priced atioial area that has commercial potential. consumption accruing to somcone at the aver- The shadow price of capital is usually re- age level of consumption (c.). Therefore, v may flected in the discount rate or accounting rate be roughly thought of as the premium attached of interest (ARI), which is defined as the rate to public savings, compared to "average" pd- of decline in the value of the numerair;: over vate consumption. Under certain simplifying time. Although there has been much discussion assumptions, b = I/v. If MU(c) denotes the concerming the cboice of an appropriate dis- marginal utility of consumption at some level c, count rate, in practice the opporunity cost of then d = MU(c)/MU(c.)Assuming that the capital (OCC) may be used as a proxy for the marginal utility of consumption is diminishing, AR, in the pure efficiency price regime. The d would be greater than unity for "poor' con- OCC is defined as the expected value of the sumers with c c c8, and vice vera animal stream of consumption, in border prices A simple form of marginal utilhty function net of replaceient, which is yielded by the in- could be vestment of one unit of pubLic income at the margin M1U(c)=cĥ A simple formula for the social-priced ARt which also includes consumption effects, Thus, d = MU(c)/MU(cJ = (c/c)n - is given by Making the fiuther assuwmption that the dis- ARI =OCC [s + (1 - s)w/b] tribution parameter n =1, gives where s is the fiaction of the yield from the d =cc= ifi ongnnal invetent wnll be saved and reinvest-- where ijf is the ratio of net incomes, which Usjally, the rigorous esimation of shadow may be used as a proxy for the corresponding prices is a long and complex task. Therefore, consumption ratio. the energy sector analyst is best advised to use The consumption term (b-w) in the expres- whater shadow prices have already been cal- sion for SWR disappears if? at the margin (a) culated. Alteratively, the analyst would esti- society is indifferent as to the distribution of in- mate a few important items such as the come (or consumption), so that everyone's con- standard conversion fctor, opporunity cost of sumnption has equivalent value (d=I), and (b) capita, and shadow wage rate. When the data private consumption is considered to be as so- are not precise enough, sensitivity studies may dially valuable as the uncommitted public sav- be made over a range of values of such key na- ings (b=l/v)- tional parameters. The appropriate shadow value placd on land depends on its location. Usually, the mar- ket price of urban land is a usefiil indicator of its economic value in domestic prices, and the application of an approprate conversion factor (such as the SCF) to this domestic price, will yield the border-priced cost of urban land in- puts. Rural land that can be used in agriculture may be valued at its opportnmity costs-the net benefit of foregone agricultural output The narginal social cost of both urban and rural land should reflect the value of associated envi- ronmental assets (see main text). Examples 2 Environmental Economics and Valuation in Devdlopment Decisionmaking -S Annex 2B: Summaries of Environmental Valuation Case Studies Change in Productivity Method' Land improvement in Lesotho Two management options for cultivation of The Fann Improvement with Soil Con- maize and sorghum have been compared mi 5- servation (FISC) project was initiated in 1985 : nancial terms. The high-input alerative im- in Mohale's Hoek district in souLthem Lesotho, plies the use of commercial fertilizer and and is grdualy being expanded.' The project hybrid seed. The 'traditional" alterative im- is now used as a model for a national trainig plies no use of ferilizer and locally saved seeds program in soil conservation. Other soil con- instead of hybrids. The less imnediate impact servation projects are already using FISC as a of soil conservation measures is left out of this model. The choice of the FISC project for calculation, which is valid for one year only. study is fiuther justified by its modern ap- Crop sampling was caried out in coopera- proach in dealing with land dcgradation: pro- tion with the FISC staff during five seasons, duction orientation, labor-intensive techniques,_ 1986-1990. The results show that fiamers us- and popular participation. Furtemore, in- ing a high-input managmt do receive hig;er formation for research could be obined at low yields on average, but that very substantial cost The project area is fairly typical for low- variations of yields make this a risky land Lesotho where most of the crop produc- investment- tion takes place. With some adjustnents the Financial calculations for maize and sor- calculations could be used for other areas in ghum show that the yield must increase by 125 Lesotho, or even for other areas with similar percent and 144 percent, respectively, in order geographic and socio-conomic features -in oth- to achieve a real rate of return of 10 percent er countries. The demands are significantly higher than the The overriding aim of the FISC project is average achieved under high-input managemn nt to raise agricultural production. It has rehabili- as sampled, which give a negative real margi- tated old terrace structures, constructed new al IRR of2l and 30 percent, respectively. oncs, and added other structures for conserva- Maintained participation in the project ap- ton; promoted hybrid maize, hybrid sorghum, pears limnited after the initial boost when con-- and fodder grasses; and planted thousands of servation efforts r.sult in in-kind paymaents. tree seedlings. It has also promoted rotational Most likely, project sales are merely replacing grazing on communa rangeland. The project alteative, less accessible sources of supply. area covers almost 26,000 hectares and reaches There are no convincing signs of a major trans- about 22,000 people- formation of the crop maement regime. The long-term impact of physical conservation Financwal analysis works may be the only net impact as fir as the The financial analysis was done from a major grain crops are concemedi Financial household perspective, using market prices, budgets for fmit and fuelwood trees show more T'his casestudy is derivd from Bojo 1991. 52 promising returns, however, and have also met ways to arrive at real economic costs, howeer. with greater interest among local people. Potential benefits of the project include increas- Possible explanations for the lack of firM- ing production of: er response to the promotion of high-input * -Maize, sogum, and crop residues due to management are discussed, including, among the use of fbrtilizer, hybnd seed and con- other things, land tenure, credit for agricultmual sevation of soil and nutrients investments, and risk pertaning to agricultural * F ( investnents. Most serious is the problem of risk. Crop yields are very unreliable in Lesot- * Fuelwood from pine and other tree species ho. The farers are quite awae of this and will * Fodder grasses such as Eragrostis and (informally) calculate the chances of losing in- Bana grass vested resources. Demands for yield increases * Vegetables fom communal gardens spon- for maize and sorghum have been shown to be sored by the project considerable m- order to reach an acceptable level of financial return (10 perce real rate). It is certainly not irrational of the fncr to g gn adopt a carel approach in the face of these Additional benefit items to consider are: * Training of personnel, and the introduction Economic LGlWsis of improved communal maag with Overall project performance has been re- Po °mfl wtIa-purOect i4pacs corded for the period March 1985 to Dcember - * Off-site physical impacts, such as less silt- 1990. Curren firm plans fir work untl ation of dams, less maitnance cost for mid-1992 have been incorporated, and extrapo- roads and bridges, improved water qualhiy, lations have been made frm past performance etc. in relation to the availability of fiture *fiancial *Seady beefits for the cmmanity at means. -- large as a result of the increase in mcome The analysis distinguishes between produc- - i-m1agrCUlie tivity impacts due to increased use of conuner- QuangtiJftion of costs and benefits. Not cial inputs and to improved soil conservation, all costs are readily available in monetary The distribution, sale, and use of cmmercial terms. Exampies are te temporary loss of soil frtilizer and hybrid seed in the project areas fiom new ungrassed taces, "'loss" of land to has been monitored. In te short term, distibu- -r1aces and other structures, and iceased tion of these mputs has mcreased somewhat as for roads due to er use. La- they are used as in-kind payments for con- bor cost for soil consation is not inluded servation work on individualy controlled land. since it was reported to be negligibk. This Convincing signs are lacking, however, for a view is, however, controversial. Monetary cost Iasting impact in terms of commercial sales or data were taken from project- accounts and their use. Farmers are known to save project- complemented by the executg company's distributed inputs for several years, and the daa for costs paid by the donor agency directly level of use in the project areas is not signifi- toth c candly differt from use in nonproject areas. Mm crop benefits due to hybrid seeds and Therefore, the project cannot be credited with a fi use is assumed to be negligible. Only nse in produtivit y due to inreased use of small quantities of te inputs have been distrib- inputs. uted through the pmjrect and the financial !denitication of costs and benefits analysis showed their use to be questionable. Financial costs are identified through project For those that wanted the inputs they were and donor accounts There are also costs of soil available through odter channels. conservation works- These are borne by the To detami the benefits fom soi con- farmers and have to be estimated scpamtely. s a mber of factors need to be con- Financial costs have to be adjusted in several sidered. First is the issue of whether soil loss Z Environmental Economics and Valu;aon in Devdopment Decisionmaking 53 actually affects crop production at all in this Road and bridge maintenance in the area particular arca. Second, if it does, to what ec- will not be significantly affiected by the im- tent soil loss occurs, and third, how this rate proved land management and soil retention, affects productivity. and no major dams for hydropower or irriga- There is reason to believe fthat soil loss im- tion in the project areas will be affected. mediately affects the average crop production Valuation in economic prices. The Loti - area since the average topsoil depth is esti- (pl. Maloti) is fixed on a par with the intema- mated to be 25 centimeter, a level at which the tionally convertible South African Rand. There water retention capacity is reduced. Research is no black market for Rands or Maloti. This in the area indicates that the annual soil loss is indicates that the distorting impact of foreign roughly 15 tons per hectare on poorly managed exchange restrictions is not significant. The of- soils and 5 tons per hectare for the areas under ficial exchange rates have therefore been used, project-influenced conservation management. based on International Monetary Funci data. In Through comparisons with other studies on 1990 the exchange rate was set to rougbly 3 loss of productivity due to soil loss, a 1 percent Maloti per U.S. dollar. annual decline in yield on nonconserved land Since there is a transfer of income to Le was assumed. sotho of about 20 percent of the import value, The impact of soil conservation on crop inclusive of duties, the value of imported comn- production can be expressed as ponents is multiplied by a factor of 0.8 when going from financial to econonic pnces. - IQ= dYi x AYi xXF-AC1 x PI x CSi Skilled and seni-skilled labor has access to the larg South African labor market and is where IQ = incremental production priced at the financial wage. The project pay- (kilogram), ment of M3.5 per day for unskllled labor is dY = relative crop decline avoided shadow priced at M2.5 which is the generally due to conservation, accepted local wage for daily laborers. AY = the base level of average Local financial pnces were used fir the (14-year) yield for the district crops. -Based on a comparison of the protein (kilogram per hectare), -and energy content of maize residues as com- ZAC = accuuated area under the pared to five substitutes, the approximate price conservation managent (hectare), of M 60 per ton of maize crop residues was PI = project impact the share of AHl derived. The same figure will be applied to sor- - ... -affected by the proect's actions ghum residues. (percent),. The value of fuelwood is based on the val- CS = the share of maize and sorghum ucs and calorific content of its substitutes, dis- respectively of the cultivated land played in Table 2B-1. - - (percent), - - The actual substitution value wrill vary de- t = time index (year 1 ... T), and pending on the household's situation. A rough, i = crop indx The assumption here weighted average could be calculated as is that P1 (project impact) equals 1, as follows: conservation activities are assumed to * 25 percent will substitute dung (0.25 * be nil in the absence of the project.- MlOO61) Up to 1992, more ftan 18,000 seedlings of -75 percent wil substitute brushwood apple and peach tes were to be delivered to * M24.33) farmers in the project area. The survival rate was estimated at 50 percent The number of * Weightedaverage: M43.0 survived fiaelwood trees by 1992-is estimated at As for the period after 1992. a number of roughly 130,000. The fbder benefits are rath- assuptnons have been made. The level of er small due to the small areas planted and the costs is assumed to remain as the average for -opportunity costofthe-land-used - 1985-92, but with all costs associated with 2 This is supported by observations duing field work in aie area before project initiation. 54 Table 2ZR1: Economic Value of Faelwood Substitutes per Cubic Meter Megajoules Equivalent Malot per per kilogram kilograms kilogram Maloti (1987) (1990) Brushwood 16.0 550 0.03 2433 Dung 12.9 682 0.10 100.61 Crop residues 14.0 629 0.02 1&54. Table 2B-2: Discounted Share of Total Benefit Benefit Shares Item I percent 10 percent Sorghum 30 25 Maize . . -3- Fruit 29 38 Fuelwood 11 12 Fodder 1 2 Total 100 110 expatriate (nondomestic) services, such as con- Results and sensitivity analyses sultant fees, back-stopping by the executing company, external evaluation missions, and so- base ase e an NPV of tha out lis rnples aleve of 4 pecs -base case are an NPV of -M7.0 nilion at a- on, taken out This implies a level of 54 percent discount rate of 10 percent and -M5.6 million of the previous average. As for benefits, it is assumed that the projinteal rat of return is .8 ect continues to add newly conserved crop land percent The qualitative interpretation is that the at a perfonnance rate of 50 percent of the pro-incealsttissgfcntnco- vious evel.Howevr, fisn- oe umuaui -project malces a loss thiat is significant in com- vious level. However, fr-om toe accumulated area under conservation manaement reaaicd -panson to the resources invested The present to previous year, a decline rate of 2 percent value of costs is only M9.9 million at 10 per- per year is applied. Another assumption is that cent and Pd 33.6 milion at 1 percent En toxns there will be no lasting impact on theIwevl of of the overiding target for the project as de- use of fiertiizer and hybrid seed Furtbermore, fined by the donor-to raise agricultural pro- it is assumed that distibution of fuit t duction among -snallholder fanning continues, but declines to a level of 50 percent hU s-to project cannot be shown to be of project maximum. For fuclwood it is as- suessful when to benits are related to the sumed that tree planting declines to a level of costs. Table 2B-2 summarizs the benefit cate- 50 percent of project average achievement in gones, wich provide some overvew of the 1986-91. For fodder grasses, the assumption in. relative impoce of various benefit items. the base-case will be that fodder growing s - Each benefit ctegory has been discountod by I lizes at 50 percent of toe level achieved in and 10perentrespectively I1986-91 . - : - - Table 2B-2 shows that fruit is a significant benefit item that deserves more torough * 2. Environmental Economics and Vaiuation in Devdopiment Decisionmakig _ 55 monitoring in the future. This, and conserva- considerably due to the extent of the market tion benefits, make up the bulk of beneft, and the real price would increase for Basotho farm- deserve closer scrutiny in tenns of sensitivity ers as incomes would fall. The value of the testing. project is thus correlated to the size of the GNP The robustness of this base case result is of Lesotho. From both a macro and a micro tested using sensitivity analysis of altemative perspective, the soil conservation program is an assumptions with regard to discount rate, insurance against hard tines. This leaves the post-1992 project perfrmnance, frnit income, decisionmakers with a partially quantified erosion impact on crop yields, future grain problem: is the present value of the options pnces, and distributional wcights. higher than the negative NPV of the stream of The qualitative impression of the sensitiv- costs and benefits that have been valued? ity analysis is that, if the rate of 10 percent is However, soil conservation is not. the only. used, the base case result remains robust, al- possible insurance against declines in migrant though the size of the deficit is changed. If the labor incomes. Lesotho needs food securi%, lower rate of 1 percent is considered acceptable not necessarily more domestic production of as a standard, howeer, the base case result gram. Establishing an economic capacity to could be qualitatively altered by several fitc- buy grain on the world market through other trs, makig the project perform better than ex- development projects may be a more efficient pectec- With the lower rate, the project oould altmative. Leotho has a comparative advan- also be justified using a heavy distibutional tage in inexpensive labor, not in good agrcul- weight refletn the higher margnal utilities of tural land and a beneficial climate. Screening the recipients as a group as opposed to the do- available development project options for their nors as a group. This weigh, however, should profitabily is therefore a usefiu exercise. Fur- then be consistently applied to alteative thermore, the capacity to ensure food security projects. . is dependent on the size ofthe population, a ne- D - -,iscuzssion - -. glected mate which needs urgent attntion in Lesotho. In the quantification of costs, the labor re- While nonagricultural investments may be quiremnt for maitance of conservation more efficient from a macroeconomic,point of structures was omitted. However, if a higher view, the majority of the populaton, and the cost estimate is acceptd, the size of conserva- poorest part, live in the rural areas. If the pri- tion benefits change drastically. It can be mary value of a project like FISC is not so shown that the break-even point for maize is 15 much to raise production immediately but rath- person-days, given a time horizon of 50 years er to protect to a sigaificant degree the land and a discount rate of 10 percent Given a dis- base in the long term, more substantial subsi- count rate of I percent, the maximum labor in- dization of cover crops such as fodder grasses put is raised to 34 days, before the net present at the expense of trditional cropping could be value of conservation benefits approximate justified. Even if fodder could not be commer- zero. Thus, it is possible that this benefit item cially sold, the grower would provide an insur- has been considerably overestimated. Only em- ance service while protecting the- land for pirical measurement in project areas could re- fihre potential uses. The economics of this op- solve this issue. tion need to be worked out. An additional point is option value. Lesot- This study cannot conclusively provide an ho is extremely dependent upon migrant labor answer to the future value of the FISC ap remittances. A high price scenario is designed prmach, but has given some rasons why mas-. to build into the calculation the possibility of a sive, full-scale replication involving a substantial increase in relative prices as a substantial number of expatriate personnel proxy for the somewhat unlikely, but not im- should be avoided. Continued efforts should be possible, event that a mass of migrant workers subject to close monitoring of their effciency are forced to retum to Lesotho. While the in order to justify any further fimding. The nominal price level in fact may not be changed original report also contains a discussion of the 56 income distributional impacts of the project, minimum wage rate, US$2.50 per day.4 Based which is not analyzed herc due to lack of space on earlier studies the transport cost for fruit and direct relevance to the topic of environmcn- and latex was estimated at 30 percent of total tal valuation. market value while extraction cost for timber * . - - - ~~~~~~~~was set at 40 percent of tDtal valuc.. Valuationt of an Amazonian rainforest Most rinancial appraisals of tropical fo- Resls rcsts have focused exclusively on timber ri The market value of the fruit production in sources and have ignored the market benefits of the sample area was almost US$650 per year. nonwood products. This has given a strong in- Annual rubber yields amount to about US$50. centive for destructive logging and widespread Deductng collection and tsportation costs forest clearing. gives net annual revenues from fruits and latex This valuation was based on a systematic of US$400 and US$22, respectively. The net -botanical inventory of I hectare of Peruvian presenlt value (NPV) of this production,.at 5 rainforest along the Rio Nanay near the small percent discount rate and assuming that 25 per- village of Mishana, 30 kilometers southwest of cent of the fruit crop is left in the forest for re- the city of Iquitos. Annual precipitation in the .Dation, is estimated at US$6,330. region averages 3,700 millimeters; soils art T Ihhectare of forest also contains 93.8 cu- predominantly infertile white sands. The inhab- bic meters of mntable timber. lf liquidated itant of Mishana are detibalized indigenous in one felling, this sawtimber would generate a people who make their living practicing shift- net revenue of US$1,000 on delivery to the mg cultivation, fishing, and collecting a wide sawmill. A logging operation of this isity, variety of forest products to sell in the Iquitos however, would damage much of the residual market . - stand and greatly reduce, if not eliminate, fLu- Method and data arte rwenues from fruit and latex trees. The ne financial - from timber eaction The inventory showed 50 ibmilies, 275 would be reduced to zero if as few as 18 trees species, and 842 trees of at least 10.0.centime- were damaged by logging. ters in diameter. Of the total number of trees on Periodic selective cutting. would yield a the site, 72 species (26.2 percent) and 350 indi- maximum of about 30 cubic meters per hectare viduals (41.6 percent) yield products with an every 20 years. With a weighted average price actual market value in Iquitos. Annual produc- of UJS$17.21 per cubic meter and deducting tion rates for fruit tees and palms were either harvest and transport costs, the net revenue is measured from sub samples or estimated from about US$310 at each cutting cycle. The net interviews with collectors. Latex yields were present value would be US$490. tak*en from the liteure. The merchantable The combined NPV of fruit, latex and se- volume of each timber tree was calculated us- lective cutting would be about US$6,820, with ing published regssion equations relating di- logging contnrbuting to about 7 percent of the amter to commercial height total. Timber management appears to be a mar- Average retail prices for forest fiuits were ginal financial option in this forest, especially collected in monthly market surveys. The offi- considering the possible impact of logging on cially controlled rubber prices were used. Four frit and latex trees. independent sawmill operators were inter- viewed to detemine the miU price of each in- Comparisons ber species. The labor investment associated The NPV of this piece of rainforest com- with fruit collection and latex tapping was esti- pares well with other uses of rainforests. Using- mated in person days per ycar based on inter- the same discount rate, 5 percent, the NPV of views and direct observation of local collecting the timber and pulpwood obtained from a tecniques. The harvest cost was based on the 1-hectare plantation of Gmelina arborea in - This case study isderived frm Peters-et al 1989 - All prices giYVC in 1987 U.S. dollars using an exchange rate orfwenty intis to the dola. 2. Em'ironmental Economics and Valuation in Development DecisionmakinfA 57 Brazilian Amazonia is estimated at US$3 184, dairy, and beef products. Approximatcly 90 or less than half that of the forest. Similarly, - percent of the provincial population of 1.2734 gross revenues from filly stocked cattle pas- million (1987) live in rural areas, and 66 per- tures in Brazil are rcported to be US$148 per cent live in cominunal lands. Over 50 percent hectare per year. TIis gives a present value of of the populations under 15 years of age. Ovcr US$2,960. Deducting the costs of weeding, 65 percent of the economically active popula- fencing, and animal care would lower this fig- tion in Manicaland are working in agriculture, urc significantly. Both these estimates are either as subsistence fiarmers or as pernanent based on the optimistic assumption that planta- or seasonal laborers on commercial farms. tion forestry and grazing lands are sustainable The objectives of the program, within the land-use practices in the tropics. framework o& Swedish and Norwegian support Tropical forests perform vial ecological to Zimbabwe's health sector, are to improve services, they are the repository for an incred- living conditions in the communal areas of ible diversity of germplasm, and their scientific Manicaland through: value is irmmeasurable. The results fiom this * Improving existng and constructing new study indicate that tropical forests can also water supplies that ensure an acceptable generate substantial market benefits if the ap- proprate esouces re exloitd an proerlyquantity and quality of water for domestic propriate. resources -are exploited and properly - ueadta r eibeadacsil o use and that are reliable and accessible for managed. the nnity Loss of earnings method-economic * Improving sanitation conditions by con- analysis of a water supply and health structing latrines, and program in Zimbabwe5 * Giving hcalth education to improve hygien- The purpose of the paper is to evaluate the ic practices and instigapt behavioral Manicaland Health, Water, and Sanitation pro- changes. gram in Zimbabwe utilizing social cost-benefit analysis. Data was gathered on cost-benefit analysis, water and sanitation projects, and Domnestic prices were used as a numeraire. health statistics in Zimbabwe. Because the That is, local currency is expressed in Zimbab- health statistics do not fully reflect the mortal- wean dollars (Z$) and foreign currency is ex- ity and morbidity rates, benefits of the pm- pressed in U.S. dollars ($US). posed. water and sanitation program will be Foreign exchange rate. Harberger's for- underestimated. mula is utilized to calculate the shadow price Only communal lands in rural Manicaland of foreign exchange. Since Zimbabwe is a price are studied. No consideration is given to re- taker, the supply and demand elasticities of for- settlement areas or commercial finn areas. The eign exchange can be replaced by the import study considers only two major bene- demand and export supply elasticities. Assum- fits-health improvements and change in con- ing that the export supply elasticity is set to sumer surplus for water. zero, and that no quantitative restrictions exist, the Harberger's formula reduces to Background Tle Manicaland Province, organized in R'=nM(I +T)R=(I +T)R seven districts, is situated i the east of Zimr- n M babwe. It is characterized by a relatively high altitude and a diverse relie, implying vrying where R' = the shadow price of foreign ex- patterns of rainfall, temperature, soils, and change and R the official exchange natural farming regions, and has the highest rate, rainflil in the country. The province is agricul- M=ci.valueofimportsintermsof turally rich and produces forestry, fruit, maize, foreign currency, goundnuts, sunflower, tea, coffee, coton, T = import duties, and * Ths cae study is derived from Fredriksson and Persson 1989. 58 n = elasticity of demand for foreign constructor's payments arc zero. The construe- exchange. tor is paid the shadow price of skilled labor Given the assumption that the export supply $Z50 for a double latrine, and $Z250 for a elasticity cquals zero, the Harberger's formula multicompartment latrine. now approachcs the standard United Nations Benefits Industrial Developmnent Organization tUNIDO) * Change in consumer surplus. The prce of guidelines approach. Using 1987 data, the au- Ca g is c incUatd by mean kCals of energy thors determine that one extra unit of foreign used in walking to and carrying water, the en- exchange can buy goods worth 1.1799 units on - e cosL per.SZI, and the time cost with shad- the domestic market Because of the existence ow wages set at 50 percent, 75 percent, and of capital restrictions in the fonn of quotas in 100 . The change in consuner surplus Zimbabwe, import duties are increased sub- per year is equal to stantially so that tie demand for foreign ex- change equals the supply. Therefore a shadow 1o0[q(p., - pJ) + pi)g]) (pw, - price of foreign exchange of 1.75 is used. 2 Shadow price of labor. Ir the dry season, + 185[ql(pdl - pn3 + &q, (p, - Pa)] there is underemployment in the agricultural -.2 sector, so the shadow wage for labor is set whee , = qut carried home in I day, zero. Given that uncertainty exists. in the peak before the project harvest season, when there is a shortage of la- = quantty carried home in 1 day, bor, the authors conduct a sensitivity analysis, after the project where the shadow wage is set qt 100 percent, p., = pnceJI in the wet season, before 75 percent, and 50 percent of the marrket wage e project of Z$ 0.46 per hour. For skilled laborers, it is. p = priet in the wet season after assumed that the shadow wage meets the mar- the project ket wage. The. opportunity cost of children's Pd = price/I in the dry season, before time is set at zero. . te project The social rate of discount Using Hel- Pa = price/l ihe dry season, afr mer's approach (taking the rate of return i the the project private sector), one would derive a real social The change in consumer surplus per per- discount rate of 4.86 percent If the World son,whenfiillcoverageisreached,isshownmin Bank discount rate of 10 percent were used, Table2B-3. adjusted for foreign exchange rate changes, the real discount rate would be 7.24 percent. There Table 2B-3. Change in consumer surplus per is no official discount rates for govement person, when full coverage is reached projects in Zimbabwe, so the authors con- ducted a sensitivity analysis with the social dis- ... shadw wge count rate ranging from 2 percent to 4 percent- so 1 7S o I I Time horizon. A time horizon of 40 years Om 1.18 1.37 is used in the analysis, given hat benefits are expected to remain as a result of reinvestnent. * Hcalth i cx, d ille esti- Calculalions mates consist of treament costs, costs of lost production, and cost for extra transportation. Cost. The materials supplied by the proj- Tie authors point out that as willingness to pay et are valued at market prices. Communi in- is not talen- into account, the values arrived at put, except for the constructor's payment in maysevlyndrtitetcvalues. case of latrines, is valued at the opportunity For treatment costs, the costs of private cost of labor. Since the shadow wage of un- For treent co sts the ospo-tfl proste A skilled labor has been sd at zero during the dry consultation witi a physician costs $Z10.80, season (when construction takes place), the andthecostofnusingis$Z3.00perhour.The costs of conmnunity input except for the tmpion cost used is $t 0.50 per single .~ ~ ~ ~~~~~~~~~asotto cos usdi Z00prsnl 2. Envirnmengtal Economics and Valuation in Dewdopment Decisionmaeking 59 joumey for all inpatients. The lower bound of VA = (275.00 - 129.41)0)1 + i)' the total value per year of reduced morbidity when full reduction is reached, is shown in Ta- with i social rate of discount. ble 2B-4. In Table 2B-5, the values for different dis- Table 2B-4. Total value in SZ of health im- count rates are shown as. lower bound provements per year after the year 2005 at a so- estimates cial discount rate of 4.86 percent with different * disease reductions and shadowr wages-- disease rdtTable 2B-5. Lower bound gain to society from a .............................. .......................... -saved life in SZ at different discount rates Disease reducLim: S-udo1wage 40 70 100 (peret) X SO 01>36 4,38JSb 6;2S4.S3 . ............. ... _.... _...... ...... .... ... ... ... ......... j;.... .. ... .... percent percent - I - . -.-. diomtme pacnt 50 2,501,9361 4,372,88 6,254,839 ____ Sd!dson ae(aet ___ 2 4.26 7.24 I9 75 2.505,760 4,325,081 G,264,401 f R __ , . (n~~~~~~~~~~~~~~~Cult &8.941l 2k13.77 1,094-56 934.35. 100 2,509,585 4,391,774 6,273,962 C 249 2137 , . Adulk 1307.77 1,131.89 1,011.3i 419.04 The value of alif saved would be ewy (>Syrs) _ _ _ , high to the individual concerned or his/her In the base case, with an estimated social fanily. The authors attempt to establish a mini- discount rate of 4.86 percent, shadow wage mum value by adopting society's viewpoint to 100 percent of the market wage, and a 100 per- detennine the net output gained by saving a cent health improvement, the intemal rate of r- lift. They use the human capital ap- turnwas greater than the social discount rate, proach-based on the stram of average in- . and the net present value of the project was come minus average consumption, discounted strongy positive. At a social discount rate of back to the time of avoided death. 7.24 percent, esimated in the alteratve ap- The average household income m conm- proach, the project is not found profitable even nal lands is SZ 550. If the household consists if a shadow wage of 100 percent and 100 per- of 2 adults and 4.5 children on averge, the in- cent disease reducto is achieved. came received by each adult is $Z275. The au- While most of the benefits come from dis- tors estimte that each child consumes $Z ease reduction, it must be pointed out that 64.71 per year, and that each adult living on benefits are probably underesti-nated because comnmaunal lands consumes $Z 129.41. the value of a saved life has been understa- Assuniing an urban migration rate of 20 ted-through the use of a lower bound. Sec- percent, and an unemployment rate of 20 per- ondly, the paper does not take into account cent, the authors use the weihted average of other benefits such as local industy that would the wage adjusted for unenployment and the benefit from improved water supply. Third, as income in communal lands as an estimate of income distribution wili probably improve as a the child's fiue production: SZ 1307.21. Fu- result of the project benefits could have been tare consumption is estimated as final house- given a greater weight on social equity hold expenditure per capita, $Z 795.13. grounds. Therefore the authors conclude that The net present value of output gained by the project is likely to be a success from a so- preventing a child's death, taking into consider- cietal point of view. ation only production and consumption as- pects, is: E~~~~~~~~ravel cosl and conlingent valziation methods Vc=-64.71I/( + i) + (1307.21 - 795.13)/ Domestic consumer surplus from visits to (1 + i a rainfores: reserve in Costa Rica. This study' measures the value of domestic ecotouism to The corresponding net present value of the privately owned Monteverde Cloud Forest output gained by society, for each adult saved, Biological Reserve (MCFR). The. MCFR is: 60 straddles the continental divide in Cost Rica significant. Higher population densities rusulL and consists of 10,000 hectarcs of rugged tcr- in more trips, which is expected since people rain, the vast majority of which is virgin rain- living in less dense cant6ns probably have forest. Tourism to the rcscrve has increased ncarby rainforests to visit. The higher the illit- markedly over the 18 years of its existence, eracy rate the lower the visitation ratc, which both in tcrms of domestic and foreign visitation indicates that visitation is positively correlated rates, despite the relatively remote locality and with education, and probably permanent difficulty of accessing the site. income. * Data. method, and results. In 1988, 755 The visitation rates predicted in Table out of approximately 3.000 domestic visitors 2B-6 are lower than actual rates, since they left their addrcsses at the rescrve for the oppor- only predict the visitation observed in the sam- tunity to win wildlif photographs. Te sample ple. Adjusting this for the vrhole sample showed a similar geographical distribution as a (3000/755) yields an accurate per capita visita- control samrple and is assumed to be repre- on rate. The linear demand uncti esti- sentative of the true domestic visitor mated in Table 2B-6 suggest that visitation population. would drop to zero only at distances of 328 ki- Costa Rica is divided into 81 cant6ns. lometers and 347 kilometers, respectively, for Each cant6n is treated as an observation. Vis- the two regressions. At the presumed US$0.15 itation rates (number of visits per 1000ooo resi- per Iilometer, this implies a maximumn price dents) were calculated for each cant6n by per visit of US$49 and US$52, respectively. dividing observed numbers of trips by census The consumer surplus for each cant6n is populations. Populations, densities, and illitera- f o uner te demand f antond I beh- cy rates for cach canton were taken from a tween the actual price fr this cantn and the 1986 census. Distances were measured along maximum price. The results are sumnmed the most likely roads between the major pop- across all cant6ns, yielding an annual consum- utabon center of each canton and MCFR. The er surplus estimate of US$97,500 and travel cost was estimated at US$0.15 per kilo- US$116,200, respectively. Given that there are meter. This includes out-of-pocket coss, a about 3,000 Cosh Rican visitors per year, the raction of fixed costs, and the value of travel site is worth about US$35 per domestic visit time. Assuming the real value of tis recreational The demand function for visits - flow remais the same over time and using a (VISITATION RATE) was assumed to be lin- real interest rate of 4 percent, iLe present value ear and to depend on the travel cost of domestic recreation at this site is between (DISTANCE), the population density US$2.4 million and US$2_9 million- (DENSITY) and the illiteracy rate * Discussion. This consumer surplus esti- (ILLITERACY). mate of about US$100,000 per year does not include foreig visitors. Foreign visitors out- VISITATION RATE = a0 + aDISTANCE + numbered domestic visitors by four to one in a,DENSITY + aILLITERACY + e 1988. Foreign visitation is liely to be worth far more than domestic since foreign tourists where c is an error term assumed to be ind- with higher incomes and lack of nearby substi- pendent and normally distnibuted. The model tutes probably value the site more than domes- was estimated using multiple - gressions. The tic visitors. The present value estimate is semi-og fimctional form could not be used on probably too low when considered that visita- this data because the visitation rate from many tion has been growing at 15 percent a year for cant6ns was zero. Two specifications were es- the last five years. Still, if we use the same timated;- with and without illiteracy rates. The US$35 for all visitors and also for the future, results are presented in Table 2B-6. that would mean a net present value of All coefficients have the expected sign. The US$1,250 per hectare. The price that the re- coefficient on price is negative and statistically serve currenty pays to acquire new land is ' This case study is derived from Tobias and Mendelsohn 1991. * 2. lEmvronmengl Economics and Valuation in Kent Declsianmakg 61 Table 2B-6: Willingness-to-pay Bids for Public Standpost and Private Connectiens. Dependent variable: Probability that a household's WTP falls witin a specified interval For a public standpost For a private connection Independent variables: Co- t-ratia Co- I-ratio efficient efficient Intercept .841 1.350 -.896 -1344 Household wealth .126 2.939 .217 4.166 index Household with .064 foreign income .232 .046 .194 (1 if yes) Occupation index (1 if farmer) -.209 . - -.597 -2.541 - Household education 1.818 level .157 2.113 .090 Distance from existing water source .001 5.716 .000 1'949 Quality index of existing source -,072 -2.163 -.099 -2.526 Sex of respondent (1 if male) -.104 -5.41 -.045 -207 Log-likelihood - -206.01 -173.56 Restricted log- likelihood -231.95 -202.48 Chi-square 51.88 57.83 (freedom=7) Adjusted likelihood ratio .142 .177 Degrees of freedom 137 120 62 Table 2B-7: Land, Airfare, and Travel rlme Costs (all moeetary units US$) Region Land Total travel Air Travel Hourly r7mie price cost fare time wage Weighlt cost North America 1,465 1,900 40 hrs 22.50 030% 270 3,635 Europe 957 1,300 18 hrs 2250 0.30% 121 2,378 between US$30 and US$100 per hectare. This In 1988 there were 63,000 visitors from suggsts expansion of protected areas near this North America and 350,000 from Europe reserve is a well-justified investment. Normalizing for population diflrenes gave Finally, it should be noted that the recre- 0.2316 and 0.9826 visitors per 1,000 popula- ational value of sUnding forests is but one of tion for North America -.and Eurape, its potetial benefits. The total value of the for- respectvely. est icludes benefis from renewable harvests The price of safari is defined as the sum of Of many commodities, biological diversity, eco- land travel costs, air fare, and travel fime costs. logical services, and sites for scientific These are summarized in Table 2B-7. Land research. costs were estimated by creating a qualiy Value of viewing elephans on safaris in weighted price index from the tour operators' Kenya --survs. The air fre and travel time were esti- * Travel cost approach. The travel cost mated from visitors' surveys.8 Average annual method can be used to esimate a demand func- income was US$45,000, which gives an esti- dion for going on safari in Kenya. The con- mated hourly wage of USS22.50, is weighted mimer surpluls (CS) is the difference between at 30 percent to reflect that vacation time is what people actaally pay and the maximum less valued tan gross wage rate. they would be willing to pay. This net econom- We now have the mmum requirment to ic benefit from a safari does not show up in estimate a demand curve-two observations of market obsevafions but would be lost to the price-quantity. Expecting demand to be a linear international society if safaris were prohibited - function we have: The analysis is based on a survey taken from samples of the approximately 80 percent P 4,023 - 1,674 Q of the tourists to Kenya who camne either from North America or Europe. The other 20 per- where P is the sum of land and air travel time cent of the tounsts were assumed to have the costs, and Q is holiday visits per 1,000 popula- same average consumers' surplus as the tion. Note that we bave yet to address the prob- sample. lcm of the percentage of visitors on holiday to K-rya that actually goes on sanri. - This study is derived from Brown, Jr., and Hemy 1989. ' The survey contained 17 questons and was distributed at some lodges and given to tourists during paru of May and June 1988. There were 53 respmodents I2 Envirasmentda Economics and Valuadon in Devdopment Decio:mainr 63 Question 10 in survey. Special Fees and Permits- Suppose that the current population of elephants can be maintained if additional foot, vehicle and acrial patrols are operated on a sustained and regular basis in the parks. If these patrols can be supportcd by a special 100 dollar annual permit (or included in each visitor's safari cosL), are you willing to support this permit fee? [iS] NO, I am not willing to.pay $100 for this permit.- [34] YES, I am willing to pay $100 for this perniiL - [ ] I am willing to pay a maximum of for this'permit Given a linear demand curve, where 4,023 photoaphing, d learning about the wildlifr" dollars is the price at wbich demand is driven -made up 50 perce of the plasure according to zero, per person consumer surplus for North to ihe answer In a follow-up question con- America is cening only the enjoyment of the wildlife, te iteiemwes attnrbud 25 perc oftheir wild- C& = 0.5 *(4,023 - 3,635) = US$194 lie pleasumr to seeing African elepbants. Ap- plming the sha of 12.6 perce, atibuted by For Europe it is the vistr to elephants, to the estmated eco- normc value of a safi yields a viewmg value CS = 0.5 * (4,023 - 2,378) = US$822.50 for elephants of 23 to 27 million dollars per Year. it seems reasonable that a safar, a once-in- Contingent valuation approach The tou-- a-lifetime adventure for most North imna, rists' survey contains a seies of contingent most of whom had a very satisfactory experi- vuation questions. One of the questions (see once, would be worth 5 percent more than it bo asks people topay in the fo n of a special costs- It seems plausible that a similar expri- amual permit (or increased safiri cost) of 100 me at less cost would be worth 35 prct dollars which would maintain the ephant pop- more hn the oost for a Europeanm ulation at curent levels thrugh increased en- * Results frm travel cost appmrach. Tfe forcement ac . Sof the weighted average consumers' surplus is about respondents said they would pay 100 domars. US$725. Based on discussions with tour opera- The average was 89 dollars while the median tons and with peel in the economic section was 100 dollars. of the U.S. Embass in Kenya, the number of Some rondets disike translating it- adults going on safari each year was estimated porant quliative expriences into a dollar at between 250,000 and 300,000. This gives a metric and respond with a zero response There total consumer surplus for those on safari in were a substanal number of zero responses. the range of US$182-21 8 million annully, de - However, to maintain a shor qestia, no pending on the assumed level of visiation follow-up questions were askedto distinguiSh To idertify the contribution elephants make "prottingr responde from "gemiine" zero to the value of a safari, tourists on safari were respondents. To diminish the importance of the asked in the tourist survey to allocate the plea- zems, the median value, 100 dollars, has been sure and enjoyment of their tnp over four used instead oftbe aveg. stipulated categories of experience. "Seeing, 64 Respondents could have a strategic bias to on the value placed on different lcvels of ser- give large values if they thought the result vice, and tariffs must be designed so that at would lead to policy decisions they like but least operation and maintenance costs (and pre- would not have to pay for. Respondents may ferably capital costs) can be recovered. A key also put in large values if they regard the ques- concept in such an improved planning mecthod- tion as a sort of referendum in which they vote, ology is that of "willingness to pay" (WTP). as it were, for a broader, perhaps moral issue. Two basic theoretical approaches are However, the largest response to this question available for making reliable estimates of was 500 dollars, less thn 1 percent of the ro- households' WTP. The first, the aihdireeCt ap spondent's income and about 3 percent of the proach, uses data on observed water use be- cost of his safari- There was therefore no havior (such as quantities used, travel times to "trimming" of data. collection points, perceptions of water quality) Starting point bias was not tested for due to assess the response of consumers to difiercnt to inadequate sample size. As to the credibility characteistics of an improved water system. of the median value, 100 dollars, it seems mod- The second, or "direct" approach, is simply to est inasmuch as it is 3 percent of-the total-cost ask an individual how much he or she would be of a safri. If one thinks introspectively about willing to pay for the improved water service. the value over and above te cost of a very sat- 'This survey approach is termed the "contingent is-ring moderately expensive experience, 100 valuation method7 and is the focus of the case dollars does not appear to be a suspiciously study. high numnber and some think it somewhat low. ' The stuWdy area. In August 1986 the re- - Results from contingent valuation method. search team conducted a contingent valuation Combing the median value of willingness-to- survey and source observations in Laurent, a pay of 100 dollars with the estimate of 250,000 village in southern Haiti.9 At the timne, United to 300,000 adult safaris per year yields an States Agency for lInternational Development annual viewing value of elephants of between (U.S. AID) was funding a rural water supply 25 to 30 million dollars- If the mean value of project designed to provide services to about 89 dolars per person is accepted, the viewig 160,000 individuals in 40 towns and villages. value is decreased to between 22 million and The project was executed by the international 27 milion dollars. agency CARE. The affiLiation with CARE pro- Note that both methods produce anmal: vided access to villages and justified the pres- values of around 25 million dollars for viewing ence ofthe team to the local population. elephants. Although the esmates are rougH, Haiti, with two-thirds of the population at they are almost certiy a good guide to the an annual per capita income of less than order of magntude of value. The viewing value US$155 in 1980, provides a field setting simi- of elephants is more likely 25 million dollars Jar to the situation in much of Africa and some annually than 2.5 million or 250 million dol- parts of Asia. In such poor areas, an accurate lars. I does not seem pmdent for Kenya's understinof the willingness of the popula- 1988 Wildlife Management and Conservation tion to pay for rural water services is likely to budget to be under 200,000 dollars when tens be particularly important for sound investnent of millions of dollars in viewig value of ele- decisions. phants alone are at stake. The population of Laurent is about 1,500, Willingness to pay for wate services in predomiaty small farmers with a few people southern Haiti. In rural areas, many of those having regular wage employment. Remittances who are in the service area of new water sup- from relatives and friends are common. More ply systms have chosen to continue with their than 80 percent of the population is illiterate traditional practices. If rural water projects are and malnourishment among cbildren is to be both sustainable and replicable, an im- widespread- proved planning methodology is required that There are seven sources of fresh water includes a procedure for eliciting information within approximately 2 kilometers of most of 'Ihis case stud is derived from Whittington et al. 1990. Z En'irosmcnal Economics and Valfation in Development De sionmak nug 65 the population: one protccted well and six service if the responses from the village are spnngs in dry river beds. The springs provide positive, but that someone else will ultimately only modest amounts of water, and individuals pay for the service, there will be an incentive to often wait more than an hour to draw supplies. overstate WTP. On the other hand, if the indi- The average 3 kilometers round trip to a water vidual believes the water agency has already source can sometimes take several hours. The made the decision to install public standposts preference for clean drinling water is strong, in the village and that the purpose of the survey and people sornetimes will walk considerable is to determine the amount people will pay for distances past alternative sources to collect the service (in order to assess charges), the in- drinking water that is considered pure. dividual will have an incentive to understate the * Research design. Economic theory sug- te WTP gests that an individual's demand for a good is An attempt to estimate the nmagitude of a fimction of the price of the good, prices of the bias was made by dividing the sttdy pop- substitute and complementary goods, the indi- ulation into two groups One group was read vidual's income, and the individual's tastes. an opeg state i to Maximum WTP for a new water system will strategic bias. The group was tDld that CARE vary from household to household and should - - decided to build the new system be positively related to income, the cost of ob- and'ta people would neither have to pay maining wvaxer from existing sources, and tie CARE for the system nor pay money at the education of household members, and negative- public fountins. The second group was read ly correlated with the individual's perception of anoter stment that was accurate but lef The quality of water at the traditional source more questions about the purpose of the study used before the construction of the improved unwe especially con the role of water supply systenm The authors hypothesize the mteflneW m designng a water*e. - that the WTP bids of women respondents The hypothesis was that if individuals would be higher than those of men because aced srWategicaly, then bids firom he second women carry most of the water, but alteraive grup would be lower than bids from the first interpretations are oertaiiCy possible. because the former would ftar that a high bid The research design attmpted to test would result in a higher charge by the cornmu- whether WTP bids are systematically related to nity water committee. die variables suggested by economic theory "Starting-point bias" exists if the initial Different ways of posing the qCcons were pnce m a bidding-game affects the individual's tried. The bidding-gam format worked better final WTP. This could, for example, be die than direct, open-ended questions. The bidding- case if the respondent wanted to please the in- game was very famiiar and easily undstood rviewer and interprted the inital price a because it was similar to the ordinary kind of clue to the "correct bid. To test for starting- bargaining that goes on in local markets of ru- point bias, three difliit versions of the ques- ral Haiti. Tests were also included for the exis- tionnaire were randomly distb each with tence and magnitude of seveal types of threats d inial pces in the bidding game. to the validity of the survey results, such as "Hpdical bias" my arise for two stategic bias, strtn point bias, and hypo- kinds of reons. First, the respondents may thetical bias. not understand or correctly perceive the char- " bStrategic bias" may arise when respon- ateristics of the good being described by the dents believe they may influence an imresmses - interviewer. This has been a particular problem or policy decision by not answering the inter- when the contingent valuation method has been Viewer's questions truthfully. Such strategic used to measure individuals' WTP for changes behavior may influence answers in either of - in quality. For example, it may two ways. Suppose an individual isaskedhow. be difficaut for people to perceive What a much he would be willing to payto have a pub- change in sulfur dioxide or dissolved oxygen lic standpost near his house. If he or she thinls means in tems- of air or water qualit. is the water agency or donor will provide the bias is not lkely i the present ease. The 66 rcspondents were familiar with public water consistent with the source observation for 101 fountains and private water connections and households (85 percent). photos of public standposts built in nearby vil- * Aalysis of contingent valuation bids. lages-were shown during the Iinterview. Fourteen percent of the households answered The second source of hypothetical bias is an answer of "I don't know" in response to the possibility that the respondents do not take WTP question for public standposts; there was the questions seriously and will respond bygiv- a 25 percent nonresponse rate for e WTP ing whatever answer first comes to niind. The question for private connections. The mean fir test for this is the same as for the applicability the bids for the standposts, .7 gourdes per of consumer demand theory: were bids system- monthi (US$1. 14), seemed realistic. - per atically related to the variables suggested by The tesz for strategic bias showed the an- economic theory? -ticipated higher bids for those who had re- * Field procedure. Fieldwork in the village ceived the neutral satement, but the difbence consisted of two parts: household surveys and was not statistically significant (i-statistics of source obsermtion. The majority of households 1.1 and 0.5, respectively, for bids on stand- in Laurent were interviewed (170 questionnair- posts and private connections). On the basis of es completed out of approxinately 225 house- this test, the hypothesis tha respondents were holds). The household interview consisted of not acting sttegically when they answered the four seceions. The fist dealt with basic oc- WTP questions cannot be rejected. cupational and demographic data on the family. The test for starting-poin bias showed The second consisted of a number of specific that the bids did not vary systematically with water-related questions. In the third section the the stating-point The null hypothesis that the enumerator rcad one of the statements used to the samples are from the same population test for strategic bias and showed the photo- cannot be rejected, although the confidence in- graphs of public standposts in other villages. tervals are wide. The respondent was then asked to present bids On the basis of these results, there was no per month for (a) public standposts (assuming reason to attempt to adjust the WTP bids for no pnvate connection) and (b) for a private strategic or starting-point bias. The mean of connection (assuming public standposts were WTP bids fbr the public standposts was 5-7 already installed). The fourth section was a se- gourdes per household per month. Assuming an ries of questions on the health and education of average annual income in Lauent of 4,000 family members and the household's assets gourdes (US$800), the mean bid is about 1.7 (such as radios or kerosene lamps). The latter percent of household income and is significant- was used, along with observatons about the ly lower than the 5 percent rule of thumb often quality of the house itself, as a substitut for used in nrual water supply planning for maxi- expenditure questions, to form a household mum "ability to pay" for public standposts. wealth index. . The mean of WTP bids for private connec- The second part of the fieldwork consisted tions, 7.1 gourdes, was not much higher (2.1 of observing the quantities of water collected percen of household income), but these bids by individuals at all the sources used by the were based on. the assumption that the public population of the village. The objective of these standposts were already in place. observations was to verify the infornation indi- The variations in the bids for public stand- viduals provided in household interviews on the posts and private connections were modeled as sources they used and tie quantities of wate a fimction of the identified explanatory vani- collected. All sou-cs were observed on the ables. The dependent vanable obtainea from same day from sunise to sunset The analysis the bidding game is probably not the maximum of the source observation data for Laurent in- amount the household would be Willing to pay creased the confidence in the quality of the but rather an interval witbin which-the true water-use data obtained from the household in- willingess to pay falls. Linear regression is terviews. Out of 119 observations of trips to not an appropriate procedure for dealig with water sources, the interview responses were such an ordinal dependent variable because the 2 Enviion nentai Economics and Valuation in Developmeu Decisionnalking 67 assumptions regarding the specification of the Aylward, B. 1991. The Economic Vaiue of error term in the linear model will be violated. Ecosystems: 3 - Biological Diversity. An ordered probit model was instead used to LEEC Gatekeeper Series No. GK91-03, explain the variations in WTP bids. IIED/UCL London Environmental Eco The results of the estimations can be seen A dnon.ics Candre. in Table 2B-8. The coefficients -ir all the inde- Eynwa v ., and Et. Balbicr. i992. i Valeling , . . . . ~~~~~~~~~~~EnvimnmcnWa Functions in Developing pendent vanables are im the direction expected. Countics." in Biodivenity and Con- The u-statistics indicate that the variables for servation, Vol. 1, pp. 3-50. household wealth, household education, dis- Barbier, E.B. 1989. The Economic Value of Eco- tance of the household from the existing water systems: 1 - Tropical Wetlands, LEEC source, and water quality are all significant at Gatekeeper Series No. GK89-02, the 0.05 level in both models. The sex of the -IIED/UC London Environmental Eco- respondent was statistically sijgificant in ft nomics Cent model for public standposts, but not in the - 990. The Economics of Controlling model £or private connectront The results - Degradation: Rehabilitating Gum Arabic clearly indicatthatheTbidsSstems in Sudan, LEEC Discussion Pa- -clearly indicate dwthat WTP bids are nm ran-peNoDO3,S/ULonnE- dom numbers but are systematically relaxcd to p No. D oon3 c CL n - vironmental Economics Centre. the variables suggeted by economic theory. 9. 19i. The Economic Value of Eco- T-he ordered probit model can be used to sytsems: 2 - Tropical Forests, LEEC predict the number of households in a commu- Gatekeeper Series No. 0K91001, nity which will use a new source if various IIEDIJUCL London Environmental Eco- prices were charged. Such demand schedules nonics Centre. are precisely the kind of infonnation needed by Baibier, E.B., W.M. Adams, and K Kimmage. planners and engineers to make sound invest- 1991. Economic Valuation of Wetland - -t decisions. Benefits. The Hadejia-Jama'are Flood- ment decisions..-- -- plain, Nigeria, LEEC Discussion Paper * Conclusions. The results of this study No. DP91-01. IIEDMUCL London Envi- suggest that it is possible to do a contngent ronmental Economics Cente valuaton surve among a very poor, ille Barde, J.P, and D.W. Pearce, eds 1991. Valuing population and obtain reasonable, consistent the Environment: Sir Case Studies. Lon- answer. The results strongly suggest that c-ions L ________199 1. "'A Note on the Economic tingent valuation sunles are a feasible method C and B s of Bn ToacoPo- for estimating individuals' willingness io pay - Dt d 5110/9 1. for improved water servces in rural Haiti It j P., E. Lu1 and S. Schwinfest 1992. may also prove to be a viable method fbr col- - 'nate Envizamnent and Eonomic lecting information on individuals' willingness ACcoumingI A Case Study for Papu New to pay for a wide range of public infrastructure -Guia, World Bank Envirmnmnt De- projects and public services in developing partinent Working Paper No 54. Wash- countries. ington, D.C. Bibliography Bartelmus, P., C. Stahmer, and J. van Tongeren. -o1989. hIntegrated Environmentav and - - - - ~~~~~Economic Accounting--Framewoxc for Anderson, D. 1987. The Economics of Afforesr-a-EcnmcAoutg-rewkfr an SNA Satellite System," PReview of In tion - A Case Study in Afiica. World Bank Occasional Paper No. l/New Series. Bator, FJ.. 1957. "General Ecrziibrium, Welfare Baltimore, Md: The Johns Hopkins Uni- snd ntion," American Economic Re versityPres. vcrsit~~~~~~ Press. ~viw,March, pp. 22-59. Atavirj, P 1990. -Soil Erosion and Land Degra- Binswanger, IL 1989. "Brazilian Policies that En- dation in the Norther Thai Uplands, in courage Deforcstation in the Amazon,' J.A Dixon, D.E. James, and PB. Sher- World Bank Envfronmint Departamnt man, eds-, Dhyland Management: Eo,-WolBakEvrnetDptmt rnomc Case Studies. Londo&ana : E S -hWorking Paper No. 16. Washington, D.C. nomic Cas-e Studies. London: Earthscan Brsl ,adD hzlrlg. Publications Ltd. ~~Birdsall, N., and D. Wheeler- 1991. "Opennes - - Publications Ltd. - Reduces Industrial Pollution in Latin 68 America: The Missing Pollution Haven Dasgupta, P., and K.G. Maler. 1990. -The Effect". paper prcsented at the Sympo- Environment and Emerging Development sium on International Trade and the En- Issues," Conference Proccedings of the vironment, International Economics Annual Conference on Dcvelopmen EEco- Departmcnt, World Bank- Washington, nomics. World Bank. Washington, D.C. D.C. November. Dasgupta, P., S. Marglin, and A.K. Sen. 1972. Bojo. J. 1990. 'Benefit-Cost Analysis of the Farm Guidelines for Project Evaluation.: New Improvement with Soil Conservation York: United Nations Industrial Develop- Project in Maphutseng, Mohale's Hock ment Organization (UNDO). District, Lesotho.' in J-A. Dixon, DE. Dixon, J.A. R Carpenter, L. Fallon, P. Sherman, James. and P.B. Sherman, eds., Dryland and S. Manipomoke. 1988. Economic A-Management: Economic Case Studies. Analysis of the Environmental Impacts of London: Earlhscan Publications Ltd. Development Projects London: Earth- 1990. EconomicAnalysis ofAgricul- scan Publications Ltd., in association with tural Developmnent Project& A Case the Asian Development Bank. Studyfrom Lesotho- EFI Research Report, Dixon, LA., and MNMI Hufschmidt, eds. 1986. Stockholm School of Economics. Economic Valuation Techniques for the 1991. The Economicsof LandDegra- Environment: A Case Study Workbook dation: Theory and Applications to Lesot- Baltimore, MD: Johns Hopkis Unvem- ho, Dissertation for the Doctor's Degree ty Press. in Economics. Stockholm School of Eco- Dixon, 1A., DE. James, and P.B. Sherman. 1989. nonics, pp. 259 - 350. The Economics of Drvland Management Braal. LC., and J.B Opschoor. 1990. "Risks in London: Earthscan Publications Ltd. fte Botswvana Range-Cattle System," in . eds 1990. Dryland Management LA. Dixon, D.E. James, and P.B. Sher- Economic Case. Studies. London Earth- man, eds, D)yland Management. Eco- scan Publications Ltd. nomic Case Studies. London: Earthscn Dixon. JA.. and P.B. Sherman. 1990. Economics Publications Ltd. of Protected Areas Washington, D.C.: Braden, JIB., and C.D. Kolstad, eds. 1991. Mea- Island Press: Siring the Demad for Environmental Durojaiye, B.O., and A.E. Ipi. 1988. "The Mone- Quality New York: Elsevier, chapter 2. taly Value of Recreational Facilities in a Bmomley, D.W. 1989. -Entitlements, Missing Mar- Developing Economy: A Case Study of kets, and Environmental Uncmainty," Three Centers in N eria, Natural Re- Journal of Environmental Economics and sourcesJournal, 28(2):3 15-328. Management, VoL 17, 1181-194. Energy Journal. December 1998. Special Issue on Brookshire, D.S, L_S Eubanks, and A. Randall. Electricity Reliability, Vol 9. 1983. -Estimating Option Prices and Ex- Finney, CE., and S Westen. 198& "An Econom- istence Values for Wildlife Resources," ic Analysis of Environmental Protection Land Economics, VoL 59, No. 1. and Management An Enxmple from ihe Brown, G, Jr., and W. Henry. 1989. Thle Econom- Philippincs, The Environmentlist, VoL ic Value of Elephants, Discussion Paper 6, No. E. 89-12, London Envircnmenal Economics Fisher, A., and K Hanemann. 1987. "Quasi Centre. Option Value: Some Misconceptions Dis- Capistrano, D., and C.F. Kiker. 1990. "Global pelled" Journal of Environmental Eco- Economic Influences on Tropical Closed nomics and 1Management, 14, pp. Broadleaved Forest Depletion. -183-190. 1967-1985.7 paper presented at the In- Fleming, W.M. 1983. 'Phewa Tal Catchment ternational Society for Ecological Eco- Management Programme: Benefits and noniics Conference, World Bank. Costs of Forestiy and Soil Conservation Washington, D.C in Napal." in L. S. Hamilton, ed., Foresr Codldin, C. 1989. 'Mathematical Programming and Watershed Development and Con- and Resources Planning I: The Limita- servation in Asia and the Pacific. Boul- tions of Traditional Optimization," Jour- der, CO: Westview Press nal of Environmental Management. Vol. Fredriksson, P., and A. Persson. 1989. The Mani- 28,pp. 127-141. calandh . Health, Water, and Sanitation Program in Zfmbabwe-A Social Cost 2 Environmental Economics und Vuluutiorn in Deveiopment Decisipmmaking- 69 -BenefirA:lnalvxi. Departmenl of lnterna- opnient Economics. World Bank. April tional Economics and Geography. Stock- 26 and 27. hol m School or Economics. Lopez, R. 1991. "The Environmcnt as a Factor of Gigengack. A.R. ct al. 1990. "Global Modellinig of Production: The Economic Growth and Dryland Dcgradation. in J.A. Dixon. Trade Policy Linkage." mimeograph, D.E. James. and P.B. Sherman. cds. op. Economics Department, University of cit. Maryland. College Park. MD. December. GilberL A. 1990. "Natural Rcsource Accounting: Lucnberger. D. 1973. Introduction to Linear and CA Cas Study of Botsvana. in J.A. Dix- on-linear Programming. Reading, MA: on. D.E. Jamcs. and P.B. Sherman. cds., Addison-Wcsley. op. cit. Lutz. E., and M. Munasinghe. 1991. "Accounting Grandstalf S.. and J.A. Dixon. 1986. "Evaluation ror the Environmcnt,j Einance and De- of Lumpinee Park in Bangkok. Thai- velopmilent. Vol. 28, March. pp. 19-21. landL in J.A. Dixon and M.M. Hufsch- Lutz1 E.. and M. Young. 1990. "Agricultural Poli- midt. cds.. op. cit. cies in Industrial Countries and Their En- Hanlcy. N.D. 1989. "Valuing Rural Rccrcation vironmcntal Impacts: Applicability to and Bcenfits: An Empirical Comparison of Comparisons with Dcveloping Nations, Twvo Approaches. Jcurmal of Agricul- World Bank Environmcnt Department tural_Economnicv, Vol. 40(3). pp. 361-374. Working Paper No. 25. Washington. D.C. Harberger. A.C. 1976. Project Evatuation: Cot- MacRae D.. Jr.. and D. Whittington. 1938. lected Papers. Chicago: University of "Assessing Preferences in Cost-Benefit Chicago Press: Analysis: Reflections on Rural Water Hartunian- N.S,. C.N. SmarL and M.S. Thompson. Supply Evaluation in Haiti?- Journal of 1981. The Incidence and Economic Costs Policy Analysis and Management, of* Major Health Impairments. Lexington. 7(2):246-263. MA: D.C. Heath. Mahar. D. 1989. "Govermnent Policies and DeCfor- Hodgson- G.. and J.A. Dixon. 1988. Logging Ver- estation in Brazil's Amazon Region," a s*us F-isheries and Tourismr in Palawan. World Bank publication in cooperation Paper No. 7. East West Center. Honolulu, with the World Wildlife Fund and the HIL Conservation Foundation. Washington. Holmberg- G. 1990. "An Economic Evaluation of D.C. Soil Conservation in Kitui DistricL Keny- Markandya, A., and D.W. Pearce. 1988. a." in J.A Dixon. D.E. James. and PRB. -Environmental Considerations and the Sherman, eds., op. cit. Choice of the Discount Rate in Develop- Kim. S., and J.A_ Dixon. 1986. "Economic Valua- ing Countries," World Bank Environment tion of Environmental Quality Aspects of Departnent Working Paper No. 3. Wash- Upland Agricultural Projects in Korea." ington. D.C. May. in J.A. Dixon and M.M. Hufschrnidt . 1989. Social Costs of Tobacco Smo- eds., op. ciL king- British Journal ofAddiction, Vol. Kramer. RA., M. Munasinghe. N. Sharma. E. 84. No. t0. pp. 1139-1150. Mercer, and P. Shyamsundar. 1992. McConnell, KZE., and LEL Ducci. 1989. Valuing -Valuing and Protecting Tropical Forests: Environmental Quality in Developing A Case Study of Madagascar? paper Countries: Two Case Studies. Prepared presented at the IUCN World Parks Con- for session on -Contingent Valuation Sur- gress. Caracas. Febuarey. -es in Developing Countries." Liebman. J. 1976. -Some Simple-Minded Ob- AEA/AERE Annual Meetings, Atlanta, servations on the Role of Optimization in Georgia. Public Systems Decisionmaking." Inter- Meier. P.. and M. Munasinghe- 1992. faces. Vol. 6. pp. 102-108. -Incorporating Environmental Concens Litle, I.M.D.. and J.A. Mirrlees. 1974. Project into Energy Sector Decisionmanling." Appraisal and Planning for Developing World Bank Environment Department- Countries, New York: Basic Books- Washington, D.C. January. 1990. Project Appraisal and MitcheblL RC.. and R[T. Carson. 1989. Using Sunr- Planning Twentv Years On? paper pres- veys to Value Public Goods: The Contn- ented at the Annual Conference on Devel- gent Valuation. Method. Washington, D.C.: Resources for the Future- 70 Munasinghe, M. 1990a. Electric Power Econom- forest,' Nature, Vol. 339. 29 June, pp. ics. London: Butteworths Prcss. - 6554656. _ ________ .1990b. Energy Analysis and Policy. Petly, F. 1990. "Who Is Afraid of Choices? A London: Buttenvorhs Press. Proposal for Multi-Criteria Analysis as a 1 1990c. 'Managing Water Resources Tool for Decisionmaking Support in De- to Avoid Environmcntal Dcgradation," velopment Planning," Journal ofInterna- World Bank Environment Department tional Development, VoL 2, pp. 209-231. Working Paper No. 41. Washington, D.C. Pezzey, J. 1992. Sustainable Development Con- . 1992. WVater Supply and Environ- cepts: An Economic Analysis. World mental Management, Boulder, CO: West- Bank Environment Paper No. 2. Wash- view Prcss. ington, D.C. Munasinghe, Mt, and E. Lutz. 1991. Phantumvanit, D. 1982. -A Case Study of Water "Environmental-Economic Analysis of Quality. Management in Thailand," in Projects and Policies for Sustainable De- YJ. Ahmad, ed., Analyzing the Options velopment,' World Bank Environment United Nations Environment Programme Department Worldng Paper No. 42. (UNEP) Studies, 5. Oxford, England: Washington, D.C. Pergamon Press Ltd. Muzondo, T.R, K.M. Miranda, and A.L. Boren- Phelps, CE. 1988. "Death and Taxes: An Oppor- berg. 1990. "Public Policy and the Envi- tunity for Subsfitution," Journal of ronment," International Monetary Fund Health Economics, Vol. 7, pp. 1-24. (IM) Working Paper. Washington, D.C. Randall, A., and J.R Stoll. 1982. "Existence Val- Norgaard, RB. 1991. "Sustainability as Intergen- ue in a Total Valuation Framework," erational Equity: The Challenge to Eco- Managing Air Quality and Scenic Re- nomic Thought and Practice," internal sources at National Parks and Wilderness discussion paper, Asia Regional Series, Areas Boulder, CO: Westview Press- Report No. IDP 97, World Bank. Wash- Ray, A. 1984. Cost Benefit Analysis. Baltimore, ington, D.C. MD: Johns Hopldns University Press. Pearce, D.W. An Economic Approach to Saing RepefBo, R. May :988. -Economic Policy Reform the Tropical Forest, LEEC Paper 90-06. for Natural Resource Conservation," London Environmental Economics World Bank Environment Departnent Centre. Working Paper No. 4. Washington, D.C. _______- 1991. "Economic Valuation and the Repetto, P, W. Magrath, M. Wclls, C. Beer, and Natural World," draft subnitted for inclu- F. Rossini. 1989. Wasting Assets: Natural sion in World Development Report 1992: Resources in tfe National Income Ac- Development and tie Natural World. counts. Washington, D.C.: World Re- Washington, D.C.:. World Bank sources Institute. Pearce, D.W., EB. Barbier, A. Markandya, S. Ramero, C., and T. Rehman. 1987. "Natural Re- Barrett, RKL Turner, and T. Swanson. source Management and the Use of Mul- 1991. Blueprint 2: Greening the World tiple Criteria Decisionmaldng Economyj London: Earthscan Publica- Techniques: A Review," European Jour- tions Ld nal of Agricultural Economics, Vol. 14, Pearce, D.W., and A. Markandya. 1989. Environ- pp. 61-89. mental Policy Benefts: Monetary Valua- Samples, KC., J.A. Dixon, and M.M. GoweL dion. Paris: Organization for Economic 1986. -Infotmaion Disclosure and En- Co-operation and Development. (OECD). dangred Species Valuation," Land Eco- Pearce, D.W., and K. Turner. 1990. Economics of nomics, VoL 62, No. 3, August Natural Resources and the Environment. Sassone, P.O. 1977. -Shadow Pricing in CBA: Harvester Whcatshca- Mathematical Pmgrauing and Eco- Peskin, KM., with E. Lutz. 1990. "A Survey of nonic Theory,' The Engineering Econo- Resource and Envimonmcntal Accounting mist, Vol. 22, Spring, pp. 219-33. in Industrializ-d Countries," World Bank Schulze, W.D., D.S. Brookshire, E.G. Walther, K- Environment Deanment Working Paper Keley MacFarland, KA. Thayer, R.L. No. 37. Washington, D.C. Whitworth, S. Ben-David, W. Malm, and Peters, CSvM. A.H. Gentry, and RO. Mendelsohn. - Molenar. 1983. The Economic Bene- 1989. "Valuation of an Amazonian Rain- fits of Preserving Visibility in the 2 Environmental Ecolnoimi-;c and Valuation in Development Decisioniakinug 71 National Parklands of the Southwest." Walsh, RCG., J.B. Loomis, and R.A. Gillman. Natural Resources Journal. Vol. 23, 1984. WValuing Option. Existence and January. Bequcst Demands for Wilderness," Land Scbastian, L. and A. Alicbusan. 1989. Fconomics, Vol. 60, No. 1, Fcbruary. "Sustainable Dvlvcopment: Issues in Ad- Wang, G., W. Wu. and F. Li. 1990. 'Economic justmcnt Lcnding Policics," World Bank Evaluation of Land Levelling in Weibei Environment Dcpartment Working Paper Dry Upland," Shaanxi, China, in J.A. No. 6. Washington. D.C. Dixon, D.E. James. and P.B. Sherman, Society for Promotion of Wastclands Dcvclop- cds. op. cit. mcnt. l990a. "Economic and Social Wang. G.. and L. Han. 1990. "Economic Evalu- Change in a Small Rural Community in ation -or Dryland Peanut Growing with the Dcgraded Lower Shivalik Hill Range Perforated Plastic Mulch," in J.A. Dixon, in North India," in J.A. Dixon, D.E. D.E. James, and P.B. Sherman, ads.. op. James. and P.B. Sherman, eds., op. cit.. cit. . 1990b. tDryland Management Options Whittington, D., I. Briscoe, X Mu, and W. Bar- in Wastelands Development: Jawaja ron. 1990. -Estimating the Willingncss to Block. Rajasthan (India)," in l.A. Dixon, Pay for Water Scrvices in Developing D.E. James. and P.B. Sherman, eds., op. Countries: A Case Study of the Use of cit. Contingent Valuation Surveys in South- Squire, L., and H. van der Tak. 1975. Economic crn Haiti,- Economic Development and Analysis of Projecrs. Baltimore, MD: Cultural Change, VoI. 38, No. 2, January. Johns Hopkins University Press. Whittington. D., A. Okorafor, A. Okore, and A. Thomas, V. 1987. -Evaluating Pollution Control: McPhail. 1989. "Strategy for Cost Recov- The Case of Sao Paulo," in G0S. Tolley ery in the Rural Water Sector A Case and V. Thomas, ads., The Econonmics of Study of Nsukca District, Anambra State, Urbanization and Urban Policies in De- Nigeria.- Draft. World Bank Infiasuc- veloping Countries. Washington, D.C.: ture and Urban Developmcnt Department World Bank. Washington. D.C. Tobias, D., and R Mendelsohn. 1991. "Valuing Willis, KRG., and G.D. Garrod. 1991. "An Mdi- Ecotourism in a Tropical Rain-Forest Re- vidual Travel-Cost Method of Evaluating serve," Ambio. Vol. 20, No. 2, April. Forest Recreation,- Journal of Agricul- Turner, R-K. 1988 -Wetland Conservation: Eco- rural Economics, Vol. 42, No. 1, pp. nomics and Ethics," in D. Collard, D.W. 3342, January. Pearce, and D. Ulph, eds., Economics, World Commission on Environment and Develop- Growth and Sustainable E.vironments: menL 1987. Our Common Future. Ox- Essays in Memory of Richard Lecomber. ford, England: Oxford University Press. London: Macmillan. World Bank- 1992. "Brazil: An Analysis of United Nations Statistical Office (UNSO). 1990. Environmental Problems in the Amazon," Report SNA Handbook on Integrated Environ- No. 9104-DR, Vol.2, Annex 5. Washington, D.C. mental and Economic Accounting, pre- liminary draft, October. van Tongeren J., S. Schwuinfest. E. Lutz, MK Go- mez Luna, and F. Guilien Martin. 1991. 'Integrated Environmental and Econonic Accounting: A Case Study for Mexico," World Bank Environment Department Working Paper No. 50. Washington, D.C. Poverty, Resources, and Fertility: The Household as a Reproductive Partnership Parffa Dasgupta This article applies the theory of resource allocation to rural households in poor countries to see what one may mean by a 'opuzration problem ". It is argued by an appeal to evidence that there is a serious popuation problem in these parts, and that it is related syn- ergistically to poverty, and possibly also to an erosion of the local environmental-resoirce base. One aspect of the problem lies in thot very high fertility rates are experienced by wom- en bearing risks of death that should now be unacceptable. An argutment is sketched to show how the cycle of poverty and high fertility rates can perpetuate within a dwzasty. The one general policy conclusion which may he novel is that a population policy in these parts would not only contain such measures as family planning programmes and increased female education and outside-work opportunities, but also measures directed at the alleviationi of poverty, such as improved credit and savings marlets, and a ready availability of basic household needs, such as water and fueL 74 Income, fertility, and food: the Table 3-1. Fertility Rates in the Late 1980s environmentalist's argument- Totailiiyrae Excepting for sub-Saharan Africa ovcr the India 4.2 past 25 years or so, gross income per head has China 23 grown in nearly all poor regions since the end Sub-Sabun Africa of fte Second World War. Morcovcr, growth Japiu andWestern 1.5-19 in world food production sincc 1960 has indusrial democracies exceeded the world's population growth rate; Souric: World Baik (1990). by an annual 0.8 percent during 1.960-70, by an annual 0.5 percont during 1970-80, and by Cross-sectional curve-fitting on data fiom an annual 0.4 percent during 1980-1987. (See 98 so-called developing countries displays a World Bankl 1984, Table 5.6; FAO 1989, An- declining relationship between the fertility rate nex, Table 2.) This has been associatedl with and national income per head (see Figure 3-1). increass in the value of a number of indicators China, Sri Lana, land and South Korea of well-being, such as the infant survival rate, are outliers, with fertility rates much lower at life expectancy at birth, and literacy. All this their levels of income than would be predicted has occurred in a regime of population growth by the statistical relationship. So are most na- rates substantially higher than in the past. Ex- tions of sub-Saharan Africa outliers, but they cepting for parts of East and South East Asia, all lie on the other side of the curve. Neverthe- modem-day declines in mortality rates baven't less, ceL par. There wvould appear to be a link been matched by reductions in fertilitv. More- between income and fertility. A regional break- over, in a number of places which did experi- dowI of even the Chinese experience displays ence a decline in total fertility rates for a while the general patten: fertility is lower in higher- (for example, Costa Rica, Tbailand, Indonesia incomc reions. (See Birdsall and Jamnison and India), they have stabilized at lvels well 1983.) With the notable exception of China and above the populaiton replacement rote (bhe Sri Lanka, poor counties are a long way from fertility rate at which population would be ex- the so-called demographic transition, that is, pected to stabilize in the long, run; a figure a bit transition to a regime of lonw fertility and low over 2.1). Table 3-1 presents total fertility mortality rates. rates in several countries (and groups of coun- Both time series and, as we have just o- tries). In the late 1980s, total fertility rate in served, cross-section dat are suggestive of a the World Bank's list of low income countries broad "inverse" (more accurately, declining (excluding China and India) was 5.6. The fig- relationship between rilit (and mortit) ures for China and India were 2.3 and 4.2, rates and national income per head. That respectively.2 morulity rates would decline wiffi increasing income is sometiing to be expected. Rising m- come usually carries with it more education, improved diet, better health-care and sanita- tion.3 However, a decline in the fertility rate This article has been adapted from Chapter 12 of the author's book: An Enquiry into Well-Being and Destituion (Oxford: Clarcedon Press), forthcomiing, 1993. In writing this article the auithor has gained much from members of the Environnmental Studies Forum at the institute for International Studies, and of the Workshop on Population and Envi- ronment at the Morrison Istitute for Population and Resources, both at Stanford University, and most especially from discussions with Paul Ehrlich and Marc Feldman. For their comments on the earlier version he is gratefil to Kenneth Arrow, Tony Atkinson, Tommy Bengtsson, Justin Yifu Lin, Karl4GOran MAler, James Mirlees, Sheilagh Ogilvie, Rolf Ohlsson, and Agner Sandmo. 2 Totalferiiliiv m:ae is the number of live births a woman would expect to give were she to live tlhough her child-bearing years and to bear children at eacd agc in accordance with the prevailing age-specific 'erility rates. The measure per- tains to the number of live births, not prcgnancies. To place international figures in perpective, we should observe that in western industrial countries todav, total fertility raLes lie between 1.5 and 1.9. 3 Preston and Nelson (1974) and Pieston (1980, 1986) are illuminating dissections of the evidence bearing on the causes of the declinc in mortality rates during the 20th cerntury. (See Preston, Keyfitz and Schoen 1972, for world tables.) To mc their most striking finding from intemational data has been that reductions in respiratory diseases (e.g. influeza, 3. Povert e siowre," and Fenilit: The HeJwrosfas a Raepro,nd vcd,e Pa,ernhip 75 Total fertilitv rate .~~ ~ ~~~~~~ I 1;f)˘ n . . . . 7 sangIadh Cl * Ethiopla . 6 T Zania SalvadorMe%ICO 6 NHtal Pilippinets _ ._-________-_ . . IZaire1 or fr 92 Io \ . - '. -- 1 g ;tissyp \ wenezueA Norm for 92 demiloponk countris 1972 3 kFiu 3C 1 rtiyn reloooandi to i _ _rie and _ t se . A to uTha iland n.Ae in e ooisof s2e tChina .e d h teden ef do o Htg Xa r t so many oth-r societi.s -evn whilethemotal-umpopossibilities.SNofor9ldevelopin 't quite - .r D $1000 h20a $3000 l $4000 of DOc p 6D0 - . - - k~~~~~~~~~~~nome per capita (198D dollars)=. --Soierce: World Bank (1984), Figure 4,4. -ar Africa), Figure qr1: Fertility in reation to imyme ib developitg countries, 1972 and 1982 psn'tselfievident. Andrtto undevs twhyithars coroinentiallya nelresced in t s ofecnora £ wl b d in bnany societes, and why at the, poor countries. But g iss lcome and its at- sam tim e vdedine has been so sluggish in reted bea fits do nft capture future .- so iany opher sodceties (even wc ile dit mortal- somption poeobioies. So b y don't lage bear ity rat has bem fialing as in modem su- on iht conflutMch of ct dcems people have Shon AfiCa) headc lrulles ito On e of my about population -gre wt ane : te.g purposes in this areicle is to dohelopume theoly enircmnenrtal-esnce basis of 197; wdll- a bit. bein& both of which are in large rneasme di- b teg we have be en eUdng abcr t at e com- rected at the fihtuu. modity ph oductman , not the commodity basis of inhle economists by and large have ne- producl onh Now, sbale stics concs mig past glefocopaes t tadem tes d ogaphers andi ldo- movewints of inome and agsicultural produc1 gists haven'ta Much of dwis discussln n has been tion per head can 9ul) us itm evecnpang whe conducted at aw inreae in he sie og., environmental-resource basis of human scil- Eprliah, ti th and Holdrene 977; Bd iona beinig. 'Rey can evena encourage the thougt- 1988; KeHey 1988; Brown, et al. 1991.) Thus, that human ingenuity can' be guante tD in the liUre on {e envirownent it is a com- solve tfie problems resourcer stress poses for monplace to read.estnates of declining land- growing populations (see Sinnon 1981; Sknion man ratios and of dwidndlg naua-resource and Kahn, eds. 1984). It nmay even explin why - bases associatd wfit inscreases in the size of cnirwena resources have been- so populations at both the global ana- region pncumonia and bronchitis) and infectious and parsatic diseases (cg. tuberculosis, the diarrheas, whooig cough, ca- laria, cholera, diptberia, measles and typhoid) have contibuted equally (about 2516-309%) to tiis decline. (There are ex- cptioms, of course, such as Sri LakL) See Dasgupta (1992b, Chapter 4) for furthecr dismssion of the epideniological , - . tanstio. -' ', " ." .- - ,, .. ' ' .' - 76 levels. That rising numbers in the fac of a fi- policy one needs to pear into the locus of deci- nitc environmental-resource base will have. sionmiaking ovcr. fertility matters. As human catastrophic. effects has been rged eloquently society is currently organized, this locus is in by Ehrlich and Ehrlich (1990), among others. lreeauetehscod5I conclude that it Thc assertion hcrc is that.beyond a point there is, the houschold we must study. Formal ac- arc no substitution possibilities between envi-. counts of fcrtility behaviour haven't placed the rorncental. rcsources (c.g., genetic diversity, household explicitly in the conte-xt of rural pov- fresh watcr, breathable air, and so forth) and crty, and it is poor rural populations that are manuf-actured goods. Now policy-miakers often grwig the fastest. Nor has the household aren't sensitive to ecological processes, which bcen analysed in the context. of enviromnmental- in any case are for the most part 'ill- rsuc use. We widll see ta h otx 4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ understood.4 Moreover, urban decisiomnmakcrs matters. can be relied upon to underestimate the dangers In this article I shall -put resource alloca- of environmental degradation. The social value tion theory to work on findings in applied de- of additional know-%ledge about ecological pro- mography so as to identify what we could cesses is therefore large, -certainly larger than today call the population problem in the poor- its private valuc, and possibly a good dcal larg- est regions of the world, namely, the Indian er than is officially acknowledged. This calls subcontinent and sub-Sahara Africa.. Both frgreater expenditur othacusinad teryand evidence indicate that current rates dissemination of such knowledge. Moreover, as of population growth in these regions are over- long as unicertainty about the value of environ- ly high, and -we will try and identify the direc- mental resources is large and enviromnmental de- tions in which public polic needs to bepu structon is at least partially irreversible, there into effect. In addition to the ecologist's con- is a case for awarding special wemight to keep- cemns,.the population problemn in these parts ngorfuture options open. In the. presenit co- displays itself starkly in the formn of unaccept- tex-t this means we ought to pursue wrhat one able risks of maternal death for poor, illiterate could call a "conservationist" approach to en- women, and newv lives doomed to extreme pov- virornmental policy: we should preserve more erty. This identification has the air of banality. than what standard cost-benefit analysis of the about it. but it needs doing. I have not found in use of~ e nviromnmental resources will wmarrant, the literature an ex~ploration of hotusehold beha- (See Arrow and Fisher 1974; Henry 1974.) viour which yields anything like the 'population This is at the intellectual heart of the ecological problemn we will uncover here. For example, movement. It is a convincing chain of even in the influential and informnative report arguments. oppulaton and devel'ipment by the World Bank (see World Bank- 1984) there are only 3 7E~ ~ Th prbe page - -- 56 -- The problem pages (pp. 54 devoted to the question why Such estimates as are currently on offcir households may be producing too many chlci- about population and resources at the aggre- dren- Moreover, the answer it provides to the gate level are valuable. They alert us to prob- question is rough n limiting: the report says lems. But on their own they aren't a guide to that households can get locked in a Prisoners' action. The reason is that neithr the ratre of Dilemmua gaue over fertility decisions, and that population growth ..nor the intensity of poor folk are typicaliy ignoraint of fa nily plan enviroo nental-rcsource use is given from ning measures. outside. Thecy are deterr aned jontlyr by a cor- More generally, in their search for tht POP- plex combination of opportunities, human mo- ulation problem deNographers have in ghreat tivation,s ecological possibilities, and chance part attempted to locate varieties of externali- factors. To entifyt and affect. population ties in household reproductive activities (se ofIt i reiarkable how little avit caologisi know abcbut ecological matters of importance. s au gratcful to Professoh Paul resslis fortre-itfraen tis tomein. .SLeEhrlic and Ehrlich (1991) foranaccountofthevaluof biological diversity. r thanowver, wo will sy below that fo# fertility analysis in sub-Scon Africa and the Caribbean, thefhousehold isn't a veiy unseful catchoryt 3. Povenp, Resources, and Feitv: The Household as a ReproductiveParinership 77 e.g., Lee and Miller 1991). However, as we dgadation (eg, vanishing sources of water, have just noted, the kinds of externalities they receding sources of fodder and household fuel) have mostly studied are those which yield the is likely to be both a cause and an effect of an Prisoners' Dilenmma. This has been limiting, increase in the net reproductive rate. To the because there are forms of reproductive exter- best of my knowlodge, this last has not been nalities which do not lead to the Prisoners' Di- given any attention in the applied demographic lemma (see Annex 3A). They have been much literatureY But for a weak empirical substanti- neglectod. In any event, we will find that there ation in World Bank (1991), 1 have not been is a lkt more to the population problem than able to locate any empirical work which tests run-of-the-mill externalities, the Prisoners' Di- the thesis. It isn't known how powerful the lemma, and ignorancc (or fear) of birth control mutual feedback is in rural communities we techni-ques. Thus, consider as an example know. Should it prove important, the policies gender-differences in the costs of reproduction. govemments will be urged to pursue for bring- Assume that each successful birth involves a ing down fertility rates will be of a different na- year and a. quarter of pregnancy and - are .from the ones usually espoused (see breastfeeding. (It can be longer if really ex- below). ft would also go some way towards ex- tended breastfeeding is practiced, as in sub- plaining the fact that fertilty rates in sub- Saharan Africa) On making the obvious Saharan Africa have not responded to a decline corrections, we can then conclude that in a so- in infant mortlity rates. The theory deserves ciety where feiale life expectancy at birth is investigation, and it explains why I shall lay 50 years, and where the total fertility rate is 7 stress on it (this is approximately the average figure in sub-Saharan Africa today), well over a third of The household a woman's expected adult lifei would be spent In the standard theory of resource alloca- either canrying a child in her womb, or in tion the household for the most part is treated breastfeeding it And.we haven't allowed for as a unit of analysis. This obviously makes unsuccessful pregnancies.6 Now this assumes sense for a single-membered household, it she survives all deliveries. In most poor coun- doesn't necessarily make sense odterwise. The tries matemal mortality is the largest single standard theory interprets a household's utility cause of death among women in their reproduc- function to be a numerical representaion of the tive years, nutritional anaemia playing a central ranking of options on the basis of which its role in this. In parts of sub-Saharan Africa choice is actually made. This makes for a per- (e.g., Ethiopia), I woman dies for every 50 fectly consistent analysis, but it puts strain on births.7 We y conclude that, at a tota fertil- the normative significance of household choice. ity rate of 7 and over, the chance a.typical Choice may reflect a rnking which systemat- woman entering her reproductive years will not ically favours some members (for example, make it through them is about I in 6. The re- males), and it may discriminate against others productive cycle in a woman's life there in- (for example, higher-birthiorder girls and elder- volves her playing the Russian Roulette. This ly relatives). Household choice assumes strong is one manifestation of the populaton problem.. normative significance only when the underly- It is unacceptable. ing rning of options is based upon a defend- I shall argue that wrong relative prices of able aggregate of each member's weill-being household. goods and. services is as much a But this may not be comunon; at least not when source of the population problem as are eider- the fimily is impoverished, and the stresses and nalities. Among other things,.we will find that strin of hunger, illness and physical weak- for poor rural households enviromnental ness make themselves felt 6 In Bangladesh about 60 percent of a woman's reprductive life is spent in pregnacy or lactation. nTe corresponding figure in Pakistan is 50 percent. See McGuire and Popkin (1989). 7 By way or contrast, we should note that the matenial mortality rate in Scandinavia today is I per 20,000 (see World Bank 1988). For example, Johnson and Lec ads. (1987) has nothing on iL Exceptions arc Dasgupta and Miler (1991) and Neulove and Meyer (1991), which have explored the thesis analytically. 78 The matter is a delicate onc. For the State levels. This is what I shall be involved with to interfbre with household choice and probe its here. inner workings systematically can have cata- At least two parties are involved in a fcrtil- strophic consequences, if recmt evidence in a ity decision. So we may to begin with think of number of poor countries is any guide. Never- the household as comprising two menbers who theless, there are indirect levers the State can may have concern fr each other, but whose pull which do not amount to direct interference motivations aren't the same. As We are study- with what happens inside the household, but ing household decisions, we can ignore children which protect vulnerable members and enable from our analysis. Parents will be taken to act them to reach a stronger strategic position. In on their behalf, to a greater or lesser extent, this article I shall be much concerned -with and for better or worse. As household re- identifying such levers. But modelling house- sources are inevitably scarce, individual well- hold fertility decisions isn't for the faint- being interests can never be entirely congruent. hearted- the motivations of the two central par- However, it is the members' motivations with ties to a decision may well be quite diflhrentz which we are concerned here; and the pair's Fortunately, I will lose nothing qualitatively in utility functions over allocations of goods, ser- what follows by not specifying the household vices, and responsibilities could in principle be model in any sharp detail. I want to avoid do- the same.'0 Whe they are, the household can mng the latr in any case. These are early days be regarded as a team, in the sense of yet in our understanding of the exact springs of Marsebak and Radner (1972). What then re- household behaviour, even in poor countries mains to be analysed is the coordination of So it is better to squeeze as much out of a par- thir asks when the two know different things. iaily unspecified model than to start with a At the other extreme is a household com- precise model and work through its implica- posed of two with totally opposed orderings tions.9 In any event this is how I will proceed. over allocations- That thev rema together The current debate on how best to model should be a cause fir concem, but we may be the household isn't about the reasonableness of unable to do anything about it. If they are viewing it as an 'optimizing agent". All current forced to stay together, the couple is involved constructs of the household are optimization in something alin to a zero-sum game. There is models. They nevertheless differ. One class of no scope for cooperation here. models sees some household agent consciously Neither model is of any relevance. Interest optimizing; for example, by taking reasonable lies rather in those situations where both par- account of the claims of all household members ties perceive that they add something by acting (Gary Becker's "altruistic dictator"; see Becker together (there are gains in forming a house- 1960, 1981; Mirrlees 1972; Cain 1981; Ner- hold and cooperating), where the pair can love, Razin and Sadka 1987a,b; Becker and break up should either party be placed under Barro 1988; Lee and Miller 1991), while- undue stress (there is an outside option for each another interprets household choice in terms of party), and where the parties' motivations an 'as if' optimization (as in the axiomatic aren't the same. This is not to say that an indi- theory of bargaining invoked by Manser and vidual's rankcing of household allocations may Brown 1980; McElroy and Homey 1981; not- reflect interpersonal comparisons of well- Sundstrom and David 1988; Haddad and Kan- being, or interest or advantage. Mutual care bur 1989; McElroy 1990). Given this common- and concern will not be absent, but we would ality, all the formal apparatus of optimization expect identification to be only partiaL This is theory can be brought to bear on analyses of the most commnon type of situation analysed in household decisions; for exaunple, household the theory of games, of which the two-person fertility responses at various prices and income version has traditionally been called ' Singh, Squirc and Strauss eds. (I1986) havc pured this route to good effect But they were not concerned aith ferility decisions. The urs we havc jutA studied about differences in rqxoductive costs between mcn and women infonn us, however, that they aren't identical. 3. PovertyReors oand Fertility: The Househodw epoutv Parnnruluip 79 bargaining dheory, the right basis upon which determinants of economic performnance; they to build models of households. misconceive the conduct of the peoples of ldes; To appreciate why bargaining theory 'is and they employ criteria of welfare so inap- relevant, consider the costs of bearing and rear- propriate that they rgster as deterioration ing children. -Pregnancy involves foregone changes which are in fatimprovements in the work-capacity for womencr and, as we have conditions of people. seen, it can involve a considerable addition to This is not convincing. Even whien mien and the risk of dying. After birth the offspring have women at the household level ratonally prefer to be fed, clothed, taught and cared for. Each large numbers of children to small numbers,. it of these tasks involves time and mateia re- doesn't fbllowv there is no population problem, sources. Reproductive costs diff-er enormously a problem they themselves may.acknow.ledge between men and women. Conflicts of interest were they to be asked about it. As in every oth- between the genders arising from this is there- cr field of individual choice we need here to ask fore a key ingredient of. the population prob- as well if a collection of reasoned decisions at 1cm. We will also see that in some regions the individual level may be sub-optimal at the (e.g., sub-Saharan Africa) the cost of rearing collective level. Putting it in a sligtly ditfftrent children is shared among kith and kin. This can way, we need to ask if there can be a "resource lead to reproductive free-riding, most especial- allocation"t falure here. *ly by men. Understanding and modellinAg the There are two broad reasons for a possible cost side of reproduction doesn't pose any seri- dissonance between household and societal 1ev- ous intellectual challenge."1 cls of decisionmaking in the field of procre- Formal accounts of fertility behaviour ha- ation. The first is that the relative prices of van't pLaced. the household explicitly in the various goods and services that households context of ruali poverty, -and it is poor rural face, may simply becwon" for whatever rca- populations -that are growing the fastest. Nor son. The second is.provided by the ubiquitous has the household been analysed in the context phenomenon of ex-ternalities. In this section and Of environmental-resource use. We will see thiat in Annex 3A.we will study the latter. We will the contest mnatterst identify the former in the sections "Children as consumption and insurance goods" and Population externalities: household "fnviromnenml degradation and children as versus societal reasoning producer goods". That many parts of the globe are even nowTwsoreofxtnaiesuesthm expeiencng svereresurcescarity s wiely selve. The first is simple eniough, it has to do recognized. This doesn't mean individual wihtefneesofpaeIcrsdppu- houshold aremakig iratioal o abnrmal tion size implies greater crowding, and we do choices. n an othrwise mot iIl~~~not typically expect households, acting on their essay Bauer (1 98 1, pp. 61, 64) goes farther. OWIto"nealz"cwdgexraiis Arguig tha it wuld e wrog of s ~ th~k his isn't a precious argument. The human epi- large famnilies in poor counties necessariydnooia nionetbcmsmr n pose a problemn, he writes: "The comiparatively mr rcnu scmucto n oua high fertility and large famnilies in many ldcs tion densities rise. Packed centres of population (less developed countries) should not be re- provide a fertile ground for the spread of vi- gardd asirraionl, anorml, ncomrehesi- ruses, and there are always new strains in the ble or .unexpected. They accord with the making. That enviromnmental resources are usu- tradiion f mot cuture andwiththe ~ptsally common property. is also cited as a harbin- of religitous and political leaders... Allegations gro xenlte.Tepithr sta or aprehnsions of adverse or even disastrous because households. have. access to. common results of Population growth are unfounded. poet eore,prnsdntflyba h They rest on scniously defective analysis of the csofraighlde,nd-teyproduce "This isn't to say that estimating the cost of having children isn't formidable. See Lindert C1980. 1983) and Lee and Buatalo(1983) for uscrul discussions of this. so too nany. Admittedly, local common property being replaced immediately by new institutions resources in poor countries have in recent stu- to soften the oDsts. Demographic tansition is dies been found to be a good deal less of a potentially a prime example of this. When this source for fre-riding than it has traditionally is the situation, only a concerted effort been taken to be (see Dasgupta and Heal 1979; (through a massive literacy drive and dtrough Dasgupta and Miler 1991); nevertheless, the inprovements in employment oppormtnities for static misallocation, however small, can cumu- women) can dislodge the economy from the latively have a large effect on population, as rapacious hold of high fertility rates. In Annex we will see presently. 3A we will study a fonnal account of it .The second source of exerality lies else- There is an additional force opeting in where, and is more subtle. Procreation is not rural communities of poor countries which en- only a private matr, it is a social activity as couages high ferilky raes. It has to do with well. By this I mean that household decisions the relative pnces of altematve sources of a about procreation are influenced by the culthual number of vital household needs, which are no- milieu. In many societies there are norms en- where in line with what they ought to be if hu- couraging high fertility rates which no house- nian well-being is to increase there. From the hold desires unilakerally to break. These norms - household's perve, the local environ- may well have had a rationale in the past; mental-resource base of£lrs the relatively cheap when mortality rates were high, rural popula- sources of such needs. We will see that this can tion densities were low, the threat of externna- encourage high fertility rates and unsustainable ton from outside atack was high, and mobility resource-use. But in order to develop the argu- was resticted. But norms often survive even ment we need to study the deteminants of fer- when their purpose has disappeared. It can then tlity. So we do this first be that so long as all others obey the norm no household on its own wishes to deviate from it; Birth control and female education eve though we all other households to re- All societes practice some form of birth strict their fertility rtes, each would desire to controL Fertilty is below the maxium possi- do so. Thus, there can be multiple social equi- ble in all societies. Even in poor countries, fir- libria, and a society can get stuck at one which, tility is not unresponsive to the relative while it may have had a collective rationale in resource cost households face.'2 Exteded thC past, does not have one any more. br-astfding and postpartum female seXual This doesn't mean society will be stuck abstinence have been a comnon practice in with it forever. As always, people differ in the Afric&a In a noted study on Kung San foragers extent of their absorption of social norms, teir in the Kalahari region, Lee (1972) obseved readiness to digest new information, and to act ta among them the nomadic, bush-dwelling upon new information. There are inevitably women had an inter-birth intval of nearly 4 those who, for one reason or another, eWeni- years, while those settled at cattle-posts gave ment and take risks. They are the norm- birth to children at much shorter intervals. breakers, they often lead the way. In the con- From the perspective of the individual nomadic text of fertility, educated women are among the Kung San woman, it is significant that the so- first to make the move towards smaler fami- cial custom is for-mothers to nurse their chil- lies. (See, e.g., Farooq, Ekanem and Ojelade : dren on demand, and to carry them during their 1987.) Female education is therefore a potent day-long trips in search of wild food through force in creating norm-breakers, as are employ- the children's fourth year of life. Anything less mnent opportunities for women. (See the section than a four-year birth interval would therefore "Birth control and female education".) Special increase moters' caryig loads enomnously, costs are inevitably bome durig transitional impose a threat on their own capacity to sur- periods, when established modes of operation vive, and reduce their children's prospects of are in the -process of disintegration without survival. In contast, cattle-post women are See Coale and Trssell (1974) for an attempt at constucting a measure ofthe extent members of a society consciously oontrol their fertility. 3. Povern Resources, and Ferilitv: The Househ old as a Reptodedtive Partnership - 1 sedentaLy, and are able to wean their children 7.1.) These large variations across regions not earlier.'3 - f only reflect a divergence in the public provision Excepting under conditions of extrme of family planning and health-care services, nutritional stress, nutritional status does not they reflect variations in demand as well. Sur- appear to affect fecundity. (See Bongaarts veys indicate that women themselves perceive 1980; Menken, Trussell and Watkins 1981.) an umnet need for access to methods for reduc- During the 1974 famine in Bangladesh the ru- ing their fertility. However, the extent of this ral population lost over 1.5 rmillion additional felt need varies across regions substantially. children. The stock %as replenished wiutin a (See World Bank 1984; Chomitz and Birdsall year. (See Bongaarts and Cain 1981.) Of 1991; Hill 1992.) Successful family planning course, undernourishment can still have an cf- programmes have proved more difficult to in- fect on sexual reproduction, through its im-- sfitute than could have been thought possible at plications on the frequency of still-births, first. At one extreme (South-East Asia, Sri maternal and infant mortality, and on possible Lanka, and the state of Kerala in India), house- reductions in the frequency of sexual inter- hold demand and State comrnmitrnent to Emily course. The central questions in economic de- planning programmes and public-health ser- mography are then these: what determines vices have merged in a successful way. China fertility, and what grounds are there for our in particular has pursued an active policy of thinking that there is a population problem? In limiting family size. Total fertility rate was the following section and in Annex 3A we will brought down, though not monotonically, to a look at the latter question. TheM rmainder of the rmarkable 2.3 by 1978 from a high 5.9 in body of this article studies the two together." 1960 In many places there has been a pauci- The first and most obvious determinant is ty of demand. In Thailand, for example, the the nature of the available technology of f:rtil- population growth rate has fallen from an aver- ity control. (Bongaarts 1984, and World Bank age of 3.1 percent per year during the decade 1984, contain good discussions.) Traditional 1960-70 to 1.9 percent during the decade mecthods have consisted of abortion, abstinence 1980-90, but now shows signs of having stabi- or rhythm, prolonged breastfeeding, and coitus lized there. At another extreme (sub-Saharan interruptus. These options are often inhumane, Africa), there has been next to nothing done at and usually ineffectual and unsafe. Conacep- the State level im the way of a supply of such iives are superior on all three counts. Neverthe- services- The population growthi rate has in- less, their use has been uneven withir poor creased in these decades from about 2.5 per.- countries. In East Asia over 65 percent of mar- cent per year to something lIke 2.9 percent per ried women in the age range 15-49 years use year. We will see below that the absence of contraceptives as against somewhat under 10 conjugal bond (in particular the practice of percent in sub-Saharan Africa. In South Asia polygyny) as a norm in sub-Saharan Africa has as a whole the figure in the early 1980s was something to do with such high rates, although about 25 percent, but in Sri Lanka it was a it has little to do with thefict that fertility rates high 55 percent. (See World Bank 1984, Table haven't declined there.16 " See also Blhron Jones and Sibly (1978) and Lee (1980j. Child-spacing in sub-Saharan Afica is the subj&t of empri- cal inquiry in Page and Lestha.glec eds. (I 981). For illuminating emirical analyscs of the determinants of fertility, see Leibenstein (1974), Easterlin (1975, 1978), Birdsall (I 977, 1988), Bongaarts (1978), Preston ed. (1978), Cochnane (1979), Freedman (1979), Easterlin cd. (1980), Cain (1981, 1982, 19S4). Bongaarts and Potter (1983), Bulatao and Lee eds. (1983), and Easterlin and Crimmins (1 985). " Sce World Bank- (1980, 1988). However, ratber draconian measures would appear to have been in usc, and the cost has not becn negligible. Hull (1990) has collated data reflecting the growing increase in the male/female sex ratio at birth in China. Thc 1987 One Percent Survey in China shows in addition dramatic pattens of high sex ratios for second and higher-order births. There are threc possible explanations for this, all of which may be presumed to be o g fe- male infanticide, gender-specific abortions, and concalments of births Tomich, Kilby and Jo'Lnston (1991) provide a fine discussion of the possibilities. -The currcnL tragedy ovcr the rapid spread of AIDS in sub-Sahamn Arica is also in part a consequence of this. See Caldwell (1991). Sec also a periodic feature article, titled A Continent's Agony", in the New York Times respecially 82 We should not be SLLrprised that in those et al. 1986; Bourne and Walker Jr. 1991; Tho- regions where family-planning progranunes mas, Strauss and Henriques 1991), and have had an impact, it has occurred mostly in children's height (Christian et al. 1988; Strauss the initial stages. Couples would be expected to 1990; Thomas, Strauss and Henriques 1990, adopt new methods of birth control to satisfy 199 1O.7 These studies confirm that education unmet needs. However, over time it is the net helps mothers to process infonnation more cf- demand for children which vould be expected fectively, and enable them to use the various to dominate household decisions. Here is a sub- social and community services which may be stantiation: Starting in 1977, 70 "treatment" on offbr more intensively. Among other things, villages were serviced by a programme of birth education appears to impart a degree of self- control in the famous experimnent in Matlab confidence on one. to avail oneself of whatever Thana in Bangladesh, while 79 "'control" vil- new facilities that may.be on offer. This is in- lages were offered. no such special service. The valuable for rural populations living dtrough contraceptive prevalence in the treatnent vil- changing-circumstances." lages increased from 7 percent to 33 percent The links between feale education, espe- within 18 months, and then more gradually to a cially secondary education, and reproductive level of 45 percent in 1985. The prevalence behaviour are varied; (Cochrane 1983, is an il- also increased in the control villages, but only luminating study.) The acquisition of education to 16 percent in 1985. The di&ference in total delays the age of marriage, and this vould be fertility rates between the two groups reached a expected to reduce fertility. Moreover, literacy figure of 1.5. (See Phillips et al. 1988; Hill and receptiveness to new ideas complement the 1992.) effOrts of family plnnning programmes. Fur- Even in the imtial stages, howlever, family- thermore, education increases wyomen's oppor- planning progranmes on their own do not do: it .tnities for work and so their opportunity cost matters greatly if women have a measure of of time. (The cost of child-rearing is higher for education and autonomy. This has been a cen- educated mothers.) And finally, educated. moth- tral conclusion of a large number of empirical cis would be expected to value education for studies. Indeed, the beneficial effects of pa- their children more highly. so they are more rents' education, particularly mothers' educa- likely to make a conscious tradeoff between the tion, have been observed more generally on the quality and number of their children. (See well-being -of their children. For the most part, below.) the studies have explored the effct of some 6 Set againhese -is an effct on fertility to 7 years of schooling, no more. Where they which runs the other way. Taboos against post- have differed is over the measurement of well- partum fenaie sexual activity, where tiey ex- being. Some have looked at the ",input" side: ist, may well be weakened through education. for example, household consumption of nutri- In sub-Saharan Africa, where polygamy is ents (Behrman and Wolfe 1984a,b) and the use widely practiced. postpartum female sexual ab- of contraceptives (Cochrane 1979; Satahr and. stinence ca. last up to 3 years after birth. It is Chigambaram 1984). Others have looked at the also not uncommon for wmomen to practice total "output" side, for example, at child health in abstnence once they have become general (Cochrane et al. 1980; Cochrane, Les- grandmothers. The evidence is curious: in Latin lie and O'Hara 1982; Wolfe and Behrman America and Asia, increased female enrollment 1982, 1987; Strauss 1990), at infant and child . in secondary education has had the effect of survival rates (Caldwell 1979, 1986; Hobcraft, lowering fertility rates, while in parts of sub- McDonald and Rutstein 1984; Mosley 1985; Saharan Africa there is evidence that the effect Mensch, Lcntzncr and Preston 1986; Victoria has been the opposite.'9 16-19 September, and 19 and 28 October 1990). " Howevrr, not all the studies I have Qcid here are methodologically immune to criticism. Indeed, in a few studies en- dogenaus variables are treated as though are exogenous. Strauss (1990) has a good discussion of such failings.. ' Here is an indication of ordes of magnitude. The infiant mortality ratc in households in Tlailand where the mother has had no education (resp. has had primay and secondary education) uas found to be 122 pcr 1000 (resp. 39 and 19 per 1000). Sce Jamison and Mosley (1990). 3. Povenp Ruources, and Fertility: Thse Household as a Reproductive Partnership 83 There is then a strong. complementarity be- explain why they haven't responded to declin- twen family planning programines,. and litera- ing mortality rates. The cult of the ancestor cy and numeracy taken together. At a more may prescribe reproduction of the lineage, it general level, social and community services does not stipulate an invariant fertility rate. and female autonomy are complementary Tfc- Even in sub-Saharan Africa total fertility rates tos. Neitheron its own will do.. In the remain- have been less ta the maximum feasible der of this chapter I shall take this for ganted rates. and try and identify the rmotivations for having. The second kind of motivation stems from children. Understanding household motivation the old-age security children can provide in an is a key to locating desirable population economic environment where capital, or annu- policies. ity, markets are next to non-existent. One way of formalizing this is to assume that parents are Children as consumption and interested in some form of household welfare insurance goods subject, however, to the condition hat the -Two broad types of reproducive motiva-. chance of there being an offspring to care for tions have figured proniinently in analyses of themn in old-age (i.e., providing sustenance, population growth in poor countries. The fir time and attention) is no less than some small stems from a regard for children as children amount. In many societies this translates itself Not only are children desirable in themselves, to a requirement.that the chance of there being they carry on the family line or lineag, and asona;ivewhentheparetsare oldisnoless they are the clearest avenue open to what one than some small amount As a numerical exam- may call self-transcendence (see Dasgupta ple, we may consider the simulation study by 1992b, Chapter 13; Heyd 1992 We are May and Heer (1968), who estimated tat an genetically progmnuned to want and to value average Indian couple in the 1960s needed to themL In short children are durable consump- have 6.3 children in order to be 95 percent sure don goods.2 This provides the broadet type of having a surviving son vhen the father of motivation. It comprises a disparate set, inraches the age of 65 yeaTs. This is a high fig- ranging from the desire to have children be- ure, about the same as the total firtility rate in cause they are playful and enjoyable, to the India during the decade of the 1950s. This dictates of injunctions emanating from the cult - -"safety first" model of fertlitv decision has re-. of the ancestor, which sees religion as essen- cently been much explored in a series of ar- tially the reproduction of the linage. This lat- ties by ain (1981, 1982, 1983, 1984)? ter motivation has been enphasimd by : Here we should note that a preference for sons Caldiwell and Caldwell (1987, 1989) in es- - leads parents. to discriminate against higher- plaining why sub-Saharan Africa has for the birth order girl children, a not-ifrequent prac- most part proved so resistent to fertility reduc- tice in China and in the northern parts of the tion. But it isn't a good argu .ment. It explains Indian subcontinent (see, e.g., Dasgupta why fertility rates there are high, it doesn't 1992b, Chapter 1 1). In much of sub-Saharan How powerful this coumtailing force has proved in sub-hr Africa is a controversial matter, and it is possible that the increased fertility respone to increased education seen in some of the data reflect aggregation biases But see Hess (1988) for a time series analysis which attests to there being such an effect in parts of sub-Sahar Africa. See also Barro (1991), who analyses data from over 100 countries to show that during 1960-85 counteris writh higher a hu- man capital base (as evidenced by school enrollment figures) had lower fertiliLy rates. Models with ibis general motivation have been explored in Beck-er (1960. 1981). Dasgupta (1969, 1974), Mirrices (1972), Bccker and Lewis (1973), Willis (1973, 1987), Becker and Tomes (1976), Behnnan, Pollak and Taubman. (1982), Caldwell and Caldwell (1-587, 1989), Nerlove, Razin and Sadlka (1987a), Barro and Beck-er (1989), Cigno (1991), Lee and Miller (1991), and in a powerful philosophical essay by Heyd (1991). Note that in evolutionary biology phenotpic cosls and benefits of reproduction are important anly to the extent that they are corrclated with reproductive measures. Offspring in this theozy are valued in terms of the end of increasing fit- ness. This isn'L dhe point of view in economic demogiaphy, where instead children are valued as means, as durable con- sunption goods, or as produccr and investment goods (see below). Preston cd. (1978) is a usefu colecwtion of essays on the effect that have been observed on fertility rates ofreductions in rates of infant mortality. 84 Africa even today, rural women lose something are exceptions of course (e.g., Mucller 1976), likc a third of their offspring by the end of their but on the whole this motivation has been ne- reproductive years. This provides a strong rea- glected in the demographic literature. son for pro-natalismLY - Poor countries for the most part are Old-age security as a motivation for having biomass-based subsistence economies. Rural children in poor countries is intuitively appeal- folk there eke out a living from products ob- ing. The question remains if there is anything tained directly from plants and animals. Pro- to it in the world as wc know it. In a significant duction throughput is low. Households there do study Nugent and Gillaspy (1983) used Mesi- not have access to the sources of domestic en- can evidence to show that old-age pension and ergy available to households in advanced in- social security do act as a substitute for chil- dustrial countries. Nor do they have water on dren. This doesn't mean fertility rates must in- tap. (In the semi-arid and and regions water evitably decline when fully-finctioning capital supply isn't even close at hand.) This means markets are introduced. They may well rise if that the relative pnces of altemative sources of parents display a niixed motivation for having energy and water faced by rural houscholds in children: viewing them both as investment and poor countries are quite different from those durable consumer goods. (See Nerlove, Razin faced by households elsewhere. Indirect and Sadka 1987a, b.) sources (e.g., tap water nearby) are often pro- Old-age security provides a -potentially hibitively expensive for the household. As we strong motiv. In 19W people aged will see presently, this provides a link between 65-and-over in South Asia formed about 4 per- high fertility, degrdation of the environmenal- cent of the total population. The sex composi-- resource base of a rural community, and an ac- tion among the aged is far from even, being of centuation of hardship among its members. the order of 80-85 men for every 100 women From about 'te age- of 6 years children in am3Og the elderly. In South and South East poor households in poor countries mind their Asia female life expectancy at birth is 59: siblings and domestic aninmals, fetch water, and years, while that of males is about 54 years. At collect ficiwood, dung and fodder. Tthese are age 60, however, they are approximatly 15 complmentary to other household activities. years and 14 years, not much less 'than lie ecx- They are needed on a daily basis if the house- pectancy at age 60 in advanced industrial coun- hold is to survive. As many as 5 hours a day tries. (See Trease and Logue 1986.) In the may be required for obtaining the bare- Indian subconftient the proportion of the elder- essential firewood, dung and fodder. (One ly who live with their children (for the most should contrast this with the direct time spent part, sons) is of the order of 80 percent or by households mi acquiring water and fuel in more. (In the United States the corresponding advanced industrial economies, which is nil.) figure is about 15 percent.) Sons are an abso- All this may be expected to relate to the lute necessity in these circumstances. A poor high fertility and low literacy rates in rural widow with no sons in northem parts of the In- areas of most poor countries.- Poverty, the thin- dian subcontinent is faced with a near-certain ness of markets, and an absence of basic ame- prospect of destitution. nities make it essential for households to engage in a number of complernentary produc- Environmental degradation and tion activities: cultivation, cattle grazing, fetch- children as producer goods ing water, collecting fodder and fuelwood, In poor countries children are also usetiti cooking food, and producing simple marketable as income-earning assets; that is, as producer products.; Each is time consuming. Labour pro- goods. This provides households in these parts ductivity is low not only because capital is with a 'tiird kind of motivation for having scarce, but also because environmental e children. It has important consequences. There sources are scarce too. Children are then 2 Leibenslein (1957) is an carly exploration of the old-agc security hypothesis regarding fcrtility behaviour. See also Neher (1971), Schultz (!974), Willis (1990), Sundstrom and David (1988), and Cremer and Pestieau (1991). Nugent (1985) provides an assessment of the then existing literature n the siibjecL .3. Povenr Resources, and Fenhlity: The Household as a Reproducdive PanAeahi, 8S continually needed as workers by their parents, adult. Using a zero (caloric) rate of intcrest, ho even when parents are in their prime. A small estimated that male children compensatc for household simply won't do. Each household their own cumulative consumption by the ago needs many hands, and it can be that the over- of 15. This may not be typical in Bangladesh. I all usefulness of each additional hand increases cite it, nevertheless, to show the vast difference with declining resource availabiliy.24 In their in the motivation for having children between study of work allocation among rural house- households in rich countries and poor house- holds in the foothills of the Himalayas, C.S.E holds in poor countries. (1989) recorded that children in the age range It appcars then that the transfer of material 10-15 years work 114 times the number of resources over a life cycle in poor households hours adult males do, their tasks consisting of in poor countries is from offipring in the ag- fuelwood, dung, and fodder collection, grazing gregate to their parents. The qualification is domestic animals, household chores, and mar- important I have seen no study which includes keting. Now, a high rate of fertility and popula- mi the calculation of resource transfers the val- tion growth fiuther damages the environmental ue of time foregone in die rearing of children, resource base (to the extent hey are unpro- nor the risks bome by the mother during the tected oommon-property), which in tum in a process of reproduction These amount to re- wide range of circumstances provides further source transfers: from parents to their children. (private) incentives for lage families, which in There is nevertheless a sense in which children turn fiurther damages the resource base,..., and are more valuable to parcnts as producers of so on; until some countervailing set of factors income within poor households in poor coun- (whether public policy, or somc form of Mal- tries thad they are m. rich communities. So I thusian checks) stops the spiralling process. shall take it that the flow of resources there is But by the time this happens millions of Iives from the offspring to their parents. However, it have usually suffirtd' Such an explosive pro- isn't mere poverty which leads to this direction- cess can be set off by. any number of factors. al flow. If mobility on the part of young chil- Government or private usurpation of resources drcn is routme matter (and this was the case in to which rural communitis have had historical early 19th Century Eng-lan) poor parents are access is a potenIal source of the problem. So not able to affect this transfer readily. In tbese is the breakdown of collective agreements circumstnces much of the motivation for hav- ainong users of common property resources a ing children is absent, and even a poor society tnggering mechanism. Indeed, even a margial may display a move towards the demographic decline in compliance can trigger the process of transition, that is, the transition from high to cumulative causation. The static efficiency loss low fertility.26 But this isn't so in the Indian associated with minor violations is, to be sure, subcontinent and sub-Saharan Africa, and it small; but over time the effect can be larg.- makes for a strong parental motivation for hav- As workers, children add to household in- ing large families. come. They are often costless to rear by the In many societies (e.g., in the Indian sub- timne tey are adolescents. This line of argu- continent) daughters are a net drain on parental ment has been emphasized by Mueller (1976) resources (dowries can be bankrupting), and and Lindert (1980, 1983). Cain (1977) has stu- this goes some way towards expWlaining the died data from the village Char Gopalpur in preference parents show for sons there. Bangladesh. He showed that male children be- (Sopher ed. 1980.) It also helps explain why come net producers at as early an age as 12 daughters in their childhood are expected to years, and work as many hours a day as an work relatively harder for their parents. All this Cooking in a poor household is a vertically-integrated activity: nothing is processed to be begin with. It is time-intensivc. 2T This can happen cspccially if households discount the future at a high ratc. z For an account of this kind of spiralling process, see Nerlove and Meyer (1991) and Dasgupla and MAler (1991). In an important empirical document, World Bank- (1991) has provided partial confi9AIl Ot wthesis in the context of sub- Saharan Africa. X I am gralzrul to Sheilagh Ogilvie ror this poinL 86 is in sharp contrast with advanced industrial fertility ought to be subsidized. I have not seen nations, wherc material resources are trans- this implication advocated by its proponents. fered on average from the adult to the young. In a long sequence of writings, Professor Ca - .. Some special features of well (see e.g, Caldwell 1976, 1977a, b, 1981,. sub-Saharan Africa 1982) has argued that whether a society has Even among poor regions, sub-Saharan made the demographic transition is related to Africa is a special case on fertility matters, and the direction of the intcrgenerational flow of it.is worth aking why? To be sure, sub- .resoturces.. Saharan Africa is not a homogeneous. social The motivation for fertility we have been -entity. There are regional differences in the emphasizing in this and the previous section ethos within this large land-mass (see Coehrane springs from a general absence of certain basic 1991). But for our purposes we, will lose little needs in rural parts of poor countries: public- by thinking in aggregate tenns. There are cen- health services, old-age security, water ad - ral tendencies in any broad cultral group, and sources of fuel. Children are born in poverty, a numer of features within sub-Saharan Ai- and they are raised in poverty. A large propor- ca would seem to be pertnent when we seek to tion suffer from undemourishment They re- loca them. When we have identified a few, we main illiterate, and are often 1,0th stunted and will contrast them with those prevailing in the wastedL Undernourishment retards their cogni- Indian subcontinent, another region experienc- tive (and often motor) development (see, e.g., ing a population problem. Dasgupta 1992b, Chapter 14). Labour produc- Unike. the Indian subcontinent, the tivity is dismally low also because of a lack of household" is not a very ml organi- infrastructure, such as roads. In this back- ing unit for production, consumption and fertil- ground it is hard to make sense of the oft- ity decisio in Africa (See e.g., Dow 1971; expressed suggestion (e.g., Simnon 1977, 1981) Caldwell ed. 1975; Cain 1984; Blec 1987; that there are mcreasirg-returns-to-scale in Caldwell and Caidwell 1987, 1989; Kamarek population size even in poor countries; that hu- 1988; Goody. 1990.) Often, there is no comnmon nman beings are a valuable resource. They are udge for a man and wife. Polygamy is idey powentially . valuable as doers . of things and practised. In the late 1950s some 35 percent of onginators of ideas, but for this they require in- all married men in sub-Saharan Africa were puts of the anls for development. H i storical -polygamists (see Goody 1976). In rurl areas evidenceon th 1 pressuenoftpopulaton led - tatters would not appear to have changed to 1jj~ng~ III thr '-anization of production, xmuch (see Farooq, Ekanem and Ojelade 1987). pTIRP-Prty, Owy l"Fays of doing things, By "polygamy' we really mean "polygyny" which is w 1w'Y 1981) studied in - here: it is the men who have more than one her far-reaching 1, iso does not seno spouse. Women in sub-Saharan Africa, just as speak to the population jtlem as it exists in vwomen among non-caste Hindus in India, often sub-Saharan Africa and the Indian subconti- have more than one relationship at any given nent today. Admnttedly, the central message tnr. But they are exta-martal relationships, one reads 'M these writings is that the spectre of n m s.3 Recall also that the dowmy sy- the Malthusian trap is not to be takn serious- t is virtually non-existent in sub-Saharan ly. But we should be permitted to ask of tiese Africa- It is for the most part men who have to oder wters what policy flows from their vi accumulate wealth to obtain a bride. This af- sions. The Boserup-Simon thesis implies that fords. a reason why the median age difference households confer an external benefit to the .bet spouses i Arca- is lag; a ct of community when they reproduce. This means cnsiderale importance, since it enables n. Caldwcll and Caid,vbll (1992) contains a good accwunt-of ertilitytrends in sub-Saharan Africa. 2' Among somc non-caste Hindus in India exltr-marital relationships offer a means of fertility-control. Abortion of a foe. tas resulLing fronm a marital union is frowned upon, whereas the abortion of a foetus resulting from an extra-marital relationship is pretty much mandatory. I am grateful to Paul Seabright for this observation concerning nonns of beha- viour in rural South hndia. 3. PovJtv. Resoames. and Ferfiiv: The Housekold as a Reproduedve Partnership 87 women to have, spouses despite the prevalence their kin at any given time. Nephewvs and nieces of polvgvny: w'omen on average become wid- have the sanie rights of accomodation and sup- ows at an earls age, and widow-remarriages port as biological offspring. Sub-Sahalan Afri- are not onlv pennissible, they are a ca has often been characterized by strong conunonplace. descent lineage and by weak conjugal bond. The sexual division of labour is powerfil, For the most part, as in much of the Indian but it assumes a different form from the Indian subcontinent, descent is patrilineal and resi- subcontinent. Even though women do not n- dence is patrilocal. (An important exception are hernt land. the primary responsibility for raising the Akan people of Ghana-) subsistence crops for the household usually rests with women, who as a consequence have Modelling fertility decisions greater power and control over food distnbu- Household fertility and savings decisions tion than their counternarts in the Indian sub- are interrelated Indeed, children as old-age se- continent. Some 40 percent of rural households curity identifies the two. Nevertheless, it will are headed by woomen. (This goes a little n ay proves useful to consider them separately. It tow-ards explaining why the dowry system is makes for expositiorml ease absent-) Women are as a consequence much Earlier, we identified three different mo- more independent of their husbands than tey tives behind procreation. Hybrid models of fer- are in the Indian subcontinent. Nevertheless, tility decision contain all of them in varying women's sexuality and labour power are firmly strengths. They enable us to analyse the matter under control by the husband's lineage. Among in a unified way. The thing to do is to consider other things it is the extent of kinship control of the "reduced fboni' of household utility as a wvomen which differentiates societies in sub- flmction of, among other things, tte number of Saharan Africa from those in the Indian sub- chidren they have. Should the child mortlaity continent. Until relatively recently, the Milow rate decline, the frtility rate would be expected land was comrunally owned within kinship to follow -suit among households averse to fac- groups. All this nakes for a considerable dif- ing risks m the number of children who sur- ference in the resource implications of having vive. But even when mortality rates decline, it offspring. Children aren't raised by their par- takes time for households to recognize this. De- ents in the way they are in the Indian subconti- mographic transitions in the past have dis- nent. Rather, this responsibility is more diffuse played shar1? declines in fertility rates following within the kinship group, affording a form of only some time after steep declines in mortality insurance protection to be expected in semi- rates?' arid regions. In much of West Africa about 113 The effect on fertility of changes in income of the children have been found to live with and the cost of reanng and bearing children is 3 I am assuming hat the cost of rearing children remains approximately the same despite fal1nig mortality rates. Thus, let n be the number of children produced, and lC(N,An) the probability that N will survive, where OcN<>n and where a is an underlying parameter of the distributiomL Now no matter what is the motivation for having children, we can al ys represent the rCduced form of ex post parental utility (net of the ax pos. cost of child rearing) as V(N). For vividnLsi., we may supposc that V(N) is increasing in N bortow values ofN and decrasing in N for high values of N. If the von Ncumann-MAorgcn!cren axioms are satisfied, ex aete parental utility is TJ(n,a)=r 7rcN,n)V(N). Parents choose n Suppose, to takle a sharp example, 7C(.) is the binomial distribution, with a representing the probabilit that any one child will survive. This means that t(N, n= CCNCC N(Z l-a)t. It can then be shown that if V(N) is concave in N, an increase in leads to a decline in the utility-maximizmg choice of n. The argument generalizes to the case where there are child-bearing costs, provided they are increasing and convex in z4 and it generalizes to the case where reproductive decisions are sequcntial. See Sab (I 990). Ihe critical assumption is the concavity of V(N). Since V(N) is a reduced form (what we may call indirect utility) we should ideally derive it from postulates on primitives, and not merely stipulate it Preston ed (1978) contains valuable empirical cvidcnce on the matter. It offers a varied picture. See in particular tle essays by Ben-Porath and by Chowdhu- ry, Khan and Chen. In the former study (based on Israeli data) child replacement is shown to be a significant phenome- non, and it occurs quickly. In the latter study (based on data from Bangladesh and Pakistan) the influence of child mortalitv on subsequent fertility was found to be of no great significance. But see Bongaarts and Cain (1981), who re- cord substantial replacumcnt through incr-sed fertility following the Bangladesh famine of 1974; and Caldwell (1991) on sub-Saharan Africa. 88 more complex. (Thc pioneering work here is rising income brings in its wake greater access that of Becker 1960.) Letting c denote current to the capital market. (For an empirical parental consumption, n the number of chil- exploration of this last link, see Rosenzweig drcn, and z an index of the quality of each and Evenson 1977.) child, we may assume household utility, U, to Inprovements in labour productivity are be a function of these three variables. So we often associated with urbanization. This accen- write U(ctnz). It is appropriate to assume that tmates the directional changes we have already U is an increasing function of both c and z.30 identifiedL Urbanization tends to disintegrate We should viewv U(.) as a reduced form" of households into -nuclear" units. This raises the household utilitv, in that in those circumstances cost parents have to bear in rearing their chil- where children are seen at least panially as in- dn. (lhe contribution of grandparents, aunts vestment goods. z would in part reflect future and other kin is in this situation greatly re- parental consumption. duced.) Growing, urbanization in a growing The quality of a child depends on the economy also offers children better employ- amount of time and resources devoted to hiim it ment prospects, which improve their bargain- depends as well on the time and effort devoted ing srength relative to their parents. This in by the child in acquiring education and skills. turn lowers the gross return on children as in- In a hybrid model, z itself would be an aggre- vestment goods, since children become less de- gate of different characteristics. The problem is pendable as a source of income to their parents that parents are faced with a tadeoff betveen c in their old age. (Sundstrum and David 1988, and z, and between n and z. Moreover, c is ic- deploy this argument in the conwtext of ante- self up to a point an incrnasing function of n in bellum United States.) Moreover, St i' legis- poor households, since children are valuable as lation on elementary schooling (for example, producer-goods. This accentuates the tradeoff makig it compulsory) and increased private parents fhce between n and z2 returns to education (arising from general in- Earlie we observed that improvements in dustializaion) make children relatively seak- education appears to have a salutary effect on ing less usefil as producer goods. household circumstances. It is also useful to Compoundmg all these considerations enables study the effec of an increase in parental in- us to have a glimpse of those forces at work come on fertility. (This leads us back to the ag- which relate fertilitv to household income. The gregate data we studied in the first section, broad "inverse" relationship between income -income. fertility. and food: the environmental- and fertility does not require of us to postulate ist's argument".) To begin with, we may trace that children are 'inferior goods", in the sense the increase in parental income to a rise in their made familiar in consumer-demand theory. In- labour productivity: as a consequence for cx- creased parental income, especially maternal ample, of rural investment, or more generally, income, raises the cost of children: children be- better employment opportunities for men and come more expensive relative to other goods. women in the labour m-arket- Now a rise mi The causal chain is therefore more complex women s labour-productivity implies an in- and somewhat reinforcing, and the inverse" crease in the opportunity cost of rcaring chil- relationship between household icome and the dren. At the same time, with increasing desired faraily size holds for a .vide class of parentWal income children are needed less as household utilitv functions.3" producer goods, and possibly less also as in- Both the costs and benefits of having chil- vestment goods, since onc would imagine that dren are experienced in different proportions by 34 Children erc taken to be identical, and are assumed to be identically treated I do this simply in order to pose the prob- lem in its pristine form. We will also ignore the externalities which social norms create. " Increascd life expectancy incrcases the return on children's education, and so human capital formation could be ex- pected to assume a largger share of total investmenL The Indian data conforns to this expectation. See Raam and Schultz (1979), who in addition estimated that during the 1970s humn capital accumulation in India was of the order of 55 percent of physical capital accumulation. 32 A formal model along the lines sketched in the text can easily be constructed. A simplified version of this is in Beckaer (I960, 1981), Neriovc, Razin and Sadka (1987a), and Cigno (1991)- See also Easterlin, Poilakl and Wachter (1980). 3. Poirrtv Resourcesv and Fertility: The Household as a Reproducdve Parmership 59 the parents. The cost of bearing and nursing a most of Asia. Parents together bear the bulk of child is inevitably bome by tie mother, but the the cost of raising their children, even tiough cost of rearing ckiildren is culturally condi- the sexes tvpically do so unevenly. The extent tioned. In some cultures. as in much of sub- of economnic dependency of wiromen on men Saharan Africa, these costs arc diffused across now assumes a central role. This dependency is the kinship. (That women are .rigorously en- enormous in the Indian subcontinent, most es- gaged in agriculture presumably has somehiing pecially the North. In these tvpes of patriarchal to do writ½ this.) When this is so, there is an al- societies women rightly perceive sons as hav- location faihilre, in that neither parent bcars the ing especially high value as insurance against full cost of their decision. In societies charac- personal calamities, such as widowhood and tenzed by w%eak conjugal bond (as in much of abandonment. But so.s can't be guaranteed. So sub-Saharan Africa and the Caribbean) male one has to keep frying. In East and South East parents often bear little of the cost of rearing Asia (and also southem India and Sri Lanka) children. Here, then, we would expect to see women's economic dependency is less. Among another source of as mmetrv in parental mo- those that were the world's poorest countries in tivations over reproductive decisions. Men's the early 1970s, fertility rates have fallen most desire fior childrcn would on balance be ex- dramatically in this part of the world. In a pected to exceed that of women by far, and this wide-ranging essay on the old-age security hy- would brings in its wake all the attendant im- pothesis, Cain (1984) used the median age dif- plications we observe in the data. Patriarchy, a ference between _pouses as an index of female weak conjugal bond, and a strong kinship sup- economic dependence in patiarchal societies to port systm of children taken together are a demonstrate a remarkably high correlation be- broad charactertc of sub-Saharan Africa. It tween this and the total feTity rate in a crots- provides a poweiful stimulas to fertily. section study of nations. That kinship support of one's children pro- There are thus forces at work which move vides the basis of a free-rider problem (yet the fertiliy goals of women relative to me in another exrtenality), leading to too many chil- opposite directions. In poor societies marked dren being borm has been noted in the demo- by gender-based asymmetry in employment op- graphic literature (see e.g., Cain 1984, for portunities and power, women's reproductive references-) But I have not seen it being recog- goals don't differ noticeably from those of nized that this is only so if the benefits of hav- men. (See Mason and Taj 1987.) But professed ing children are not equally diffused across the desires are sensitive to the extent to which kinship. An excessive number of children is the women are educated. educated women tend to outcome only when parents appropriate a desire smaller fmilies th illiterate ones. A greater proportion of the benefits their children preference for sons is nearly universai in the provide than the proportion they incur of the Indian subcontinent. In sub-Saharan Africa this cost of rearing therm Professor John Caldwell isn't so. (See Williamson 1976.) Indeed, in has told me that this would appear to be die parts of Africa for which data on utene sib- normal case in sub-Saharan Africa.' ling groups have been obtained, there is no evi- Matters are different in societies where ihe dence of stopping rules which would reveal an conjugal bond is powerfial. This is the case in 33 To see that there is no distortion were the proportions the same-, suppose c is the cost of rearing a child and N the nun- ber of couples within a kinship. For simplicity let us assume that each child makes available y units of output (this is the norm) to the entire kinship, which is then shared equally among all couples, say iu their old age- Suppose also that the cost of rearing each child is shared equally by all couples. Let nl be the ntunber of children each couple other than the one under study chooses to have. (We will presently endogenize this.) If n were to be the number of childrem this couple produces, it would incur the resource cost C=[nc+(N-l)n*cJ/N, and it eventually would receive an income from the next generation equalling Y=[ny+(N-l)n*yJ/N Denote the couple's aggregate utility function by the form U(Y) K(C). where both U(.) and K(.) are increasing and strictly concave fimctions Letting n be a continuous variable for simplicity, it is easy to confim that the couple in question will choose the value of n at which yU'(Y)=cK'(C) The choice sustains a social equilibrium when n=n*. (This is the symmetric non-cooperative Nash equilibrium of the social system.) It is easy to check that this is also the condition which is met in a society where there is no reproductive free-riding. 90 implicit sex-preference. (See Goody ct al. infrastructure as we have identified requires re- 198 Ia. b.)34 sources, and standard resource allocation theory informs us why it is the State's obliga- Allocation Failure and Public tion in poor countries to mak-e attempts at en- Policy suring their supply. The aim should not be to A good bit of the motivation underlying force people to change their reproductive beha- analytical wvork- on fertility behaviour has been viour.6 Rather, it should be to identify policies in identifying circumstances where there is no which would so change the options men and allocation failure. that is. situations in whichl omen face that their reasoned choices would individual household decisions unencumbered involve a lowering of their fertility rates to re- by any public population policy yield Pareto- placement levels. The evidence, some of which efficlent outcomes.35 The literature is a picture I have put together here, tells us that poor par- of economists busilv burying any "distortion" ents in poor countries do calculate when mak- theyr can imagine, and heaving a sigh of relief ing such decisions as those which bear on when they have fashioned a waorld in whief household size and assets, even though what none is visible. In this article wve have identfied emerges at the end is a greatr number of yet reasons for rejecting this point of view- Never- another cohort of poor, illiterate people. It theless- for the moment let us leave aside social would certainly be unjust of governments to in- ternalities. Let us leave aside as well the fact sist on parents sending their children to schools that a person's utility function doesn't neces- for so many years if this requirement further sanly correspond to the function which reflects impoverishes poor households. But it Nvould her well-being. I want to argue that even then not be unjust vere the complementary house- there is a need for public policy. Public policy hold production inputs made available through does not end when well-being efficiency is in te provision of familyv-planning and sight The distribution of benefits and burdens public-health service, and infiastructural can be unjust. We have noted the importance of Imvesmnent, and for governments to then make children as producer and insurance goods in fi school attendance compulsory-' Reasoned poor households. The aviilability of household choice at the household level would be ex- infrastrucure in the form of cheap fuel and po- P to respond to this through an alteration table water makes children less important as in t:rtihty goals. income eaming assets. By a similar tok Taxes required for the provision of such availability of credit and access to the capital household infiastructure and family-plarming mauicet lowers the impotance of children as in- and public-health service as we are discussing surance goods. Children as producer goods and here can only be obtained fiom households that children as old-age securty yield somehat are better off. In the absence of such public different implications for public policy mprovisions the dynamics of a social system can But theey bothl entail public policy. The pro be horrifing. For reasons we have identified, vision of such patterns of housleold the bulk of the very poor in poor countries havc 3 Stopping mrles based on sex preference provide a different type of infonnation regarding sex preference than sex ratios within a population. The reason is that in a steady state stopping riles have no effect on the sex ratio- To see ihis, sup- pose that in a society where sons are prefered, parents continue to have children until a son is bom, al utich point they stop. Assume for simplicity that at each try there is a 50 percent chance of a son being conceived. Now imagine a large population of parents, all starting from scratch. In the first round 50 percent of the parents will have sons and 50 per- cent will have daughters The first group will now stop and the second group will try agauL Of this second group, 50 percent will have sons and 50 percent will have daughtcrs. The first sub-group will now stop and the second sub-group will have another try. And so on. But at each round the number of boys bom equas the nunber of girls. The sex ratio is 1. 's See for example Becler and Barro (1986), Nerlovc, Razin and Zadk-a (1987a), and Willis (1987). Fora critiquc of this, see David (1986). -B The draconian measures employed in India during the Emergency period 1975-77 as regards sterilization are an cxam- ple of the kinds of activities governments must avoid. 37 The school-year in rural United States until the beginng of this century was shorter than in urban areas, and it took accoumt of seasonal labour requirements. I am gratefih to Professor Gary Becker for giving me an account of this. 3. Ptvnrt, Resourca and Fert;lity: The Household as a Reproductive Partnership 91 continually to aim at large household sizes, Let us sum up. We have identified three making it in turn much more difficult for them broad categories of policics for alleviating the to lift themselves out of the grip of poverty: population problem: (i) increasing the costs of household labour productivity remains abys- having children. (ii) reducing the benefits of re- mally low, investment credit is for the most production, and (iii) improving the information part unavailablc to them, and the avenue of base concerning the technology of reproduc- savings is consequently that much constrained tion. and afficbting the locus of household deci- for them. Thc mattcr is differcnt for those with sionmaking Now (i) and (iii) have found much a greater access to resources. Thev are, as al- expression in the demographic literature: un- ways,. in a position to limit their household size happily, at the neglect of (ii). It is (ii) which and increase the chance of propelling them- raises the most interesting economic issues; it selves into still-higher income levels. I have not tells us that among the most potent avenues been able to locate published data on the mat- open for easing ffie population problem are ter, but my impression is that among the urban those which involve social coordination (see middle classes in India, the demographic tansi- Annex 3A), the provision of infra-structural tion has already been achieved. This doesn't goods, and measures of social security. But mean there is an inexorable vicious circle of these services are desirable in themselves. and poverty". People from the poorest of back- commend themselves even when we don't have grounds have been known to lift themselves out the population problem im mind. It seers to of the mire. Nevertheless, there are forces at me, this consonance among desirable social po- work which pull rich and poor groups awvay licies is a most agreeable fact. from one another in tenrs of the quality of life. Admittedly, in saying all this we are look- The Matthew Effect ("For unto everyone that ing at matters wholly from the perspective of hath shall be given, and he shall have abun- the parents. This is limiting)'2 But identifying dance. but from him that hath not shall be tak- the right basis for population policies is ex- en awvay even that which he hath") works tremely difficult. (See Meade 1955: Narveson relentlessly in poor countries. 1967; Dasgupta 1974, 1988. 19899 1992b; I am puttng matters in stark terms so as Heyd 1992.) What I have tried to argue in this only to focus on the idea that parents would article is that there is much we can establish have a different set of fertility goals vere the even were we to leave such conceptual difficul- relative prices of environmental and infiastruc- ties aside. Population policy involves a good uare goods different, and wvere the economic de- deal more han making fimily planning centres pendencv of women on men less. As regards available to the rural poor. It also involves the latter, fbmale education is now widely rec- more than a recognition that poverty is a root ognized to be a key propellent But so are in- cause of high fertility rates in a number of soci- creased employment opportunities for females eties. The problem is deeper, but it is a route to greater autonomy. identifiable. There are, of course, other measures which should be thought about in parallel. Compulso- ry schooling, for example, makes children pro- hibitively expensive as assets for generating current income, so it reduces their attractive- ness as a comnu.-dity. Making available alter- native sources of basic household needs improve the well-being of poor households via an altered set of fertility goals. t Enke (1966) is a notable explortion in the valuc of prevented births when the worth of additional lives is based entire- ly on their effect an the current generationL As a simplification, he took the valuhe of a prevented birth to be the dis- countod sum of the differences between an additional person's consmption and output over the person's lifetimc. 3. Poverty. Resources, and Fertility: The Household as a Reproductive Partnership 93 Annex 3A: Strategic Complementarities in Fertility Decisions In the section 'Population extemalities: 832Uk(xkxk)/Ixxkx > 0, for all ktn. (1) household versus societal reasoning", we iden- tified the social milieu as something which in- When individual utility functions satisfy condi- fluences fertility decisions. It is a source of tion (1), we will say that the social system in- extemality. We i%ill call it an atmospheric ex- volves strategic complementarities among ternality (see Meade 1952). Here we will for- individuals' motivations. (See Cooper and John malize a simple version of the idea. But it is as 1988.) In game theory it would be to say that well to note that the formulation has wide ap- payoff fimctions are supermodular. (See e.g., plicability, and it isn't restricted to fertility be- Milgrom and Roberts 1989, 1990)2 haviour. Our purpose here will be to We will find it usefil to appeal to a speci- demonstrate that social externalities often lead fication which is somewhat sharper than condi- to multiple Nash equilibria. This means that tion (1). It will be assumed that a person's history matters, and that societies which are utilty is a function of her own action and of simlar in terms of technology and utility the average action of all others (hence the tenm fimctions may gravitate in the long term to atmospheric externzlity); and that the marginal quite dissimilar stlteS of affirs. utility of her own action is an increasing func- There is a wide variety of actions we all tion of the average action of all others. are regularly engag,ed in which are numerical; E-xamples aboundL Ih-v include how hard that is, they are scalar in dimension. (The we work at ours tasks, what wage rate we ac- analysis can be generalized to the case wvhere cept as reasonable, how much we spend on a decision variables are vectors.) They also share particular type of consumer-durable (through, the Thature that any given individual's goal- for example, what Duesenbery 1949, chris- maximizing choice of action is an increasing tened the "demonstration effict", or through function of the choice of action of any other in- what in moder industrial organization litera- dividual. This isn't quite the primitive one tit are called "network externalities", see Far- wants (choice is to be explained, not assumed), rell and Saloner 1986), how much education and we look for properties of utility functions we allow our daughters to attain, at what age which will yield this. So, supposing there are they get marrid, how many children we aim M people (k,n=1,_,M), let X4k (a real number) for (albeit this only allows for non-integer val- be individual k's action. Assume too that k's ucs), and so forth. In sociological parlance utiW function is of the forn Uk(Xli,..,xm); these examples reflect peer-group emulation which we also write as Uk(xl,xk).' Tbe primi- and norm-guided behaviour, and in economics tive which will do the job is that k's marginal they are often called atmospheric external utility of her own action is an increasing economies. fimction of the choice of action by any other Earlier, we discussed household reproduc- person. Formally, this means: tive decisions, and the concomitant private , = (x11.. ,xt.i,xt.ir.Xu). 2 This iSn't strictly correcL The inequality in (1) is taken to be weak in the definition of sapernodularity. We :ose noth- ing here by assuming strict inequality. 1 It is also on occasion called Marshallian external economies. The idea has a long and distinguished pedigree (see Rosenstein-Rodin 1943; Scitovsky 1954), and it has found rich expression in recent years (see eg. Schelling 1978; David 1985, 1987; Stiglitz 1987; Cooper and John 1988; Hahn 1990) An underlying idea is to find a general condition 94 0 A Figure 3A-1 costs and benefits of having children. These de- of activity, not just twvo; which is to say, there cisions are also influenced bv aspects of shared may be several (non-ooperative) Nash equilib- values within a community. A woman on her ria, with their associated norns of behaviour. onvn will not find it desirable to break out and That we find ourselves at one equilibrium re- assert her autonomy when no one else is doing flecting a particular custom is no reason for in- so. (Among other things, the psychic costs may sstig there isn't another equilibrium be too high.) But she may find it permissible do (reflecting a different custom) at which mdi- so if all others (or at least most others) are do- vidual wvell-being would be higher. ing so as well. Both are self-enforcing situa- To see this, imagine a numnber of identical tions. and they sustain quite different personal households, each of whose stategy (say, the norns of behaviour. We often refer to this as a level of some activity) is, vwithout loss of gen- society's cutslom, (ceg. "The custom in rural erality, a number between 0 and l. (See Figure India is to have mant children"). Of course, 3A-1.) I denote the representative household's therc may well be several self-enforcing levels activity level by z, and the average value of the (e.; condition (1)) under which a social system can possess several non-cooperative (Nash) equilibria, at least two of which can be ranked on the basis of the Pareto criterion. 3. Poverty. Resourcem. and Fertility: Thte Household as a Reproductive Parnership 95 activity level chosen by all other households by games, the latter is the material of social Z. Let U(z.Z) denotc the typical household's sciences.4 utility function, net of costs of choosing z. For To illustrate their differences, it is simplest each possible value of Z there is a correspond- to think of a situation where the activity level ing value of z. sav z* at wuhich U(.) is maxi- has to be chosen each period (day, week, year, mized. (We ignore multiplc houschold optima or whatever). Imagine that decisions are cost- for simplicity.) We write this as z*(Z). ft is the lessly reversible from period to period, so that household's reaction curve. We are assuming everything starts afresh each period. If in any that OaU(z7Z)/zZ o>c 0. This means z*(Z) is period each household expects every other upwad-sloping, as shown in Figure 3A-I household to choose a particular equilibrium Since households are identical, the house- action, each housthold will find it in its interest hold whose reaction curve we have drawn is a to choose this action as well. Expectations are represcntative one. We now simplifv for expo- not only self-fulfilling, it is expectations which sitional ease and pick a representative house- do all the wvork. Under an eductive mode a hold from all the remaining households. This Nash equilibrium is held up by its own boot- would be anchored most firmly were we to as- straps, so to speak (A charismatic leader may sume that when the average choice of all other come along and persuade people to change households is Z. it is a consequence of each of their expectations to a set of self-flilfilling them having chosen this average. So we do so. ones.) There is no analytical reason why ecx- We are now interested in points at which eZ() pectations must necessarily be based on histori- intersects with the 450 line. Each intersection is cal experience. When they are not so based, a non-cooperative Nash equilibrium point of history has no role to play in the detmination the social system. Each household's level of ac- of the final outcome, although of course, it is tivity at an intersection is utiliy-maximizing on special contingencies (eg., the emergence of a the assumption that all other households choose charismatic leader) which do the picking from this same level. The intersections are social the set of equilibria. On the other hand, ex- equilibrium points. In Figure 3A-1, there are 5 pectations are typically influenced by history. equilibrium points, P, Q, R, S and T. Notice In this case history has a strong role to play that we have not assumed utlity fimctions to even under an eductive mode.5 be concave. Equilibrium is guaranteed by vir- An extreme altenative has histry assum- tue of the assumption of strategic ing the dominant role Suppose, for example, in coomplementarities. each period households base their expectations Because households behave identically at of what others will do on what was done on av- equilibria, the latter can be ranked by the Pare- erage in the previous period This is sometimes to criterion. Equilibria can also be ranked in called the Marshallian tatonnement- Its ratio- terms of household well-being. So there is a nale is provided by substantial sunk costs in- best and a worst equilibrium, and there are curred in each period's decisions, so that from ones of intermediate goodness. period to period only marginal changes in indi- The question arises as to which equilibrium vidual decisions are undertken, if they are un- point gets selected. The answer depends on the derLaken at all. Consider the simplest form of relative importance of a number of things, all this: each household in each period expects the of which can be succinctly captured by two no- average level of activity to be equal to the aver- tions: that of the expecrarions individuals have age level of activity during the previous period. about one another (they have on occasion been It is now an easy matter to check that under the called "eductive" considerations), and that of dynamics generted by such adjustments in be- the history of the social system (they have fre- haviour, P, R and T are stable equilibrium quently been called "evolutiver _;onsiderations). points, while Q and S are unstable. The former is the stuff of the pure theory of 4 See Binmore and Dasgupta (1986) for an elaboration ofthis distinction ' Note that individuals in the society we are studying in the tex are not locked in a Prisoners' Dilemma. Thcer are multi- pic equilibria here, and no dominant strategy for any household. 96 To illustrate. imagine that through a se- Becker, G. (1981): .4 Treatise oin 1lfw Iwnsldv quencc of chance events households find them- (Cambridgc. MA: Harvard lhJiivursaly selves at R (vhcre. sav the avcragc age of Prcss)} marriage for girls is low) and that household Becker. G. and R. Barro (1i86): Altrwsrti :trid ult %vell-being would be greater at P (%vhcre the av- Economic Theot or Fertilif'. pOPu- cragc age of marriage for girls is somewhat ean and Development Review (Supple- higher). There is then a precise sense in which Becker mCnt). 12. B ckrG. and R_ Barro (1988): "A Reformulation houscholds are engaged in an excessive level of of the Economic Thcory of Fcrtility". activitv at R (i.e.. a rapid turnover of daughters QuarterlvJournal of Economics. 103. in houscholds)- It is possible each household Becker, G. and H.O. Lewis (1973): Interaction know%%s this. But no household on its owvn has an between Quantity and Quality or Chil- incentive to deviate from its chosen pattern of drcn-. Journal of Political Economtyv. 81. behaviour. This is a social dilemma, and only a Becker, G. and N. Tomes (1976): -Child Endow- coordinated policy can ovcrcornc itements and the Quantity and Quality of Children". Journal of Political Economv This is an extreme example. By this I don't mean the example's technical features, for ex- Behman. L. RA_ Pollak and P. Taubman (1982): ample that households are identical. (This can WParent Preferences and Provision for be relaxed easilv without our losing anything Progeny". Journal of Political Economy. other than simplicity.) I mean something else. 90. In the world as we knowv it norms over time are Behrlan. J.. and B.L. Wolfe (1984a): "The So- intnalized, and so brcaking out of established ciocconomic Impact of Schooling in a De- patterns of behaviour even when others are do- veloping Country'. Review of Economics ing so can prove hard for an individual. This and Statisrics. 66. gives a certain additional stickiness to estab- Behnnan, J. and B.L. Wolfe (194b): "More Eui- -- . . -- ~~~~~~dence on Nutrition Demand: [ncome [ished equilibrium behaviour. Social equilibria Sees Overrated and Women's School- are vet more hlstory-dependent. Our present ing Overedpandc. Journal ofe- ing Underemphasized- . Jounal of Devel- example has not reckoned wth this additional oppnent Economics. 14. reason xvhv history matters. It has been de- Bimmore. K and P- Dasgupta (1986): "Game signed only to explore the possible multiplicity Theory: A Survey" in K1 Bimnore and P. of Nash equilibria. Dasgupta eds.: Economic Organizations as Games (Oxford: Basil Blackavell). Bibliography Birdsall. N. (1977): "Analytical Approaches to the Relationship of Population Growth and Development. Population and Develop- Arrow. K.J. and A. Fisher (1974): -Preservation, ment Review. 3. Uncertainty and Irreversibility". Quarter- Birdsall, N. (1988): "Economic Approaches to lv Journal of Economics. 88. Population Growvth". in H. Chenery and Barro. R. (1991): "Economic Growth in a Cross T.N. Srinivasan eds.. fandbook ofDevel- Section of CountriesC. Quarterly Journal opment Economics. Vol. 1 (Amsterdam: of Economics. 106. North Holland). Barro. R. and G. Becker (1989): "Fertility Choice Birdsall. N. and D. Jamison (1983): 'Income and in a Model of Economic Growth". Econo- Other Factors Influencing Fertility in metrrca. 57. China". Population and Development Re- Bauer. P.T. (1981): Equaliyv. the Third Wf"orld and vWew, 9. Economic Delusion (London: Weidenfeld Bleek. W. (1987): "Family and Family Planning in and Nicolson). Southern Ghana-. in C. Oppong ed Becker. G. (1960): -An Economic Analysis of Fer- (1987) tility-. in G. Becker ed.. Demographic Blurton Jones, N.G. and RM. Sibly (1978): and Economzic Change in Developed 'Testing Adaptiveness of Culturally De- Coufntries (Princeton. NJ: Princeton Uni- termined Behaviour: Do Bushmen Wom- versity Press). en Maximize their Reproductive Success by Spacing Births Widely and Foraging Seldom?". in N.G. Blunon Jones and V. 3. Poverty, Resources, and Ferdility: The Household as a Reproductive Partership 97 Rcynolds eds., Human Behaviour and CaIdwell, IJC. (1977a): "The Economic Rational- Adaptations (London: Taylor and ity of High Fertility: An Investigation II- Francis). lustrated with Nigerian Data-, Population Bongaarts. I (1978): "A Framework for Analyz- Studies, 31. ing the Proximate Determinants of Ferti- Caldwell, J.C. (1977b): The Persistence of High litv". Population and Development Fertility: Population Prospects in the Review, 4- Third WYorld (Canberra: Australian Na- Bongaaris, 1. (1980): -Does Malnutrition Affect tional University Press). Fecandity? A Summary of the Evidence", Caldwell, J_C (1979): -Education as a Factor in Science, 208. Mortality Decline: An Examination of Bongaarts J1 and lvi Cain (1981): Demographic Nigerian Data", Population Studies, 33- Responses to Famine (New York: Popula- Caldwell, J.C. (19K1): -The Mechanisms of De- tion Council). mographic Change in Historical Perspec- Bongaarts. J. and RG. Potter (1983): Fertility, live", Population Studies, 35. Biology and Behavior: An Analysis of the Caldwdel, IC (1982): 7heory of Fertility Decline Prorimate Determinants (New York: (New York: Academic Press). Academic Press). Caldwell. JC. (1991): -The Soft Underbelly of De- Boserup, E. (1965): The Conditions of Agricu- velopment Demographic Transition in tu-ral Growth (London: George Allen & Conditons of Limited Economic Unwin). Change, Proceedings of rhe World Bank Boserup. E. (1981): Population Growth and rech- Annual Conference on Development Eco- nological Change (Chicago: Chicago nonrics 1990. University Press). CaldwelL I.C. and P. Caldwell (1987): "The Cul- Bourne. K[C. and G.M Walker Jr. (1991): "The tural Context of High Fertility in sub- Differential Effect of Mothers' Education Saharan Afica", PopIlatfon and Devel- on Mortality of Boys and Girls in India", opment Review, 13. Population Studies, 45. CaldwelL J.C. and P. Caldweil (1990): "High Fer- Brown, L.R_ et al. (1991): Srate of the World (New tility in Sub-Saharan Africa", Scientific York: W.W. Norton). American. 262 (May). Tulato, RA. and RD. Lee eds. (1983): Determi- Caldweil, J.C. and P. Caldwell (1992): nants of Fertility in Developing Coun- "Popuation Growth, Physical Resources, tries, 2 volumes (New York: Academc and Human R urces in sub-Saharan Press). Africa", in P. Dasgupta and K.-G. Maler Cainr, M (1977): "The Economic Activities of eds. (1992). Children in a Village in Bangladesh", Chomitz, KM and N. Birdsali (1991): Population and Development Review, 3. 'Incentives for Small Families: Concepts Cain, Mi (1981): "Risk and Insurance: Perspec- and Issues". Proceedings of the World tives on Fertlity and Agrarian Change in Bank Annual Conference on Develop- India and Bangladesh", Population and ment Economics 1990. Development Review, 7. Chrian, P. et al (1988): "The Role of Maternal Cain, M. (1982); -Perspectives on Family and Fer- Literacy and Nuiion Knowledge in De- tility in Developing Countries", Pop;,_?- termining Children's Nutritional Status", tion Studies, 36. Food and Nutrition Bulletin, 10. Cain, MI (1983): "Fertility as an Adjustment to Cigno, A. (1991): Economics of the Family Risk", Population and Development Re- (Oxford: Clarendon Prss). view, 9. Coale, A.J. and I.T. Trussel (1974): "Model Fer- Cain, A (1984): -Women's Status and Fertity in tility Schedules: Variations in the Age Developing Countries: Son Prefrrnce Structur of Childbearing in Human Pop- and Economic Security", World Bank ulations". Population Index. 40. Staff Working Paper No. 682. Cochrane. Sit (1979). -Fertility and Education: Caldwell, J.C. ed. (1975): Population Growth and What Do We Really Know?' (Baltimore: Socioeconomic Change in West Africa Johns Hopkins Univesity Press). (New Yodk: Columbia University Press). Cochrane, SH. (1991): "Comment", Proceedings Caldwell, JC (1976): "Toward a Resatement of of the World Bank Annual Conference on DemogWphic Theory", Populatior and Development Economics 1990. Development Review, 2. 98 Cochranc. S.H. (1983): Effects of Education and David, P.A. (1985): -Clco and the Economics of Urbanization on Fcrtility", in RA. Bula- QWERTY", American Ecoromic Review, lao and RD. Lec cds. (1983), Vol. 2. Papers and Proceediags. 75. Cochranc. S.H. ct al. (1980): -Effccts of Education David. P.A. (1986): -Commcnt". Population and on Hcalth". World Bank Staff Working Development Review (Supplement), 12. Paper No. 405 (Washington. D.C.). David. P.A. (1987): -Some Newv Standards fior the Cochrane. S.H.. J. Lcslie and D. O'Hara (1982): Economics of Standardization in the In- -Parental Education and Hcalth: Intra- formation Age". in P. Dasgupta and P. country Ev.idencc" Health Policv and Stoneman eds. (1987). Education. 2. Dow, T. (1971): -Fertility and Family Planning in Cooper. R. and A. John (1988): -Coordinating Africa", Journal of k1odern Studies, 8. Coordination Failures in Key;nesian Mo- Desenwbey. J1S. (1949): Income. Savings and delst. Ouarterlv Journal of Economics. the 7heory of Consumer Behaviour 103. (Cambridge. MA: Harvard University Cremer. J. and P. Pesticau (1991): -Bequests, Fil- Press). ial Attention and Fertility". Economica, Easterlin, RA. (1975): -An Economic Framework 58. for Fertility Analysis", Studies in Family C.S.E. (1990): Human-Nature Interactions in a Planning, 6. Central Himalayaon illage: A Case Study Easterlin. RA. (1978): -The Economics and So- of Villae Bemru (Ncw Delhi: Centre for ciology of Fertility: A Synthesis". in C. Science and Environment). Tilley ed., Historical Studies of Changing Dasgupta, P. (1969): -On the Concept of Optimum Fertility (Princeton: Princeton University Population, Review of Economic Studies, Press). 36. Easterlin, RA. ed. (1980): Population and Eco- Dasgupta. P. (1974): "On Optimum Population nomiiic Change in Developing Countries Size", in A. Mitra edo: Economic Theory (Chicago: Chicago University Press). and Planning: Essays in HIonour of 4-K Easterlin. RA. and E. Crimmins (1985): The Fer- Dasguspa (New Delhi: Oxford University 11liW Revolution: A Supply-Demand Press). Analysis (Chicago: University of Chicago Dasgupta, P. (1988): -Lives and Well-Being". So- Press). cial Choice and Welfare, 5. Easterlin, RA., RA. Pollak and M.L. Wachter Dasgupta, P. (1989): "Population Size and the (1980): -Toward a More General Model Quality of Life, Proceedings of the Aris- of Fertility Determination: Endogenous iotelian Society (Supplementary Volune Preferences and Natural Fertility", in LXIII). RA. Easterlin ed., Population and Eco- Dasgupta, P. (1992a): -Population, Resources and nomic Change in Developing Countries Poverty", Ambio, 21 (Chicago: Chicagr University Press)} Dasgupta, P. (1992b): An Enquiry into Well-Being Ehrlich, P and A. Ehrlich (1990): The Population and Destitution (Oxford: Clarendon Explosion (NCw York: Simon and Press). forthcoming- Schuster). Dasgura, P. and G. Hedl (1979): Economic Ehrlich, P. and A. Ehrlich (1992): "The Value of Theory and Exhaustible Resources Bio-Diversity'. in P. Dasgupta and IC-G. (Cambridge: Cambridge University Mfler eds., (1992). Press). Ehrlich, P.. A. Ehrlich and J. Holdren (1977): Dasgupta, P. and K.-G. Miter (1991): "The Envi- Ecoscience: Population, Resources and ronment and Emerging Development Is- the Environment (San Francisco: WA sues", Proceedings of the World Bank Freeman). Annual Conference on Development Eco- Enke, S. (1966): "The Economic Aspects of Slow- nomics 19._O. ing Population Growth". Economic Jour- Dasgupta, P. and K.-G. Mfller eds. (1992): The nat, 76. Economics of Transnafional Commons FAO (1989): The State of Food and Agriculture (Oxfo.d: Clarendon Press), forthcoming. (Rome: Food and Agricultuial Dasgupta, P. and P. Stoneman eds. (1987): Organization). Economic Policy and Technological Farooq, G.M., I.I. Ekanem and S. Ojelade (1987): Performance (Cambridge: Cambridge "Family Size Preferences and Fertility in University Press). 3. Poverm Resourmes and Fertilit: The Househlold as a Reproductive Partnershtip 99 South-Western Nigeria". in C. Oppong Kamarck, A.M. (1988): "Thc Special Case of ed. (1987). Africa", in D.E. Bell and M.R. Rcich Farrell. J. and G. Saloner (1986): 'installed Base eds: Hlealth, Nutrition and Economic and Compatibility: Innovation, Product Crises (Dover. MA: Auburn House). Preannouncements and Predation"', Kelley, A.C. (1988): -Economic Consequences of .American Economijic Review. 76. Population Change in the Third World", Freedman. R. (1979): -Theories of Fertility Dc- Journat of Economic Literalure, 26. clinc: A Rcappraisarl, Social Forces. 58. Lee, RB. (1972): "Population Growth and the Be- Goody. J. (1976): Production and Reproduction: A ginnings of Sedeniary Lifc among the Comparative Studv of the Domeslic Ter- !Kung Bushmen". in B. Spooner ed.: rain (Cambridgc: Cambridge Univcrsity Population Growth: Anthropological Im- Prcss). plications (Cambridge. MA: MIT Press). Goody. J. (1990): $Futures of the Family in Rural Lce. R.B. (1980): -Lactation, Ovolution. Infanti- Africa"' Population and Development cide. and Women's Work: A Study of Review (Supplement). 16. Hunter-Gathercr Population Regulation", Goody. J* ct al. (1981a): -On the Absence of Im- in M.N. Cohen. RS. Malpass and H.G. plicit Sex-Prefcrence in Ghana". Journal KIClein eds.: Biosocial Mechanisms ofPop- of Biosocial Sciences. 13. ulation Regulations (New Haven: Yale Goody. J. et al. (1981b): "Implicit Sex Preference: University Press). A Comparative Study", Journal of Bioso- Lee, RD. and RA_ Bulatao (1983): "The Demand cial Sciences, 13. for Children: A Critical Essay", in RA. Haddad. L and R Kanbur (1989): -Are Better Off Bulatao and RD. Lee eds. (1983), Vol. 1. Households More Unequal or Less Un- Lee, R-D. and T. Miller (1991): -Population equal? A Bargaining Theoretic Approach Growth, Externalities to Childbearing, to "Kuznets Effects" at the icro- LcweV', and Fertility Policy in Developing Coun- mimeo. World Bank, Washington, DC. tries", Proceedings of the World Bank- Henry. C. (1974): "Investment Decisions under Annual Conference on Development Eco- Uncertainty: the Irreversibility Effect". nomics 1990. .Inerican Economic Review. 64. Leibenstein, H. (1957): Economic Backwardness Hess, P 4. (1988): Population Growth and Socioe- and Economic Growth (New York: Join conomic Progress in Less Developed Wiley). Countries (New York: Praeger). Leibenstein. H. (1974): -An Interpretation of the Heyd. D. (1991): Genethics: The Adoralitv of Pro- Economic Theory of Fertility: Promising creation (Los Angeles: University of Path or Blind Alley?", Journal of Eco- California Press) nomic Literature, 12. Hill, K. (1992): "Fertilit and Mortality Trends in Lindert, P. (1980): "Child Costs and Economic the Developing World". Ambio, 21. Development", in RA. Easterlia ed. Hobcraft, J-N., J.W McDonald and S0. Rutstein (1980) (1984): "Socio-Economic Factors in In- Lindert, P. (1983): "The Changing Economic fant and Child Mortality: A Cross- Costs and Benefits of Having Children", National Companson", Population Stu- in RA. Bulatao and RD. Lee eds. (1983). dies. 38. Manser. M.L and M. Brown (1980): "Marriage and Hull, T.H. (1990): "Recent Trends in Sex Ratios in Household Decisionmaking: A Bargain- China-. Population and Development Re- ing Analysis". International Economic view. 16. Review. 21. Jamison, D.T. and W.H. Mosley (1990): Marschak, J. and R. Radner (1972): Economic "Selecting Disease Control Priorities in Theory of Teams (New Haven. Conn: Developing Countries", in D.T. Jamison Yale University Press). and W.H. Mosley eds.. Evolving Health Mason, KIO. and AM. Taj (1987): -Differences Priorities in Developing Countries between Women's and Aen's Reproduc- (Washington, DC: World Bank). tive Goals in Developing Countries". Johnson, D. Gale and R Lee. eds. (1987): Popula- Population and Development Review, 13. lion Grow th and Economic Development: May, D.A. and D.M. Heer (1968): "Son Survivor- Issues and Evidence (Madison: Universi- ship Motivation and Family Size in India: ty of Wisconsin Press). A Computer Simulation". Population Studies. 22. Joo McElroy. M.B. (1990): "The Empirical Content of Ncrlovc, M., A. Razin and E. Sadka (1987a): Nash-Bargained Household Bchaviour". Hlousehold and Economy: Wetfare Eco- Journal of Human Resources. 25. nomics of Endogenous Ferfility (Ncw McElroy. M.B. and M.J. Horney (1981): "Nash York: Academic Press). Bargained Household Dccisions: Toward Ncrlove, M.. A. Ra7in and E. Sadka (1987b): Pop- a Generalization or the Thcory of De- ulation Policy and Individual Choice: A inand". Intel7national Fconomric Reviewv, Theoretical Investigation, Research Re- 22. port 60 (Washington. DC: Inlernational McGuire. J. and B.M. Popkin (1989): "Beating the Food Policy Rcscarch Institute). Zero-Sum Game: Women and Nutrition Nugent, J. (1985): "The Old-Age Security Motive in the Third World: Part 1", Food and for Fertility", Population 6nd Develop- Nutrilion Bulletin. 1. ment Review, 11. Meade. J.E. (1952): Extemrnal Economics and Dis- Nugent, J. and T. Gillaspy (1983): "Old Age Pen- economics in a Competitive Situation", sion and Fertility in Rural Areas of Less Economic Journal. 62. Developed Countries: Some Evidence Meade, J.E. (1955): Trade and W'elfare (Oxford: from Mexico", Economic Development Oxford Univcrsity Prcss). and Cultural Change, 31. Mcnkcn. J.. 1. Trusscll and S. Watkins (1981): Oppong, C. ed. (1987): Sex Roles, Population and -The Nutrition Fertility Link: An Evalu- Developmpient in West.4frica (Portsmouth, ation of the Evidence", Journal of Inter- NH: Heinemann). disciplinary History. 11. Page. H.J. and R Lesthaeghe eds. (1981): Child- Mensch, B., H. Lentzner and S. Preston (1986): Spacing in Tropical Africa: Tradition and Soci-economic Differenlials in Child Change (London: Academic Press). Mortality in Developing Countries (New Phillips, J. et al. (1988): "Detenninants of Repro- York: United Nations). ductive Change in a Traditional Society: Milgrom, P. and J. Roberts (1990): "The Econom- Evidence from Matlab, Bangladesh", Stu- ics of Modern Manuflcturing: Technolo- dies in Family Planning, 19. gy. Strategy and Organization", Preston, S.H. (1980): "Causes and Consequences Atmerican Economic Review. 80. of Mortality Declines in Less Developed Mirrlecs, J.A. (1972): "Population Policy and the Countries During the Twentieth Cenu- Taxation of Family Size", Journal of ry", in RA. Easterlin ed. (1980). Public Economics, 1. Preston, S.H. (1986): "The Decline of Fertility in Mosley, W.H. (1985): -Will Primary Health Care Non-European Industrialized Countries", Reduce Infant and Child Mortality? A Population and Development Review Critique of Some Current Strategies, with (Supplement), 12. Special Reference to Africa and Asia", in Preston, S.H1 ed. (1978): The Effects of Infant and L. Vallin and A.D. Lopez eds., Health Child Mortality on Fertility (New York: Policy Social Policy and Mortality Pros- Academic Press). pects (Paris: Institut National d'Etudes Preston, S.H., N. Keyfitz and R. Schoen (1972): Demographiques). Causes of Death: Life Tablesfor National Mueller. E- (1976): "The Economic Value of Chil- Populations (New York: s eminar Press). dren in Peasant Agriculture", in RKG. Preston, S.H and V.E. Nelson (1974): "Structure Ridker ed.: Population and Development. and Change in Causes of Death: An In- The Search for Selective Interventions ternational Summary-, Population Smu- (Baltimore: Johns Hopkins University dies, 28. Press). Ram, R and T. Schultz (1979): "Life Span, Narveson. J. (1b67): 'Utilitarianism and New Health Savings and Productivity", Eco- Gencrations"`. iMLind. 76. nomic Developnent and Cultural Neher, P. (1971): "Peasants, Procreation and Pen- Change, 27. sions", American Economic Review, 61. Rosenstein-Rodan. P. (1943): "Problems of Indus- Nerlove, M. and A. Meyer (1992): -Endogenous trialization in Eastem and Southeastern Fertility and the Environment A Parable Europe", EconomicJournal, 53. of Firewood", in P. Dasgupta and K;-G. Rosenzweig, MIt and R Evenson (1977): Miler cds., The Environment and Emerg- "Fertility, Schooling and the Economic ing D)evelopment Issues (Oxford: Claren- Contnbution of Childen in Rural India: don Press), forthcoming. 3. Poverr. Resouzces. anud Fertili;r: Thre Household as a Riproductive Partnenship 101 An Ecwnomctric Analysis". lgconomnetri- Thomas, D.. J. Strauss and M.-H. Hcnriqucs ca, 45. (1991): "How Does Mother's Education Sah, RIK. (1990): Thc Effects of Mortality Affccr Cbild Heigiht?". Juurnal of Human Chanigcs on Fcrtility Choicc and Parental Resurces, 26. Welibrc". .Journal ojf JPolitiCicl kconomw. Toniicih. T.P.. P. Kilby and B.F. Johnston (1991): 99. glgriculture and Structural 7Trainrforma- Satahr. Z. and V. Chigamibaram (1984): tioii: Opporltuilies Seized, Opportunities -Diffcrcntials in Contraceptive Use'. Alissed. (Washingtoni. DC: World Bank). 11World I ertility Suve Sidiew. 36. Trease. J. and B. Loguc (1986): "Economic Dcvel- Schclling. T. (1 978): Aficrnmnvesex ajnd Alacrohe- opment and the Older Population". Pop- haviour (Ncw York: W.W. Nortoni). ularion and De elopment Review. 12. Schultz. T.W. (1974): Economnics of Jhe Family: Victoria. C.G. ct al. (1986): "Risk Factors for Mal- Afarrige., Childhren annd Human Capital nutrition in Brazilian Childrcn: The Rolc (Chicago: University orlliCicgo Press). of Social and Environmental Variabics". Scitovsky. T. (1954): -Two Conccpts of External Bullelin 111/0. 64. Economics" ..Journal ojfPolilical Econo- Williamson. N.E. (1976): Sons or Daughters? A my, 52. Cross-cultural Surney of Parental Prefer- Si.on, J.L (1977): The Economics of Population ences (Bevcrly Hills. CA: Sage Growvth (Princeton: Princeton University Publications). Press). Willis, R. (1973): "A New Approach to the Eco- Simon. J.L. (1981): The Ullimiiale Resource nomic Thcory of Fertility", Journal of (Princeton: Princeton University Prcss). Politica! Economny (Supplement), 81. Simon. J.L. and H. Kahn Wds.: The Resourceful Willis, Rt (1980): "The Old-Age Security Hypoth- Earth: A Response to Global 2000 (Neov esis and Population Growth", in T. Burch York: Basil Blackwcll). cd.. Demographic Behaviour: Interdisci- Singh L.. L. Squire and J. Strauss cds. (1986): Ag- plinarv Perspectives on Decisionmaking ricultural Household Models: Extensions, (Boulder, CO: Westview Prcss). Applications and Policy (Baltimore: Willis, RI (1987): "Externalities and Population", Johns Hopkins University Press). in D. Gale Johnson and RD. Lee eds. Sopher. D.E. ed. (1980): 4n Exploration of India: (1987). Geographical Perspectives on Society Wolfe, B.L. and J.R. Bchnran (1982): and Culture (Ithaca. NY: Cornell Univer- "Determinants of Child Mortality, Health sity Press). and Nutrition in a Developing Country", Stiglitz. J.E. (1987): -Learning to Learn. Local- Journal of Development Economics, 11. ized Learning and Technological Pro- Wolfe, B.L. and J. R. Behrnan (1987): "HOw Does gress". in P. Dasgupta and P. Stoneman Mother's Schooling Affect Family eds. (1987). Health, Nutrition, Medical Care Usage Strauss. J. (1990): "Households. Communities, and Household SanitationT', Journal of and Preschool Children's Nutrition Out- Econometrics. 16. comes: Evidence from Rural Cote World Bank (1980, 1984, 1988, 1990): World De- d'lvoirc", Economic Development and velopment Report (New York: Oxford Cultural Change, 38. University Prcs). Sundstrom. W.A. and P.A. David (1988): "Old World Bank (1991): 7he Poprdalion, Agriculture Age Security Motives. Labor Markets and and Environmnent Nexus in Sub-Saharan Farm Family Fertility in Antebellum Africa (Washington. DC: World Bank), America", Explorations in Economic His- mimeo torv. 25. Thomas, D., I. Strauss and Nt-H. Henriques (1990): "Child Survival, Height for Age. and Household Characteristics in Brazil". Journal ofDevelopmnent Economics, 33. 4 The Question of a Public Sewerage System in a Caribbean Country: A Case Study Arthur H. Darling, Christian Gomez, and Mario E Nilditschek The objective of this study was to evaluate the economic feasibility of the construction of a public sewerage system. This paperfocusses on two techniques: contingent valuation to estimate the value of preventing the deterioration of coastal water to the point that it be- comes unswimmable, and risk analysis to reflect the possible margin of error in the estimates of benefits and to assess the robustness of the conclusion. In the contingent valuation study, people were directly asked what they would be will- ing to pay for specfic improvements provided by a project. In risk analysis, the paper ana- lyzes the implications of the likely error ;- the benefit estimates. The results of the risk analysis are used in a number of ways. First, they give a gener- al impression of whether or not the project is feasible with the range of assumptions used. Seconct they permit an estimate of the expected loss of making the wrong decision, Le, building the project when perfect information would have shown that it was not economically feasible, or alternatively, not building when perfect information would have shown that it was feashiblez Third, they identify which information is most important to the decision to un- dertake the project, and permit the design of the terms of reference for cost-effective feasibil- ity studies. 104 The public sewcragc project under analysis tributc to the project The consultant had taken is designed to run parallel to a denscly pop- a total of forty samplcs at ten beaches. The ulated beach area on a Caribbean island whose samples indicated that the coastal waters com- primary foreign exchange carnings are from plied with U.S. Environmnental Protection tourism scvices. The scrvice arca includes Agency (EPA) mean and extreme standards for commercial (including hotels), residential, and class A waters- The study projected that con- institutional users. At prcscnt, the island's citi- tamination water would not exceed the mean zens providc thcir own sanitary services. The standard in the foreseeable future and that it mnajority havc wells that allow excrcte TO filtcr would not excecd the cxtremc standard until the slowly through the limestone fornation to thc year 2000. Unfortunately, the sample was not sea. There are some septic svstems and a few large enough to be dcfinitive, so it was not cer- package treatment plants. tain that coastal waters met the standard. In general, the population of the potential service area is fairly satisfied with the existing Issue Studied systermr Thc govemment fears, however, that The objective of the study was to evaluate the filtration of excrema into the coastal waters the econ-omic feasibility of the construction of a may be having serious consequences. It is be- public sewerage system. The results of a pre- lieved that filtration of sewage nay be damag- vious sewerage project in the country suggested ing the fringing retfs: (a) which are important that willingness to pay of private users was not to the food chain of the fishing industry and (b) sufficient to cover the costs of the project In which help protect against beach crosion. More that previous case, potential users preferred to importantly, the government fears that contin- continue to use their owvn facilities rather han aued filtration of sewage will make the waters pay the connection fee and the costs to adapt unsafe for swimming by nationals and will their plumbing to use the public system. There cause tourism to decline, was reason to believe, however, that significant Despite the implications of these concerns, benefits might be received by people who the technical studies that accompanied the proj- would not be connected to the system, and that ect did not make it easy to estimate project it was worth analyzing the projec benefits. The studies did not analyze a number The additional benefits were difficult to of other sources of pollution (including agricul- measure with precision, and the information tual run-off, storm rmn-off, and filtration from provided was incomplete. The analysis solid waste dumps) that could be affecting the presented here developed complementary fringing reefs. It also was no: clear that the de- infonration on the value of preventing the cline that had been observed in commercial beaches fiom becoming unswirnmable, and fishing would be solved by the project, sincc at structured the analysis to reflect the impreci- least part of the problem was caused by over- sion of the estimates. fishing and pesticides in the food chain. Final- ly, at least some of the erosion problem was Methodology Used caused by man-made civil works that had This paper focuses on two techniques, changed the ocean currents. In short, no techni- one-the contingent valuation method cal data were given to show the relative role of CVM-to estimate the value of preventing the the sewage in the context of the more general deterioration of coastal water to the point that problem. it becomes unswimmable, the other-risk The technical sludy of water quality also analysis-to reflect the possible margin of made it difficult to know what benefits to at- rhe authors are economists at the Inter-American Development Bank (IA%B). Jorge Dueci now Director of Planning at the Ministry of Public Works in Chile, was the IADB economist who worked on the contingent valuation study, and Professor KIE. McConnell of the University of Maryland, was primarily responsible for the execution ofthe subcontract The Question of a Puiblic Sewerage System in a Caribbean Countr" A Case Study *OS error in the estimates of bcnefits and to assess paying and go without the public sewerage sys- the robustness of the conclusion. tem? Contingent Valuation Risk Analysis To detcrminc the valuc of maintaining Every survey has a margin of error, and coastal waters in 'swimmablc condition," the CV survey estimates are no exception. Tnc CV analysis uses the direct questioning technique cstimates, howevcr, are only one area of uncer- (contingent valuation[CV]).1 The mcthod esti- tairty in the decision whether to undertake the mates a price that reflects the value of cleaner project. Bccause of the absence of technical beaches despitc the fact that thcrc is no mark-et studies on the impact of sewage on fishery to reveal a pricc for clcaner beachcs or better yields, beach erosion, and the tourism industuy, seweragc systems. It does this by presenting the estimates of other benefits are rough. The consumers with a hypothetical market in which analysis had to take into account this impreci- they maklc the decision of whether to buy the sion and reflect how certain the country could good in qucstion- be that a large investment in public sewerage Pcople were directly ask-ed what they was worth undertaking. would bc willing to pay for specific improve- To do this, we worked with ranges. We es- ments provided by a project The value that timated the range for each benefit in a way that peoplc gave depcnded (was contingent) upon makes it unlikcly that the true benefit is outside the description of the good, its provision, and the range. To mak}e systematic use of this in- the way it would bc paid for. Hencc, the name formation, we used risk analysis. This method "contingent valuation." describes the range of each cost and benefit as The typical interview contained three parts a probability distribution. A computer is pro- (see Annex 4A): grammed to select one estimate from the proba- 1 A detailed description of the good being bility distribution that describes each valued and the hypothetical circumstance assumption about benefits and costs and use under which it would be made available to the set of selected assmnptions to calculate one the respondent internal rate of return. Having finished one eco- nomic calculation, the computer selects a se- 2. lQuestos wi elictedthe rpdt cond set of assumptions and repeats the willingness to pay for the good calculations. The computer calculates 500 re- 3. Questions about the respondent's charac- sults, each one based on the relative likelihood teristics, prefcrcnccs, and uses of the goods of the assumptions. The probability distribu- being valued tion of the results characterizes the range and The cconomic literature suggests that the relative likelihood of different results. CVM can givc skewed or invalid results if the This information can be used in a number willingness-to-pay question is incorrectly of ways. First, it gives a general impression of phrased or if respondents can guess the pattern whether or not the prcject is feasible under of questions. To reducc the potential for these most sets of assumptions. Second, it gives an biases (called hypothetical and strategic estimate of the expected loss of making the biases2), the valuation questions were asked in wrong decision, that is, building the project a dichotomous (yes/no) format-for cxample, when perfect information would have shown Would you be willing to pay S15 morc on your that it was not economically feasible or, alter- water bill each quarter to have public sewerage natively, not budlding when perfect information or would you prefer to pay what you are now would havc shown that it was feasible. Third, it This method was chosen in prefercnce to the travel cost mcthod (an indirect approach based on obscrvcd behavior) be- cause there was vcry little variation in travcl distances from the residcnces to the beache. 2 lypothetical bias refers to the possible diffecrcnce in behavior when a consuncr confronts a hypothetical, rather than a real, market Strategic bias occurs when a respondent trics to answer a survey qucstion to influence the project decision rather than to reveal his personal valuation of the scrvices provided by the projecL ' We uised the PROPLAN project analysis and Monte Carlo simulation computer program written by Information for In- formation Dccisions (lID) oaWashington, D.C. 106 identifies which information is most important The specific dollar amounts that a respon- to the decision to undertake the project dent was asked about were assigned to the questionnaires on a random basis. For the Description of the Study households located in the area to be connected to the sewer system, the proposed method of Contingent Valuation payment was a sewerage charge that would The first step in contingent valuation is the cover both the services of the public sewerage design and testing of the questions to be asked. system and cleaner beaches. It is well known that the form, wording, order, In each case, the interviewer read a and context of a questionnaire can affect the preamble to the contingent valuation question answers received in unexpected ways. We that focused the respondent's attention on the therefore pretsted the concepts and questions issues that were relevant for the economic val- investigawd with focus groups' and a pilot uation. Households were told the potential im- survey. pact of disposing of wastewater into the One of the issues explored in focus groups ground, and the potential for avoiding beach and the pilot survey was the likely upper bound pollution and damaged reefs by construction of of willingness to pay- As the questionnaire was the sewer system. The interviewer also used an designed to ask only one price on a take-it-or- illustration (see Annex 4A) to reinforce the dif- leave-it basis, it was important that high prices ferences between the situation with and without wxere included so that benefits were not under- the project estimated, but it was also important that the The use of closed-end questions ("yes or prices were not so high that a large percentage no" to a specific dollar amount) made it impos- of respondents would indicate they were not sile to determine direcdy the willigne-to- wl,ling to pay the price proposed- If this were pay. Therefore, the willingness-to-pay re- the case, the size of the saunple would have to sponses had to be transformed into benefit esti- be very large and the survey would be more ex- mates. To do this, we exercised the standard pensive, practice of using an econometric model to re- The final CV survey covered 277 house- late the probability of a 'yes" response to the holds that would be connected to the sewerage amount of payment and household charactens- system and 433 households that lived outside tics to obtain mean values for the sample's the area to be connected to the sewerage system willingness to pay. The exogenous variables but used the beaches that would be affected by used were: whether the respondent used the the project. The total population of these two beaches, the respondent's age, whether the re- areas was 3,268 and 53,041 households, re- spondent saw the television coverage about spectively. beach pollution, and the amount ofthe random- A different questionnaire was used in each ly assigned sewer charge. All of the coefficients area Households who lived outside the sewer estimated were significant at the 95 percent district were asked only about the enviromnen- level The estimates of mean wil!ngness-to-pay tal benefits of the sewer systerr .hey were of- were USS178 and UJSSII per year for house- fered two choices: pay a certain dollar amount holds inside and outside the area to be con- in their quarterly water bill for the constuction nected, respectively. The larger mean benefit and maintenance of the sewer system, or not derived for the catchment households captured pay and not receive thc corresponding services. £ Focus groups are discussion groups comprised of people selected from the population of potential beneficiaries and led by someone familiar with market research and social psychology. Their objective is to leam how respondents conceptu- alize and talk about the topics being investigated. Focus groups are used to cxplore hypotheses and to fornulate specific questions for quantitative research. Often the vocabulary used in the discussion is adopted to phrase the questionnaire- The econometric estimation is set up as a logit function. F(S) = I (probability of saying yes to the purchasing question) I+Cxp(-8) Furthermore, 5=f(P,x,) where: P = price and x,= exogenous variables Maximum willingness to pay is derived by setting 8 = 0 and solving for P. The Question of a Public Sewerage SYstem in a Caribbean Country: A Case Study 107 both the private and public benefit while the yields would increase to 1980 levels in the ten smaller mean benefits derived for the outside years after project completion. We valued the catchmcnt houscholds measured only environ- fish catch at the cost of importing fish, and we mental or public benefits. assumcd that the costs of fishing were half the Risk Analysis value of the fish. Most of these assumptions have no empirical basis, and appeared-if any- In the feasibility analysis, we assigned thing-to be generous. To reflect the invented probabilitv distributions to all cost and benefit nature of the assumptions, we posited a uni- flows. This paper will focus on a few. For the form distribution-any estimate within the invcstment and operating costs, we used nor- rangc cqually likely-and a margin of error of mal probability distributions, with a mean 50 percent6 equal to the engineer's cstimate plus contingen- cy allowance and an 80 percent confidence in- Coastal Erosion terval defined by a range 15 percent above and The value of preventing beach erosion was 15 percent below the engineer's estimate. difficult to conceptualize. We used the cost of The benefits of particular interest were (a) beach replacement ("beach nourishment") as the contingcnt valuation estimates, (b) the the measure. This cost comprises the costs of fishery benefits, (c) the coastal erosion bene- obtaining sand by dredging or quarrying, trans- fits, and (d) the tourism benefits lost if a health porting it to site, and dumping it. The process crisis occurred. is a continuous one. A consultant estimated the Contingent Valuation Benefits likely frequency and volume of replacements with and without the project. The estimates Istead of using the statistical confidence themselves were far from precise: "wtout the interval around the contingent valuation esti- project you will have to nourish beaches every mate, we used a normal distribution that as- five to ten years, with the project every ten to sumed that the 80 percent confidence interval twenty years." To use this information, we es- was within plus or minus 40 percent of the esti- tablished the extreme ranges: No difference mate. We used this wide estimate because we with or without the project (that is, every ten believed that the errors possible in contingent years in any case) and a reduction from once valuation include more than just sampling every five years to once every twenty years. variation- This potential problem is treated in We annualized the costs associated with this more detail in the section on assessment of the range and assigned a uniforn probability dis- methodology, tribution to it Fishery Benefits Avoiding a Decline in Tourism The data on fisheries were poor There was The government's concern that the absence a historical series indicating a significant de- of public sewerage might cause tourism to de- cline in kilograns landed since 1980. Hearsay cline was based on two problems. First, it evidence suggestcd that this process had started seemed that when hotels had problemns with long before 1980. There was no information to their package treatment plants or with delays in indicate what a maximum sustainable yield septage collection, they pumped septage into might be in perfect conditions, and no inforra- the ocean late at night This action often pro- tion on other factors that might be contributing duced odors or visual evidence the next day. In to the decline (for example, pesticides, over- addition, it was believed that some septae tank fishing, agricultural and storn run-off). After cleaning companies were dumping their loads consultation with a local university research illegally rather than hauling them to the teat- center, we assumed that, without the project, ment plant. While the public sewerage project the fishery yield would decline to zero in a peri- would solve these problems, it seemed that oth- od of ten years, but that with the project, the cr, more cost effective measures, could be '6 Thc risk analysis programn did not afford the possibility of making tc declinc and recovery periods stochastic, tbat is, shortcr or longer than ten years. 108 tak-en. Sccond, the system of letting excrcta The U.S. EPA requires that the geometric filtcr through the limestone formation and flow mean for fccal coliforms bc less than 200 per to the ocean could raisc the lcvel of pathogens 100 milliliters, and that a ratio of 400 fecal co- to a lcvel that could causc an outbreak of dis- liforms per 100 milliliters in any 30-day period ease. Any such outbreak-, traceable to the water be exceeded no more than ten percent of the on the beaches, would have an impact on tour- time. The feasibility study found that the mean isrn. was withir. EPA standards. The study had sam- Although the problem was clear, it was not ples that exceeded 400 per 100 milliliters but easy to estimate the probability that there there were not enough to conclude that the 10 would be a health problem or how large the im- percent of the samples exceeded the standard pact would bc if there were one. The consulting for a 30-day period. The feasibility study pre- study that supported the project had attempted dicts that coastal waters will not exceed the to deal with the problem. The consultants sent standard for the mean in the foreseeable fiture the statistics oi the mean and extreme coliform and will not exceed the extreme standard until readings to experts in public health and asked the year 2000. them to estimate the probability that fifty peo- Unfortunately, the study is not definitive. ple (arbitrifily assumed to be the critical level The level of colifonns found in the coastal wa- that would trigger a crisis) would get sick. ters is a fanction of three things: (a) rinfall When we reviewed the question that had been (rain flushes out sewagc wells and bacteriologi- askled the public health experts, we concluded cal mn-oft), (b) time of day (bright sunlight that there was no basis for them to provide an rapidly kills coliforns), and (c) tidal conditions accurate answer. First, the data on contamina- (incoming tides keep coliform near shore). The tion did not specify the frequency with which consultant's sample did not control for these coliformns exceeded safe levels. The mean coli- factors. The sample was small-four observa- form was within the norm and, while the cx- tions at each of ten beaches. This number of treme exceeded the norm, nio frequency was observations can not adequately characterize associated with it. Second, the question did not the contamination of coastal waters. Four of specify how many people had contact with the the consultant's six extreme readings were col- water. Because of these problems, we had to leeted on the same day, and it is not known rejcct the estimate and try to make our own. whether they were after a rain fall, at high tide, The amount of "tourism benefits" that on a cloudy day, or all three; it is also not might be saved by the project depends on a known whether the other readings were taken at number of factors. They are: (a) the probability high noon or at low tide. It is impossible to that water has reached a critical le--el of con- conclude whether the sample is representive. tamnination, (b) the probability that enough The sample seems to suggests that there is no swimmers get sick to make contamination a problem at present, but we cannot be certain. If perceived problem, (c) the number of tourists there is, it is probably not too serious yet that stay away once the problem is perceived, To assess the probability of a health prob- and d) the economic value of each tourist lost. lem given the level of contamination, we ex- There is no basis to make a rigorous statistical amined the epidemiological literature. The estimate for thcse factors, so wc used a range literature indicates that, although a number of of cstimates that appear reasonable. Because infectious diseases can be contracted by swim- the assumptions and the cvidence behind them ming in sewage polluted water,8 research has are critical to the decision, they are presented only been able to establish a quantitative rela- in detail. tionship for gastroenteritis, the most common of the diseases.9 The present EPA standard is 7 Such mcasures could havc included greater supervision and fines for violation, investment in back-up truck faicilities for emcrgencics, and changcs in the structure of incentives to drivers, paying thcrn for loads delivcred to the treahment plant rather than for loads picked up. ' Thc diseases includc hepaLitis A, gastroenieritis, Lyphoid, shigellosis, and cholera 9 Elpidemiological rescarch reportcd several oulbreaks of shigellosis from swimming in contaminated waters (Rosenberg, ct al. 1976). nesc cases wcre reported in frzsh water, and in one case the mean coliform counL was 17,500 fccal coli- The Question of a Public Sewerage System in a Caribbean Countr: A Cost Study 109 Table 4-1: The Probability and Cumulative Probability Distributions of a Perceived Health Problem Resulting frc:n Contaminated Ocean Water Cumulative pmobability of Probability of a at least one crisis crisis in this year by this year Yea- (percent) (percent) 1 5 5.0 2 6 10.7 3 7 16.9 4 8 23).6 5 9 30.5 6 10 44.3 7 11 51.0 S 12 57.4 set consistent with the risk that 19 of 1,000 written (1 - p). The probability of no crisis in swimmers will oDntract gastrcenteritis. If the either the first or the second year is p*p or p, standard (mean fecal coliforms 200 per 100 and the probability of at least one crisis in ei- milliliters) were increased four times to 800 fe- ther the first or the second year is i-p' The cu- cal coliforns per 100 milliliters, the incidence mulative probability of a crisis increases each of gasroenteritis would increase 1.8 times.'0 In year, as the probability of not having a crisis in the consultant's sample the mean was well be- any of the first, second, third, fourth, etc- de- low 200, and there was only one observation creases: p'p'p*p*.. The probability of a crisis above 800. in a specific year can increase over time if the The coliform data for the country suggests sewage load increases. Thus the probability of that the probability of a health crisis is not not having a crisis in any of the first, second, high, but that it could grow over time as the third, fourth year can be written p3 "p,0 p *p4... sewage load increases. We assumed that if a We arbitrarily assuned that there was a 95 crisis occurred, tourisml would decline until the percent chance of no crisis in the first year of sanitation problem was fixed. The probability analysis, that is, p,=-95. We assumed that each of this type of event is well desenbed by the year, as the sewage load increased, the proba- cumulative probability distribution that results bility of no crisis would decline by one percent- from a binomial. A binomial distribution has age point, that is, p2=-94 in year two and two conditions: (a) crisis, and (b) no crisis. The p3=.93 in year three. Table 4-1 shows the re- data suggest that in any given year the sulting assumptions with respect to the probability of 'no crisis" (p) is relatively large. probability of a crisis having occurred by a The probability of at lcast one crisis can be specificd year." forms per 100 milliliters. ° EIvcn vcry high Ievls of readings do not appear to have dramatic effects. A study of Hong Kong beaches with contam- inauions ranging lirom 411 to 3,200 fccal coliorom pcr 100 milUiliters showed an illness rate or 22 per thousand to 12 per thousand on beachcs under 400 coliforms per 100 milliliters (Holmes 1990). The economists who worked on this project believe tlar these probability estimates overestimate the risk of a crisis, 110 The probabilities imply that there is a 51 high of 26.8 percent The use of risk analysis percent chance that a hcalth crisis will have al- indicates that, with the range of estimates that ready occurrcd by the year the consultant prj- we believc reasonable, the project could be ei- ects that coastal waters cxceed EPA's extreme ther feasible or not feasible. The best estimate standard. We assurned that once the project is of the project's rate of return is 8.2 percent'3 completed, the probability of a crisis is reduced The expected net present value of the project is to zero. -USS1 1.7 million. The second factor that is impossible to pre- If the country is risk neutral, it should not dict with certainty is the amount of decline that undertake the project If it does, it can expect will take place once a problem is perceived, to lose USSi 1.7 million. This is not the end of We assumed that, in the absence of problems, the story, however. Risk analysis allows us to tourist visitor days would grow between 0.2 Icok at the cxtremes. If the country carries out percent and 2 percent per year. If a problem the project, the worst loss it could face would should occur, we assumed that these tourist be S45 million (a little less than the capital cost days would declineforever'2 by somewhere be- of the project). If it does not carry out the proj- tween 0 and 50 percent. We assumed a large ect, the worst loss it might fhce is S127 million. decline was extremely unlikely, because the A country that is risk averse might decide that type of health problems that are likely to result it is worth taldng a probable loss by building a from the contamination now occurs at other in- project that is not needed currently, tO avoid ternational resort areas. We believed that it the possibility ofa very large loss if a crisis would take a true crisis (typhoid or cholera) to should occur. In this sense the project would be provoke anything larger than a 20 percent de- viewed as an insurance policy. dline, and we assumed that there was only a 15 It does not make a lot of sense to look only percent probability of this. More probable (85 at the extremes. The probability of either of percent) are declines of 20 percait or less. these extreme losses is I in 500 or 0.2 percent. The third factor not known with certainty It is better to take a more detailed look at the is the value of expenditure lost Based on stu- probability of losses and gains, their range, and dies of tourist expenditures, the import compo- the expecttion (average to be expected). Risk nent, and the shadow price of labor, we analysis allows us to do this. estimated the net foreign exchange value of Table 42 indicates that the probabilitv of tourist expenditure at between USS30 and 540 gains (that is, the probability that the project is million per year. The economic benefit of the feasible) is small-only 21.2 percent This is project is computed as the difference between the same as saying the probability of large net axpenditures (valued in economic terms) losses, if the country does not undertake the with and without the projec project, is small Even a risk-averse country might not want to pay a large insurance pre- Results of the Study mium to insure against losses wvdh remote Figure 4-1 shows the range and probability chances of occurring. Another way of looking distribution of the results simulated. The rates at this is to calculate the cost of making the of return range from a low of 0.8 percent to a wrong decision: building the project when it since the mean-on which all epidemirlogical calculations axe based--will still bc far beiow 200 fecal coliforns per 100 milliliters in the years shown. This specific estimate was the result of a negotiation witnin 'bc Inter-American De- velopmnct Banlk that considered the inadequacy of the sample and the reluctancc to nm the risk or rejecLing a projcct when it should be accepted. Z It should also bc noted that it is highly improbable that a significant crisis would go on forever, as has becn assumed. If a significant decline in tourism occured, hotelierm would take measures and the govermnent would regulate the dump- ing of wastc. The government could also carry out the projctL Thus, the true upper bound benefit that should be as- signcd is the surm of four elements: (a) the cost of the project, (b) the lost cxpcnditure from thc beginning of te crisis Lo the cnd of the construction period, (c) the cost of advertising and promotion to bring tourism up to the levcl it would havc attained withouL this crisis, and (d) the lost expenditure during the recovery period. This analysis does not trun- catc the benceit so the extreme rate of relurz calculaions are substantial overestimates, wnd the average results are pulled up by these extremes T -his corresponds to the mcdian rate of return (fiftieth percentile). The 1 ~~- *#*44 *#&4 "P* * _ 944* t *tt4ti 4+. or *+ . it T 44i44+ V, . 1* *4T -tbi 4*4*' 0-~~~ + it 4..4. 4*4* i n4.4 rzd + 4i**~~ *~~*** 144'i 4 tt*ftt t-t+t iri4i4 *44*4 ff*04*4* t +tt*44 4t444 ni.* tittt *4**4 414+4 *+44* Tiitti tf+44 itiii -titit 1*44. 1*.4- *iTi ------ ;- i . I I i .8-.E -.-.8.:?P 2&.78 Internal Rate of Return Figure 4-1: Frequency Distribution of Simulated Rates of Return turns out not to be feasible, or not building it that the project is feasible. The analysis indi- when it turns out to be feasible. The cost of ei- cates that the most important variables are ther errc.r is called the expected opportuni4y those ftat have to do with (a) a decline in the loss. Table 4-3 shows that the expected oppor- tourist industry (the probability of the outbreak tunity loss of building the project when it of disease, the amount of decline in tourism, should not be built is three times greater than the amount of net benefits from tourist expen- not building the project when it should be built ditures), and (b) the contingent valuation esti- The information on which the calculations mates. The estimates of benefits from the have been based is flimsy at best Under some recovery of the fishing industry and slowing the assumptions believed possible, the project is erosionof beaches are not important sources of feasible. The cost of uncertainty, because of error. Fishing and erosion benefits could be imperfect information, is the expected oppornu- three and four times higher and our conclusions nity loss the country incurs when it takes the about feasibility would not change. It is not best decision: in this case to reject the project. worth spending any money getting data on If the country could buy perfect infomiation, themL the most it would rationally pay is S5i8 million, Both the tourism estimates and the contin- which is the loss it could expect to avoid if it gent valuation estimates are related to the bac- had perfect information. The risk analysis tech- teriological quality of the water. If the water nique makes it possible to identify what in- were perceived as risky, nationals' willingness formation could make us change our to pay and tourists' adverse reactions would conclusion, that is, might lead us to conclude probably be much greater. There was no 112 Table 4-2: Probability of Losses and Gains if the Project is Undertaken (dollar amounts are present values) Possible loss Probability (USS millions) (percent) 0-9 9.8 9-18 15.0 18-27 26.4 27-36 23A 36-45 4.2 Probability project not reasible 7S8 Possible gains Probability (USS miltions) (percent) 0-25 11.4 25-51 7.4 51-76 1.2 76-102 0.8 102-127 0.4 Probability project feasiblc 21.2 evidence that people, in fact, thought there was of the study was less than US$100,000 (much anything wrong with the coastal waters, and less than the cost of uncertainty) and should re- the limited sample that was carried out con- duce much of the uncertinty. The study is be- firmed that people were . .rect in their percep- ing carried out beach by beach because Bank tions. staff think that there may be more limited solu- Alhough a couna, inat was highly risk tions to a specific problem that could postpone averse might still want to carry out the project the need to invest in a large public systecm. proposed, the Bank-in view of the existing in- fonnation-did not want to encourage the Assessment of the Methodology country to do so. The Inter-American Develop- and Lessons ment Bank concluded from the risk analysis We are generally satisfied that the results that it was critical to be sure that there was not, from the contingent valuation study are believ- in fi&ct, a bactenological problem. It recom- able and that the risk analysis method produced mended a statistical study to test the null hy- results sufficiently robust to justify the decision pothesis that coastl water quality at each that was made (that is, not to go forward with beach complied with EPA standards. The cost The Questions of a Public Se werage System in a Caribbean Country: A Case Study 113 Table 4-3: Expected Opportunity Loss of a Wrong Decision (dollar amounts are present values) Erpecied opporiuniy loss (USS n72illions) Build project when project is not feasible 17.5 Not build project whcn project is feasible 5.8 Cost of uncertainty 5.8 t.he project without clear statisti evidence the present system compared with that which that a violation of standards exists). would be provided by a public system. There Our first conclusion, however, is that risk was only one valuation question, and no oh- analysis is not a substitute for information. The vious way to separate these two elements. This terms of reference for fture studies should posed a problem because the consultants had carefully specify the environmental relations made a separate estimate of the cost savings to that the analyst will subsequently need to do prospecive customers, and using both the CV his work. The study that supported the project estimate and the cost savings would have prob- in question did not deal with reefs, fish, beach ably involved a double count Although a erosion, or willingness to pay. Its primary con- group outside the service area was asked only cern was the quality of receiving water for sew- about keeping coastal water usable, the means erage discharge, not the quality of coastal of the two samples could not merely be sub- waters where people swam. Increasing empha- tacted, because the choice of a residence sis on environmental effects of projects and the proximate to the beach may indicate that these increasing number of projects whose primary residents valued water and its quality more objective is to improve the environment will re- highly, and the groups may have been system- quire a broader approach that must be built in adcally different. from the beginning. The risk analysis method A second problem with the CV question does, however, help identify what infomation was that it did not specify when the water is worth collecting (thlt is, could affect the de- would become unswimmable. It merely speci- cision), and it can be useful to do risk analysis fied that it would eventually become unfit for even with very preliminary data, in order to de- swimming. In addition, there was no informa- sign a cost-effective study. tion on the views of the respondents on how Our second conclusion is that more atten- contaminated the waters were or what their im- tion should have been paid to the way the con- plicit time frame was. Clearly the perceived im- tingent scenario was conceived. With the mediacy of the problem should have some benefit of hindsight some improvements could impact on the amount people would be willing have been made: The question to people who to pay to solve it were to receive sanitary services really com- The problem of urgency and consequences prised two elements. One had to do with the raises a more general problem with the method value of keeping coastl waters clean; the other The problem is how to express the implications had to do with the quality and cost of service of of the problem. If the interviewers had stressed 114 the inmediacy of the problem or dramatized the possible consequences, they may have started a panic-which could have set off the tourist crisis that the project was to help avoid. If the interviewers had put the question in a low key way, the respondents might have assuned that the problem was remote and not worth spending much money on. Because dte results of the contingent valua- tion were ultimately used in Monte Carlo simu- lation, it would have been useful to have had the information necessary to simulate the confi- dence interval around the willingness-to-pay esmates. It is probably worth noting that the true margin of error is probably greater than that which would have beern estimated from the statistical results. Bibliography Feachem, Richard 0a, David J. Bradley, Henda Garelick, and D. Duncan Mara. 19813 So- nitahion and Disease Aspects of Excrela and Waste Water Management. New York John Wiley and Sons. Holmes, Paul. 1990. "Bathing Beach Pollution, Research and Control in Hong Kong," UK Journal Institute of Water and Envi- mrnmental Management. London. Kassalow, Jennifer, and Diane CameronL 1991. "Testing the Waters, A Study of Beach Closings in Ten Coastal States?' Natural Resources Defense Council, New York Mitchell, Robert, and Richard T. Carson. 1989. Using Surveys to Value Public Goods: the Contingent Valuation Method. Washington, D.C.: Resources for the Fu- ture. Rosenberg, M.t, K.K. Hazlet, J. Schaefer, J.G. Wells, and P.C. Pruneda. 1976. "Shigel- losis from Swimming," Jounal of the American Medical Association, 230 (16), pp. 1849-52-.Annex A. 4. Questionnaire Suawev for Houselhold ConnectiOns 11s Annex 4A: Questionnaire Survey For Household Connections PREAMBLE 2.4 How many people live in your house? Good momninglaflemoon. The Departnent 2.5 How many people I8 or over? of Health is conducting a survey on the Basi- 216 And what number best describes the bility of installing a public sewer system in this total income of all persons in the commnuity. The interview will take just a few household for 1990? number minutes, and your views are important for the 1. Less than $3,000 study. Your answers are confidential. Before I 2. 3,001 to 5,000 start, may I ask how long you have lived in this 3- 5OO1 to 7,500 house? 4. 7,501 to 10,000 Years/Months 5. 10,001 to 15,000 6. 15,001 to 25,000 1. DENTIFICATION OF RESIDENCE 7. 25,001 to 35,000 1.1 Address 8. 35,001 to 50,000 9. 50,001 to 75,000 1.2 What is the tenancy of your l0. More than 75,000 residence? 11. No response or don't know 1. The dwelling is 2.7 Have you seen the recent TV cover- 2. owned () age on the potential for marine pollu- 3. rentd/leased i) don here on the island? 4. otber () L yes() 1.3 The land is 2. no () 1. owned () 3. CONDITIONS OF HOUSE 2. rented/leased () 3.1 Is your water piped into your house? 3. other () 1. yes() 1.4 Type of residence 2. no() 1. chattel () If yes, skip to 3.3. 2. blockwall () 3.2 Into your yard? 3. wood and wall() l. yes( 1.5 Number of roonis 2. no () 2. RESPONDENT CHARACTERISTICS 3.3 Where does your sink and bath water 2.1 Sex drain (water from kitchen sink, laun- 1. male() dry, and shower/bath)? 2. female() 1. sewage well() 2.2 Age 2. septic tank() 2.3 Does your work or that of members 3. garden() of your household depend directly on 4. street () business from tourists? Examples: 5. pit () iaxis, restaurants, hotels. 6. public drain () 1. yes() 3.4 Where does the sewage from the toi- 2. no () let drain? 116 1. sewage well() 2. 1 to 15 times per ycar() 2. septic tank() 3. morc than 15 times pcrycar() 3. pit toilet() In order to keep beach water clean for 3.5 Do you share toilet flcilities with swimming and to eliminate the potential threat other households? to coral reefs, we need to build a central sew- 1. yes() erage system. This system would collect waste 2. no () and sewage water from households and estab- 3.6 How long has the septic tarik (pit or lishment, carry it though a pipe to a sewage sewage) been in use? treatment plant where it would be treated to 1. 0 to 5 years () remove damaging pollutants. It could then be 2. 6 to 10 years ( ) discharged safely into the ocean far from the 3. more than IO years () shore. However, this is expensive. Substantial 3.7 Have you had any problems with construction is required. One way to pay for your sewage system in the iast three part of the system is with a quarterly sewerage years? charge covering hookup and continuing oper- 1. Problem Yes No ating costS. 2. Overflow () () 3.12 Which of the following would you 3. Flooding () () choose? 4. Blocked () () 1. Pay a quarterly sewer charge of 5. Other () () S for public sewage dispos- 3.8 Do you think you will have to replace al and clean beaches. () your septic tank (pit or sewage) with- 2. Not pay the $ quarterly in the next year? sewer charge and continue with 1 yes() the current system. () 2. no () If yes, skip to 3.14. If no. Within the next five years? 3. yes() 4. no () 3.9 How much have you spent in the last year for operation and maintenance or repair of your sewage system? o PREAMBLE TO CV QUESTION Currently much of the community's waste water goes into the ground. This water drains through the ground into the ocean. The wastes in this water will eventually contaminate the beach water, making it unfit for swimming. This contaminated or polluted beach water might also discourage tourism to our island. There is aLso a possibility that polluted ocean water can damage some of the coral reefs. 3.10 How important are clean beaches to 3.13 What was the main reason you said you? NO or would not answer? (Ask and 1. very important() then categorize yourself) 2. important () 1- Does not use public beaches C) 3. not importat() 2. Does not want a higher watcr bill 3.11 Do you or members of your house- 0 hold visit beaches? 3. Can't afford it( 1. never() 4. Wants to pay another way () 4. Queson;naire Sumeyv for Household Conneedons 117 5 Not enough information () 3.15 By installments ahded to your quar- 6. Does not understand question () terly sewer rate? 7. Cannot decide () 1. yes() 8. Other () 2. no () 3.14 If thegovernment installed the house 3.16 If the public sewer system is con- connection on your property daring structed would you construction of the sewer system, 1. connect wvith the system immedi- would you be prepared to meet the ately ( ) cost of connecting by a long-terrn, 2. connect eventually () low-interest loan? 3. not connect to the system() 1. yes() 2. no() 5 Environmental Economics and Natural Resource Management: The African Development Bank Experience Anil Markandya and Carlos Mufioz This paper reviews the environmental isswes of greatest importance on the continent of Afric, and examines how projects carried out by te African Development Bank have tred to address them. Methods of evaluation are di&scued avdseveralprojectsrewewed ret- rospecfively. These include direct invesnent projects, as well as some sectoral loans. The paper concludes ta more can be done to value enviromnental impacts systematically, and that such valuaion needs to be extended, especially to cases where there are environmental damages and to sectoral loans. 120 The environmental crises of Africa are appraisal and sectoral planning, as carried out among the gravest in thc world. The continent by the African Development Bank. The focus is suffering fiom population growth of 3 per- is on the economic valuation and planming, and cent per year, which is creating pressures on its the interfice between that and the environment natural resources and resulting in them being The paper first reviews the ADB's formal used in an unsustainablc way. Forests are be- environmental policy. Next, it elaborates on the ing lost primarily to agricultural encroach- problemns integating the economic and environ- ment, at the rate of 3.6 million hectares a year, mental dimensions of projects; it lays out the or 0.6 percent of the stock. Soils in some rc- basic questions that arise with regard to nation- gions are bcing degraded to an cxtentthat is al and sectoral macroeconomic policy and its alarming and lands unsuitable for cultivation implications, both for the success of individual are being finmed from the sheer need to expand projects, and for policy relevance to the design output under declining soil productiviLy and in- of sectoral lending programs. Thhe paper re- creasing population. The problem is particular- views twelve of the ADB3's projects in which ly severe in the drylands of the there was a significant environmental dimen- Sudano-Sahelian region. Biodiversity from its sioni Some are still at the planiing stages, and tenrsial and marine resources is being irre- others are in the process of being implemented. versibly lost. They concentrate on land-based activities (as This dismal environmental picture is not opposed to industrial projects, where different brightened when one looks at the economic de- issues arise). Projects covered include a dam velopment of the region as a whole. Economic project, rural road construction projects, agri- growth has barelv kept pace with population cultural expansion, a resource conservation growth, so that living standards, even as con- project, and rangelands improvement projects. ventionally measured, are not rising. If one This paper then reviews ADB's activities that were to construct a measure of sustainable in- are more sectoral in nature, and lastly, it pro- come (that is, the consumption level that the vides some overall conclusions and combination of natural physical and human recommendations. capital that the continent can support in perpc- tuity), it is certain that the figure would have Environmental Policy Paper been declining over the last twenty years. So it In June 1990, the ADB presented an Envi- is not the case that the environmental degrada- romnental Policy paper (ADB 199 la) that: tion is buying' economie development through * Assessecd the state of the environment in the creation of other forms of capital. Africa In the face of this monumental challenge, * Ouied policies for its main areas of lend- the African Development Bank (ADB) has * Outlinedpthesoriromneason- been maldng a serious attempt to arrest the inc that addressed the enviroemnenal con- losses and to contrilute to the move to a more cenis described above. sustinable development process for the conti- * Presented a set of procedures for the con- nent. To some extent, its main foous of activity duct of projects that took proper account of (agriculture) has always been involved in the envirornental concems. promotion of agricultural practices that would * Discussed ways in which regional member result in sustainable increases in living stan- countries (RMCs) could be assisted in imn- dards. In recent years, along with the rest of the plernenting environmentally sound policies, development community, it has become even and in setting up legislative frameworks more aware of the complex ties that link activi- that improved the management of natural ties in one sector with the use of resources in resoures in their respective countries. another, and with the need to allow for these That documcnt includes many rcom- linkages when planning its involvement at the mendations regarding policies that need to be project or sectoral level. followed for a sustainable use of resources in This paper looks at the integrtion of envi- each of the sectors. Broadly spealing, the re- ronmental concens in the methods of project commendations can be divided into those 5. Emwironmenial Economics and Natural Resource Management 121 dealing with projects and those dealing with environmental benefits. This is a point that sectoral and economywide policies. Each is emerges fron a detailed review of some of the considered in turn below, projects that the ADB has recently undertaken. The ADB, in its attempt to achieve maxi- Project Appraisal and Environmental mum benefit fiom its resources, needs to ad- Issues dress both negative and positive impacts, and The ADB environmental policy document to integrate the analysis of the environmental is primarily concemed to ensure that projects and economic impacts. This is not an easv task, undertaken by the ADB are not environmental- but tools for the monetary quantification of en- ly damaging. It is proposed that projects be vironmental costs and benefits are available classified into four categories.' namely- and could be more exploited. Recent projects e Category I-projects that mav have sig- have started to do this butLthere is scope for nificant environmental impacts and that more integration of the two areas. will require a detailed Enviromnental in- The underlying philosophy behind the pres- pact Analysis (ELA) ent approach is that where there are Cunacceptable7 environmental benefits, they Catagory Im-proaectst o at ted environ- should be mitigated and the costs of mitigation included as part of the cost of the project. IE gated tlirough routine measures and gtes t hrough rouine m alter allowing for these impacts, the project meets the rate of retum and other criteria, it * Category Im-projects for which environ- may go ahead. However, this strategy has three mental impacts are not expected and that important implications First, it imposes a re- should not need a detailed environmental quirement on projects that may raise their costs analysis to such an extent that they will become unvi- * Category [V-projects that have an envt- able. (This cannot always be discerned by ronmental sustainable development as a fo- looking at project reports, because ADB man- cus and for which "separate environmietal agers tend to select for consideration only those analysis may not be required" (ADB projects where there is a strong prima facde 1991b)- case that the project will meet the required rate A similar classification has been esab- of returm) It is reasonable to ask, are the miti- lished by other development barnks, such as tie gation requirements always justified in terms of inter-American Development Bank (IDB), and the damage avoided? Or should one impose a serves a useful purpose in screening projects so lower level of mitigation requirement? This is tat the most important ones are allocated the not merely an academic issue. Many project scarce professional resources at the bank's managers can recount cases where, in their disposal. opinion, projects have had to be abandoned be- HoNwever, it is clear tiat, in arriving at the cause the mitigation costs are too high. How allocation, the emphasis is on the avoidance of should one decide on what level of mitigation negative environmental impacts, rather than on to demand? This can be answered partly by re- the active canvassing and development of proj- course to the quantification of the value of the ects with environmentally positive aspects. The damage avoided. But such monetarv quanti- allocation of projects concemed with the con- fication is not always possible or, if possible, servation of resources to category IV means gives answers tiat are too uncertain to be use- that they will not have to run the gaundet of en- fill. In such cases one is obliged to use rules for vironinental scrutiny that projects in category I deciding on mitigation, but these rules need to will have to. Yet they may be precisely the ones be set out more formnaly, wihin a framework that need careful examination, to see how they of a set of sustainabilitv criteria. Both these are can best be designed to maximize Subsequently, the ADB has decided to merge categoies U? and m into one, vith all proiects under either categoly now not requiring detailcd enviomental analysis, but requiring attention to project design on envomental grounds. and an environmental mitigation plan. 122 discussed fiurter in the next section of this development The links between macrocconom- paper. ic and sectoral policies and these goals are only The second difficulty with the prcsent ap- beginning to be understood but from what proach is the opposite-that is, that where knowledg,e is available, they can be extremely mitigation is not required, or not possible, the important.? The ADB environmental policy projcct may still be acceptable, but may cause document rightly alludes to the importance of considerable envirornental damage. In other policies in the areas of pricing of forestry re- words, the residualr environmental damage sources, agricultural pricing, roads and rural may be quite high. For example, a project that development, and others where the economic involves the clearing of forest land with many strategies chosen have an impact on the envi- indigenous species may be "mitigated" by the roment in ways that are not alwvays obvious or planting of a single species of equal land area direct. It also mentions the importance of prop- elsewhere. One is not a replacement for the er accounting of environmental assets as im- other, however, and some damage remains. Un- portant for the sustainable use of natual less there are ways of accounting for it, the assets. present approach could lead to poor choices Macoeconomic policy has to be fonau- and unsusuinable practices. lated in the areas of Third, the point made earlier about the cur- * Pricing of natual resources (water, forcsts, rent emphasis on negative impacts needs to be land, energy). repeated- Where one is seelg to justiufy eM * Pricing and subsidization of agricultura ronentally positne projects, such as resource inputsand suts. conservation, or dune prtecon, it is hard to Eputs and outputs. quantify the benefits. Choices are then made on * Exchange rates andtrae controls. the basis of a limited quantification of benefits, * Govermnent taxest which in murn means that there is much more * Govnment d between urban arbitraniness m the decisions made. or that the and rual areas. project is justified only on the basis ofthe mea- * Taxes on polutants. surable benefits, which would result in an Tis policy must be designed in ways that under-allocation of resources to such projects. rthe pipacts that such polices have on The essential messages from this discus- t he im pats tbats polion abon sion are: (a) there is a need to extend environ- the environment. This needs inoirmation about mental cost and benefit estimation and (b) tnhe lages at a quanttive level, tbat is by where this is not possible there is a need to for- and large missing for most counthiesc Thus it malir the use of sustinability critria to es- must bea prioritytoassist RMCsint e collec- tablish gidelines for mitgation action and for tion of the relevat information on which to resource conservation projects These are is- base sound polices for sustainable develop sues that all development banks and bilateral ent. However some indications ave emerged donor agencies are grappling wim and the Afri- ofWte importnce of certain areas. can Development Bank is not alone in facing Work by the World Resources Institute, ther. Auaby, wweer, ts eed inooning and otChers (Repetto 1988, Panayotou 1989, themn. Arguably, however, its needs in coming and Markandtya 1990) has shown th;at often it up with solutions are as great, if not greater, is Mae undei of key cown that is than those of any other ag,ency involved in in- is the underpricing of ka:y commodities that is temational lendin resulting in their excessive use. This applies to water, tmber and firelwood (with low stum- Macroeconomic Policy and the page charges), fertilizer and pesticides, and Environment food and energy subsidies to urban consumers. As important as a proper framework for If such polices are in place, there will be a con- the analysis of projects is the issue of how the tinuing pressure on certain resources that will ADB can encourage maeconomic polices remain even if new projects are initated to in- that are consisten with the goals of susinable crase forest area, or supply more water for 2 For a survey of the issues involved, see Mrkidya and Richardson (1990). £ Envrvenbal Economics and Natural Resource Manraeenrg 123 irrigation etc. Resolving them however, runs (BCA) is that it does not specify the dimension into several political and social conflicts, of the problem in a sufficiently wide mamner to created by vested interests. Someimes under- encompass all the environmental ramifications. pricing is justified on the grounds of income Projects such as penetration roads or hydro- support fbr the poor, but this is rarely proven. dams could change the use of land and the na- Even where this is a valid reason, altematives tirc of economic activities in an entire region, that are less damaging environmentally need to with secondary and indirc effects fhat swamp be considered. the direct ones. A benefit cost analysis, or even It is not just macroeconomic policy that is a cost effectiveness analysis, based on a nar- of importance but also the sociolegal context m row interpretation of the costs and benefits which it is carried out Where traditional sys- would be misleading in such a situation. In tems of resource management are breaking broad terms this concem is correct but what it down, and new ones not replacig iern, there points to is not a rejection of BCA as such, but is the contining danger hat any attenpt to a call for the use of a wider and more efifctive conserve existing resources or to expand those BCA. Indeed, one of the criticisms of many of that lie in disputed ownership, will fi&l It is not th traditional applications of BCA that are easy to address this issue at has bedeviled made in the next secton are precisely along many projects, particularly those concerned these lines-hat is, that the analysis does not with reafforestation and rangelands improve- look at the implications of the project widely mnent However, a failure to rake a realistic and deeply enough. The fact that development view of what can be achieved may result m re- programs and investments have many indirect sources being wasted in projects that do not impacts makes decisions regarding tien more achieve their desired goals. complex, and may even require the use of gen- ,eal equilibrium methods or macroeconomic Integrating Economic and models to obtain reasonable solutions. In these Environmental Concerns circumstances it would appear evn more nn- portant to document and value the envionmen- Role of Benefit Cost Analysis tal impacts, so that dthy can be assessed along As was explained in the last sctin, the wih the other changes. monetary valuation of ninomnl impacts is Sepcrabilfty of/the Environmental importait fir the rional allocation of re Dm- sources to the envrnmnent hI recent years many techniques, which are described later -WVhen carrying out a BCA, it is nonnal to have been developed for such a valuation, requie fitat ea independent component of the However, the results from the application of proqect sati the net present value (NPV) or these techniques have to bu credible and ac- intenal rate of return (IRR) test Some envi- ceptable if they are to be adopted. Some of the ronmeists argue that this is unasonable issues and doubts tat anse from the monetary when applied to environmental compont of valuation of environmental benefits and costs projects, essentially because the benefits are are as follows: not properly measured. An example would be * Specifying the dimenions of the problem. the sugar plantation projects where the eviron- * Searability ofthe environmental problem. mental components are water and soil con- servation measurcs and replanting of some lost * Discounting of costs and benefits f * Sensitivity analysis and the environment Where the environmental component is * Aeptability of the monetary vahues separable in a technical sense, it makes sense to treat it separately if the benefits and costs asso- Specifying the Dimensions of the Problem ciated with it are adequately and filly mica- sured. If tat is so, there is no reason why the Oneconernexresed yevionmnta- nviomna component should not be subject ists about the use of benefit cost analysis e m- ca mponent s. to the same rules as the other component. 124 However. the component is not always separa- environment (see Markandya and Pearce ble, and if separable, measurability is frequent- 1991), where the thrust of the argument is that ly nct possible. Therefore, it is necessary to use (a) it would be better if a lower discount rate some altemative rule to evaluating the compo- could be used for all projects, but (b) it is like- nent. One rule could be based on the notion of ly to result in more problems and resource mis- sustainabilitv. This has been referred to earlier allocation ift as some environmentalists have and is based on the concept that, if develop- suggested, a lowver rate is used for ment is to be sustainable in the long term, key "environmental" projects and components. environmental resources need to be protected In fact, the rate of return on enmironmental irrespective of their current value which, in projects need not be so low. The ADB's recent any event cannot be properly assessed im many projects, many of which have a major environ- cases- Exampics would be critical zones, such mental component, manage to meet the IRR as wetlands, vuhich provide a multitude of eco- standard. The key to this lies in quantiting as logical functions. For such resources the sepa- many of the benefits as possible. rability assumption in project valuation is inappropriate. This does not resolve the issue &tZhLOyAnalysisand theEnvironment of the timing of the investment, or the difficult Conservationists have argued that the use problem of deciding Which resources come un- of sensitivity analysis is an indequate substi- der this category However, there is an urgent tute for a proper teatnent of risk, especialy need to define some sustainability criteria so wher emiromnental dangers are concerned. that decisions on what environmental compo- For example, a project may involve dumping nents should be required components of a proj- hazrdous waste in a site One possibility is ect can be defined. It may well be the case that that the dumping may generate groundwater the sustainabilitv criteria should not apply at cortmination, ading to many deaths and se- the project level at all, but rather at the sectoral rious illnesses. This nsk may be reduced by or reaional leveL The issues involved need to project design but cannot be eliminted. In the be clarified further, but some recent thinking event that the leakage occurs, the project would on this subject has stressed the need to meet the obviouslv have a massively negative net value, sustainabilitv criteria at least coSt (see Pearce, but how is one to give that evenualty due Markandya, and Barbier 1990). This will prob- weight? Even a filll risk analysis using an ex- ably requr, some rationalization of pected utility criterion, or a standard decision "compensatng projects so that costs are kept nde such as risk benefit analysis or the maxa- to a minimum. mi rule (see Pearce and Markandya 1989), Discounting Costsand Benefits would not resolve the dilemma, in the sense that the resulting choice regarding the desi- To many individuals concerned with the ability of the project would command universal envirornment, the fict that projects are selected agreement (for example, it migit still be the using an NPV criterion with a discount rate of case that the above "project" would be consid- 12 percent in real terms (as is the case at the ered acceptable). Although the process of ar- ADB), or using an IRR criterion with a similar riving at ajudgment can be better infomied and required rate, is a reason why the enviromnent improved by using decsion analysis methods is inadequately protected by such processes. when possible, there is, unforunately, no seri- They argue that the high discount rates do not ous alternative to a judgmental evaluation of give enough value to futire enviromnenral the risks of a project, whether they are environ- benefits or costs and that the emphasis of such mentl risks or any other risks. alloa:tion rules is to go for short term projects, without paying enough attention to the long Acceptarily andl Coo fidenceofthe term implications. This issue has to be talen Monetary Values of Envronmentoi seriously, but it is more complex than this ar- Benefits gument would suggest. It has received some Although the methods for deriving the attention in the economic literature in the costs and benefits of environmental inpacts S Eavironpentat Economics and Natural Resource Managmeen 125 described in the next section are relatively es- Indirect procedures for benefit cstimation tablished in dcveloped countries (most have do not constitute a method of finding willing- been in use for around twenty years), they are ness to pay (WTP) for the enviromnental bene- less familiar to decisionmakers in developing fit-or the willingness to accept (WTA) countries. This raises the problem that the fig- compensation for environmental damage suf- ures derived may not havc the credibility and fcred. What they do is to estimate the relation- confidence they need to have to be acceptable. ship betwee the "dose" (pollution, for There is no simple answer to this; time and ex- example) and the nonmonetary effect (health tensive use of a technique are needed for it to impairment, for example), and only then is be considered acceptable. In Europe and North some measure of preference for that effect ap- America, environmental benefit estimation has plied. Although they have this obvious weak- not been equally effective in all situations, nor ness, they are, nevertheless, frequently used in have all the methods found equal acceptability, valuing environment changes, especially in de- but the technique is becoming accepted as part vweloping countries, where the infbrmation of the decisionmaking process Furthermore, available often only pennits the use of such the techniques described below are now begin- methods or where data for the calculation of ning to be used in project appraisal in develop- preference based esfimates is limited. ing countries (see below). Direct Valuaton: The Hedonic Price Techniques for Monetary Valuation Approach This is not intended to be a survey of val- The value of a piece of land is related to uation techniques for enviromental costs and the stream of benefits, but to be derived from benefits, but rather an indication of areas that land. Aicltural output and shelter are where particular techniques may be used suc- the most obvious of such benefits, but the envi- cessfilly and where there is scope for their ex- romnental quality of the neighborhood in which tension in the ADB's appraisal activities. the land is located is also important The prop- The approaches to the economic measu erty value approach to the measurement of ment of environmental benefits can be broadly benefit estimation is based on this simple un- classified as direct and indirect techniques. derlying assumption. Given that different loca- The former considers environmental gains-an tions have varied envirommental attributes, improved level of output in agriculture, better such variations will result in differences in levels of air quality or water quality, and so property values. With the use of appropriate forth-and seeks direcdy to measure the money statiiucal techniques, the hedonic approach at- value of those gains. This may be done by tempts to (a) idenify how much of a property looking for a surrogate market or by using ex- diffirential is due to a particular environmental perimental techniques. diflrence between properties and (b) infer how The surrogate market approach looks for a much people are willing to pay for an improve- market in which goods or factors of production met in the environmental quality that they face (especially labor services) are bought and sold, and what the socal value of the improvenent and observes that environmental benefits or is costs are fiequently attnrbutes of those goods The identification of a property price effiet or factors. Thus, piped water is an attribute or due to a dififrence in pollution levels is usully feature of a house, a nsky environment may be done by means of a multiple regression or a feature of certain jobs, and so on- The exper- similar techmique in which data are tak either imnetal approach simulates a market by plac on a small number of similar properties or ing respondents in a position in whch tey can pieces of land over a period of years (time se- express their hypothetical valuations of real ries), or on a larger number of diverse proper- ipovemens in specific environments. The ties at a point in time (cross section), or on aim is to make the hypothetical valuations as both (pooled data). In practce almost all prop- real as possible erty value studies have used cross section data, 126 as controlling for other influences over time is context in which it would be provided, and the much more difficult. way in which it would be financed. Hedonic price techniques have been suc- One major attraction of CVM is that it cessfully used in estirmating the costs of air and should, technically, be applicable to all circum- noise pollution, and of changes in amenities, in stances. Its aim is to elicit valuationsr developed countries. Their use in developing bids-that are close to those that would be re- countries has been more limited, although the vealed if an actual market existed. The hypo- World Bank and the Inter-American Develop- thetical market-the questioner, questionnaire, ment Bank have used them for the valuation of and respondent-must therefore be as close as improvements in sites and services (sewerage, possible to a real market The respondent must, potable water). There is a considerable body of for example, be fimiliar with the good in ques- literature on the issues arising in the applica- tion. If the good is improved scenic visibility, don of these techniques (see Markandya 1991). this might be achieved by showing the respon- The question that is at the center of this dent photographs of the view with and without discussion is, are hedonic price valuations reli- particular levels of pollution. The respondent able and accurate? The difficulty, of course is must also be familiar with the hypothetical tlere is no absolutely correct yardstick aga means of payment-for example, a local tax or which to measure the reliability. Lf there were it direct entry charge-kmown as the payment would not be necessary to enrgge m hedonic vehicle. pnce approaches! It is thus in the nature of The questioner suggests the first bid (the non-market valuation that accuracy and reh- starting point bid [price]) and the respondent ability have to be tested by other means. The agrees or denies that h or she would be willig main tests are: to pay it An iterative procedure follows: the * Consistency of results in similar contexts. strting point price is increased to see if the re- * Consistency of results with other benefit spoadent would still be willing to pay it, and so estimation techniques. on untl the respondent declares that he or she estimation techniques. is not wiling to pay the extra increment in the * Consistency of results with 'real maret" bid. The last accepted bid, then, is the maxi- cxperiCince. mu wiingness to pay (MWT1). The process On the basis of these tests, considerable wors in reverse if the aim is to elicit wiing- evidence exists to show that hedonic prce val- ness to accept (WTA): bids are systematically uation, properly executed, provides reasonably lowered until the respondent's minimum WTA reliable beaefit estimates in areas where there is reached. are well-fumctioning land and property markcets. A very large part of the literature on CVM Direct Valuation: Contingent Valuation is taken up with discussion about the accuray of CVM. Accuracy is not easy to define. Blut The contingent valuation method (CVM) since the basic aim of CVM is to elicit arear uses a direct approach-it asks people what values, a bid will be accurate if it coincides they are willing to pay for a benefit, or what (wthin reason) with one that would result if an they are willing to receive by way of com- actual market exist But since actual markets pensation to tolerate a cost, or both. What is do not exist (otherwise there would be no rea- sought are the personal valuations of the re- son to use the technique), accuracy must be spondent for increases or decreases in the tested by seeing that quantity of some good, contingent upon an hy- * The resulfing bid is similar to that achieved pothetical market. Respondents say what they by other techimques based on surrogate would be willing to pay or wiling to accept rfa markets (house price approach, wage stu- market exsed for the good in question. A con dies, and so orth). tingent market is taken to include not just the esul2mg so forth). - . * - . ~~~~~~~* The resulting bid is similar to onIe achieved good itself (an improved view, better water quality, A s ft, b a t i by introducing the kinds of incentives that exist in real markets to reveal preference. 5. Environmental Economics and Natural Resource Management 127 One significant feature of the CVM litem- Indirect Valuation: Dose-Response ture has been its use to elicit the different kinds of valuation that people place on environental The procedures for valuing the use of envi- goods. In particular, CVM has suggested hat ronmental resources thus far have been based existence values may be very important. on individual preferences. However, other CVM has been used extensively to elicit mehcods exist tiat do not seek to measure such values of improvements in water quality, the prefrences explicitly. Instead btey calculate a benefits of less air pollution, and the option and adose- epffet, relationship between pollution existence values of species and sites. Until re- o effect and then they apply a measure cently, the direct use of such techniques in de- of preference to that effect. Such methods are veloping countries was believed to be very reoerred to as indirect methods. Examples of difficult, if not impossible, due to the sophis- dosresponse relationships include the effect tication of the "as if' experiments involved of pollution on health, the effTct of pollution on However, some recent work carried out aquatic ecosystems, and the evact of soil er- through the World Bank on the valuation of sion on agricultual productivity and thee values water benefits in Pakistan :4 Nigeria, and the oftropical forests. For exacple, there are now expericnce of the IDB on water projects, on several fhirly detailed stdies valuing by this tourism valuation, and on sites and services has metod the damage done by soil erosion.3 shown that the technique can indeed be effic- The dose-response tends to be used par- tivelv employed in valuing the benefits of alter- tisniarly for two situations. The first is when it native water supply systems. Views on is thought that people are unaware of the ef relevance of this method are therefore chanmg feel that pollution causes. The second is when fast and it is quite likely that their use will ex- ecitmg preferences by any one of te direct fsteand ito the valuiteikely tof t e iromnena wmpa efmethods is not possible for reasons of data, or tend to the valuation of enviroanental impacts lack of "market sophistication" in the popula- in ag&icuhtnM areas. tion, or both. The second situation applies es- Direct Valuation: Travel Cost pelly in developing counties, where price Travel cost models are based on an emen- and expenditure data are gerally poor and sion of the theory of consumer dmnand,m where, at least until now, the use of contigent which special autenion is paid to the value of . techniques has been limited becuse it is believed that the answers would suffer tim Inthedevloingcoutris, om from strtgc, hypothetical and opera;tional are being used to estimate benefits fiom tour- sesr ism development in countries with game parks beases. (suc as enya Or PCC1 UCLng aeasWhere envirnmental benefit esbimatim (such as Kenyra). or special trekking areas ~ b onre,i (suc as epal- Ante.raofapiain has been undertaken in developing countries, it (such as Nepal). Anlother area of applictions ~ ~ ~ tedrc plcto has bee tD vau beeft of fulwo supl has been mamly through the direct apphlcation has eento vluebenfitsof uelwod uwWof the dose-response fimetion method. Some (or the supply of replacements such as kero- oi toe to ut The usef of tis thnique in sene), where households pay" 'for the fuel- the agrncultural sector (which is a partcular fo- wood by spending time collecting it (see Boyo, ths aDS pect actiit arthefllowing Mailer, and Unemo 1990). Finally the World cus of ADB proect actvity) ax the followmg Bank and the IDB have also been supporting * Ewviromnental costs and benefits are esti- the use of travel cost data as part of the valua- mated mainly for changes in agricultural tion of vater benefits referred to above as well output following land use and land man- as of recreational benefits from improvanents agement programs. However, the dose- re- in beaches. Thus, travel cost models have a sponse relationships on which these are role to play in benefit esimation in developing based are fi-epently quite crude, with no countries and, moreover, a role that should in- allowance being made for the fact that mdi- crease in the futur viduals adapt to changes in their environ- ment. Thus, for example, as soil conditions See, for exmple, the pap on "National Economic Cost of Soil Erosion in Zimbabwe," by David Noise and Resma Saial, in tins volume. 128 changc, different inputs will be used and provide "orders of magnitude' to the size of the different crops grown. Not allowing for benefits, and that some inaccuracy is inherent that would result in an underestimate of the in the nature of the task being attempted. Nev- benefits of such changes. ertheless, the values obtained are useful in * Many cnvironmental impacts arc not val- reaching rational decisions with regard to in- ued in these cxercises. The reasons range vesinents involving such benefits. from a lack of data to an unwillingness to Although there would appear to be little di- use the appropriate techniqucs. The former rect relevancc of nonuscr values to ADB proj- include benefits such as increased crop ects and policies, such an impression is residues and the spillover effects of proj- misleading. Nonuser values such as option and ects. The latter include benefits of con- existence value can, for some natural re- servation pcr se, sources, be so large as to influence conserva- & Where environm-iental costs are involved, tion decisions. These in tum can have a major the impact is ofien dealt with through a re- impact on sustible development. quirement that certain standards be met Sectoral and Macroeconomic Policy and certain levels of protection be Formulation and the Environment achievedL In these cases no cost or benefit Both developig country govements and estimation is carried out The use of such donors are becoming icresigly aware of the methods is sometimes retered to as in tween macroeconomic polices pursued responses- In these cases, the costs of for stablization and growth, and the environ- meeting the standard are considered as part men* As stated in the previous section, these of the costs of the project and often a lnks are nfectly understood but they are equated with the "environmental costs-" important and need to be accounted for in both However, the later is a fallacy, as the in project planming and sectoral lending. Broad- costs of mitigation may or may not be ly, the polices that affect the environment can equal to the costs of the environmental be divided into the short-term stabilization poh- damage. Nevertheless, the use of gated re -des and the long-term structrl adjustment sponses can be inportant in project ap- . * - c praisal, particularly m connection with the p E i criteria of sustainability. How they should The short-term policies seeking fiscal and monetary balance can affect the environment he designed and carried out is discussed in by cutting govrnnt progm that are con- the-next section. cerned with environmental protection. In times * In other instances, the issue of environmen- of crises, it is often the environmental budget, tal benefit or cost estimation is avoided by small as it is, that gets cut. The same pres- comparng different means of achievimg a sures, however, can be used to cut expenditures given goal. The goal has envinmental that are being environmentally damaging. For benefits that are assumed to be desirable. example, the need to cut subsidies, or the need The analysis only concerns itself with the to raise charges for water, will be greater. way of achieving this goal at least cost These can be environmentally beneficial Re- Conclusions on Methods for Valuing the cent reviewvs of govenment spending, taxatior, Condsironsont Mehos orVauig he and public enterprise pricing point to the im- Environment portance of addressing government policy fail- There is considerable scope for the use of ures in these areas if environmental damage is some of direct techniques in the valuation of to be reduced (Panayotou 1989; Bishop, Mar- environmental benefits in developing countries, kandya, and Richardson 1991). Other policies, and in their rural sectors, but this strategy has such as changes in import and export taxes, ex- yet to be substantially exploited. Particular change rates, and direct and indirect taxaanon, promise holds for contingent valuation methods can have ambiguous effects, although one re- and travel cost approaches. However, it is es- cut survey has shown that the inpact of sential to note that such models really only Woild Bank sectoral lending and the polices S. Environmental Economics and Natural Resource Management 129 that it has supported have not been particulariy never quantified in monetary terms. Some damaging (Sebastian and Alicbusan 1989). quantification in physical terms is carried out In the longer term, the more fumdamental but even this is quite limited. The most frc- question of what constitutes a sustainable de- quent approach is to evaluate the project by es- velopment path has to be faced. There is a rec- timating its rate of return in the absence of the ognition that some enviromnental damage may environmental impacts, and then to say that the be tolerated if the economic benefits are large latter are "additional unmeasurable benefits." and if the damage is not irreversible. In later This is the way, for example, that reforestation years, when the country has more resources, it benefits associated with dam construction, sug- can start to tackle these less urgent environ- ar plantations, and so on, are evaluated. Alter- mental issues. This notion of an environmental natively, a project may have many components, Kuzncts Curve is being examined by a number including "environmental" ones, and the ap- of environmental economists (Bishop, Markan- praisal is carried out for the package as a dya, and Richardson 1991). It suggests that whole and not for the specific contribution of policy makers need to distinguish betwem en- the environmental component This is done for vironmental damrage that can be tolerated in the rangelands projects, and for resource conserva- interests of general economic development and tion projects. damage that camnot. At present, however, no Viere projects have negative unpacts, guidelines exist on how one might distinguish there is a tendency to underplay these; and in beween the two some cases a detailed environmental impact If the ADB is to design its sectoral and analysis would provide some much needed in- project lending to be more conscious of the ma- formation of the wider environmental effrcts of croeconomic famework, it wiU need to face a particular project Where a negative impact is these questions, and to collect information on acknowledged, there is an attempt to include an what the links between the environent and element for mitigation. In no case is the poten- macroeconomiic policy are. On a piecemeal ba- tial environmental damage valued in money sis this infbrnation is beginning to be avail- terms. able. In the rLeview of projects and progams All projects reviewed were judged on the that folows, where issues of macroeconomic basis of a 12 percent real discount rate. In gen- policy are relevant they are highlighted, so that eal, the authors take the view that this rate one can see when they matter most should not be adjusted for projects with a sig- nicant environmental dimension. However, if Analysis of Projects: Recent the benefits cannot be quantified, it may be Experience necessary to do so. Otherwise it will be diffi- The object of this section is to analyze the cult to get such projects accepted. way in which environmental issues are beng The twelve projects reviewed here cover a integrated into the economic valuation of proj- wide range of activities, from projects that po- ects receiving consideration for fding from tentially could have a major negative impact on the African Development Bank (ADB) and the the quality of the environment, like dams or African Development Fund (ADF) A group of roads, to projects whose aim is precisely the twelve projects was selected, for which the Ap- opposite-to protect or recover the naural re- praisal Reports were examinedL All but one of sources. Table 5-1 gives the country, the activ- dte projects were initiated in 1990 or 1991, and ity, and the main characteistics of the projects many are stfill under consideration. Hence this reviewed. review would be usefil in comuenting on cur- 1. SUDAN: Kenana Sugar Plantation rent ADB practices, but not of much use in un- Expansion derstanding how these practices have evolved over the recent past. Descrption of the project Environmental impacts of projects may be divided into positive and negative ones. As far Kenana Sugar Estate began working in as the positive impacts are concemed, these are 1975, and has been a successful enterprise. It 130 Table 5-1 ____II COUNTRY J TYPE OF PROJECT ENVIRONMENTAL FEATURES | Sudan Sugar Plantation Reforcstation, Soil and Waxer Conservation. 2 Malawi Smallnolder Sugar Cane Sugar Cane Production & Is lPos.sible Environmcnial Impacts. 3 Malawi Rural DcclopmenL Expansion of Agricullurc Combined with Forest Protection. 4 Tanzania Irrigation and Rice Reforestation, Soil and Water Consemvation. 5 Burkina Faso Irrigation Reduced Water Loss. 6 Ghana Foresty and Paper Mill Substitution of Natural Forest With Commercial __________________ ~~~~~~~~Species. 7 Morocco Rangelands Improvement Soil Preparation. Sowing, Small Danms Dune FLxation. Improved Management of Commons. 8 Ethiopia Rangelands Improvement Resource Restoration. Improved Management or Commons. 9 Mozambique Family Livestock Potential Forest and Rangclands Impacts. 10 MoroCCo Dam Construction Flooding of Land, Resettlement, Watcr Deviation. Reforestation of Catechment Area. I L Tanzania Road Rehabilitation Opening New Areas to Cultivation. Subsidence. Disposal of Spoil. 12 Morocco Natural Resouroe Reforestation Program. Forest Management. Consealion Sustainable Use of Forest Products covers 35,000 hectares, and processes all its total costs are 39.4 million FUA4 and it has an production. The expansion project has six economic internal rate of return of 15.6 objectives: percent. 4. Plant 3,000 additional hectares wit su gar The rate of retun was calculated by look- cane mg at the expected increase in yields of sugar .. Rmse productivity and yields over time and comparing them with the direct costs. No environent benefits were included 6. Create an enterprise fhat uses byproducts for items 5 and 6 above. Hence the marinal to produce charcoal fuel benefits of these components are not estimated. 7. Increase power generting capacity by 10 In fict, the project was appraised as a single megawatts with thermoelectricity entity and no attempt was made to look at the 8. Reforest 1,800 hecte additional costs of the components parts. This 9. Carry out soil and water conservation may be justified on the grounds: measures * Either that the outputs are genuinely "joint The report was submitted to the ADB products" and if the soil and water con- Board in October 1990, the starting date for servation measures were not carded out, the project was April 1991, and the project is the other benefits would not be realized expected to be completed by March 1996. The 4 The Fund Unit of Account (FUA) is the miut of accomnt used by the ADB Group for its projects. 1 FUA USS1.2. It al- lows for dhanging exh rates betwe thL donor cmrencies. I Environmental Economics and Natural Resource Management 131 Or that it is a requirement for the sustain- included in the overall benefits, but they are not able use of resources that if new land is compared with the direct costs. Furthermore it brought under cultivation, some degraded as is argued, the project would reduce defores- land should be conserved, so thal in net tation by reducing the demnand for fuelwood, terms there is limited loss of forest cover. the latter should be quantified and included as Similar reasoning would be used to justify a component. It might, for example, sugget the charcoal fuel production unit. that this component be expanded. Environmental Issues 2. MALAWI: Study for Expansion of A number of environmental issues arise in Smallholder Sugar Cane Production comection with this project. One is the issue of Description of fte project water valuation. Although water is available for the project, its value will depend on its op- The Smallholder Sugar Authority in Mala- portunity cost, which could be higher than the wi is responsible for coordinating sugar cane official price. There is no discussion of the production in smallholder plots. A study on the shadow price of water resources used in the feasibility of the expansion of its activities is project. proposed. It will look at land productivity in A second issue is that of the environmental the prospect areas, the availability of process- and other benefits lost from dte land that is ing facilities and the situation of markets. The convented into sugar production. The benefits possibility of introducing irrigation is also be- of the project are the difference between the ing considered. The sudy will design a full in- two. ft is possible that there are few benefits to tegated program, including livestock be lost fiom existing uses, but it would be use- production. No estimates of the costs and bene- fi to know what they were. No information is fits of the project are available at this stage, given on wh.at kind of vegetation or cultivation but the ADB has allocated around 1.3 illion it is displacing. FUA to a detailed study, including an envimron- Third, the project makes some passing ref- mental impact assessment (ELA). erence to possible negative impacts, particular- _ ly increased breeding grounds for parasites and ThwronmenialIses vectors from the increased inigation, and the As stated above, part of the preparation of depletion of soil fertility if insufficient phos- the project will involve an EIA of the sugar phates are applied, as is currently the case. For cane production expansion In the project's do- both these impacts, the project team has made sign, the consultants are aske$ to bear in mind some suggestions of mitigation. It is not clear, the likely efiects of recommended echnological however, whether these measures will be car- packages They should suggest ways to avoid ned out, and what the consequences are of environmental damage. Refe must be them not being carried out. made to the use of pesticides, the disposal of The project has two environmental compo- waste, and other kinds of air and water con- nents, both of which are stand-alone items, but tamination. Thus the approach is to put togeth - neither is quantified. One is the afforestation er a mitigation package but without refece component, which will produce fuelwood and to the potential benefits of that package or to protect the soil frm erosion. It will be done on its costs. It vould be useful if guidelines as to land not suited for cane, and wili provide a new what is acceptable for these impacts were made wind barrier. This component is justified on the available. They could then be reviewed from need to comply with the new regulation in Su- ime to time, as more information about poten- dan, which nakes it compulsory to have at tial damages becomes available. least 5 percem of the estates with forest More importantly, the project needs to look Hence, no separate analysis is required of its at the implications of this expansion for the use costs and benefits. The same does not apply, of unpriced (or inadeqaty priced) resources, however, to the production of charcoal fuel. such as water. The increased demands may i- The private benefits of this conponent are volve a shift away from other uses, and fthse 132 should be taken into account. The project's Also, the cattle fattening program is said to terms of reference do indicate that the issue of avoid overgrazing. water will be examined, but it is the efficient None of the extemal environmental bene- use that is difficult to plan for in this context. fits (for example, the protcction of forest re- serves, the reductions in soil depletion, and the Project, Mwanza reductions in overgrazing) is measured in monetay terms. Again, as with other projects, Description of the project all components are treated as a unified whole, and no attempt is made to appraise the parts: Mwunza is a resource-iich district in the effect of internalized enviromnental invest- southem Malawi. Having a low population ments, like the soil and water conservation density, the area is now receiving new settlers techniques, is not separated from the efibet of from neighboring districts. The aim of the proj- the rest of the investments. Furthermore, many ect is to improve agicultural and livestock pro- of the suoposed benefits depend on changing duction and to help manage land resources farmer behavior, and on tnany previous occa- effectively in the region. It covers a wide range sions projects have been overaopimistic about of activities: extension, training, rnral credits, such changes. A better appreciation of their and the construction and repainng of rural perceptions of their self-interests is required of roads and water supplies. project plaimers. The date of the report is July 1991, and the The potential negative impacts of the proj- proposed starting date is February 1992. The ect are underplayed (for example, the expan- project would be conducted by the Ministry of sion on nonproteted forest land). The loss of Agriculture (Blantayre Development Division). forest land and its ongoing uses and values It's costs are 10.87 million FUA and the eco- needs to be taken into account nomic internal rate of return, supposing 27 per- 4. TANZANIA: Study for the Irrigation cet of the frrmers adopt the new practices and Project, DAKAWA H use the credits. is estimated around 33 percent Enironmental fssues Descriptfion of the project To avoid the tansformation of part of the Dakawa I was a project that began in the forested areas into new ag-icltural and live late 1970s It cleared 2,110 hectarcs of land in stock ranching land, one of the prerequisites for the Morogoro region, devoting 2,000 of them the loan is that the government of Malawi must to rice production on irgated land. The project designate, gazette, and maintin the forest re- also built a rice mill corplex, roads, and sever- serves in the Mwanza area. At present they are al functional and residenfial buildings. The pro- vulnerable to encroachment (although whether ject's perfrmance has fallen short of this is tainlg place is not known). expectations. The current proposal is to con- Another component of the project deals duct a study that identifies past errors, pro- with the land not suited for agriculture. Some poses corrective measures, and designs an such land has in fact been used in this way exnsion of the project area (Dakawa HI) to The project will suggest nonagricultural aker- another 2,000 hectres. natives to the farrmers. Other land that is still Environmental issues forest, but that will not be protected as re- serves, will belong to the taditional autorites When Dakawa I began, her was no re- who will be advised on how best to allocate it quirment to produce an environmental impact most efficiently, assessment. The curret sudy's main aim in The report states that the project will have the environmental area is to assess the actal other positive environmental effects. The im- state of resources. The proposed ex post eco- proved agricultural practices and the use of nomic valuation of Dakawa I, however, does mineral fertilizers will avoid soil depletion. not include an ex post environmental 5. Environmenrala Economics and NVatural Resource Management 133 valuarion, which would bc useful to gain cx- lcads to bctter usc of water. In the second puricic for Dakawa 11. place, rising productivity reduces prcssure on It is laudable that project preparation for marginal land. so helping to avoid an unsus- Dakawa 11 will include an cnvironmental im- tainablc developmcnt path. No calculation or pact analysis of "hcalth hazards. pollution, valuation of the probable benefits in those two ecology, etc.' Such informnation can hclp in de- issucs is given. signing a better project. Howvevcr. it would bc The qucstion here is whether it is possible useful to provide the team carrying out the to evaluate such a project in any more quantita- EIA, with an indication of what altcrnative tivc a manner. Clearly, one can set ccrtain tar- project designs are considered In other words, gets for the project to achieve, and ask whether the environmcntal analvsis can both serve as a the proposed measures will achieve those tar- guide to the mitigation mcasures and assist in gets. This provides a discipline that can act the design of the projcct itself. The project it- both to help idcntify the linkages between the self is rather vaguc as to what policies vill be management changes and the targets, and to implemented to achieve the improved mdnage- serve as a goal for the agencies involved. Tar- ment of the natural resource base. gets have not becn identified in this project, There are several suggcstions about w%hich could make it less efiftive as well as "compensating projects" for reforestation, en- more difficult to evaluate cx post. ergy from agricultural waste, and water and soil conservation. These need to be appraised 6. GHANA: Industrial Plantation individually, and with more monctary quanti- Project, SUBRI fication than is indicated from the project docu- . o ment Such quantification is possible, given Description of fieproject resources. The appraisal report for this project dates If these environmental costs and benefits from 1984, when the expected starting date are not integrated into the analvsis, a situation was 1985. The project consisted of the conver- could be created where mitigation subprojects sion of 4,000 hectares of natural fbrest inside are carried on even if they are not cost effec- the Subri River Forest Reserve, into a Gmelina tive, or, worse, wh%Iere the whole projcct is car- arborea (fast-groving tree) plantation. A ried on just because it "manages the resources plantation of 1.000 hectares was already exist- well,' even if residual environmental damage ing at the time of the project. makes it unjustifiable. While the plantation was young, an agro- 5. BUIRKNA FASO: Institutional silvicultural project would be developed, sup- Support for Irrigation Management plying food for labor and local sales The tim- ber would be directed to a pulp and paper mill Description of the project in Daboase, seven kilometers away, a project sarted in parallel. The plantation project There is a great pressure over natural re- would be managed by a parastatal fbrmed for sources in Burkina Faso. The irrigated sector that purpose. The costs were estimated at comprises 13,000 hectares and it faces a con- 40.45 million FUA, and the benefits would give tinuous mismanagement and poor maintenance. an economic internal rate of return of 26 The project is aimed to give institutional sup- percent. port to the government agencies dealing with E irrigation. The costs are estimated at 2.76 mil- Environmental Issues lion FUA. The project is due to begin in 1991 The appraisal report states that because and the appraisal report was completed in Au- only 69,000 of the 1.7 million hectares covered gust 1990. by forests in Ghana had been defores,ed by that Enviromnental issues time, and the plantation would convei. only 83 of the Subri River Forest Reserve, the efftcts The report states two positive efi:bts of the of the project would be "not alarning from the project. In the first place, good management environmental point of view." However, no I34 environmental impact assessment or valuation of a regional development project encompass- of the forgone forest benefits %ras made. Thus ing many related agricultural and environmen- the rate of return is certain to be an overesti- tal impacts is a good one. The project clearly mate. The loss of a small land area, such as identifies two major issues: one is that the re- that described here, however, can be associated source degradation is coming from the environ- with major losses in ecological and environ- mental stress caused by droughts, and the other mental terms; it depcnds on the sensitivity of is that there is a problem of the managernent of the ecosystem to loss of forest This is not to the coummons. say that the project w as necessarily a bad deci- It is not clear, however, how successful the sion, but simply that a more careful evaluation proposed management of the rangelands com- of the selected project would in general be a de- mons, which is critical for the sustainable use sirable feature of an environmentally sensitive of rangelands, will be. The effectiveness of flrestry policy. Incidentally, a post evaluation governments in "organizing" communities to of this project would be interesting, to see manage such commons has never been con- what, if any, real impacts have been felt from vincing and one might doubt its success in this the forest loss (including any secondary efects case. from the opening up of an area that lies within Although the project is rightly concerned a forest reserve). with the management of the rangelands, it does 7. MOROCCO: Pastoral and Livestock not devote the same attention to the manage- 7.veMoROCCO PnthEastorland givtoc ment of open access forests. The project takes Development in the Eastern Region the view that the mere process of increasing Des,piption of the Project^forest resources through reafiorestation should Descri-tion of the Project reduce the depletion from existing forests. With The main activity in the Eastem region of open access, however, that may not be the Morocco is sheep farming. It is a region se- case. verely affected by drought and overgrazing, but Tlhs project is one where the benefits of it has still some forests. The land is mainly in the environmental components have been quan- communal ownership, with some private and tified using dose-response relationships. The state land existing alongside. The state land is benefits are measured as the revenues from the mainly in forest areas increased production of marketed goods and The project seeks to restore the physical the shadow price of self-conumption food. milieu of the region. The existing rangelands Thus the project offers an example of how such would be improved by soil preparation, sow- techniques can be used to value environmental ing, dune fixation, and the building of small benefits. dams. Agricultural plots would receive mcen- 8. ETHIOPIA. South-East Rangelands tives to produce fodder crops. Cooperatives Prjec would be organized and receive the rights to manage the pastoral areas; tr-ining and veteri- Descrplion of the Project nary services would be provided to these cooperatives. This project is similar to the Moroccan one The project's total costs are 41.28 miillion considered above, with the differrnce that there FUA. The benefits are calculated using a dose- is less of an issue of forest management here. response model and valuing the yield increases The southeast of Ethiopia is one of the poorest at average current prices. The estimates show and driest regions in the country. The droughts that in eight years the internal econonic rate of have affected it heavily and, combined with the return would be 15 percent. The project report overgrazing, the natural resource base is being was submitted in April 1990. degraded at a fast pace. The project includes a series of actions to restore and upgrade the Envinronmental Issues rangeiands capacity to sustain livestock. Most of the expected effects of this project One of the components is the construcion on the enviromnent are positive, and the notion of infiastructure: the creation of sto;k routes, S. Enriruvnme rutml Eenonaics und Natural Resource Managemen! 135 access roads. grain stores, and new water import of cattlc from other countries to sell ponds. The project aims also to develop Pas- them to the family farm sector. Also, research toral Associations. They would be responsible on resource use, marketing, and other areas for the management of communal lands and the will be conducted. would receive veterinary and agricultural 'M- The report was submitted in May 1990 and tension, training, and assistance. A commercial the project's starting date was set for January effort would be to set up a marketing 1991, with a proposed duration of five years. intelligence service and new infrastructure. The total costs are 16.64 million FUA and, al- The project's total costs are 31.98 million though there are some data problems, the best FUA. The project report was submitted in July estimate of the economic internal rate of return 1989 and the project was expected to begin in is around 20 percent. The latter is based on January 1990. The benefits are calculated us- much less data regarding the dose-response ing a dose-response model. The prices of the relationships than are the Moroccan and Ethio- expected increase in skin, meat, and milk pro- pian estimates. duction of the different species are based on the 1990 averages. In a period of twenty years, the Enviromental Issues project reaches an economic rate of return of The war has brought down the number of 18 percent The sensitivity analysis, simulating livestock, so the report states that the cunrent draughts every seven years and rising costs, situation is one of underutilization- of pastures yields a minimum rate of return of 14 percent, and water resources to well below carrying ca- so the project is considered a solid one. pacity. With the most optimistic results the Environmental Issues program would enlarge the current stock just to prewar levels. Because at tat time there was Most of the expcted effects on the envi- no evidence of overgrazing, deforestation to in- ronment are positive and the quantification of crease grazing lands, a lowering of the water the environmental benefits provides a useful table, or other environmental problems, this guide to the desirability of the project Similar time the govenmment is not expecting any. issues arise here, however, as with the Moroc- Changes in management regimes have occurred can projectr The key to the success of the proj- in the intervening years, however, and the as- ect is the effective management of the sumptions of no eflfcts may be overoptimistic. rangelands. Even when the new roads and wa- Part of the research wili develop technolo- ter availability expand their carrying capacity, gies to ensure soil conservation and make opti- the element that confers snstinability is the mal use of resources. Those results will be management of common property that the Pas- incorporated into the extension services. No es- toral Associations are expected to provide. timate of the project's benefits (if adopted by the farmers) is incorporated into its valuation, 9. MOZAMBIQUE: Rehabilitation of although there is frequent mention of the notion Family Farming Livestock of sustanbility. Indeed, it is a commonly re- Description of the Project peated idea in many of the projects that certain conservaton measures do not need to be val- Expecting that the civil war can be settled ued, because they will conf£r sustinability to and peace can be achieved in the near future, the rest of the project. the government of Mozambique wanted to set 10. MOROCCO: El tachef Dam the basis for quick econormic recovery. This program is intended to support the family DescriptCon of the Project farming livestock sector in that recovery. It will concentrate in the ten safest rural districts and This dam will store the waters from the hopes to reach 150,000 families. seasonal river El Hachef to satisfy the growing The project will build watcr points, provide requirmts of Tangier, Asilah, and surround- veterinary services, and carry out extension ing centers up to the year 2015 An area of programs. One of the main elements is the 2,000 hectares will be deforested and flooded 136 for that purpose. The land will bc subject to would be stopped: the loss of positive bcnefits compulsory purchase and the inhabitants re- is not mcntioned. settled upstream. The projcct includes a devel- The appraisal report does not mention if opment program for the catchmcnt area the whole 2.000 hctares to be flooded are pri- (including the new settlements) that consists of vate farm land. If some of there were public technical support for agriculture in flat private forest land, the loss of value wvould be ignored, land, development of fruit plantations in steep as no value is attached to such land. private land, and reafforestation of all public The project mainly considers the environ- lands mental problems that could affect the perform- The economic analysis of the project is ance of the darn. not the effects that the dam more sophisticated than that of many other can cause on the surrounding region. Thus. ADB projects: The estimated cost of the proj- special works on the catchment area are con- ect is 81.2 million African Development sidered as an environmental -investment' for Bankcs Units of Account (BUA) The cost in- the long-term viability of the dam. These in- cludes the price paid in the compulsory pur- clude -sound agricultural practices,j the chase of land, the building and operation costs plantation of fruit trees, and the fixing of river of the damn and a measure of the forgone agri- beds. It Nvould be desirable to decide on the er- cultural products in the flooded area. The bene- tent of such environmentally positive measures fits are calculated as the revenues from water by looking at all their benefits. so that a larger sales at 1991 nominal prices and the value of budget may be allocated to such item- the fruit produced by the development project Finally. the project has a con.,&nent for m the catchrnent area. The balance gives a rate the reafforestation of an area equal in size to of return of 15 percent. The best altemative, a the area that is being inundated. This is in- dam in the Maharauf area would have a rate tended as a compensation in environmental of return of less than 1 percent. terms-a putting into practice of the notion of Environmental Iss.es sustainrability. It is not clear. however, how good a compensation this is. It may be perfect- There appears to be a double counting of 1v adequate. but more information is needed some of the costs of the project and no count- about the ecolaog and economic values. both of ing of other costs. Double counting occurs the lost forest and of the reforested areas. when the compulsory purchase price of land is 11. TANZANIA: Road Rehabilitation added to the forgone benefits of that land: the Pjt former is in part at least based on the latter. If Project Fe compulsory purchase price were equal to Descriorz of the Project the market value, which in tum fully reflected the commercial value of the land. then the proj- The project seeks to rehabilitate and up- ect is exactly double counting the costs. How- grade 389 kilometers of roads linking Chalize, ever, there is no certaintv that the price paid for Segera, Tanga- and Arusha. in the rich land in the compulsory purchase is the price Chalize-Arusha comdor, and those linking the market would give. Therefore the extent of Mutukula, Bukoba, and Lusahunga in the Lak-e the error in taking both figures is unknown and Victoria Circuit After fifteen vears of neglect. it would be better to take only the estimated they are in a very poor condition and are in- frgone benefits. On the other side, the project creasingly becoming a bottleneck to develop- errs in not including many of the enviromnental ment efforts in the region. costs The existnce or nonexistence of original The report -as submitted in April 1990, ecosystems or wildlife in the area to be flooded and the starting date of the project was April is not reported. Also, deviation of water would 1991. The cost of the project is 45.01 million surely alter original ecosystems and economic FUA The benefits are calculated using a mod- activities downstream. The only mention of this el which simulates the economic activity in the issue is that the damage of seasonal flooding region over the next fiftv years. The main bene- fits come from savings in time, vehicle S. Enrironmental Economnics amd Natural Revource Managerent 137 opcrating costs. and reduced road accidents. protection of water eatehment areas, strength- Also, some effects on the economic activities ening of governmcnt forestry services, and ceco- carried in the influence zone of the roads arc logical studies. talken into account-for cxample. "high quality The appraisal report dates from March fruits and vcgetables can be grown on the 1990, and its expected starting date was June slopes of the Kilimanjaro and Usambara" 1990. Different government agencies would be (page 56). A high economic rate of return is involved. The costs are 26.93 million FUA, expected. It varies between 24.5 percent and and the project's economic rate of return is cal- 28.5 percent. depending on the sensibility culated at 18 percent. analysis. The benefitr come mainly from marketed Environmental Issutes forest products: timber, fuelood, and fodder. By law, the recipients of the benefits are the In general. the project rcport looks carefiul- communes. This creates problems of cost re Iv at the direct environmental consequences of coverv because the govermment does the in- the road. Issues such as effbcts of the road vestment, managing. and works. There are works themselves potential land erosion, and problems of continuing the project once the pe- damage at the material extraction sites are all riod of the loan runs out, because even when properly addressed. Land erosion would be the law states that 20 percent of the forcst curbed by improvcd drainage of the roads- and carnings must be reinvested the communes ac- material extraction sites would be restored. tually reinvest around 10 percent. However, it is the indirect impacts of such a Environmental Issues rehabilitation scheme that reed to be examined. Although the appraisal report states that a con- As with the other project in Morocco re- sultant would be ask-ed to conduct an EIA. viewmued in this report (see project number 7 which should lookl at some of these issues the above), there is an implicit assumption here final decision does not seem to depend on its hat the very process of reafforestation will re- outcome. The Report argues that no significant duce manis pressure on forests. But if incen- environmental degradation would occur in the tives to deforest are established, and these surrounding areas because the project just re- incentives remain in force, reafforestation habilitates existing roads. it does not build new would only delay depletion. In fact, the report ones. Such an argument ignores the significant recognizes that poor coordination in the man- negative enviromnental changes that will result agement of common land hinders the efforts of from the development of the region, as the getting back to a sustainable path. Such poor overall costs of access to the areas affected are coordination may well result in the project not reduced. The transformation of surrounding achieving its aims uncultivated areas, with their attendant ecologi- Nevertheless, it is encouraging to see that a cal functions, into agricultural land, could have project such as this, with many diverse envi- some important consequences. These need to ronmental benefits, can be justified from the be looked at. This does not mean that the marketable benefits alone, assuming that 'the scheme is necessarily undesirable: it is merely institutional issues can be resolved. The project part of a prudent strategy in any development manages to achieve an internal rate of return of program. IS percent-without consideiing in quantita- tive terms such benefits as the protection of wildlife, the reduced siltation of dams. and the Conservation prevention of desertification and soil erosion. As the document righdy recognizes, these un- measured Impacts are the "most important This is a general project, acting widespread part' of the project Including such benefits in ihz country and in sevcral different ways. could justify a higher level of spending in the It's main features are: reafforestation. forest sector. The constraint may well be the inability management, silvopastoral managemcnt, of the govemrnment to recover the costs of the '3, initial investments; doing so would requirc a understanding of the linkages. In this section, program of fiscal change that would be part of three sectoral projects with somc potential en- a drive to a more sustinable use of resources. vironmental implications are reviewed: the So- malia Agricultural Sector Adjustmnent Conclusions on the Review of Progmram the Ghana Health Service Rehabilita- Projects tion Program, and the Malawi Study on the So- ADB projects in recent years have taken cial Aspects of Adjustment the environment seriously, as can be seen from 1. SOMALIA: Financing Agricultural these reviewvs However, there is considerable Sector Adjustment Programme room to extend and integrate the environmental analysis into the economic analysis. The mone- Description of the Project taiy estimation of environrental benefits and costs are confined to a fev cases where a dose- The Agricultural Sector Adjustment Pro- response model has been used. These have grainme (ASAP) is a medium-term adjustment yielded useful results of benefits in the areas of program. The reviewed loan proposal is for its conservation and should be extended. In addi- second phase. The report dates from November tion, there is scope to apply other techniques 1989 and it was expeced to be effctive from referred to in here in ADB project appraisal; January to December 1990 particularly, as mentioned at the beginning of The project seeks to liberalize controls and this report, the treatment of positive benefits institutional bottlenecks troughout the sector. needs to be strengthened. Privatizing or closing parastatals, eliminating Where mitigation programs are recom- state monopolies on commodities trade, and mended, a need exists, in some cases, to look at liberating inputs and outputs from price control the wider impacts of the project, so that its in- are some of its measures. The effects of these direct consequences can be allowed for Tnis is measures, combmed with good climate, have particularly the case where the ADB's involve- resulted in a fast recovery. The loans will be mewa is going to change significantly the land used to import inputs to develop agticulture, use patterns in a region. livestock, fisheries, and forestry. Finally, the success of many of the projects is dependent on instutional changes and alter- ations in land management systems, where ex- One of the stated objectives of the Agricul- pmience would indicate that there is a real tural Development Policies is to preserve the possibility of failure. There is not much that environnent at a susainable level Somalia's the analyst can do about this, except to scale fragile ecology makes this goal difficult to down the expectations of the changes in yields, achieve. Because this is a sectoral program, and so forth, that the project will result in, and few details are given on the specific projects to to assist in setting the policy context for its ac- be carried out. The only infornation in the re- tivities. In the next section the relation between port on environmental matters concerns the use these policies and environmental impacts are of fertilizers, pesticides, and herbicides for discussed. which Somalia has one of the lowest use rates in Africa To encourage their adoption by Environmental Issues in Sectoral farmers, these inputs are currently provided Bank Aciities with credit and price subsidy. Five dangerous As indicated in the bginning of this paper, pestcides are banned from imports. On the (a) the African Development Bank rigtly rec- livestock sector, the current rising stock is said ognizes the importance of the broader policy to be causing overgrazig and rangelands context in formulating its lending program, but degradation. (b) there is a geat need to gather the relevant The main issues here are the linkages be- information on the linlcages between gover- tween changes in agricultural prices and land ment policy parameters and environmental deg- use. How will the revised prices affect the sup- radaflon, and to establish a better ply of fruits, the growth of wheat and sorghum, 5. Environmental Economics wnd Natural Resource Management 139 and the holding of livestock? That there are movement but the combined impact of several strong relationships through price elasticities programs could be significant. It should be part has been established (Gammnage 1990). How- of the ADB's responsibility to look at the over- ever, the linkage backwards from the changed all environmental consequences of its programs supply to changes in land use is less clear. The in a particular country. Such a strategy would program should have access to a set of studies require a separate exercise into which each of that would allow it to predict the environmental the programs such as this one would feed. consequences of the macroeconomic policy changes it is proposing. At present this cannot ALAWI: Study on the Social be done and it would be too much for one pro- Aspects of Adjustment gram to undertake the research. Related to this is the need to introduce DescriptionoftheProjec additional measures specifically to reduce land The project seeks to study the effects on degradation and misuse For example, a pro- the lower-income groups in Malawi of the ma- gram which raises the prices paid to frmers croeconomic adjustment program. No environ- for annual export crops could increase the mental issues are part of its objectives. pressure to bring land into cultivation for this Although some of its consequences may have purpose, land that is not suited for this pur- indirect efftcts (for example, through poverty pose But the fact that such pressure exists reduction on deforestation), they are not con- does not mean that the original measure is in- sidered relevant to this study. The estimated valid; rather it points to the need to introduce costs are 860 thousand FUN The report was proper inentives for sustainable land use and submitted in August 1990 and the starting date management. of the study is fiscal year 1991 (beginning 2. GEHANA: Health Service April 1, 1991), with a duration of two years. Rehabilitation Environmental Issues Description of the Project A study such as this would offer an ideal occasion to integrate the linkages between pov- The project seeks to rehabilitate the health erty/income distribution and enviromnental service in the whole country, strengthening hos- degradation. Orne of the hottest debates in the pital maintenance and management. Support is environmental field is the extent to which envi- also to be given for primary health care. The rournental damage is increasing and sustainable report was submitted on December 1989 The development is being arrested, because of in- total cost amounted to 18335 million FUA, and creasing poverty in some countries, where the no estimate of benefits or cost-effectiveness poor have no option but to degade their envi- analysis was made. ronnent This is disputed (Jagannahan 1989) Environmental Issues but the debate remains an important one. The World Bank has been conducting a study of the The report states that the project has no macroeconomic-environment linkages, and pro- negative environmental impact. As the re- grams such as this one from the ADB would habilitation will make disposal of waste more benefit in design from the results of that study. hygienic, some (unmeasured) positive effects are expected. The wider implications of a bet- Conclusions and ter set of health facilities are not addressed, Recommendations however. These include the attractiveness of This paper has examined the issues arising, urban centers where such facilities are usually for the Afi-can Development Bank, in the in- located relative to the rural areas, a point that tegration of the econormic and environmental has been made in the ADB's Environmental aspects of sustainable development The ADB Paper. is dealing with these problems on a continent It is unlikely that any one program such as with the gravest environmental crises. Hence this will have a major impact on population 240 the concems could not be more important. In On the secboral/policy lending side, stress rccognition of this, thc ADB has begun to for- is laid on the importancc of the links between mulatc an cnvironmental policy that deals vith macroeconomic policy and the environment. both project and sectoral policy aspects of its Such links are not fully understood, let alone activities. The essential conclusions on the quantified. but some of the research has shown project side are: (a) there is a need to extend that often it is the underpricing of key conrno- environmental cost and benefit estimation and dities that is resulting in their excessive use. (b) where this is not possible. there is a need This applies to water, and to timber and fuel- to formalize the use of sustainability criteria wood (with lowv stumpage charges). Such poli- to establish guidelines for mitigation ac/ion cics result in continuing pressure on certain andfor resource conservation projects. Howv- resources that will remain even if new projects ever, these are issues that all development are initiated to increase forest area, or to sup- banks and bilateral donor agencies are grap- ply more water for irrigation, and so on. Re- pling with and the African Development Bank ducing this pressure, however, runs into several is not alone in facing them. Hence, there may political and social conflicts, created by vested be some sense in a joint effort in addressing interests. Also, information is needed about the manv ofthe issues raised here. linkages at a quantitative level that is, by and The reviewv of some of the ADB's recent large, missing for most countries. Thus, it must projects reveals that it is takling the environ- be a prnoritv to assist countries in the collec- ment seriously. However. there is considerable tion of the relevant information in which to room to extend and integrate the environmental base sound policies for stustainable develop- analysis into the economic analysis. The mone- ment. tary estimation of environmental benefits and From the review of programs, it appears costs are confined to a few cases athere a dose- that there is a strong need to collect more in- response model has been used. These have formation on the macroeconomic policy- yielded useful results of benefits in the areas of environment linkages, and to develop aframe- conservation and should be extended. In addi- workfor disseminating this to program teams tion, there is scope in ADBs project appraisal so that they can use it Again. there is scope to apply the other techniques referred to earlier. for inter-Bank cooperation in this regard. As mentioned at the beginning of this repor, the treatment of positive benefits needs to be Bibliography strengthened. This will require the use of such techniques. African Development Bank. 1990a. Environmental There is not much to be gained fLom a ret- Poliuv Paper of the ADB Group, rospective analysis of the projects reviewed ADB/BDIWP/891108. Abidjan. here, except in one or two cases where a African Development Bank. 1990a. Environmental change in the operation of the project can be Assessment Guidelines for the African made when account is taken of the environmen- Developmnent Bank. Abidjan: Haskoning tal impacts, or where a compensating project Royal Dutch Consulting Engineers. mi.ht'be jutified-Onthcse grouds, the Mo Bishop. J.. A. Markandy-a and J Richardsoni might be justified. On these grounds, the Mo- 1991. Afacroeconomic Adjustment and rocco Dam project and the Tanzanian Road the Environment. London: London Envi- Project could be worth re-investigating. Other- ronmental Economics Center and Univer- wise. effort needs to be devoted to projects sitv College. coming up in the near future- Boyo. J. K-G Maler. and L. Unemo. 1990. Envi- W4here mitigation programs are recom- ronment and Devetopmnent An Economic mended, there is a particular need to look at Approach. Kurwer Academic Publishers, the wider impacts ofthe project. so that its in- AM Dordrechtv dirct onsquecescan be allowed for This Gammage. S. 1990. Enviromznental E:conomircs in direct consequences ca ealwdir hsthe Developing Wforld, Report to U.S- is particularly the case where land use patterns Age D orel ntrnationl Delopment in a region will change significantly. (UAS AIDo) Washington, D vC. S. Environmental Econgomics and Natural Resource Manaement 141 Jagannathan, N.V. 1989. "Poverty, Public Policies Pearce, D.W., E.B. Barbier, and A. Markandya. and the Environmcnt$' World Bank Envi- 1990. Sustainable Development: Eco- ronment Department Working Paper No. nomics and Environment in the Third 24. Washington. D.C. World. London: Earthscan Publications Markandya, A. 1991. The Economic Appraisal of Ltd. Projects The Environmental Dimension. Pearce, D.W., and A. Markandya 1989. Environ- Washington, D.C.: Inter-American De- mental Benefir Estimation: Monetary velopmcnt Bank. Evaluation. Paris: Organisation for Eco- Markandya, A., and D.W. Pearce. 1991. -)evelop- nomic Co-operation ard DevelopmcnL mcnt, Environment and the Social Rate of Repetto, R. 1988 Economic Policy Reform for Discount,- WYorld Bank Research Observ- Natural Resource Conservation.' World er, Volume 6. No. 2, pp. 137-152. Bank EnvironmenL Department Working Markandya, A, and 1. Richardson. 1990. -The Paper No. 4. Washington, D.C. Debt Crises. Structural Adjustment and Sebastian, L, and A. Alicbusan. 1989. -Sustain- the Environment," mimeo. Department of able Development Issues in Adjustment Economics, University College, London. Ler. ting Policies," World Bank Environ- Panayotou, T. 1990. The Economics of Environ- ,uent Department, Policy and Research mental Degradation: Problems, Causes Divisional Woring Paper No. 6. Wash- and Responses. Harvard Institute for In- ington, D.C temational Development Discussion Pa- per No. 355. Cambridge, MA: Harvard University. 6 The Rural Development and Environmental Protection Project in the Day Forest in Djibouti: A Case Study Nessim J. Ahmad The case study covers selected issues in the economic analysis applied in the IFAD appraised Environmental Protection and Rural Development Project in the Day Forest in Djibouti Using the project as a substrate for discussion, generic opportunites and con- .sraints in the economic valuation and evaluation of environmental changes in marginal areas in developing counties are identified It is argued that opportimitlies existfor the rou- tine application in project appraisal of economic valuafion melhods such as the surrogate market approach, the change in production approach, and the opportuniy cost of time ap- proach, while contingent valuation methods (CGM and the travel cost approach can be con- sidered less suitable. In all cases, judicious use of sensitivity analysis is advocated In terms of valuation of natural resources on which poor groups depend, it is stressed that the treat- ment of intra-generationaf equity is as important as the inter-generational issue. The case study concludes that the routine application of environmental economics in project appraisal is constrained not so much by methodological gaps, but by difficulties in ensuring that physical estimates of environmnental change in "with" and "without" project scenarios are generated in a timely fashionm As such, a case is made for undertafing pro- active environmental assessments well in advance of appraisaL 144 The Intemational Fund for Agricultural likely to suffer irreversible degemeration. On Development (IFAD) has a specific mandate to the basis of trends detennined through the alleviate rural poverty. Its target groups consist analysis of available studies and the interpreta- of small farmers, landless laborers, artisanal tion of aerial photograph time-series, the Juni- fishermen, nomadic pastoralists, indigenous per stand is projected to disappear within ten peoples, and the women who make up these years. Linked to tOe disappearance of the for- groups. As an outcorne of the Fund's efforts to est, rangeland degradation is likely to con- sharpen its focus on the poorest groups, an in- tinue-threatening the livelihood base of creasing proportion of IFAD projects are lo- pastoralists who subsist in the vicinity of the cated in marginal, resource-poor environments. forest. In these areas, it is often the case that poverty In response to a request by the government processes and environmental degradation are of Djibouti, the Rural Development and Envi- closely intertwined. In recognition of this, ronmental Protection Project in the Day Forest WFAD projects are increasingly addressing the was identified for possible inclusion in IFAD's poverty-enviromnent linkages directly. An ex- portfolio by an WFAD General Identification ample of this trend in IFAD's portfolio is the Mission in March/April 1987 (Report No. IFAD-appraised Rural Development and Envi- 0231-D1) In April 1988, an IFAD ronmental Protection Project in the Day Forest "pre-preparation' mission was fielded to syn- in Djibouti thesize additional infonnation for project de- This case study is based on the economic sign generated by various pre-investnent analysis applied in the appraisal of the Day conservation activities undertaken by the Food Forest project. The case study does not address and Agricultr Orgaization of the United Na- the myriad issues associated with the sustain- tions (FAO) and United Nations Development able nanagenent of common property and Program (UNDP) in the project area. This was open access resources in dryland areas that are followed by a preparation mission fielded by central to the design of the project. Instead, the FAO Investment Centre on behalf of IFAD certain valuation aspects of the economic in November/December 1990 (Report No. analysis of the project are discussed, with a 3/91/IF-DJI5). Finally, an IFAD/United Na- view to drawing some preliminary conclusions tions Capital Development Fund (UNCDF) concernng the -econonuc valuation and evalu- mission visited Djibouti in April/May 1991 to ation of environmental changes in the context appraise the project for possible financing by of marginal areas." In line with the objectives IFAD. of the CIDIE Workshop on Enviromneal The objective of the proposed Rural Devel- Economics and Natural Resource Management opment and Enviromuental Protection Project in Developing Countries for which it was pre- in the Day Forest is to alleviate the poverty of pared, the study focuses specifically on the pastoralists in the Day region of Djibouii relevance of environmental valuation tech- through the sustainable nanagement of the Day niques for investment decision-making. Forest and surrounding sylvopastoral areas on which the pastoralists depend. The project as The Project appraised by IFAD would have a total cost of USS4.37 million over a project life of seven Project Background years. The financing plan as originally pro- The Day Forest is a unique 1,000-hectare posed is as follows: World Food Programme remnant of an ancient Juniper forest that once (WFP) would provide a grant of an amount covered the higands in the Hom of Africa. As equivalent to US$480,000 (11 percent of total such, it constitutes, m a predominantly hyper- project cost); UNCDF would provide a grant arid country, a recognized national patrimony of an amount equivalent to US$1,587,000 of international biodiversity significance. How- (36.3 percet of total project cost); [FAD ever, in the absence of timely intervention to in- would provide a higHly concessional loan (1 troduce more sustainable sylvopastoral percent interest, ten-year grace period, fifty resource management practices, the forest is years repayment) of an amount equivalent to 6. The Rural Development wad Environmental Protection Project in the Day Forest in Djiborni 145 US$1,388,600 (31.8 percent of total project Zone 3, 75 percent of rainfall is concentrated in cost); and the government of Djibouti would the hot season. contribute an arnount equivalent to The brown tropical soils and fluvisols US$915,100 (20.9 percent of total project found in the area offer some potential for pas- cost). The estimated economic internal rate of toral productivity if properly managed, but are return at appraisal is 15.23 percent. extremely susceptible to erosion in the absence of adequate vegetative cover. Areal Due to high temperatures, low and uneven 'The project area, covering an estimated rainfall, sparse vcgetative cover, and the fis- 25,825 hectares, lies to the northwest of the sured nature of the bedrock, water is an ex- GFulf of Tadjourah and e.xtends northwest from tremely scarce esource. the Day Forest towards Lake Assol. The proj- ect area has been defined to include the Day Land Use Forest and all other sylvopastoral lands under With certain exceptions, rural land in Dji- the direct communal control of the Sekoh bouti is de jure property of the state. Tis land Aboussa tribe. In addition, the project area in- is de facto under the collective control of the cludes an extensive area of untapped potential tibes and clans who occupy the land. These de for grazing, known as the Mak'arrassou, which facto land rights are granted by the traditional is under the imformal control of the Sekoh authorities and are restricted to usufiuct Aboussa but which is currently underutilized Pastoralism is the predominant activity in due to an inadequate distribution of water the project zone. The principal sylvopastoral points. The project boundaries also include cer- iry utilized by the Sekoh Aboussa tribe tain encampments of the Songo Goda and Bar- consists of the Day Forest and its fringes, an gak Adbara tribes t1hat border the grazing lands irtrmediate zone (Zone 2) and the of the Sekoh Aboussa. Mak'anassou (Zone 3). The Day Forest is un- The project area can be divided into three der the control of three lineages of the Sekoh major geomorphological zones: Aboussa, while the intenreAi-te zone is divided * Zone 1. The zone covers an area of about into several parcels each under the control of 1,790 hectares. Peaking at an altitude of one of the five lineages of the Sekoh Aboussa. 1,784 meters, this zone includes the Day With the exception of a small portion of the Forest. Mak'arrassou under the direct control of one of * Zone 2. A severely desertifled zone of the Sekoh Aboussa lineages, the Mak'arrassou about 4,035 hectres covering plateaus and is an open-access territory shared with other hills on the northwestern fringe of the Day tribes-notably the Songo Goda. However, the Forest. Altitudes range from 900 to 1,400 Sekoh AbDussa lineages exercise a pre- meters. eminence over a portion of the open-access * Zone 3. Covering 20,000 hectares, this Mak'anassou that is included within the proj- ect boundaries and would exercise a right over zone consists of an extremely fro d topo- at least 30 percent of available pasture. This gwahical unit quite different feom the Oeter area is shared with those Songo Goda who re- two zones. Alttude vares between 500 to side within the project boundaries, as well as 1,200 meters. with approximately 1,300 Songo Goda based The climate is tropical semirniad to arid otiehpoetra characterized by irregular inter- and inr- outside the project area. amualminzedfbya ll highte era-um, and gha- The territory directly available to the Se- amnual rainalll, high temperatures, and high koh Aboussa is not sufficient to support their potential evapotranspiration. ZCone 1, lying Iicestck throughout the year, especially be- within the 300 millimeters isohyet, receives the cause throughout thn ycar, of be- highst ainfll n th contry Zoe 2 iesbe- cause the current configuration of livestock highest rainfall in the country. Zone 2 lies be- water pomn does not allow efficient use of ex- tween the 300 millimeters and 250 millimeters isting resources. Access agreements with other isohyets, while Zone 3 lies between the 250 tribes allow the Sekoh Aboussa to transhume millimeters and 200 millimeters isoh'ets. In to other zones, essentially in the vicinity of 146 Lake Allol and Lake Assal as well as in Ethio- 1,906 metric tons of dry matter per year would pia. Howevcr, transhumancc is becoming less be available for thc. livestockl of the Songo ftequent as the population shifts towards a se- Goda. In the context of inevitable sedcntariza- dentarized way of life. As an outcomc of this, tion, this resource takes on a critical the Sekoh Aboussa incrcasingly provide their importance. herds with supplemcntary feed to cover fodder deficits. Target Group: Income and Economic In recognition of the central role that the Activities Day Forest plays in their lives, the three Sekoh The projcct target group consists of ap- Aboussa lineages that control the forest subject proximately 3,700 seminomadic pastoralists its use to strict management. Only dead trees who rank among the most disadvantaged in are allowed to be cut. All livestock are ex- Djibouti. Per capita income has been estimated cluded during certain periods. In periods when to lie in the region of US$160 or 34 percent of access is allowed, only cattle and some camels gross national product (GNP) per capita are permnitted to graze within the forest. Under (US$475). Ninety-three percent of the target traditional rules, any animals found in con- population bave a per capita income below travention of these principles are slaughtered. US$200, while the remainder fall below In the intermediate zone, where each lineage US$250- controls a specific grazing area, all animals The population is dependent on several are allowed to graze, although cattle are given sources of livelihood. A significant source of priority over other animals. income is the caravan trade with Ethiopia (40 In years of average rainfall, the sylvopasto- percent), although the raison d'etrc for the ml management regime is strict, with each line- trade appears to lie at least partly in the provi- age exercising full control over its pastoral and sion of supplementary fixd for livestock. Live- sylvopastoral resources. In years of drought, a stock provides a third of revenues, while controlled relaxation of grazing rules is ef- salaries and pensions contribute less than a fected to ensure that livestock belonging to the quarter. The remainder is made up of remit- five lineages of the Sekoh Aboussa are safe- tances from relatives living in Djibouti City. guarded. Under special circumstances and sub- 7 ject to the explicit consent of the Sekoh IncomeJfrm Livestock Aboussa, other tribes are allowed access to the Livestock sales are infiequent and take Sekoh Aboussa pasture lands. At the same place primarily to meet contingencies. Because time, the dearth of water points within the area there is currently no livestock market in the vi- controlled by the Sekoh Aboussa renders the cinity of the Day, the animals are taken to the tribe dependent on sources of water outside the district capital or as far as Djibouti City, with area of their control. resultant loss of weight and quality. As a re- The pastoral management strategy in dry sult, prices received are relatively low. Live- years consists of forced transhumance outide stock are slaughtered mainly for funerals and the project area, supplementation of lactating religious festivals. Livestock income derives females and calves, and intensive tree loppmg from sales and auto-consumption of milk and in the fbrest and its fringes. While the fodder meat less supplementary ted used for lactating balance is precariously maintained though animals. these coping strategies, a significant pastr re- I source (in the pastoral facies Acacia melliJfera, Incomefrom Caravns Rhigozum somalense, Acacia ehrenbergiana Caravans provide an important source of renains underexploited in the Mak'arrassou income for the population. Involving more than due to the lack of water points. The rational ex- half the families in the zone, these caravans ploitation of this resource could provide the Se- tansport salt collected from lake Assal to koh Aboussa with more than 1,700 metric tons Ethiopia and return with cereals and other of dry matter units (DM) per year instead of products. Although significant variations exist the 388 tons currently providedc An additional between lineages, caravans are mounted an 7he Rwi Development and Environmental Protection Project in thle Dnv Forest in Di;hout 147 average of seven times a year (the modal aver- pend on it and as a national heritage and age is four caravans a year). It has been esti- international reserve of biodiversity. mated that each caravan brings in an imputed The overall objective of thc project is to al- net revenue of approximately 30,000 Djibouti leviate the poverty of pastoralists who depend francs (DJF) (US$170). However, a proportioni on the Day Forest and surrounding sylvospas- of cereals acquired through caravans is used as toral areas while ensuring that these natural re- supplementary feed for the livestock herd. sources are managed in a sustainable way. The project would seek to achieve its objective Income from Salaries and Pensions through a strategy-encompassing economicr, In the project zone, forty-three persons environmental, and social elements-that have access to employment outside, while would set in place the conditions for sustain- twenty heads of household receive pensions. able management of the natural resource base Salaries are in the range of DJF 20,000 and relieve pressure on the Day Forest and as- (US$113 per head of family) per month, while sist its regeneration by: pensions are roughly half that figure. * Stengthening the capacity of traditional Remittances from Outiside ie Project natural resource managernent institutions Area * Enabling a more balanced utilization of overexploited and unier-exploited sylvo- Remittances from family members working pastoral resources through the strategic sit- in town or abroad provide an important source ing of livestock- water points and through of revenue for the population. These can range measures to increase the productivity of from DJF15,000 to DJF50,000 per monath Ap- sylvopastoral resources and the livestock proximately 119 families distnbuted across that they can sustain most of the encampments have. memnbers who * Meeting the immediate water-related needs work outside the area. In 45 percent of the of the population to relieve the burden on cases, remtnc are sent on a reguWar basis. women and to foster cooperation betwee Families without access to remittances are as- the beneficiaries and th project sisted, collctively. * Diversifying the range of income- Description of Project generating opportunities available to the The rationale for the project lies in the cur- target group to relieve some of the pressure rent process of degradation threatening pas- on the resource base and to increase the vi- toral, forest, and water resources ia the project ability of rural life area In the absence of intervention, and in the In order to achieve project objectives context of an inexorable process of sedenta- through the strategy outlined above, the follow- rization, the current situation will deteriorate ing components were chosen, in consultation over time, and the natural resources on. which with the beneficiaries, for implementation over the target group depend will no longer provide sCvm years. The components would be im- a means of subsistence. plemented in a phased manner to allow adjust- Tie linchpin of the sylvopastoralism sys- ments in response to the findings of a midtenn tem practiced by the Sekoh Aboussa is the Day evaluation, the changing neds of beneficiaries, Forest, which is in the process of irreversible and the evolution of environmental changes. degeneration. The multifinctional role of the Sylvopas-oral Management Activiies forest includes its use as a pastoral resource and as a source of fuielwood and minor forest This component includes the progressive products, its niicroclimatic functions, and its itoduction of a system of deferred grazing, key role in the rgion's hydrological balance. eventually involving 24,400 hectares; the cre- The rationale for conserving the Day Forest ation of 750 hectares of pemanent pasture; the lies both in its role as a supplier of good. and establishment of 400 hectares of forage plan>- services to the population who immediat:y de- tion; the introduction of improved tree-lopping 148 practices on 200 hectares; and livestock sup- nature of the project, provisions have been port activities. made for a comprehensive M&E system. Forest Conservation and Regeneration Economic Analysis The project would undertake various ac- The economic analysis of the project was tions to support the conservation and regenera- projected out over a time horizon of thirty tion of the Day Forest The project would years. Total econonuc costs and benefits promote both artificial and assisted regenera- streams are shown in Table 6-1 together with tion of Juniper procera on various sites in the the net benefit stream. For case of reference, all Day Forest (50 hectares). Windbreaks around costs have been aggregated as either investment the livestock water points would be planwed in or recurrent costs. The net present value (NPV) addition to the planting of 86 hecares of com- and economic internal rate of return (EIRR) munity woodklos. were calculated on the basis of the net benefit stream presented in Table 6-1. The E[IRR is es- Water Development Component timated at 15.23 percent while the NPV The project would provide support to the amounts to DJF295,488,200 at a discount rate beneficiaries for the construction of thirty-nine of 10 percent which reflects the opportmity drining water storage tanks and twenty-five cost of capit (OCC) in Djibouti. livestock water points. The water points form Quantification and Valuation of Costs an integral part of the sylvopastoral manage- met strategy to reduce pressure on the Forest Financial Costs resources. The financial costs of the project are Soil and Water Conservation shown in Table 6-2. The project costs were es- The project would support soil and water timated over the seven years of project life us- conservation (SWC) activities aimed at en- ing 1992 (January) as Project Year 1. The hancing the productivity of pastr resources exchange ratc pevaing at appraisal of and maintaning the life of water-related infr- DJF177 per US$1 or DJF1 per US$.0056 was structure. SWC measures (stone bunding, wa- applied in the costing. ter harvesting, terracing, and gully pluggng) In confornity with Government of Djibouti would be carried out in an area of 900 practice, the project would benefit from duty- hecures. free import of goods and equipment Only taxes on local salaries and locally purchased Diversification of Income-generating goods and services were therefore taken mto Activities account in the financial costing. All imported goods were valued at c.i.f. (cost, insurance, and Ther projct ould ppot income- freiht)-Djibouti prices obtaining at April 1991 generating activites on a pilot basis as a means -asl taigit- con oa as .. . ~~~~~~~~(appraLsal), talhng mito account local trans- of diversifyng mcome opportunites. Charcoal portation costs where relevant. production using excess deadwood tlhreatening Becau of the relatively open nature of the the forest would be supported, and provisions Djibouti economy and the fict that the D1F s have been made to promote woodcarving, avi- tied to the U.S. dollar, local inflation follows culture, and apiculture. the rate of inflation of Djibouti's principal Project Management Unit and trang partners, m particular the Unted Institutional Supporn States. The local and international inflation rate was esumated at 4.0 percent for the first A Project :Aanagemet Unit would be three years of the project and at 4.5 percent for created. The project would provide a core of the last four years. - These assumptions were locally and internationally recruited staff, hous- used in esimating price coniAngencies, with the ing, vehicles, training, and monitoring and exception of the salaries of locally and inter- evaluation (M&E) support. Given the sensitive Table 6-1: Economic Cost and Benefit Streams of Project over Time Horizon of 30 Years (thtousands of Djibouei fraclcs) Year 1 2 3 4 5 6 7 8 9 Coses Investment cost 197,196.00 88,146.00 83,041.00 59,477.00 36,936,00 22,062.00 10,531.00 - - Recurrent costs-project 25,164.00 29,240.00 31,520.00 *27,321.00 28,312.00 27,537.00 16,639.00 2,033.00 2,033.00 Total cost 222,360.00 117,386.00 114,561.00 86,798.00 65,248.00 49,599.00 27,170.00 2,033.00 2,033.00 Economic benefiu Forage production -42.00 47.00 274.00 539.00 970.00 1,674.00 2,627.00 3,407.00 4,018.00 Price fonge 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 Forage production -1,050.00 1,175.00 6,850.00 13,475.00 24,250.00 41,850.00 65,675.00 85,175.00 100,450.00 benefit Avoidance of (odder 1,227.50 2,455.00 3,682.50 4,9!0.00 6,137.50 7,365.00 8,592.50 9,820.00 11,047.50 loss Food production 725.00 1,750.00 2,775.00 3,175.00 3,535.00 3,535.00 3,535.00 3,535.00 benefit Charcoal production 7,560.00 14,840.00 14,840.00 7,560.00 7,560.00 7,560.00 7,560.00 7,155.00 7,155.00 benefit Net benerit beebreeding - 347.00 480.00 1,231.00 1,796.00 2,420.00 2,580,00 2,580.00 2,580.00 Net benefits poultrv 94.50 154.50 214.50 274.50 334.50 334.50 334.50 334.50 Net benefit woodcraft 320.00 820.00 1,543.00 2,425.00 2,977.00 3,545.00 3.545.00 2,941 00 2,941.00 Total benefit 8,057.50 20,456,50 29,300.00 32,590.50 46,170,00 66,609.50 91,822.00 111,540.50 128,043.00 Total net bencfit -214,302.50 -96,929.50 -85,261.00 -54,207.50 --19,078.00 17,010.50 64,652.00 109,507.50 126,010.00 - - - Table 6-1: Eonomlc Cost and Benefit Streams of Project over lTle Horizon of 30 Years (thousands of Djibouti francs) Cont'd Year 10 11 12 13 14-IS 16 17-30 Costs Investment cost 2,400.00 2,400.00 - 2,400,00 Recurrent costs-proJect 2,033.00 2,033.00 2,033.00 2,033.00 2,033.00 2,033,00 2,033.00 Total cost 4,433.00 2,033.00 2,033.00 4,433.00 2,033.00 4433,00 2,033.00 Economic beneftis Forage production 4,381.00 4,680.00 4,849.00 4,851.00 4,851.00 4,851.00 4,851.00 Price forage 2S.00 2S.00 25.00 25.00 25.00 25.00 25,00 Forage production 109,525.00 117,000.00 121,225.00 121,275.00 121,27S.00 121,275.00 121,275.00 benent Avoidance of fodder 12,275.00 12,275.00 12,275.00 12,275.00 12,27S.00 12,275.00 12,27S.00 toss Food production 3,535.00 3,535.00 3,535.00 3,535.00 3,535.00 3,535.00 3,535.00 benerit Charcoal production 7,155.00 7,155.00 7,155.00 7,155.00 7,155.00 7,155.00 7,155.00 benetlt Net benent bcebreeding 2,580.00 2,580.00 2,580.00 2,580.00 2,580.00 2,580.00 2,580.00 Net bencfits poultry 334.50 334.50 334.50 334.50 334.50 334.50 334.50 Net benefit woodcraft 2,941.00 3,291.00 3,291.00 3,291.00 3,291.00 3,291.00 3,291.00 Total benerit 138,345.50 146,170.50 150,395,50 150,445.50 158,44550 150,445.50 I50,445.50 Total net benefit 133,912.50 144,137.50 148,362.50 146,012.50 148,412.50 146,012.50 148,412.50 6 The Rural Devdooment anwd Environmentga Proaecdon Project in the Day Forer in Plibouti 151 Table 6-2: Fmancibl Costs by Project Component (thousands of Djibouti francs) Sylwv- Seol ad paosts N Cr Hvm- haunc Ih*c Miyd ccvsugd Pr" ctnangetnA&s msa - win- I h,JIu- diva4- Mosage-- mcat sm an ure mieu s; am, rTar Ptal Amount Poems Amt L hnne:t at A. Civil works t. Housing 0.0 to to Qo0 13.275.0 13,275.0 10.0 13277.5 Z2 2920 2 Administrative buildings 0.0 Q0 0.0 0.0 15,93Q0 15,930.0 10.0 1,593.0 22 350.5 3 Oher buildings o.o to 0.0 5Q310O 24,730. 30,090.0 10.0 3,009.0 10 9003 Subtotal tO 0.0 0. 5,310.0 5350 59,25.0 10.0 5,933 26 1,542.8 a Equipmnt and vehicles 1. Agrieultural equipment 12,89ZO to 00 0.0 0.0 1,8921 0 0.0 0.0 20 257.8 2 Hydraulic equipment QO 0.0 34.12QO 0.0 0.O 34.120 83 2830 2.2 739.0 3 Vehides (O 0.0 4,6727 0O 27,73.70 42,453.0 0O QOO 7.7 327(17 4. Offce equipment oaO QO 0LO 0.0 7,8440 7,844.0 3.8 300.0 21 162.9 5. Other equipment 99QO Z7555 0.0 4,6015 LO 8,37.0 0.6 50.0 3l 257.7 Subtotal 13J88Z0 Z755.5 4.7920 4.C*15 35C25.0 105,691.0 3.0 3.182.0 4.4 4,685.1 C Inputs and services 1. Raw mkterials and tools 0.0 0.0 431611.7 0.0 0O 43161L7 0.O 1O 13.6 5.93663 Z Seeds and plants 28,5420 0.0 0O. 0.0 0.O 2;,SSZO 0.0 0O. 16.6 4,7753 3. Qualiried manpower 0.0 0.0 168.0 0O 0.0 168.0 10. 16.8 81 13.7 4. Food rations 29.1S2.0 23,5410 3,00Q0 0.0 .O 55,6920 0.0 O 145 8,075.0 S. Compensation indemanty 1 1,343.1 0.0 0.0 0.0 0 17,9411 0.0 0.0 18.1 3,246.2 Subtoial 75,9371 23,54Q0 45.779.7 0. 0.0 1462568 0.0 16.8 I15 2Z,046.5 D. Tedmical assistace 1. Technical assistane 21.2152 0.0 21,212 15,58Q0 96o00O 14,010.4 0.0 QO 7A 11,457.6 Z Training 0.0 0.0 0.0 0.0 1,416.0 1,410.0 0. 0. 2.0 283 3. R&D and M&E support 0.0 0.0 0.O 0.0 6,915.0 69150 0.0 0O. 9.1 626.1 Subtotal 21,215.2 0.O 21,2152 15580.0 104,23L0 162,341A 0.O 0O 7.5 12,1120 Total investment cost 111,0343 26,W .5 116,7869 25,5313 193,941.0 473,5893 1.9 9,12.3 8.5 40,3895 IE Reinvent cefl A Civi wors maintenance 1. Buildings 0.0 O O.O 0.0 16,1955 16.1f95 0.0 0.0 181 2,927.9 B Equipment maintenace 1. Agricultural equipment 19,020.4 QO 0.O 0O 0.0 19.02A.4 0.0 0.0 15.8 3001.7 2. Hydraulic equipmcnt 0.0 0.0 25,117.4 0.0 0.0 25,117.4 0.0 0.0 153 3,395. . 3, Vehides (LO 0.0 9,90 0.0 19,446.7 29,379.7 6.6 1,914.7 13 3,01.2 4. Odter equipment to. 0.0 (O 0.0 4,2300 4.200D 00 0.0 15.8 662. Subtoal 19.0204 QOL 35,0504 0.0 23,646.7 77,7175 2.5 1,944.7 14.8 11,52L1 C. Pesounel salaries 30,826.0 0.0 40,360 0.0 41,160.0 112,356.0 0.0 0O 8.7 9.745.9 and fees D. Other opertiag csts OO 0.0 0.0 1,447.5 13,02Q0. 14,4673 9.0 1,3020 16.6 2410.7 Total recurrent costs 49,844 0.0 75,41S.4 1,4473 94,022 223736.5 1.5 3,246.7 120 26.59S6 Total baseline casts 16,8817 26,2955 192,353 2,979.0 287j9632 694,325.8 1.8 12375.0 9.6 66.985.1 Plysical contingencies 50 0 0.0 2848 531.0 8,945.2 1Z375.0 Price contingecies 21.1933 3,217.1 18.5712 1,869 2Z,0569 66,985.1 1.0 6393 Total prqect costs 1S2,126.0 29,5826 213.6322 29,379.7 318.9653? 773.658 1 13.0143 8.7 66&985.1 Tames 11,20.1 390.0 165713 0.0 15,855.8 44.0i9.2 IA 633.4 Foreign exchange 118,963.9 2,192.6 133.691.8 2568.2 206,434A 514.150.8 1.6 X,19&8 Values scaled by 1000.0. 152 nationally recruited staff, to which an indexa- value of the food ration is an underestimate of tion of 2.5 percent was applied. their opportunity cost of labor. To compensate Physical contingencies of 10 percent were for this, the economic price of labor was taken applied to buildings, to equipment and inputs, to be DJF500 (A conversion factor of 1.25 was and to operating and maintenance costs for applied.) equipment. Cost of Skilled Labor. The scarcity of The total cost of the project, including skilled labor in the Djibouti market implies that physical and price contingencies, was estimated its financial price reflects its opportunity cost at US$4,365 million or DJF772,072 million. adequately. Skilled labor was therefore valued The ptrcentage of costs in foreign exchange is at financial prices. However, an exception was in the order of 66.4 percent. Investment costs made in the case of masons in the project area, of US$2,950 million represent 67.6 percent of whose economic cost was estimated at 50 per- total project cost, while recurrent costs of cent of the financial cost to reflect the fact that US$1,415 million constitute 32A percent of to- this category of labor faces significant tal project cost underemployment. Total project cost is distibuted among Cost of Pasture Resting. The deferred components as follows: project management grang scheme and the forage plantations to be unit, 41-3 percent; water development compo- promoted under the project generate an oppor- nent, 27.6 percmt sylvopastomal management tunmity cost to pastoralists in the form of land component, 23.4 percent; soil and water con- taken out of pastoral production in the early servation component, 3.8 percent; and pilot years of project life The project will compen- incom-generating activities component, 3.8 sate pastoralists for this loss. While the costs percent of resting pastue as part of the rotational graz- Economic Costs ing scheme should be included in the economic analysis, it would conste double-counting if Economic costs were estimated in c this value were to be included as a separate prices (that is, net of inflation) and were based cost in addition to the compensation payments on financial costs net of taxes and transfers. already included under the costs ascnbed to the All quantified econoric costs of the project sylvopastoral component. Because the costs of were included in the analysis, regardless of compensation reflect the economic loss from whether they were grant-financed- The valua- pasture restg, no additional economic cost tion of certain economic costs that require fur- has been included in the cost stream. Com- ther elaboration are discussed below. pensation in fodder units (EU) per hectare is Cost of Unskilled Labor. In the financial shown in Table 6-3. analysis, the cost of beneficiary labor was tk- Quatificon and Valuation of en to be equivalent to the cost of WFP rations Economic Benefits provided under the project's food-for-work ac- tivities. In the economic analysis, beneficiary ForageBenefits labor both under the food-for-work scheme and for other activities was valued on the basis of As a result of the project's pastoral and the local population being willing to provide la- sylvospastoral management activities, the proj- bor in return for WFP food rations equivalent ect would increase amnual forage availability to DJF400. This can be considered a lower by 113 percent from 4,264 tons in the base bound estimate of the opportnity cost of teir year to 9,115 tons at full development These labor, because the experience under pre- figures represent 'accessible" forage only. investment projects in the area indicates that In addition to incremental fodder produced beneficianies only engage in food-for-work acti- tromgh planting and rangeland management, vites when they perceive the benefits of the benefits would accrue in the form of costs of work as accring indirecdy or directly to them pasture degradation avoided. These benefits (for example, soil and water conservation for would arse from the prevention of the productivity gain). Thus, it is clear that the 6 The Rural Development and Environmental Proeation Project in tla Day Forest in Djiboud 153 Table 6-3: Compensation for Pasture Resting Year Aclivity 1 2 3 4 5 6 7 Prairic establishmcnt 400 0 0 0 0 0 0 Forage plantaLion 100 100 5C 0 0 0 0 Junipcr regeneration 500 500 500 400 300 100 0 Re-aElorestation 400 400 300 200 100 0 0 degeneration of Day Forest resources and sir- in a price of DJF37.6 per kilogram. This in rounding pastoral vegetatioa turm was converted mto dry matter units (DM) Surrogate Market Technique. Because no using a conversion factor of 2, resulting in a market for forage exists, prices prevailing in a price of DJF18.8 per kilogram of dry matter. surrogate markcet were used to impute its eM- Taking the deficit in digestible protein of sor- nomic value. Because pastoralists are increas- ghum (evaluated at 30 percent of the value of ingly relying on supplementary feed to covcr its substitute) into account, the price of sor- fodder deficits, sorghum can be considered a ghum equivalent of 1 kilogram of dry matter substitute for forage. As such, dte price of sor- was estimated to be DJF24.5. ghum feed equivalent can be used as a surro- Change in Production Approach. In order gate for the price of forage to confirm the result provided by the surrogae The price of sorghum feed equivalent was market technique, another valuation technique, calculated in the following manner As a first known as the "change in production" approach, step, the ci. price of sorhum was computed. was also applied Under this approach, the val- * Gulf PortUS$96 per ton uation of incremental dry matter was based on * FreightUSS 1 00 per ton fodder conversion to milk and meat production. Because of data constauits and to ensure a * InsuranceUS$5 per ton conservative estimate, the value of hides was * c.i priceUSS201 perton excluded and technical production paamets When local transport costs from Djibouti prevailing in the '"ithout project" situation port to the project area of DJF5,337 per ton are were appied (in other words, projected im- added to the c-i.f price of sorghum, this results provements in the quality and quantity of the in a price of DJF41 per kdlogmm. In order to herd in the with project situation were not tak- translate this into the price of fodder equiva- en into account). lent, a coefficient of 1.09 was applied, resulting '54 Table 6C4: Livestock Production Parameters Meat Milk Species (kilograms) (liters) Sheep 4.7 30 Cattle 16.0 400 C-oats 6.6 40 Camels 14.0 600 Donkeys 0.0 0 Table 6-5: Sekoh Aboussa Livestock and Feed Requremeits Tropical Meuric tons of Sorghum feed livestock dry mnaler Type Number equivaleni standard units peryear Cattle 1,324 0.73 966.52 2,206.51 Shccp 1,244 0.12 149.28 349.51 Goals 3,331 0.12 399.72 911.75 Camcls 472 1.00 472.00 1,076.75 Donkcys 218 0.40 87.20 198.34 Total 6,589 2,074.72 4,734.34 &The Rural Delop,ment and Em'irvnueutal Protecion Project in the Da Forest in Djiboud 155 Table 6-6: Evolution of PustorA Facies without Project Year 0 Year 10 Production Prcductio,i Area in in Ions of Area in in tons of hectares dry matter hectares diy matter Jun. procera 820 873 0 0 Acacia etbaica 1,364 812 1,502 894 Aizoon canariense 5,164 2,272 5,846 2,572 Acacia mellircra 6,596 132 6,596 132 A-mell.and others 7,984 120 7,984 120 Rhingo. somalense 2,300 23 2,300 23 A. cherbergiana 1,424 32 142 432 Total 25,642 4,264 25,652 3,733 Table 6-7: Incremental Forage due to Soil and Water Conservation Activities (thousands offodder cnits) Year 1 2 3 4 5 6 7 8 9 10 11-30 Benefit 0 0 4.5 23.5 51.8 85.1 125.6 146.8 155 160 1653 Milk and mea production parwmets for circuito route was required to derive the total cattle, amels, goats, and sheep in the project value of milk production. On the basis of an area were derived from available estimates and exaustive household survey carried out at the fiom the inbmtion gerated by fte project appr ste, the proportion of lactaig fb- preparation and appraisal missions. The csti- males was established fir each type of live- mated livestock production paramts are stoclk The total number of lactatng fnales of shown in Table 6-4. each type of animal was multiplied by the milk Data on the size and composition of the production paramet for each respective type herd were collected during pre-investent ast- of animal. Total milk productioni was valued at vities and during preparation and appaisal the obseved arket price of milk of DJF150. missions. Herd size and composition fbr Sekoh The ttal value of milk and meat produc- Aboussa livestock are shown in Table 6-5. tion was estimated to be D3F56,500 million. The value of meat production was calcu- On the basis of estimated fodder requirements laWed by applying the observed prioe of meat tD of the herd in kldogamms of dry mater tal meat production of the herd. A more (calculated by converting the herd into standard 156 tropical livestock units), the total fodder re- Benefits in the Form of Forage Loss quirements of the herd were established (see Avoided Table 6-5). The valuc of a kilogramn of dry matter was Becaus the s"wit and wout promect" derived by dividing the value of total meat and framesork forms utie basis or the economic milk production by total energy requirements analysis, preduction changps expected to take of the herd. This results in an imputed price of place in the "without" project scenario should DJF 12 per kilogram of dry matter. This can be be reflected in the estimation of the net benefits considered a lower bound estimate because the generated by the project. In the pwithout value of hides, the "psychic value" of owning proect" situationg the paetoral resources are livestock, and other values have not been re expected to degrade over time. [nr particular, flected in the imputed price. In addition, no relatively productive species are expected to be provision was made for the fact that in the replaced by less productive vaneties. Notably, 'with project" scenario, the structure and pro- the pastoral species of the Day Forest which ductivity of the herd would improve as a result represent a pastoral resoure of high value If of the project's livestock support activities. aged properly, are projectd to disappear Valuation of Incremental Production Due m ten years under the current pastoral manage- lo Project The price of using the surrogate ent re. The overal change i tb com- market technique was considered more robust position and productivity of vegetative cover is and was applied to incremenal orage produc- given in Table 6-6. tion in the economic analysis. However, the As can be seen from Table 6-6, the reduc- sensitivity of the EIRR to changes in the value tions in pasture production expressed in metric of forage was explored and presented as part of tons of dry matter per year are quite substan- the analyeis (please see section on "Uncertainty tial, representing a decrease of 11.5 percent and Sensitivity Analysis" below)- over ten years As costs are benefits lost, and A basic assumption used in the analysis is benefits are costs avoided, this loss of produc- that the entire incremental production of forage tion was included in the economic analysis as a would be put to productive use. The incrase in benefit ofthe "with project" situation. forage availability due to the project would al- Benefits of Soil and Water Conservaion low the current rate of increase in Sekoh The soil and water consevaon activaes Aboussa livestock numbers of 3 percent per to be promoted under the prqject are geared to- year to contnue, while maintaning a buffer of wards increasing forage production and enend- 30 percent for dry years. Much of the surplus ing the life of livestock water ponds. The soil forage available beginning in Project Year 6 and water conseuton actiitivs wil increase would accrue in the Mak'arrassou portion of rangeland productivity through (a) mainaining the project area, where in good years the Sekoh soil depth and quality, (b) increasing water Aboussa allow other tribes to graze under re- retention and soil moisture, and (c) colltroing ciprocal arrangements that enable Sekoh the severe gullying currently taking place. Aboussa cattle to graze elsewhere in years The incremental foRage production in fod- when productivity is low in Sekoh Abboussa der units due to soil and water consevation terntory. A certain amount of tolerance for activities is shown in Table 6-7- neighbonng trbes is also evident in the non- Avoidance of Double Counting. The for- forest grazing areas directly under the control age production benefits of soil and water con- of each of the five Sekoh Aboussa lineages. servation are an integal part of the projected Accordingly, it was considcred reasonable to incremental forage production estimates. In or- assume that the forage would be put to produc- der to avoid double counting, these benefits, al- tive use. ready subsumed under forage production benefits, were not included as a separate bene- fit stream in the economic analysis. The anti-erosion tratment of the 5-hectare cateQments of livestock water ponds would 6. The Rural Development and Environmental Protection Project in the Day Forest in Djibouti 157 reduce siltation by at least 50 percent. This re- Avoidance of Double-counring. In order to duction in siltation rates would result in a sav- avoid double counting, the benefits of livestock ing of 100 man-days of maintenance water ponds were not included as a separate requirements for all livestock water ponds. benefit stream in the economic analysis. The This benefit was fully reflected in the estima- ponds are an integal part of the rangeland tion of annual operating and maintenance costs management strategy of reducing pressure on for the livestock water ponds. It would consti- the Day Forest and opfimizing the seasonal and tute double-counting if thcsc savings in labor spatial distribution of grazing pressure. These were to be reflccted in the economic analysis as benefits are therefore subsumed witin the esti- a separate benefit. mates of incremental forage availability result- Downstream Externalities. Soil and water ing from project activities. conservation, in addition to increasing on-site productivity, would also generate positive eco- Benefits of Woodand Charoal nornic extemalities tirough the mitigation of Production flooding donvnstream. Floods have occurred Direct Market Price Valuation of Wood with increasing firquency ir recent years, caus- Production. While the Day Forest does provide ing severe damage to infiastructure and to agn- a source of wood for certain lineages of the Se- culture. The benefits of mitigating flood koh Aboussa tribe, there is a need to take a damage were not calculated due to the absence long-term view providing for alternative of reliable data on downstream farm produc- sources of wod supply. The plantions and tion and maintenance costs of roads. Data con- forest regenertion activities would result in an straints of this nature constitute the norm for a incremental wc -'4 production of 145 cu- project analysis im marginal areas for which bic meters in Project Year 2, and 350, 555, and secondary sources of data are rarely available. 635 cubic meters respectively in the succeeding Benefits of Water Der!rlopmenz three years, stabilizir& at 707 cubic metes from Project Year 5 onwards. Altogether, The scarcity of drinkdng water is a severe wood production would increase by 11,760 cu- constaint in the project area and constitutes bic meters durng project life. the main preoccupation of the local population. While most of the wood produced would The drinking water component of the project be autoconsumed, the observed market price would provide important social and health for wood of DIJF5,000 per cubic meter was benefits that are difficult to value in the ab- used for the analysis. The market price of sence of epidemiological studies. wood rcflects primarily the opporunity cost of Opportunity Cost of Labor Approach. In labor in collecting and transporting wood. principle, the economic value of the drnking While the market price might overemate the water benefits could have been imputed using direct user value of wood of encampments the opportunity cost of time approach. Under close to the forest, it is a reasonable approx- this approach, the savings in time resulting imation if one considers that for many of the from availability of water would have been val- encampments removed from the forest, the ued on the basis of the opportuity cost of availabili tf fuielwood is an important con- time. This was not done due to inforrnation straint. This is especially the case for lineages constraints. Nevertheless, it is clear that a sig- that do not have access rights to the forest It is nificant amount of time is spent collecting wa- primarily in te vicinity of encamnpments re- ter fiom distant wells or springs. In the case of moved froin the forest, where fuielwood collec- many encampments, it was ascertained that lion takes plane once every two or three days women require an entire day to make one round and can occupy an entire day, that the project- tip to a well. In several cases, women spend supported woodlots would be established. Fur- up to two days per week on this task. However, thennore, the value Gf the envirommental ser- for the valuation exercise to have had creibili- vice functions provided by the trees, which can ty an,,in-dcpth. survey -of household water only be quantified-even in physical collection activities would have been requird.. terms-with difficulty, is likely to be Is Table 6-8: Effect of Time Horizon on EIRR TinJi Izonzi EIRR (years) (peirent) 10 1.84 15 11.05 20 13.82 30 15.24 50 15.57 subsuntial but is not reflected in the market Benefits of Income-generatingActivities price. Orher V. larion Approaches. Fuelwood In order to broaden the income base and could also have been valued, using the provide a more diversified portfolio of incoane- surrogate market technique, by applying the generating activites for the local population, rprce of kerosene to fiehiqood, because kero- the project would support a number of pilot sence appears to be fte cosest energy substitte mcome-geneating activties. These activites in the project area. b oetwever it wras considered would generate tangible financial and economic tha this would overestimte the beneists con- benefits. Inputs and products of the artsanal siderably. Alternatively, the o theits-cost- woodcri apiculture, and aviculture activities of-time approach could have bee applied. were valued at market prices. Labor costs were However, the costs of generating the necessary valued at the shadow price of unskilled labor. data were deemed not to justify any incremen- The net benefits of these arctvites (net of ap- tal accuracy that might have been achieved in beatio costs and labor costs) are shown in Ta- the valuation. ble 6-1. Market Price Valuation of Charcoal Pro- Olier Benefits duction An important ifctor contnrbuting to the degradation of fte Juniper stand is the pres In adtion to the benefits included in the ence of the fungus Armillana mellea. The sig- analysis, the project would generate benefits nificant amount of deadwood littering the forest that are not readily quantifiable in physical, let contibutes to te spread of this parasitical alone econoiic, terms. In addition, benefits ac- fungus. Total deadwood amounts to 30,000 cu- cn to future generations have not been tal- bic meters. Of this amount, approximately ec iato account Tese benefits can be 5,650 cubic meters would be converted into kategorized as user benefits and nonuser charcoal. The imnproved charcoal methods to be benefits. extended under the project would allow a net User Benefits. The Day Forest provides annual gain over wood sales of US$35,000 in both productive and service fimctions. While Project Years 1, 4, and onwards, vith a net somne of its productive functions are reflected in gain of US$70,000 in Project Years 2 and 3. the economic analysis, its service fimctions are These values were based on an estimated char- more difficult to idntify and to quantify in coal conversion factor of 20 percent and an ob- physical terns and value. In addition, some of served market price for charcoal of DJF200 die eavironental service flmctions translate per kilogram. into production benefits outside the project boundaries. 6. The Rural Developmcent and Environmental Protedfion Project in the Doy Forest in Djihouii 159 Tabl 6-9: Value of Forage and the EIRR Price EIRR Valuation method (Djibouti francs) (percent) Surrogate market method 25.0 1523 Produclivity method 120 0(877 Average 18.5 12.29 The likely presence of such externalities in- On an operational level, data on visitation rates dicates that the boundaries of analysis for proj- are extremely patchy, although a limited survey ect appraisal should be extended. While ihere has been carried out by the National Tounism are reasons to believe that the loss of forest Office. The year-long survey necessary for the cover would upset the hydrological balance in proper application of the method was beyond the surrounding areas and reduce groundwater the scope of project prepantion activities. On a availability downstream, the precise cause-and- conceptual level, the approach would have effect relationships would be difficult to speci- been hampered by the fliat that most visitors to Wy. The prime constraints include the lack of the Day Forest begin their journey fiom Dji- empirical data on the rate of infiltration in the bouti town. Accordingly, significant variations forest area and the difficulty in modelling in travel cost do not occur, and therefore vari- groundwater recharge in the exemely fissured able entrance costs could not have been simu- basaltic substratm characterizing the region. lated meaningfilly to derive a demand curve. In Had it been possible or cost-effective to gener- the case of international tourists, sophisticated ate the necessary data in the context of the nor- models would be required because of the multi- mal project cycle, it might have been possible purpose nature of visits to Djibouti. to include a proportion of the ret value of cur- In principle, the CVM approach could rent production in surrounding areas, which de- have bee used to value the consumers' surplus pend on water sources originating in the forest, accruing to recreational users of the forest Al- as a benefit in the form of costs avoided from though hampered by several inherent biases, forest degen.sation and the loss of vegtive CVM is particularly use.ful in situations where cover in surrounding r2ngelands market data are lackg. However, this ap- Other uses of the forest that were not val- proach, like the TCM, is hampered by the sur- ued include its role as a recreational resource veying constraint and problems in defining the for residents of Djibouti town as well as in- user group. Difficulties in its application for lo- trnational tourists. The travel cost method cal users would have stemmed from the benefi- 07CM) or contingency valuation method ciaries' lack of integration in markets, which (CVM-4) night have been applied to value these would have hampered the proper fimetioning of benefits. The TCM approach, entails deriving a any hypothetical market. Moreover, CVM has demand curve for a resource (generally a recre- not achieved the necessary credibility for appli- ational resource) us ug variable travel costs as cation in fornal project appraisal. a surrogate for entrance fees, in order to esti- Nonuser Benefits. Other benefits of pro- mate consumer's surplus. tectng the Day Forest are in the fim of non- Certain constraints militate against the ap- user benefits or so-called "intrinsic benefits.? plication of TCM in this particular situation. The existence value or bequest value of the Table 6-10: Comparison of Dry Matter Productvity in Wet, Normal, and Dry Years Productivity Area (hectares) Kilograms of dry matter Total production Grazing rate (metric tons Global production (metric tons of dry matter) of dry matter) Vegetative Wet Normal Dry Wet Normal Dry Wet Normal Dry Wet Normal Dry Wet Nonnal Dty cover Jun. procera 820 820 820 2,070 1,775 1,410 1,697.4 1,455.5 1,156.2 0.60 0.60 0.80 1,018.4 8733 925.0 Term br na no 304 na na 1,100 na na 334.4 na na .80 na na 267.5 A. et baia 1,364 1,364 1,364 960 850 680 1,309.4 1,159A 927.5 0.70 0.70 0.90 916.6 811.6 834.8 Alzoon ca 5,164 5,164 5,164 650 550 440 3,356.6 2,849.2 2,272.2 0.80 0.80 0.20 2,6853 2,272.2 454.4 A. mellifera 6,596 6,596 6,596 460 400 320 3,034.2 2,638.4 2,110.7 0.05 0.05 0.05 151.7 131.9 1055 A meli others 7,984 7,984 7,984 340 300 240 2,114.6 2,395.2 1,916.2 0.05 0.05 0.05 135.7 119.8 95.8 Rhlgo.som 2,300 2,300 2,300 240 200 160 552.0 460.0 368.0 0.50 0.05 0.05 27.6 23.0 18.4 AEhren. 1,424 1,424 1,424 520 450 360 740.5 6408 512.6 0.05 0.05 0.05 37.4 32.0 25.6 Total 25,656 25,652 25,956 13,404.6 11,589.5 9,597.8 4,972.4 4,263.8 2,724.1 6. The Rural Development and Environmental Protection Project in :he Day Forest in Liibowt 161 Tabic 6-11: Production Scenaios and the EIRR Production scenano EIRR (percent) Appraisal mission estimate 15.23 30 percent shortfall 12.90 Sigmoid production 15.00 C(yclical drought 14.70 Table 6-12: Without-project Scenario and the EIRR Without-project scenario EIRR (percent) Degradation trend 15.23 Stable production 13.96 forest as a recognized national patrimony is an example of this. There is a also an option value Risk and Uncertainty and Sensitivity (straddling the distinction betwee user and Analysis nonuser benefits) associated with maintaiing An obvious means of addressing uncertain- the forest as a unique store of biodiversity. The ty in project appraisal is to invest in more in- irreversible loss of the forest is a major issue in formation. However, this is not always possible this regard. However, techmiques for the or cost-effectivc in the course of the stadard quantification of these relatively intangible project cycle. In crsz.s whee uncertainty can values-including CVM--have not reached the be converted ino risk, sensitivity analysis can stage where they can be considered appropriate demons the effict of changes in basic as- the context of fonnal MFI (multateal fian- sumptions governing the analysis. It can also ial istution) project appraisal. The lack of serve to highlight variables that require closer famLiliarity of decisionmakers with these ap- scrutiny either in the course of furither project prwaches would offer another significa con- prepafation or dur:ig project implementation. straint to their application. in the sensidvity analysis of this project, the effect on the EIRR of changes in variables such as the imputed price of forage, the pro- duction of forage over time, the time horizon of 162 Tablc 6-13: Scnsitivity of EIRR to Chages in Total Costs and Total Benclits (perccnt) Changcs ir. total -10 percent -20 pcrcent -50 percent bencfils Changes in total 15.23 13.856 12.40 7.43 costs +10 perocnt 13.98 12.68 11.31 6.54 +20 percent 12.90 11.66 10.34 5.754 +50 percent 10.34 9.23 8.05 2.o3 Table 14. Switching Valoes (thousands of Dji-bouti francs) Appraisal present Switching present B3enefit stream value value Percent change Forage production 60,783.26 312,375.05 -&61 Wood production 25,704.47 -269,783-73 -1,14956 Charcoal production 80,907.38 -214,580.82 -365.22 Avoidance of loss 75,932.39 -219,555.81 -389.15 Apiculture 17,054.27 -278,433.94 -1,732.64 Aviculture 2,396.41 -293,091.80 -12,330.45 Woodcraft 24,0023 -271,497.98 -1,230.78 Total benefLts 833,866.41 538,378.21 -35.44 Total cost 538,378.21 833,866.41 54.88 6. The Rural Development and Environmental Prorection Project in the Day Forest in Djibouti 163 analvsis, and "without project" degradation The case for choosing the higher esimate trends were explored to offer the decisionmaker is fulrther reinforced by the likelihood that the a better understanding of the economic relative price of forage will increase over time. justification for the project. This prediction is based on the plausible as- Sensitivi.y of the ElIR to the Time sumption that the current dryland degadation Horizon trends threatening grazing resources in Djibouti will continue, diminishing opportunities for Although the time horizon used for the eco- tashumance outside the project area. nomic analysis of many investment projects is Sensitivity of the EIRR to Changes in often Icss than twenty years, in the case of this Forage Prodtiction Assumptions project it was extended to thirty years to cap- ture some of the longer-term benefits of the The sensitivity of the EIRR to changes in project. assumptions concerning thc magnitude and It can be seen in Table 6-S that the choice time profile of forage production was explored. of time horizon for this project can affect the The EIRR was calculated uaing three alterna- EIRR for periods below thirty years. When the tive forage production scenarios. The first sce- time horizon is extended beyond thirty years, nario entails a 30 percent drop in annual forage the change in the EIRR is small. While this re- production estimats , to simulate below- flects the nature of the net benefit stream, it is average rainfiall conditions and/or fdilure to fil- more directly an outcome of the fact that pres- ly adopt the proposed sylvopastoral manage- cnt values of costs and benefits occumng after meet activities (see Table 6-10 for forage thirty years are considerably reduced at any production in dry, wet, and normal years). The positive discount rate. second scenario entails a sigmoid production It should be noted that extending the curie to simulate slow adoption of sylvospas- boundary of analysis in time reduces the credi- toral management activities, which then accel- biity of'"with and without" project projections. erates over time before tapering off. The third This must be weighed against the need to re- scenario simulates the occurrence of droughts flect benefits or costs accruing in the longer at five-year intvals. The results of the sensi- term. tivity analysis are presented in Table 6-11. Sensitmty of the ETRR io the Relative The EIERR is surprisingly robust in the face Srienitivi of tore R!th eliOf plausible changes in the -nagnitude and ime Price of Forage profile of forage production. It is clear, howev- In order to check the sensitivity of the in- er, that gazing discipline would collapse in the ternal rate of return to variations in the imputed event tiat severe and persistent drought occurs. price of foage, the EIRR was calculated under Should this happen in the early years of the different price assumptions. The results of the project, before buffer plantions mature and sensitivity analysis are presented in Table 6-9 before the merits of pasture resting, and so below. forth, are demonstrated, it is likely that the suc- k is clear that the EIRR is quite sensitive cess of the project would be severely to the imputed price of forage. However, it jeopardized. should be noted that both the surrogate market Sensitivity of the ElIR to Changes in technique and the productivity method, as ap- J01hout-Proect Assmptions plied in this analysis, underestimate the value of forage. This is because the imputed price Forage production in the without-project only reflects the value of the vegetation as a situation is projected to decline over time. The productive resource and does not capture the sensitivity of the EIERR to changes in this as- enviromnental service value of the vegetative sumption is shown in Table 6-12. cover. Thus, a case can be made for choosing the higher rather than the lower estimates for the price of forage. 164 Table 6-15: Distnrbution of Incomc Percent of Per capita incone Persons total USS200-250 167 7 USS150-200 249 10 USSIOO-150 1, 235 52 USS1IO or Icss 627 26 Not available 106 5 Sensitivity of EIRR to Changes in Total Costs Fiscal Impact of Project and Total Benefits Streams Djibouti's external debt outstanding and The sensitivity of the EIRR to changes in disbursed (DOD) increased from US$96 mil- total costs and total benefits is shown in Table lion in 1985 to US$163 million at the close of 6-13. The EIRR is quite robust in the face of 1989. This rise, amounting to 40 percent of plausible changes in total benefits and total goss domestic product (GDP), is a result of costs. Drastic increases of total costs or de- heavy boownving in the early 1980s to finance creases in total benefits do affect the EIRR public investment projects. The 1991 state substantially, with the EIRR being slightly budget is an austerity budget, prepared in the more sensitive to changes in the benefi; stream. context of the Gulf crisis, the closing of the In the unlikely worst-case scenario with costs border with Somalia, the collapse in trade with increasing by 50 percent and benefits decreas- Ethiopia, and the devaluation of the dollar, ing by 50 percent, the project EIRR would be which-given the fixed parity exchange rate reduced to 3.87 percent. between the Djibouti Franc (DJF) and the U.S. dollar-was fueling inflation. One operational Switching Values for Benefit and Cost outcome of the austerity plan has been that the Variables government of the Republic of Djibouti is b- The switching value of benefit streams and coming extremely cautious in undertaking new the total cost stream were calculated using a development projects. It is increasingly seeking discount rate of 10 percent. At this discount a al grant financing for public investment- rate the net present value (NPV) is In this context, the fiscal impact of the D1F295,488.2. The results are presented in Ta- project takes on special importance. A key ele- ble 6-14. ment in the project's strategy is the gradual It is clear that forage production benefits strengthening of beneficiary institutions to al- constitute the critical benefit stream. Changes low a gradual devolution of project activities to in this stream have the greatest impact on the the beneficiaries themselves, which would economic viability of the project. At the same serve to reduce the fiscal burden on the govern- time, these benefits would have to fall by 48 ment The recurrent costs after project comple- percent for the project to be economically tion are quite low, amounting to DJF2,033,000 unattractive. or US$11,485. However, a characteristic of many environ- nmcmally focused projects in marginal areas is that the environmental improvements rarely generate significant cash flows or transfers. 6. The Rural Development and Environmental Protection Project in the Day Forest in Djibouti 165 The maintenance of the productivity of a mar- and the construction of livestock ponds in the ginal rcsource base translates-by dcfini- Mak'arrassou. The third group consists of the tion-into marginal cconomic returns. Even population of the Bargak Adbara tribe in the where the cconomic bcncfits justify the costs of encampmcnt of Mcngucla, who would bencfit intervention, these bencfits are in the form of from water tanks but whose livestock do not gains in subsistence consumption while the graze in the project arca. The fourth group con- costs are in the form of financial flows. Few sists of approximatcly 1,300 additional mern- positive financial cash flo-ws to othcr scctors of bers of the Songo Goda tribe, who would the econormy are involved. The imputed eco- benefit dircctly from the creation of livestock nomic benefits of cnviromnmntal improvement water points in the Mak'arrassou. This latter do not pay any bills. In this sensc, these eco- group has access to drinking water from wells nomic benefits are intangible. Accordingly. the recharged from the Day Forest catchment and borrowving governments arc not in a position to would not require wvater storage tanks. recover costs and might bc reluctant to borrow In addition, a large group of nontargeted funds for such projects- beneficiaries exist wvho would benefit from the Projects that are economically desirable project through the stabilization of vegetative but are not attractive in terms of government cover in the project area. This stabilization cash flow are often grant-financed. In situa- would reduce soil erosion and environmental tions where the EIRR is unattractive because damage due to flooding that would otherwise environmental benefits have not been valued, increase over time. It would also maintain and projects are also commonly grant-financed. enhance current infiltration rates to ensure However, in these cases it should be recognized availability of groundwater in surrounding that the presence of grant financing will not al- areas. The villages of Randa and Bankoucle, ter the EIRR. It will only alter the monetary rangeland in Dorra, and surrounding areas rate of retum to the government would thus benefit from the stabilization of vegetative cover. A rough estimate of the pop- Income Distribution E:ffects sulation that would benefit in this way is on the As can be seen from Table 6-15, income order of 10,000 persons. distibution reflects a modal average in the The total of targeted beneficianes and non- range of US$100-150I Within the project's targeted beneficiaries is therefore esimated at overall target group, no explicit measures have approximnately 14,000 people. Although irrele- been taken to target the poorest of the poor. vant to WFAD's mandate of rural poverty al- The nature of this project militates against such leviation, an unknown number of nonrural, an approach; targeting within the context of nonpoor would also benefit from the conserva- common property resource management is like- tion of the forest both vicariously and as users. ly to weaken already fragile traditional man- agement regimes. At the same time, as shown Income Weighting and te ETRR in the table, the entire group falls within the Conventional project appraisal as currently poorest category in the country and therefore practced does not entl valuing costs and meets IFAD's criteria for assistance. benefits differently according to the income of Distribution of Project Benefts the individuals affected. However, weighting procedures have been worked out in detail in The project's targeted beneficiaries can be the cost-benefit analysis literature, involving divided into four groups. The first group con- the use of "income or consumption weights" to sists of the entirety of the Sekoh Aboussa tribe, reflect the fact that the margnal utility of in- who would benefit from all of the project acti- come (or consumption) is not constant across vities The second group consists of five en- different income groups. In this way, equity canpments of the Songo Goda tribe, which considerations can be reflected in the EIRR, border ghe land controlled by the Sekoh Abous- and projects that benefit very low-income sa and; would benefit directly from the con- groups will appear relatively more attractive. struction of water tanks in their encampments 166 In projccts that arc intended to enhance or investigation was not considered essential for rehabilitatc the environment of the poorest project design. groups, mcasurcd bcnefits valucd using an im- Extending the boundaries of analysis re- plicit weight of unity can be quitc small be- quires grcater investments in information (see cause cnvironmcntal enhancemcnt translates below). This infonnation may not always be di- into relativelv low increases in production. This rectly rclevant to the technical design of the is especially the case in marginal areas, where project, and the costs of gathering this informa- the cconomic value of subsistence production is tion must be weighed against the benefits to be low. The economic benefits of avoiding envi- derived from it. In the case of the Day Forest ronmental degradation are thereforc equally Project, information on positive externalities low. However, the utility gains accruing to the benefiting nontarget groups is not relevant to beneficiarv population arc quite high. In the the project's poverty alleviation objectives but absence of income wecighting, the EIRR of is important for the economic justification of many envirorunentallv oriented projects in mar- the project. Howvever, it should be noted that in ginal areas is likely to be unattractive and, as a accordance with IFAD's Enviromnental Princi- result, such projects may be rejected on alloca- ples and Criteria, in cases where negative ex- tive efficiency grounds unless significant bene- ternalities might be generated, this must always fits to other sectors are involved (for example, influence project design. tourism). Another implication is that, in deter- mining tradeoffs between different groups of Valuation Methods users in mutually exclusive uses of the environ- Ecoromic valuation methods for environ- mental asset, it is the poor that are likely to mental changes such as the surrogate market lose out approach, the change in production approach, and the opportunity cost of time approach can Some Conclusions be readily applied in project analysis The ex- The Environmental Protection and Rural tent to which this can be done is subject to the Development in the Day Forest Project is an availability of tie necessary data, which are of- example of a growing number of IFAD inter- ten lacking in the context of marginal areas. In ventions that seek to achieve poverty allevi- many cases, the necessary identification and aton and environmental objectives in an physical quntification of environmental integrated manner. The case studv outlined change can be undertaken in the course of the some of the salient features of the economic project cycle. However, specific valuation analysis of this project. The following are cer- problems can be encountered when appraising tain generally applicable issues that might war- projects that focus on unmarketed environmen- rant attention in the analysis of such projects. tat goods and services in marginal areas where fisw markets exist even for surrogate goods. Extending the Boundary of Analysis in Valuation approaches such as CVM and Time and Space TCM require separate information-gathering The boundaries for project analysis should exercises over an extended penrod of time and be extended in time and space to the extent fea- do not lend themselves well to standard project sible. This would allow externalities to be iden- cycle work. Furthermore, the cost of carrying tified, quantified, and valued during the project out such exercises may not be justified, given cycle. In the context of the Day Forest Project, that decisionmakers are net familiar with these the time horizon applied in the analysis was techniques and may not accept the results. thirty years, beyond which "widt and without" While user benefits of environmental goods project projections would have had diminishing and services can be valued with varying de- credibility. The extension of the geographic grees of ease, nonuser benefits present a chal- boundary of analysis to include off-site effects lenge. The latter would require the application would have enhanced the economic analysis of of CVM and other survey-based methods not the project. However, because these externali- thus far applied in IFAD's work. ties were judged to be positive, their 6. The Riral Developmeent and Environmental Protection Project in the Day Forest in Djibouti 167 Fiscal Impact such as CVM and TCM, which require addi- While it may be true in many cases that tional survey data. The result is that the full "good ecology means good economics," the fis- range of enviromnental effects is not identified, cal impacts of projects that have a strong focus physically quantified, and valued during proj- on environmental objectives might be unfavor- ect preparation. At the appraisal stage, there is able. This is especially the case in siations litle time available to capture these effects in where economic benefits are imputed and no the analysis corresponding financial cash flow exists. In The economic analysis of the Day Forest marginal areas, these problems are amplified Project benefitted from the fict that extended. due to the absence of markets and mechanisms pre-investment activities were carried out prior for cost recovery. Governments may be reluc- to project design, generating a significant tant to accept loans for such projects, prefcr- amount of relevant data to supplement in- ring to seek grant financing, foffnation acquired during project cycle mis- sions. However, data on which to value Equity Concerns nonuser values of the forest, recreational bene- Projects that focus on enhancing the pro- fits of the forest, and downstrean externalities ductivity of the natural resources on which were not generated, because these were not di- subsistence farmers or pastoralists depend may rectly relevant to project design. As mentioned appear unattactive on strict economic criteria. earlier, a fundamental problem is that inomna- This is especially the case where economic val- tion required to design environmentally sustain- uation techniques translate enviromnental able poverty alleviation projects in maxrinal changes into changes in productivity- In mar- areas does not always coincide with infonna- ginal areas, where the value of production is lion required to justify the project on economic low, the imputed economic value of environ- gounds mental changes will be low. Incremental bene- A case can therefore be made for idenify- fts rnay be relatively small, but utilW gains mig data needs for economic valuation of envi- may be, high. romnental changes at the earliest stage of the One implication of this is that the economic project cycle. In this regard, the increased use rationale for projects that protect natural re- of focused proactive environmental assess- sources on which low income groups depend ments in the preparation of environmental proj- may, in certain cases, rest on nomnser benefits ects in lFAD's work might serve to improve and externalities. It is precisely these benefits, the quality of the economic analysis and, hence, however, that are more difficult to iden, improve decisionnaking Indeed, IFAD bas quantify, and value. embarked on a Preliminary Development and Ultimately, decisions to proceed with such Testing Phase of proactive environmentl projects may rest on non-economic criteria assessments, natural resource management su- such as intra- and inter-generational equity dies, environment-related preminvestment activi- concerns and other social objectives. ties, and operational guidelines for susuinable agriculture. Preliminary phase activities, in Data Requirements addition to supporting the design and appraisal The economic analysis of projects in rmar- of specific EFAD projects, will generate more ginal areas Is hampered by data constraints generally applicable lessons on the most cost- When projects are focused on enviromnental effective means of ensuring that data require- effects, these constraints are amplified because ments are identified and met in a timely fash- data required to specify environmental cause- ion. These activities will assist IFAD in the and-effect relationships are unlikely to be challenging task of designing interventions that readily available. In the context of the standard rehabilitate the natural resource base on which project cycle, many of these data constaints the poor depend, in an economically attractive cannot be overcome. E-.en where environmen- maimer. tat change can be predicted, further data con- straint fice economic valuation techniqacs Bibliography Bojo, L. KarlGoran MAler, and L. Unema. 1990. . 1991. Progress Report on IFAD's Environment and Development: An AEo- Evolving Approaches to Environmentally nomic Approach. The Netherlands: Sustainable Rural Alleviation. Kluwer Academic Publishers. (GC14/L.9/Rev. 1). Rome: IFAD, April. Dixon, LA., D.E. lames, and P.B. Sherman. 1989. . 1991. D4ibouti Rural Development and The Economics of Dryland Management Environmental Protection in the Day London: Earthscan Publications Ltd. Forest. Djibouti Appraisal ReporL Rome; Internaional Fund for Agricultal Development IFAD, lUne. (FAD). 1989. Environmental Sustain- _ _ . 1991. Natural Resource Management abifity and Rural Poverty Alleviation: for Rural Poverty Alleviation: Prelimi- Operational Issues for IFAD. (GC nary Development Testing Phase of Envi- 13/L.12). Rome: FAD, January. ronmental Assewnents, Pre -Investment .- 1990. Report on the IFAD Staff Train- Activities, Natural Resource Management ing Workshop on "The Economic Valua- Studies and Operational Guidelines for lion and Evaluation of Environmental Sustainable Agriculture. (EB Changes in Marginal Areas.- Rome: 911441E.80). Rome: IFAD, December. 1FAD. 7 Incorporating Environmental Costs into Power Development Planning: A Case Study of Sri Lanka Peter Meier and Mohan Munasinghe This paper applies economic valuation and multi-wttribute decision anaysis tech- niques to the incorporation of environmental considerations into power sector pkanning, us- ing Sri Lanka as a case study. Unlike most past efforts, the emphasis here is on the long-range planning stage of power system pansion, rather than assessment of project lev- el impact& Multi-attribute techniques are particularly useful in situatons where explicit eco- nomic valuation of externalities is subject to great uncertainty or controversy (as in the case of human health effects associated with thermal power generation). The paper illustrates the importance of evaluacting the environmental and economic impacts of technology, siting and pollution control options in a systems context, rather than on the basis of stand-alone com- parisons, and emphasizes the use of trade-off curves as a toolfor assisting decisionmakers. 170 Background are operationally practical. This is rcflected in the reccnt experience of the industrialized The environmecntal Impacts, of energy dc- .h .tei ex.rnCo h nutaie countncs as well. In the Unitcd States, despite vclopmnent arc widely recognized, and they have become an increasingly important topic of public debate, in both the developed and the de- latory commissions have yet to adopt formal procedures for including environmental exter- panoaroaermering conctnesInthelattern, ovr thi tei nalities into their procedures while most of proabaeerior tthose that have donc so have adopted some- relationship between economic policies for sus- wha crude expedients that seem littl related to tamable development and environmental costs; and donor agencies and the international fi noouot-eeftaalss cial ionstiationcs have becoe inremasiongl can- A second feature of the emerging literature cial instiun h- on the subject of how to incorporatc environ- cerned about implementing new policies that mental considerations into energy and power respond more effectively to environmnntal sector planning i developing countres is the The power sector has received otiul perception that somehow the analytical ap- attention, a consequence of the significant po- proaches that should be used are "different" tential impacts associated with the sector. In- tha those used in the developed countries, ,presurnablv because public sector investrnent deed, historically, many of the most celebrated instances of unanticipated environmental con- plays a much larger role in many developing countries, and, in the case of electric utilities, sequences m developmg countn'es have oc- sequencesmndeveloping countriesbecause of the substantial differences in the curred at major power system projects-such ownership and regulatory frameworks (see, for as the Akosombo Dam in Ghana and the example, Asian Development Bank [ADBI Aswan Dam in Egypt. Even though it is quite 1991) That, however, is a matter of the effec- unlikely that one would todav encounter unan- .veness of the difibrent policy instrments ticipated impacts on the scale encountered at available, and has nothing to do with the ana- these projects in the past, given the procedures now i plae, tere s genral greeent hat ytical methodology that enablcs one to quantb- fv the linkges betveen the configuration of the the most pressing need is to incorporate envi- p s e romncntal~~~~~~~ ~ cocm.noivetctpamn power system, economluc costs, and environ- mental quality.3 Whether, for example, inest- and decision making rather than simply react- . ing to enviromnental problems after they ments are made bv state-owned electric utilities occur] (as in developing countries) or by privately Exactlyu howrtis.is tobe achieed , owned utilities under regulatory supervision (as Exacdy how this IS to be achieved IS still unclear- First, there remains a huge gulf be- m the United States) does not alter the linkage, tween general discussions of the sub- say, between a requirement for flue gas ject-which typically argue for comprehensive frameworks and the like-and techniques that f The writers acknowledge the contributions to this study made by a team of Sn Lankan consultants: Pro- fessor K K. Y. W. Perera (Leader), H. S. Subasinghe, KI D. Arudpragasam, Sunith Fernando, Shavi Femando, J. Kotalawala, P. Illangovan, L.R. Sally, and T. Siyambalapitiya. For a good general discussion of the major issues, see Intcrnational Atomic Energy Agency (IAEA) 1991. Also see World Bankd 1989; or Giampaoli 1988. 2 For example, a U-S. Agency for [nternational Development (AID) report (OcL 1988) concludes 'In the near term, this calls for devising a strategy that encourages developing countries to assess the costs and benefits of environmental protection-as well as those of competing alternativesat the investment planning stage of energy sector development rather than after project investment decisions have been made." 3 For example, the ADB report argues (p. 4) that .. by contrast, a relatively stronger role is played by public investment in developing countries and this is a major factor determining the choice of approach. In fact, in the absence of efricient markcts and effective enforcement mechanissms for govemment regtulation and revcnue measures, investment decision- makdng may well be the most effective policy instrument developing country governments have to influence the nature 7. Incorporating Environmental Costs into Power Devlopment Planning: A Case Study of Sri Lanka 171 desulfiurization systems and ambicnt air quali- an environmental assessment (EA) and a bene- ty; it makes different only how a policy that fit or cost analysis at the project level can ad- would require such systems might be dress many of the most important implemented. environmental issues. The project-level EA In fact, the literature is quite limited, a re- deals well with local, site specific questions, flection of the very few attempts that have been and the options for project level mitigation made to formalize the process into operational (such as resettlement questions at dams, loss of models. The recent review of models and the wildlife habitat, local socioeconomic impacts literature by Markandya (1990), despite its associated with construction). However, the stated focus on power systems planning, found project-level EA deals less effectively with re- that most of the emphasis to date has been on gional (for example, dowvnstream water quality modifying much broader energy systems mod- impacts), national (for example, acid rain) and els to deal wvith environmental constraints or global scale (for example, greenhouse gas) is- objectives. For example, the MARKAL, an en- sues. These can really be addrcssed only at a ergy system optimization model developed a much broader scale than that of the single decade ago under the auspices of the Intema- project-namnely at the scale of the system tional Energy Agency (IEA), has been modified plan. At this level of the planning process, the to dea! with CO, emissions, and is being used fundamental question is not whether the impact in a number of deveioped countries to study of any single project is environmentally accept- tecimological options for greenhouse gas eints- able, but what are the relative environmental sions control. The only model that can be re- costs and benefits of alternative power sector garded as a power systems model for which development strategies. any attempt has been made to incorporate envi- Indeed, there are widespread perceptions ronmental considerations, and which has been that project-level EAs are nothing more than cx used extensively in developing countries, is the post justifications of decisions already made.5 Wein Automated System Planning (WASP) In Sri Lanka, the first full-scale EA for a pow- model for which a so-called IMPACTS module er project was prepared for the proposed Trin- was developed a few years ago by Argonne comalee coal fired power plant (Blaclk and National Laboratory.4 However, this module Veatch International [BVII 1988) and many of calculates only the quantity of pollutants the public comments refcted just this emitted (or "residuals'), together wi-th cost esti- perception: mates for any pollution control devices: it does ... it is therefore evident that the en- not deal with impacts per se. tire feasibility study process, teininat- Given this methodological vacuum, it is ing with the ELA [environmental hardly surprising that most of the attention to impact assessmernt], has followed a date has been given to improving enviromnen- policy known as "decide-announce- tal assessment procedures at the project level. defen&d ... The EIS [enviromnental ixn- Thus the World Bank, for example, now has pact statement] is an ex post facto extensive new procedures in place to ensure in- document seeling to environmentally dlusion of environmental concerns in the proj- justify a project which was formulated ect cycle. Nevertheless, even if done well, there without any environmental consider- are inherent limitations to the degree to which ations whatsoever... [t]his EIA is a and location of economic activity and consequent enviroimental impacts." 4 Houwever, Markandya notes that "... in practice, this model hi. not proved easy to apply, or particularly useftil. None of the World Bank staff involved in the power sector used it" (Makandya 1990: p.20). A new version of the IMPACTS module that is part of the broader energy planning package called ENPEP (for ENe-rgv and Powver Evaluation Package) has since become available, and is equipped with a modem user interface that makles it much easier to use than pre- vious versions. For a broad review of the environmental assessment procedures in Asia see, for example, A L. Brown, R A. Hind- marsh, and G.T. Macdonald 1991. Reporting on a workshop on ETA, the authors noted "... a perceived futility of some environmental assessments in terms of their ability to modifi the Project. ...JMost constnUctior projects starts before the EIA work begins. [T]he EIA work should begin at Lhe planning stage of the project" 172 post planning excrcise, whereas it facilitics proposed for financing are in the ought to have becn an ongoing excrcise 'least cost" plan. In practice this has meant the from the inccption of the project. application of sophisticated capacity expansion Although it is pcrhaps casy to dismiss such optimization models- such as the WASP conmnents as mere rhetoric, in fact we know of model that is very widely used in developing no EIS prepared for a poxvcr project that has countries-whosc objectivc function is the cver concludcd that the project should not go minimization of the prcsent value of systcm cx- fonvard. It is probably fair to assert that the pansion costs ovcr somc planning horizon and twenty vears of cxpericnce with ElSs has for some given level of system reliability. shown that while they have succeeded in forc- Yct, in the complex regulatory environment ing projcct devclopers to bccomc more sensi- of the industrialized countries, fbw utilities use tive to cnvironmental issues, and in cnsuring the results of mathcmatical optimization mod- that mitigation options are fiilly intcgrated into els to develop their capacity expansion plans. the project design, thev have provcn to be in- The most common approach is to develop a adequate as a tool for cnsuring that environ- small number of discrcte options, with revenue mental considerations are incorporated into the requirements and environmental impacts planning decision process itself.6 derived in some detaii as part of the necessary In rcsponse to thcsc issucs, the Environ- regulatory submissions. Moreover, 'least cost" mental Policy and Research Division of the planning in countries such as the United Staes World Bank initiated a case study of the power has come to encompass a much broader per- sector of Sri Lanka, onc of a scries of research spective than the traditional stress on minimiz- efforts that seek to explore alternative ap- ing the costs of supply expansion for some proaches to quantifying and valuing environ- exogenously specified level of demand, since mental impacts in different sectors The the demand itself is subject to modification by express objective of this effort is to develop so-called demand side management ap- methodologies that can be implanented in proaches.7 Indeed, as noted earlier, over te practical ways, and whose potential impact on past few years a number of U.S. state regulato- decision rules is clearly demonstrable. This pa- ry commissions have expanded the operational per summarizes some of the key findings of the definition of 'least cost" planning even further case study: the reader is referred to the detailed by introducing specific techniques for including report for a morc completc prcsentation of the costs of environmental externalities. methodological issues and research findings The problems of an excessive reliance on (Meier and Munasinghe 1992). supply-side optimization models as a basis for pianning in the developing countries has be- come broadly recognized of late, not least with- In the past, the principal planning objective in the World Bank itself (see, for example, for the development of the power sector has Crousillat 1989). Perhaps the most important been to deliver the anticipated need for electri- objection to the use of the traditional nmdels cal energy at "least cost." Indeed, kirgely at such as WASP is the difficulty with which they the insistence of the World Bank and other in- deal with the huge uncertainties currety faced ternational financial institutions, developing by the sector, ranging from uncertainties in the countries have had to demonstrate that prices of imported fucls to the impacts of ' That the EIS is an inefficient vehicle for evaluating anything more than the very local site-specific impacts is well illus- trated by the American experience: because there was little if any public and environmental input to fhndainental ques- tions on the gencration mix, E[Ss prepared for proposed nuclear plants proved to bi the only focus for such input. resulting in long delays as fundamental questions on the generation mix, on the role of energy consenration, and on nu- clear safety were endlessly re-examined. ' In its modem usage in the lJnited States, least-cost planning has been defined a process that .. explicitly includes con- senration and load management programs as capacity and energy resources; it considers environmental and social fac- tors as well as direct economic costs, it involves public participation; and it carefully analyzes the uncertaRinties and risk-s posed by differemt resninrce portfolios and by external factors" (C. Goldman et al. 1989). The Electric Power Re- search Institutc (EPRI) prefers the use of the term '"integrated resource planning" to capture the idea of a balanced con- sideration of demand and supply side options. 7. Incorporating Environmentud Costs into Power Development Planning': A Cave Study of Sri Lanka 173 construction cost overruns and droughts. In Planning Process in Sri Lanka other words, the so- callcd "Icast cost" plan This general characterization of the plan- may in fact be optimal only for the particular ning process described above applies almost set of external circumstances assumed, and perfectly to Sri Lanka. The annual planning may deviate from optimality for most other cycle begins with load forecasts, which are combinations of external factors. used in tuM by the Gcneration Planning Branch In the traditional approach to power sector as a basis for their Generation Plan, for which planning, environmental considerations have the WASP-III model is currently used. Hydro gencrally been deferred to at least the siting scheduling is done with the SYSIM model, stage of the process. In other words, decisions which was developed as part of the Electricity about the generation mix, and the basic Masterplan sponsored by the German assis- technology choices, have been made entirely on tance agency GTZ. (For a description of this the grounds of direct costs, in the case of hydro model, sec Nanthakumar 1990.) projects, however, casily quantified mitigation Both the Masterplan system expansion >tu- and opportmity costs-such as the cost of re- dies and the present CEB (Ceylon Electricity settlement, or the opportunity costs of foregone Board) Generation plan have in the past re- production from inundated land-are now rou- stricted their consideration of environmental tinely incorporated into economic analysis. issues to a few calculations of residuals: yet the Attempts at quantifying environmental fac- usefulness of such calculations and their tors at the siting stage-efforts that are quite gmphical portrayals could be improved and comunon in the United States and growing in may have very little to do with the actual envi- developing countries-have tended to be based ronmental impacts that might be expected For on some very ambiguous ranking and weighing example, in Figure 7-1, the simple horizontal schemes, with an ubiquitous reliance on 'stacking" of SO2 , NOx, and ash emissions 'judgmnent and experience of siting experts.- k provides a visual implication that tons of these is quite unclear whether such approaches have very distinct pollutants could be arffimetically in fiat improved the quality of siting decisions added together to yield some form of aggregate Quantification at this level of analysis has impact. generally meant the application of mathemat- The npproach followed in the early 1980s ical models of the fate of pollutants in the envi- with respect to the coal plants followed very ronment-such as atmospheric dispersion much the traditional pattem: enviromnental is- models for air pollutants, and thermal plume sues were realy only addressed in any detail, models for predicting ambient temperature - as part of the project-level EA, once the basic creases from cooling water discharges. The siting decision-for a coal-burning baseload thrust of such modelling studies is to demon- station at Trincomalee-had already been strate compliance with ambient standards: if made. The subsequent decision to examine a these are not met, mitigation options are ex- site on the South Coast, followed by another plored to bring the facility into compliance. An political decision to drop further consideration analysis of actual impacts-say in the case of of that site-illustrates the largely ad hoc na- air quality an estimation of the health risks as- ture of how environmental considerations were sociated with pollution exposure-is rarely being addressed. More recently, the Kukule hy- part of such assessments. droelectric project has come to the forefront of In general, not much detail is presented on public debate and controversyr, again the dan- the impacts at alternative sites, whatever may ger is of public perception increasing that the bc the regulatory requirement that aLternatives feasibility study that is currently underway will to the proposed action need to be ealenifd. inevitably lead to the plant's construction. More often than not, in an EIS of several The need for a more systenatic and early hundrad pages, there is but a section of a few approach to analyzing the environmental as- pages of discussion about alternative sites, pects of power system alternatives in Sri Lanka most of which is in the nature of undocumented is quite clear. The ad hoc procedures df the (and almost always unquantified) assertion. past, in which major projects are examined one Figure 7-1: Master Plan Portrayal of Eminsions Fasterplan for the Electricity Supply ot Si Lanka - Client..ot a A: C0 - Congultinq EnqLna.rsu Lahmuymr EXPECTED EHISSION FROM THERMAL PLANTS SEXSI LISTI SCENARIOS SUBAR029 Type of eLusuion Emission from thermal plants (kton) :Year -------------- -- -.--- --------- -------------------------------------- t02 1 NO. t Ash I 50 too 150 :…… - - -- ------ ------------------ .............- : 1991 3 I S o1 0 n 1992 3t 5° I I 1993 5 9 0 snnn I I ,1994 t 12 , Oa Is nnnn I 1995 11, 17 ' °ss. nnnnn I ,1996 IS, 24 a 1ssssW nnnnnnnn I 1997 19 29 1 0 ssoJ nnnnnnnnn I 199, 19 23 1 14 Isss"s" nnnnnnnnn csea I 1999 19 11 26 !Ssaui nnnnn aaaaaaaaa: 2000 24 t 26 t 26 tsssuu. nnnnn aaaaaaaaaa: 2001 29 33 27 I"'ss'ss s nnnnnnnnnnn saaaaassesa s 2002 30 1 IS 41 Issusuissim nnnnnnnnn eeaaaa4aaGasaaaa 2003 1 37 ' 39 1 41 Issusussu ussu nnnnnnnnnnnn aa.eaaaaauagaaaaa 2004 37 t 33 55 issshsssh ssuuhh nnnnnnnnn aaa aaaaaaaaaa: 2005 45 1 45 55 ussususussuuuus.nannnnnnnnnannnnnn aaaaaaaaamaaaaaaaaaae 2006 ' 45 I 38 69 isssssssssussSsg nnnnnnnnnnnnn *a.a.uaeaaaaaaaaoaaaaaaa 2007 ' 41 ' 36 I *2 lsssssssisissasio nnnnnnnnnnnnI aaaaaaaauaaaaaaaaaaaaaeea 2008 1 57 1 50 32 lsullssulosssssssu nnn nnnnnnnn nnnnnnIaaaaaaaaaaaaaaaaaaaaaa,a z 2009 55 1 36 1 107 Ih hs.u.uuhu.uah.u nnnnnnnnnnnn aaaamaaauaaaaaaaecaeaaaaaaaaaai 2010 64 49 109 ssSumuuguuu.umuauuuuam.u nnnnnnnnnnnnnnn a aaaaaaaaaaaasaaaaaaaaaaaaamas Total 573 555 734 s type of emission 102 n Type of smLision NOm a type of emLsuion Ash 7. Incorporoaing Enviro,nmental Costs into Power Devdopmena Planning: A Case Study of Sn Lanka 175 at a time, and rejected onc at a timc, will create * Archeology: a preliminary assessment dis- substantial problems in the fiiturc unless a covcred some unique iron smelting sites more proactive approach is taken. dating to the sixth to ninth centuries AD; ODA recommended a complete excavation Environmental Issues of Power of the site prior to inundation. Sector Development in Sri Lanka Because hydroelectricity dominates the Long-term issue present power systmn, it is logical that the prin- Need for an environmental management cipal environmental issue that the power sector plan (for both upstream and downstream has had to dealw ith in the past has been that of areas). resettlement. In recent years, the few thermal As the system gradually shifts towards a plants in the system (the diesels at Sapugaskan- greater role for thermal plants, the major issues da, and the combustion turbines and steam will also shift towards those associated with plant at Kelanitissa) have been only sporadical- such plants. Clearly, the issues related to ther- ly operated, and neither ficility is seen to have mal plants are of a quite different nature from major environmental problems. (However, dur- those at hydro plants, and how to compare the ing times of drought [as most recently in April two types of quite different impacts-primarily 1992], these thermal units may run twenty-four air and thermal discharge issues at thermal hours a day, with the remaining limited hydro plants, against land use issues at hydro resources used for peaking purposes.) plants-is one of the main thernes of this case The Samanalawewa hydroelectric project, study. currendy in the final stages of const,Uction, il- The commnents received in response to the lustrates well the kind of environmental issues Trincomnalee EIS is a ,good indicator of what associated with hydro development in Sri Lan- these new concerns are likely to be. As required ka. An earlier Overseas Development Adminis- by regulations, the document was circulated for tration (ODA) (1990) evaluation of ft public comment, and on Figure 7-2 we illus- Victoia project highlighted a number of social tate the results of an analysis of the 154 com- and environmental issues, and scveral ODA ments received. That air quality would be the missions were undertaken during construction major concem at a coal-fired plant is to be ex- of the Samanalawewa project to ensure imple- pected: perl'aps less expected is that the second mentation of mitigation measures. The issues most fiequent concern was over the process by addressed in their 1990 report were as follows: which a coal-fired power plant had been se- lected for this particular site. As noted earlier, Construction phase issues widespread dissatisfaction was expressed by * Water table drawdown during power nongovemental orgaizaions (NGOs) and nel construction. others over the way m which decisions had been reached. * Removal of biomass from the area to be- come inundated. Land Use and Resettlement * Need to ensure adequate irrigation water Sri Lanka is one of the more densely pop- supply immediately downstream of the res- ulated countries of the world, and land avail- ervoir for a 2.000-acre rice paddy irrza- ability to support a growing population is an tion scheme (and the possible need for important issue. In general, hydro sites are iP compensatory water supply). the wet zone areas where there is little avail- c Resetolemnt isy .s: (a) need to extend able land nearby for resettled inhabitants to be attention from frniliezs with propert to re-located; and such land as is available at landtiss laborers, tenants, and traders greater distances (mainly in the newly irrigated whose customers move f,om the arna; (b) dry zonc areas of the Mahaweli schemnes) is of- need to eseblish more fromal counseling ten seen by potential evacuees as undesirable semce adtangeiments. because of questions concerning the availabil- ity of adequate waer supply. Figure 7-2:- Public Comments to the Tfrincomalee EIS 50 45 35 30 E0 2 0 00 E 15 4 10 --V 5~~~~~~~~~~~1 0 ~~~Process Water Groundwater F-isherie -TSecurity Air quslity Larld use Tourism Ecology Wildlife All others 7. Icrporating Emironmental Costs into Power Development Planning: A Case Study of Sni Lanka 177 Or- the basis of the rccnt experience at volves the resettlement of only some 27 Samanalawewa, wherc some new ideas have families. been tried by the CEE, the rcsettlement issues at most of the remaining major hydro sites ap- Deforestation pear to bc tractablc. At Samanalawewa itself, The progressive loss of Sri Lanka's natural as of fall 1992, the families from the power forest over the past fifty years is well docu- station site have already becn moved and re- mented, and represents one of the country's settlcd in a tca estate. and some 365 families most important environmental concerns. Power remain to be rcsettled Of these, 68 have agreed sector projects will likely be scrutinized vcry to move to the Mahaveli area, 80 are still ne- carefully for their potcntial impact on what gotiating, and 120 have agreed to find on their natural forest areas remain, even if it is true own land for resettlement in rcturn for an addi- that the power sector per se has been a relative- tional cash payment This has the merit of giv- ly nino. contributor to the loss of forest lands. ing evacuees the opportunity to resettle in areas The main causes of deforestation in the past of their choice, and reducing the pressure on have been planncd agricultural development governnent organizations to meet the demands and settlement schemes, chena cultivation, en- of evacuees. croachment by unplanned settlement and crop- The one major hydro project where sub- ping, illicit logging, and uncontrolled fuelwood stantial resettlement issues are likely to occur and timber extraction. in the future is at Kukule. Although there are a Air Quality number of variants of this project currently un- der study, the GTZ Masterplan variant of the Relatively little is known about ambient air Kukule Ganga Hydropower Project, which at quality in Sri Lanka, except perhaps on a very 2,080 hectarcs has the largest reservoir size of general level: in most parts of the country the any of the remaining hydro projects, estimates air quality is fairly good, a reflection of the that some 9100 persons would need to be limited extent of industrialization (except in evacuated. This represents four times the num- Colombo) and the natural ventilation provided ber of potential evacuees as at Samanalawewa, by strong monsoonal wind regimes. In Colom- and will therefore require very carful bo, however, the sharp increase in automobile management. and bus traffic over the past decade has led to On the other hand, it should be noted that strong indications of increasing deterioration of this Kukule project variant will also provide urban air quality. Between 1970 and 1990 the substantial irrigation benefits, calling for some number of vehicles registered annually in Col- 68,000 hectares to be irrigated in the southeast- onbo has increased fourteen-fold. It is esti- em dry zone, of which some 95 percent would mated that vehicles produced 98 percent of be new land. The net increase in agricultural carbon monoxide emission, 79 percent of NOx, production would substantially offset the pro- and 46 percent of SO.. With 60 percent of the duction losses in the area affected by reservoir country's vehicles registered in the Colombo inundation, and would provide substantial em- area, and with about 60 percent of the total in- ployment in an area of high youth dustky, the outlook for air quality in Colombo, unemployment. while still nowhere close to the levels experi- A compreheasive feasibility study of the enced in other Southeast Asian cities, is still Kukule project is currently underway. In the poor in the absence of appropriate actions. Progress Report of January 6, 1992, prelimi- The first major effort to monitor air quality nary analysis suggests that the most a irbevegan in 1989 when the National Building Re- option may be a run-of-river project, that in- search Organization (NBRO) initiated a three- year program for monitoring air quality in Col- ombo; in a first phase, a citywide survey of Again, however, it is useful to make a comparison to the scale of projects elsewhere where the resettlement impacts have become matters of major national (and intemational) debate. The two Narnada projects in India (Sardar Sarovar and Nannada Sagar), for example, would require resettlement of 170,000 people, but provide over 2 million hectares of irrigated land (see, for example, Dixon, Talbot, and Le Moigne 1989). 178 dustfall and sulfiation rates is being conducted, emission trends of acid rain precursors (SO2 with detailed measurements of SO2, particu- and NOx) in India as in Sri Lanka itself. As lates, and NOx to follow in phase two for those suggested by the predominant wind directions areas with unsatisfictory air quality. This pro- in South Asia, the most significant source of gram is viewed as a necessay first step not just acid rain precursors to Sri Lanka would be to develop a data base for Colombo, but also to coal-buming power plants on the East Coast of assist in the formulation of national air quality India. standards, and to establish a permanent moni- In light of these observations, it is clear toring capability. that the debate over the potential acid rain con- Based upon what we do know about pat- sequences of the Trincomalee coal-fired plant, terms of energy utilization, certain inferences as evidenced by the discussion in and com- can be drawn. It is fairly certain that at present mrents to the EIS, was not very well founded. the power sector contributes only marginally to What is most relevant is the matter not of SO2 air pollution in Sri Lanka. The existing thermal emissions in North Arnerica (which was used plants in and around Colombo prcsently need by the consultant as the yardstick of compari- not be rin at all during a normal hydro year. son), but of emissions in India- Once the anticipated coal burning power plants ft is probably true that from a local acid- are added to the systen beginning in the late rain perspective, the best location for a large 1990s, however, this is expected to change sig- coal-burning power plant would be the south- nificantly, from essentially zero in 1991, the /southeast coast, since the plumes from this power sector is expected to contribute 80 per- location would be overland for a much lesser cent of SO2 and 70 percent of NOx emissions distance than at either the Colombo or Trinco- by 2010. Clearly the potential for significant malee areas. However, different intnsities of degradation of air quality is the major environ- the two monsoons that would tranz.port the mental concern associated with coal-burming plumes (about 12 kilometers per hour during power plants, and the options to minilimize such the northeast monsoon, and 19 kilometers per air pollution impacts-whether by technology hour in the southwest) make even such simple choices (say by application of some of the new- comparisons quite uncertain. er clean coal technologies such as fluidized bed combustion [FBC]), by impact mitigation op- Global Waning tions (say by the use of flue gas desulfurization Global warming (and tansnationa acid system [FGD]), or by locational choices-is rain) are conceptually different from local envi- one of e main themes of this paper. ronmental impacts, since iu the former case the Another major uncertainty is the air quality impacts will occur predominandy in other impact of the use of traditional fuels, particu- countries. If the main economic objective is to larly in the congested lower-income areas of maximize welfare in Sri Lanka, decision mak- Colombo where fuelwood and agricultmual ers in Sri Lanka would be unwilling (quite jus- wastes are still widely used as a cooldng fuel. tifiedly) to incur additional costs (for example, The release of carcinogens may be a particular- to reduce CO2 emissions) if the benefits of such ly important concern. actions aceme mainly to other nations. The as- sumption here is that Sri Lanka will be reim- Acid Rain bursed by the international commity fior the Acid rain is likely to become an increasing- incremental costs of global warming mitigation ly important environmental issue in the Asia- cflbrts, or that the Government would have Pacific region given the fact that the energy signed an international agreement committng plans in many countries, in India and China in itself to undertake certain CO. emission reduc- particular, call for rapid development of fossil tion measures. In either event, power sector energy systems. Acid rain is largely a long- planners would need to consider the impact of range phenomenon, and it is fairly obvious that altenative expansion strategies on greenhouse the extent to which acid rain is or will be expe- gas emissions. rienced in Sri Lanka is as much a fuinction of 7. incorporating Environmental Coss into Power Development Planning-A Case Study of Sri Lanka 179 Carbon emissions in Sri Lanka, both in ab- erosion and adversely affect stability in the hill solute terms and in per capita terms, are ex- country watersheds where deforesution rates tremely low, a reflection of the dominance of already represent a threat to the sedimentation hydro in the electric sector, and of low energy rates experienced by hydro and irrigation reser- intensity of the industrial sector. Beyond the voirs. Unlike many other countries in Southeast year 2010, however, CO, emissions will rise Asia, the tectonic conditions are relatively sta- very sharply as the electric sector generation ble, with little significant seismic activity of the mix moves toward fossil fuel. This has impor- type that has produced significant surfice de- tant consequences for the negotiating posture pressions of coastal areas in dte Philippines. of Sri Lanka in the event that a consensus Nor arc the major cities presently threatened by emnerges for tradable CO, emission rights; it major subsidence problems caused by exces- will be in the interests of small developing sive exploitation of groundwater? Neverthe- countries whose present electic systems are less, extensive areas of the coast, especially in hydro dominated to argue that the initial alloca- the south and southwest, are already threatened tion of emission rights be based on population, by coastal erosion (that can, at least in part, be or income, ra-her than on present fossil fuel attibuted to large-scale mmiing of coral reefs), consumption. and there are extensive areas of highly pop- Indeed, Sri Lanka may be an example of a ulated coastal areas that would be severely af- country that may be as much affected by the fected by sea level rises of one to three meters. global measures to reduce CO2 emissions, as In the north, one of the immediate conse- by the physical impacts of any global wamiing quences of sea level rise is likely to be contam- that may in fact occur. Some of the recent s- ination of the limestone aquifers that are dies of measures that might be necessary to important sources of groundwater for tht area. stabilize CO, emissions have staggermg in- Whatever may be the resolution of these plications for developed and developing coun- uncertainties, it seems clear that power sector tries aiice. At the June 1991 meeting of the planners, atthe very least, will need to be much Energy Modelling Forum, whose focus was the more aware of this debate tan in the past, and energy sector impacts of greenhouse gas emis- should be cognizant of the implications of al- sion control strategies, a number of models ternative expansion strategies on CO, emis- suggested that the level of carbon taxes neces- sions, especially in relation to any future global sary to stabilize emissions (typically defined as agreements on emissions and er financial 1990 emission levels by 2005) should begin and technical ramifications. We therefore in- now at over $1 00 per ton of carbon, rising, in clude CO, emissions as one of the enviromnen- some cases, to as much as $550 per ton by tal factors to be considered in this study. 2005. Such tax levels are pehaps not very likely Biodiversity to materialize. Yet much more modest rates of Because Sri Lanka is a small island, which carbon tax, even if imposed only in the indus- has been isolated for relatively long periods, it trialized coubnies, will have substantial in- is the home to a large number of endemic plications for relative fuel prices, and the ratio (unique) species. Among Asian countries, Sri of coal to oil prices in particular has relevance Lank has the highest levels of biological di- for fossil fuel importing countries such as Sri versity. Indeed, the U.S. National Science Lanka. Foundation's Committee on Research Priorities On the other hand, the impacts of global in Tropical Biology identifies Sri Lanka as de- wamung that might be experienced by Sri Lan- manding special attention. Biological diversity ka are also quite speculative. Sone researchers is under trat in Sri Lamka primarily from the expect an intensification of the monsoon in progressive reduction in its natural forests and tropical latitudes, which may increase soil ' The classic example is Shangtai, which subsided sone 2.5 meters between 1920 and 1965. An artificial rcharge pro- gran was, howevec, ccssful in arrsting further subsidence Parts of Bangkok are now subsiding several centimetess per year, and many other coastal cities in Soutlhest Asia face potentially serious subsidence problems. 180 other ecosystems and the selective exploitation sandy beaches and coastal estuaries and la- of species (parficularly for timber). goons, with over 80 percent of hotel rooms lo- The main biodiversity issue for the power cated on the coast. Also, the marine fishery sector is the degree to which hydro projects and indulstry provides employment to some 100,000 transmission line corridors affiect forest lands persons, and is the largest source of animal of high biological diversity, and to which ther- protein for the island. The main concern relatd mal generation projects on the coast affect the to the power sector is the discharge of heated high diversity coastal systems (particularly cor- effluents into shallow waters that contain valu- al reefs and seagrass beds). able ecosystms. It is not merely the loss of acreage that is The range of imacts will depend on the of potential concem. Since most ecosystems re- exitece of a thennocline and the depth of the quire minimum contiguous areas for stable ex- well-mixed layer. Two surveys by the Norwe- istence, fragmentation of habitats will be of gian oocanographic survey vessel Nansen give significance for long-term survival. Many of a broad idea of the temperature structure of the high biodiversity wet zone forests are al- shelf in Sri Lanka. The thermocline lies at ready highly fragmented-for example in the depths of 50 meters, differing slightly on the Matara district, the 160 square kilometers of west coast with season, and showmg greater remainig wet zone forest is fragmented into 30 variation on the east coast The well-mixed lay- patches, the largest of which is no more than er extends down to about 50 meters. Approach- 1,000 hectares. ing the shore, therefore, coastal waters will be Detailed site specific stxdies need to be well mixed. conducted at the project level EIS stage to ex- In Trincomalee Harbor, where depths well amin wheher or not a proposed project will in in excess of 50 meters exist within the harbor fact affet an endemic species. At the planning area, an effluent led ito the surface water will stage, however, information at this level of de- not reach the bottom, and major communities tail is not likely to be available. The best that will rmnain unaffected. However, in shallower can reasonably be done, therefore, is to develop coastal waters, because the weli-nmixed layer a quantitative index ta:t predicts the probabili- reaches the bottom, impacts of the thennal ef- ty that a particular prcject, or combination of fluent are likely to be quite drastic!'0 projects in a sectoral development plan, wOill have an eflet on endeic species or des Outline of the Proposed habitats ofhigh biodiversity Methodology Coastal Zone Based on the experience with the Sri Lanka case study, one can outline the following seven- As the generaton mix shifts from one that is predminantly hydro, to one in which large step appioach to pnningposta g enviromrsyst baseload fossil-fueled plants play an increasing issues mto tfe planig te of power systcns role, it is inevitable that sites wiEl need to be developmet in developmg countries: the key fiund on the coast to accommodate such dthr- stps are depicted on Figure 7-3. wal plants. Since imported v al is the most Step 1. Selection of Attributes likely fuel for those plants, the implication is Perhaps the most important criterion in de- that any inland location would mcur substantial ciding which environmental issues to address is additional transportation cost penalties. to be clcar about the scale of analysis that is The economic inportance of prevnting en- require at the systan plaing level. Many vironmnenal degradation in the coastal zone is impacts that are purely local in nature can well established (Energy and Science Authoritr safely be deferred to the project-level EIS stage 1991) Foreign tourism, an important source of without running any significant risk that these foreign excbange, is largely focused on its would ultimately stand in the way of At the Turkey Point plant in the state of Florida (United States), where water temperatures are comparable to those in Sn Lanka, thermal discharges went diectly int Biscayne Bay, which has a depth of about I meter and is semi- enclosed. Thermal effuent killed or damaged 118 hectares of benthic organisms in the first two years of opration. Figure 7-3: General Methodology for Incorporating Environmental Costs >s into Power Development Planning n C Environmental Impact (First streen) mitigatlo options /quantifcaton\ _ y_e3 oss ible ' ~~~~~~~~~~~~~~~~Conventional S. economic analysls e cost~~~~~~~~~~~~~~~ Cnvironme attribute Implied v tuailons a mitigation als strategy 11 impact valuaton 182 implementing a strategy identified at the plan- thout given to what is really important."' For ning stage. Thus, in the case of Trincomalee, example, in the scoring system used in testimo- while it may well be true that a number of very ny in a Vermont (United States) proceeding vocal interest groups raised objections based (which attempted to deal with the question of on very narrow issues (whale watching, oyster the relative environmental impacts of hydro im- beds), the issue that in fact prevailed was the ports from Canada, demand-side management view held by the broader environmental com- measures, and decentralized renevables), 36 munity that discharge of effluents into shallow different attributes wvere used (Mintzer et al. waters constituted a broad risk to aquatic eco- 1990). These included separate attributes for systms. The trade-off to be examined in sys- each of the following material uses: steel, con- t=m planning studies, therefore, concerns the crete, aluminum, silicon, glass, plastics, and cost of various cooling system options to limit non-ferrous metals.'2 Exactly what environ- thennal discharges, not the presence or absence mental concerns the consumption of these ma- of specific species at specific locations. The terials implies that is not already reflected by latter is, however, the proper concem of the other attributes is unclear. project-level EIS. A prolifration of attributes tends to make Judgments must still be made: and there weighing more difficult (and may introduce a may be some impacts that at some projects bias simply because one is reluctant to weight may turn out to be purely locaL but at others any partcular attribute as near zero). More- prove to be of national significance. Resettle- over, trade-offs become difficult to understand ment at hydro plants is a good example: at and display to decision makers in comprehensi- some projects, the resetement inpacts may be ble fonn if there are too many of them. The quite small, and would neither strain the ability guiding prmciple ought to be that one starts of local authorities to manage, nor be likely to from what are the most unportant inpact is- become a political issue at a national scale. But sues (greenhouse gas emissions, health effects at other sites, the scale of resettlement needs of fossil fuel pollution, the risk of groundwater may be such that indeed it may require the contamination, the risk posed to aquatic eco- attention of national decision makers. Certainly systens from thermal plumes, and so on), and if the high dam variant of the Kukule project is then select one attribute for each of these built, the need to resetle some 9,100 people de- concerns.'3 fines an issue of national importance; on the One of the issues at this stage is how to other hand, the run-of-river variant would dis- draw the boundaries of analysis. In this case place only L35 people in 27 fhmilies. study we have bounded the problem in such a Many applications of multi-attribute meth- way that we have examined only that set of en- ods suffer from the same malady that afflict en- vironmental issues that are likelv to be of direct vironmenaal impact statements: in their desire importance to national decision makers. Thus to be comprehensive, they become tedious reci- we have not included the environmental im- tations of all possible impacts, with not much pacts associated with the extraction of "T This was a major problem in the United States in the early years of the Environmental Protection Agency (EPA). In 1977, in response to EISs of increasingly encyclopedic proportions, President Carter directed the Coumcil on Envirn- mental Quality, the entity established a monitor EPA, to issLe new regulations .. designed to make the EIS morc use- ful to decision mak-ers and the public and to reduce paperwork and the accumulation of extaneous background data in order to emphasize the need to focus on real environmental issues and alteratives" (Executive Order, May 24, 1977). This led to the development of the so-called scoping process as a means to direct effort to real problems. 12 This study also proposed an interetg approach to the detemnination of weights, based on the reversibility of the i- pact as the most important tactor high level radioactive waste firm a nuclear power plant would be assigned a high weight because it remains radioactive for thousands of years while themmal pollution of a river during low flow periods would be assigned a low weight What is critical to validity, however, is whether or not the weights actually reflect the trade-offs that decision makers would be prepared to make in practice. " All these points have long been established in the decision sciences literature, and were validated in a large number of applications in the 1970s to power plant siting, particularly by Keeney and his colleagues (see, for exaxnple, Keeney and Raiffa [1976], and Keeney and Nair [1 977T). The 1981 EPRI report by Woodward-Clyde (Keeny et al. 1981) is one illustration of a well-founded application of decision taeoly to utility resource planning (a coa/nuclear choice). 7. Incorporating Enironmen tal Costs into Power Devdopment Planning: A Case Study of Sri Lanka 113 imported coal (that would arise in Australia or willing to incur the incremental health costs of South Africa), except in the sensitivity analysis this option (associated with exposure to fossil to coal prices: if in fact Australian policy mak- plant air pollutants) in order to avoid the health ers do decide to include the cost of environmen- impacts of EHV transmission. tal externalities associated with extraction (say The measure used for quantification is es- by imposition of rigid land reclamion require- pecially important. For example, the attribute ments on the coal mining companies), then fiat "solid waste disposal" expressed as tons of would simply be reflected in a higher coal price waste generated, is ofte encountered. Yet, pre- to Sri Lanka. cisely why the quantity of solid waste is rele- On the other hand, we have included the vant is unclear, except that it is easy to question of greenhouse gas emissions; this is estinate. Since a large part of the land required because Sri Lanka decision makers will need to may be for solid waste disposal, there may in deal with the issue as part of the intemational any event be double counting if there is also a discussions that are now underway, and any in- criterion for "land use" (expressed in acres). tenational agreements that are reached will Moreover, in fact the relevant enviromental surely have repercussions in Sri Lanka itself risk associated with ash and scrubber sludge Finally, attributes (and criteria) should be disposal is the risk of toxic leachates contan- conceptually distinct; that is, preference inde- inmnng nearby aquifers, which has very little pendentl Decision analysis differentiates be- relationship to the quantity of waste produced, tween statistical independence and preferental but rather to characteristics such as soil perne- independence; the fonner refers to the correla- ability, depth to the groundwater table, and don strdctumr of the alternatives, the latter to rainfl structure of the user's preferences-a distinc- tion that might be characterized as "cts" ver- tep 3. Preliminaor Screening/Option sus "values? Additive value fimctions-sucrh Definton as those used in weighing and summafion me- A great deal of care must be taken in any thods-assume certain type of value indepen- preliminary screening exercise. Some exclu- dence, but statistical independence is irrelevant sionary scres can be applied with great confi- to the validity of the additive fonm dence: no facilities in designated wildlife Ssancuaries, no nuclear plants in areas of Step 2. Physical Quantification known geological faults, and so on. However, To some extent; of course, the physical many other screens routinely applied by engi- quantification is not unrelated to the selection neers in a siting study involve judgmnents that of attributes: clearly there is not much point in may reflect only some personal viewpoint (even selecting criteria that cannot be quantified, if supported by personal experience)-such as That of course does not imply that only quanti- a presumption that a coal-based plant must be fiable criteria are relevant or important it "near" its coal receimng port What does this means only that such unquantfiable criteria really mean? Obviously there are costs asso- need to be considered through some other cated with intennediate truck or conveyer belt means. transport, but these have to be put in the con- As an example, the health impacts asso- text of other costs associated with particular ciated with exposure to very high voltage trans- sites. mission lines may be viewed as significant Indeed, one ofthe most important consider- There is presently no basis for quantifying that ations at this stage is to properly define the uni- health effect. Therefore, the way to include an verse of altematives to be studied. A priori assessment of the issue is to define a technolo- judgment that ertain technology and rniti- gy scenario that elimiates the need for EHV gation options need not be considered further transmission (for example, by moving the base- are particularly to be avoided. This is again load coal plants from the remote areas m the clearly indicated by the historv of the Tnco- south and Trincomalee to the major load center malee site selection process. In the early stages, in Colombo), and then asking whether one is the advantages of the deepwater harbor were 184 decmed so compelling that south coast sites one is better off not to attempt them, and sim- were not firther considered. Then, at the sub- ply use the physical impact quantification in a sequent prefcasibility stage, the only technolo- subsequent multi-attribute analysis of trade-off gy options considered for Trincomalee Bay curvas. In this study, thcreforc, we have were conventional, once-through cooled plants avoided use of dosc-response functions in the with dischargc into the shallow waters of the multi-attribute analysis, but instead- used the bay. Once the environmental community came population exposure to specific pollutants as to the conclusion that the rcsulting thermal im- the proxy for health damages. In any event, pacts were unacceptable, attention then re- since the main purpose here is to provide a ba- verted back to the south coast. Yet it is entirely sis for comparison of practical options, rather unclear that the additional costs of an alterna- tan to establish some absolute measure of im- tive cooling system at Trincomalee, which pact, the population exposure measure still would mitigate the impacts of the thennal ef- provides a great deal of relevant information. fluents, are greater than the additional costs of Step 5. Multi-attribute Analysis coal transportation at a south coast site.S In the particular case of Sri Lanka, sceuri- Multi-attribute decision analysis has been ty considerations, and a temporary reduction in developed expressly for situations where deci- the rate-of-demand growth, made it relatively sions must be made upon more than one objec- easy to suspend discussion of first the coal tive that cannot be reduced to a single plant at Trincomalec, and then of its replace- dimension. Its central focus is the quantifica- ment at Mawella. Nevertheless, as a matter of tion, display, and resolution of trade-offs that methodological procedure, the important point must be made when objectives conflict. In the is that all reasonable technological configura- case of application to the power sector, there tions should be considered for all reasonable may well be strategies that have beneficial im- sites at the stage of the initial analysis of envi- pacts on environmental and on economic objec- ronmental trade-offs. tives-most energy efficiency investments that are economically justifiable also bring about a Step 4. Economic Valuation reduction in emissions and hicrC improve mvi- Clearly the most difficult valuation issue ronmental quality as well as economic efficien- concerns the value of humnan life, and the eco- cy. In most other cases, however, there are nomic cost of illness. Indeed, one of the major economic penalties associated with investments findings of a recent comprehensive literature designed to reduce enissions or improve envi- review is that a substantial fraction of the envi- ronmental quality. It could be argued that if ronrental damages associated with particulate, economic valuation of the environmental costs SO_, and NOx emissions is related to health and benefits were in fact possible, these might impacts and visibility impairment (Pace Uni- well outweigh the short run- costs. However, versity Center for Environmental Legal Studies from a practical standpoint, expressed in terms 1990). American valuations of these impacts likely to be understood by laymen decision indeed suggest damages that are higher than the makers, the trade-off is indeed one of short-run costs of pollution control now imposed by economic costs (or the present value of readily American standards. Whether that is true for quantifiable costs) against longer-termn benefits developing countries is an interesting and im- to environmental quality. portant question, and indicates the need for Multi-attribute decision analysis, even if some country-specific contingent valuation not fonnally described as such, has been used studies to establish the willingness to pay to for some time for power plant siting and avoid illness. transnission corndor routing,'4 including in Sri In general, we are of the view that where Lanka itself. More recently, attempts have been valuatons involve a high degree of uncertity, made to use such techniques for broader 4 It is hard to evaluate whether cases in which formal models were used for siting decisions were any better than in cases in which they were not It is certainly true that the most celebrated siting controversies in the United States, froml the Storm King Mountain pumped storage project to the Shorehm and Seabrook nmclear plants would not have been re- solved sooner, or at less cost to the economy, had formal siting models been applied. But ti- c siting decisions were in 7. Incorporating Environmensal CosJ into PowerDevelopment Planning:A Case .tudy of Sri Lanka 185 studies of energy-environmental interactions the many uncertainties that arise, and that may (for example, ADB 1991). Finally, the tech- be completely independent of what economic nique has come into increasing use by state values are attributed to environmental extema- regulatory commissions in the United States in lities, or what relative weights are placed on an attempt to develop practical approaches to particular impacts. Of course it is unlikely that considering environmental objectives in the options can be found that are compietely-in power sector. other words, that they are best across all crite- ria. Neverthieless, as dcmonstratcd in our Case Step 6. Imriplicit Valuation study, ccrtain policy and iechnology options Implicit valuations can be most useful to -such as demand side management measures guide resource allocation decisions. For exam- and ecrtain ciean coal technologies-do appear ple, -we estimated the valuation of a human life to be desirabie from both environmental and implied by a hypothetical decision to install an economic perspectives over a wide range of FGD system at Trincomalee. The particular conditions. value of $1.5 nillion is important not as an es- Finally, it is almost certainly true that timate of human life per se, but what it implies whatever the shortcomings of the sort of analy- about health expenditures of the Goverunent sis conducted in this study, and whatever the as a whole. Clearly it makes no sense to spend level of uncertainty that is associated with both $1.5 million to avoid one death at a coal-fired physical quantification and economnic valuation power plant by installation of an FiD systern, of enviromnental externalities, existing prac- if the expenditure of $100,000 on improved tices are so flawed that the application of al- medical diagnostic equipment at a hospital most any rigorous analytical approach might avoid 10 deaths. Implied valuations can represents a significant step forward. also be derived from the trade-off curves themselves. Evaluating Options Step 7. Decision Rule The key step in the multi-attribute analysis is the identification of the best set of candidate Tlhere are a number of reasons to avoid the options, from the large number of candidate use of formal decision rules at this stage of en- plans, that merit close study for vironmental debate in Sri Lanka. For example, implementation. we might have ranked the policy options X- A first criterion is that of dominance. Let died usmg a weighmg summation ruie to denrve the set of such plans be denoted P,, P2, and so an overall merit score for each stategy. How- forth. Suppose for the sake of clarity that there ever, that would inevitably force attention on are only two attributes- cost, and an eniron- whether or not our weights were correct, or mentll attribute reflecting the population expo- whether we had elicited weights from the ap- sure to SQ2. Figure 7-4 depicts the solution propriate group of decision makers. Moreover, space for this problem, in which we plot the a number of methodological problems arise im values of the two attributes for each plan. the application of procedures to elicit valid The plan PI is said to strictly dominate weights, uihich can be descrbed as still being plan P2 if P1 is better than (or equal to)-that subjects of research. is, dominates-P, in tenns of every attribute. We believe that for those counties where Thus, as illustrated on Figure 7-4. P1 is better the process of setting clear enviromnental pno- ta P, in both cost and SO, exposure. PI rties are still at a very early stage, the first strictly dominates all of the plans beyond the step is to first try to focus on the nature of the boundary AP1B. By repeating such compari- trade-offs themselves, and to attenpt to identi- sons for all pairs of plans, and discarding all fy policy options that are robust with respect to donminated plans, the remaining plans constitute fact madc in the 1960s, and the use of fonnal siting models was significantly improved in the 1970s as a direct conse- quence of the problems encountered at these (and other) plants. The most important lesson from the application of for- mal siting models in the United States is the need for public participation in the decision proces: it was very quickly realized that as decision models forced ecplicit quantification of values and preferences, this could no longer be done by the self-appointed "siting cxper'- of the utilities and their consultants. Figure 7-4:The Two-Attribute Problem A Worst P12P *P11 Environmental objective P1i P .~~~ P8 plo Pi,4 P16 / * P1 c P3 P7 P1 *~~~~~~~0P5 Trade-off curve ---. P4 .---.... P6 Best Best Worst Cost objectve 7. Incorporating Environmental Costs into Power Development Planning: A Case Study of Sri Lanka 187 the set of the so-called non-inferior solutions significantly bettcr than P,, it must have a -in our case the set of plans (P6. P, P, P14, value less than Pt - e. P13, P1,5. These points, in turn, define the The knee set is the set of plans that are not trade-off curve, as indicated on Figure 7-4. significantly dominated. In the example of Fig- This procedure will in general provide a ure 7-5, redravn on Figure 7-6, the knee set is means for reducing a very large number of pos- {P1, P4, Pg4. Note that this set differs from the sible plans to some smaller number of plans-a non-inferior set, defined by the trade-off curve, short list'7 or "candidate list" that is to be which is the set {P2, Pi, P,). From the practical presented to decision makers. In the example standpoint of decision making, the klnee set is shown, using the concept of strict dominance, clearly the more useful, since the non-inferior plan P3 would not appear on the resulting short set also includes solutions where slight im- list. Yet one might argue that while P3 iS some- provements in one objective are bought only at what worse in both attributes than Pt, it is not great cost in the other-such as P5 and P2. significantly worse (as opposed to, say, Pk6 and Another way of expressing this is that the knee P., which are significantly worse in both attrib- set consists of that set of plans for which deci- utes). In particular, because of uncertainties as- sion makers representing different objectives sociated with the calculations, one may be are most likely to agree upon: in the above ex- reluctant to discard a plan that is not signifi- ample, both P. and P, are unlikely candidates cantly worse than another from a final short for agreement, even though both lie on the list trade-off curve. This idea is captured by the concept of sig- The knee set wvill be a function both of the nificant dominance (as defined by Crousillat values of the significance parameters used, and and Merrill 1992). P, is said to significantly the shape of the trade-off curve. In the situation d1ominate P2 if a;(P1) + m < a;(P) for at least of Figure 7-6, the knee set is relatively con- onei,andifa(P)-e . . .~~~~~~~~~~~~~~I Worst p P5 U P8 Non-Inferior set lP7 A1....... ......z _ Non-Infenor set criterion Al I P (Environment) ml P3 .. ....................... ;<-- el Knee set '-----uP2 * ~, -L Best e2: m2 Worst A2 SKneeset criterion Best A2-Cost) I Figure 7-7: Knee Set Equivalent to the Non-inferior (Trade-off Curve) Set Worst U PC U PB II u~~~~~~~~~~~~P? I I~ ~ ~~~~0P (Environment) Fml .. ... .. . . . . . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . . .I.. . . . .. . . .~~~~~~~~~~~~~~~~P Best~~~~~~~~~~~~~~~~~~~ Best e2~~~~~ ~ ~~ m ors I I p~~A(Cst Figure 7-8: Quadrant Criterion Worst I Quadrant IV Quadrant P2 I A1~~~~~~~~~ zI P3 Al P3 (Environment) I o P1 (=basecase) | .................................... ...... .................................................... P7 P5 P6 Quadrant III P4 * Quadrant 11 P8 Best Best Worst A2 (Cost) 192 It is in quadrants I and IV where trade-offs of low, base case, and lugh forecasm of the must be made-where an improvenent in one type that we needed to examine the robustness objective necessarily implies worsening in the of our findings to altemative load growth other. assumptions. Extension to Multiple Dimensions The System Load Curve All of the mathematical definitions of knee An analysis of the impact of demand side sets and non-inferior sets continue to be valid nagement and nondispatchable technologies when there are more than two dimensions- In requires much more detailed assumptions about this casc, one should properly refer to the the natare of the load carve thar. what is pro- trade-off surface, say S = S(C, A,, A2, -. A,), vided by the simple aggregate charactrization which is an n-dimensional noninferior hyper- as the system load factor (SLF). We, therefoe surface, where C is the cost attribute, and Ai used a linearized, trapezoidal representation of (i=I,..n) are n environmental attributes, the annual load duration curve of the form A rigorous definition of the teade-off be- shown on Figure 7-9. This curve was then ma- tween anv tw.o of these attrnbutes would require nipulted algebraically to depict the impact of that the values of all of the others are held con- measures such as reductions in transmission stant In practice that is rarely possible-for and distrnbution (T&D) loss rates," demand- example, finding a series of plans that all had side nagmnt measures, or, as depicted on the same water impact, but varied only in cost Figure 7-9, wind energjl~-which, as a nondis- and air quality impact would be hard to find. patchable technology, is modelled as negative What can be done is to simply plot the values demand. Because early field tests of wind ener- of two attributes (for example, C versus A,), gy availability on the south coast show little and to assume that the vanations in the other impact on the peak, the principal impact is on attributes may be ignored. In oter words, in intexndiate demand, as indicated. the three-dimensional case, the third dimension is in effect projecied onto the two-dinensional Fuel Prices plane of the two attributes being displayed, ig- The question of fuie pnce projections are norng variations in the third dimension)' obviously of aciical importance, beause they It might be noted, however, that while are a critical fictor in detrmining the genera- looking at attrbutes only two at a tine pro- tion nix But t is not just the absolute level of vides great insigbht, it may not in fact reveal all prices that maters. For Sri Lanka, a critical of the members of the non-minferior set This is variable is the ratio of coal to heavy flieloil discussed in some detail in the main report price, becau that ratio will detcrmine the bal- (Meier and Munasinghe 1992). ance of diesel and coal plants in the generation mix, in turn of cental importance to te envi- Assumptions for the Case Study ro impacts that follow. Again, as a Although cast as a research study, we nev- base case we use the Septenber 1991 projec- ertheless deemed it to be importnt that the re- tions of the CEB. These in tun are based on sults were properly aligned to the work of the the December 1990 projections of the World CEB. Thus, for example, rather than derive our Bank own demand forecasts, we elected to use offi- Historically, the ratio of fuieloil to coal cial forecasts of the CEB. which include a set price has shown great flucuations, ranging is very close the CEB's estimate of 5.61 US. cents per kilowatt-hobrn 16 In die literatur, there are a number of eamples of such de-off curves deaing with ecnomic-envionmental tade- offs in the power sector The first example that we encounteed is Femell (1978), who exmined the trade-off betw cost and SO, emissions. Odr examples include Meier andRuff (1979), and Amagai and Leung (1991), who looked at CO emissons tade-ofE;- I7 t is important to recogie that T&D losses are propotional to the square of the load. This means tha .IT&D losses are specified in the conventional way as some percentage of total eergy geneated, this value cannot sitnply be applied across the entire load duration cuve in a unforn manner In other words, if amumal energy losses axe reduced, say, from 13 to 12 peret, the benefit i the peak load will be much greater than this average value, and the benefit to the ott-peak period will be smaller. Figure 7-9:Impact of Wind Energy on the Load Duration Curve Lp FE L ~ ~ ~ ~ ~ ~~~~Wn enrg generat3ed LI- Load LI* hp 8760-hp = hb Hours per year ed' 194 from 3 in late 1984, to less fian in early A reasonable conclusion would be that in 1986 and late 1987: indeed, this was a time the export markets in the Indian Ocea cific when many dual-fueled plants, which had bem rim area, sulfur content is not as important a converted from oil to coal in the late 1970s, consideration as it is in Europe or in the East- switched to residual oil. cm part of the U.S.; both 0.6 percent or I per- From an environmental policy viewpoint, cent sulfiur coals would be regarded as "low" the relationship between fuel price and sulfur sulfur, that would not generally require FGD content is of particular interest. The relation- systems. High sulfur coal (3 percent and ship between low sulfur (05 percent sulfur by -above), for which a -subsuntial discount would weight) and high sulfur (3.5 percent) fueloil be expected, is simply not available as an ex- traded on the Singapore spot market is fairly port fuel in this market constant in tie three-year period 1989-1991, the average premium for the low sulfur fueloil Hydro Assumptions was about 27 percent, equal to about $4 per As a result of a close re-emination of the barrel. historical hydrological records, the assump- On Figure 7-10 we graph the low sulfur tions for hydro energy availability have bee fueloil price as a function of dte high sulfar revised downwards by the CEB. The result is fucloil price. There are two periods, one in that in the 1991 Genration planing study, to- spring of 1990, the second in October and No- tal bydro eneWr of eistig plants has been re- vember 1990, when the price of low sulfur duced from 4,070 gigawatt-hours to 3,869 fueloil is unusually high, and fills outside the gigawatt-hours, with corresponding reductions general tendline, these deviations are seen to in projected energy from new projects as welL be of short duration and can be ignored for The changes in hydro assumptions, and the long-term projection purposes. difrences m the rltiv fuel prices between The relationship between price and sulfur residual oil (the fuel used in diesels) and coal, content of intenationally traded coal is less have a dramatic impact on the optima genra- dlear. A regression analysis of recent prices of tion mix as detennind by the WASP mo- coals traded in the Pacific rim market' pro- del-as indicated on Figure 7-11. This shows duced the following relationship:1' capacity additions built in the periods throuh 2000 and 2005. As one might epect, giv P= -0342+0.003686H 277sulfurparnt-0744ashpecent; lower hydro energy, &uwer hydro plants are R2=.558 (0.0013)(9.4iX0.45L) buit in the 1991 study. But the most dramatic dffece is the mix of diesels and coal plants; where H is the heat value in Brtish thral in the 1991 study, by 2005 some 900 mega- units (Bta) per pound and P is in dollars per watts of coal fired plants would be needed, as ton. The figures in parentheses indicate the opposed to only 300 megawatts in the 1990 standard errors. The heat value and ash terms study. The differences in the en entl im- have the expected sign, but the sulfur content pacts that follow fiom tbese two very difirt term has an unexpected sign (one would expect generation mixes are obvious. the relationship to show that lower sufur con- tents have higher price, as with ash). The range The Analytical Model of variation in sulfur contmt is too small for a Although a review of the iteature reveals staistically significant relationship to be a great many difiret approaches to the prob- establisheL lem of idcenfying the non-infirior set, three methods account for the bulk of practical 1' We excluded, for example, internationaly raded cals from the east coat of the US. and Poland, smce thse would be unlikely sources for Sri Lanka 9Thme and resoc constraits did not permit a comprehensi7e analysis of likly coal costs, cAiL Sri Lnka, as a fioc- tion of all of the parameters of importance IJI addition to sulfir, ash, and heat contents, other significant characeristics include moisture content, volatile matter contents, HEngove grindty Index, and assseimnts of corrosion, fouling, and slaging potential, among othecs These were all considered in the Black and Veath Intational (BVI) Trnonma- lee Fuel Supply and Trnsportation Study (March 1985), althoug no statistical analysis was presented. Figure 7-10: Relationship between Low and High-Sulfur Fuel Oil 36 } 34 _ 32 30 Oct90 28 26 ,~24 -Fb9 22-Ja 20 17 P(.5%8)=0.40 + 1.1 P(3,50e) 1 8 1 6 1 4 1 2 10 19 1 3 1 5 1 7 1 9 21 23 3.5% 8 fueloli Figure 7-11: Generation Mix Additions In the 1990 and 1991 Studies 1.2 1.2 ~~~~~~~~~~~~~~~1991 Study' 1.1 _ / s\\\\\ Gas turbrnes 1 I tl m lti m Hydyro 0.9 Dls 0.8 X 0.7 - 0.6- 0.5 0.4 ~1990 Study 1991 StudyGstubnsCi 0.3 -Hydro I 0.2- 0.1co 0 Projected mix, 2000 Projected mix, 2005 7. Incorpratin Environmental Cost into Power Devlopment Planning:A Case Study of Sri Lanka 197 applications. The first two, the so-called con- abatement options, and technology variations: straint method, and the weighing mecthod, are the usc of a complex cngineering model such as derived from the theory and practice of multi- WASP or EGEAS (for Electric Generation objective mathematical programming; the third Analysis System) is simply unsuited for the is simply to generate the solution space of dis- sort of sensitivity analysis required here. cretely defined plan altermatives in a simulation Ir. any event, a second problem with the model, and determine the trade-off curve and existing models such as WASP concems their knee set by inspection scope: while they provide great detail on the Evcn though a number of existing opti- supply side, they have very limited capability mization models appear suited to either the of examining demand side issues beyond sim- constraint or the weighing method, we rejected ply using alternative exogenously specfied the mathematical programming approaches in load forecasts. IAEA, which supports the favor of direct simulation.' The most impor- WASP model, does provide a demand side tant reason is that we are intercsted in not just model for use with WASP called MAED: but the non-inferior set, but alSD those solutions this model is not currenly istalled in Sri Lan- that are near this curve, and in particular the ka. Moreovcr, even if it were available, the plans in the knee sets Solutions that are close MAED-WASP combination is still very to the optimum are not nornally generated by cumbersome, and ill suited to the requiremnts the comnonly available LP software packages. of our case study. Of course, all of the algorithms will pass Several other eisting models were re- through near-optimal solutions as they make viewed as possible candidates for our case their way through the solution space, but to Cx- study, including GEMIS/TEMIS, developea in amine them individually is not generally Germany;,2' ENPEP/MACTS, developed at possible. Argonne National Laboratory,= EGEASI The direct simulation model has a number COMPASS, developed by EPRI?3 and LEAP, of adv ges for pactical policy making that developed for the Beijer nstitute of Sweden?' offset the problem of not being sure that all of All of the models contain some parts of the points in the non-iferior set have in fiLct what we needed reviews of these models are been looked at Besides dte ability to identify provided in the annex to chapter 10 of the main knee sets, one may also be interested in identi- report None contained all of the necessary fea- iing very poor programs and policy packges, tur. Most imporntly, none had the direct namely those that never occur in any of the capability of multi-attribute analysis inte- knee sets. The main requirement, Therefore, is grated into a credible characteization of the for a model that can examine very quickly per- electic sector, and suitable for extensive semi- Iaps several hundred differnt expansion plan tivity analysis. And because all involved the variants, given the very large number of combi- use of compiled progran enviromnents nations of different site locations, pollution (PASCAL, FORTRAN, C, and so on), One LP model in particula that hs multi-obective capabiliy is MARKAL Developed as a mainfiame model fen years ago for the Intenational Energy Agency (WEA) to examine long-tam azerny R&D strategies, it has recently been moved to a nicrocomputer environment and has seen new use in the United States to examine the mplications of CO2 emissions reduction strategie&The model has also been used i some developmg countnes, notably kIdonesia. Whatev- er the merits of the new user-intrface, we viewed this model as being simply too unwieldy for application to oar study. 2 GEM[S: for Gesamt-Emissions-Modelln tgrierter Systems, see OKO Institutc (1989). An English version of the mod- el, known as TEMIS (for Total Emission Model for Integrated Systems) was developed in 1990 with supr from the U.S. Departmat of Energy. Thc modd promvides a total fuel cycle analysis of residuals-in which residuals are calcu- lated in a cmnistent way from the point of fuel exrAction to the point of use. See Buchring et al. (1991). The WASP model is widely used in developing countries for capacity expansion optimiza- tion-indeed it has been used for this purpose by the generation plannig branch of Sri Lanka's CEB for some yeas Argone National Laboratory has developed a micocomputer vsion of WASP that has been integrated into a broadw elergy planning pakaged called ENPEP (for ENergy and Power Evaluation Package). This includes a modle called IMPACTS, which provides the capability for cakulating pollution residuals and pollution control costs. D' For a brief description see, for example, Stone & Webster Mnagement Consultants, lnc (1989). For a more detailed description, see EGEAS Capabilities Manual, April 1989, also available from Stonc and Webster 24 SeeRasin (19S5). Figure 7-12: Major Features of ENVIROPLAN Computer Model Demand projecton D8M Measures Wnd plants T & D Losses Power plant database Load curve generation Capacity expansion Environmental Fuel costs and attributes charaotcristios _ _emissTon u . ~~~oalculatons Merit order dispatch .~~~~~~~~rd- M curve % Incorporatig Enirnmsenral Costs info Power Dedeop enr Panning: A CaseSudy of Sri La 199 modification would have involved considerable reserve margin (PRM)" is what is used as the programlming effort that was beyond fte re- expansion criterion by the "adjusted " algo- sources of our study. rithm: the next plant is added if the reserve In the end, we decided to use ENVIRO- margin would otherwise drop below this mini- PLAN, a LOTUS 1-2-3 model expressly de- mum level. signed to analyze the environmental impacts of However, the actual expansion path is it- the power sector in a multi-attribute analysis.' eratively adjusted in such a way as also to meet With the entire set of calculations in a single a maximurnm unserved energy requirement, tak- spreadsheet, including the generation and dis- en here as I percent of toal annual energy. play of trade-off curves and.knee sets, sensitiv- That is, if after merit order dispatch, the un- ity analyses that permit the examination of served energy exceeds this minimum iL any wide ranges of assumptions and policy and year, the target margin is increased incremen- program -alternatives are relatively easy to con- tally for that year, and the expansion calla- duct And as a spreadsheet, modifications and tions are repeated. enhancernents could be easily added as the ENVIROPLAN makes merit order dis- study progressed. patch calculations for each of the tree hydro The main fTatures of ENVIROPLAN mod- conditions used by the CEB for WASP simu- el itself are indicated on Figure 7-12. The pro- lations, and the `average" dispatch, which is cess starts with the demand projection, which is used to calculate fuel consumption and pollut- aligned to the CEB forcasts as described in ant emissions from thermal plants, is hen the the previous section, and also as noted above, weighted sum of the three individual dispatch adjusted to reflect the impact of demand-side conditions. This will provide a much better management (DSM), wind energy, T&D aignme of ENVIROPLAN to WASP than lsses, and so forth. the much simpler procedure of dispatching into Generation capacity expansion is based on "average" hydro conditions. the anticipated peak load plus some reserve The merit order itself is calculated in the nmari ENVIROPLAN has several available usual way by incremental marginal cost EN- generation expansion algorithms: the one used VIROPLAN may yield a marginally lower esti- in the Sri Lanka case study is based on a heu- mate of fuel consumption for thermal plants nstic such that the model builds plants in the because fiel consumption is calculaed at tihe same sequence as WASP, but adjusts the tim- full load heat rate; we do not distinguish be- ing in such a way as to maintain an exogenous- tween heat rate at minium load and the incre- ly specified planning reserve margin. This is a menal heat rate as does WASP. much less sophisticated approach than the pro- babilistic dynamic programming of WASP that Technology and Policy Scenarios provides an optimization subject to loss of load Six major policy scenarios were examined. probabilties (LOLP); however, epricnce in- Each scenario is translated into a set of specific dicates that this provides a reasonable approx- technological assumptions ncesary tD quanti- imation if the changes to the load curve are fy the impact of these policies in practice, as small. Ihus, as we perturb the original load shown on Table 7-1. Et should be noted that the curve, say tD provide for enery conservation, specific measures examined are purely illustra- ENVIROPLAN will defer the construction se- tive; for example, the replacement of incandes- quence in light of the lower peakdemands. cent lighting by high efficiency fluoresent These ooncepts are illusmated& on Figure lights is only one of a potentially qite -large 7-13. The "target 1iiargin" is the smoohffed number of DSM measures that might be ex- (5-period) reserve margin as yielded by the amined. The same observaton applies to the WASP solution. The "minimium planwg wind power example as well, since 7 ENVIROPLAN uscs version 2.4 of LOTUS 1-2-3. In contrast to version 2.2, spreadsheets of essmtiay unlimited size can be hmadled, because cell pointeas are stared in exaded mcmory rather than in conventional memory. In version 22, the size of spreadsheets was limited by the 640 kilobytes of conventional nimory, in version 2A. the size is lim- hitad only by the amount of expanded memory available. ENVIROPLAN runs comfortably on 386 macines with I me- gabyte of conventional and 3 megabyts of expalded memory. Figure 7-13: Reserve M4argin Definitions 0.8 0.7 0.6 0.5 Reserve margin 0,4 0.3 Actual margin 0.2 _ Target margin Minimum PRM 0.1 ' ' I I I l I I I I 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 7. Incorporaing Environvental Costs into Power Devdoepmen! PlannngiA Case Sndy of Sri Lanka 201 photovoltaics and direct solar might be also ex- than 0.8 percent might be used; the emission amined as renewable energy technologies. As a level is at A, the corresponding cost at K. research study, the intent here is primarily to Efficiency improvements have the general demonstrate how the methodology can treat dif- effect of shifting the non-inferior curve toward ferent types of technologies and policies, rather the origin-all other things equal, higher effi- than reach defintive conclusions. Finally it ciency means some combination of lower costs should be noted that although we examine these and lower emissions. Suppose, for emple, policies individually, they are by no means that the heat rate of the plant is improved; the mutually exclusive:. a discussion of strategies non-in&rior curve shifts firom FF to GG. All that combine some of these options together is other things equal, less fuiel is required per presented later. kilowatt-hour produced, so fuel consumption, and S emissions, decase; and the net cost to the utilW goes down (fuel costs less the One of the most important ways of reduc- amorbzed costs of the investment that pro- ing the environmental impacts of power sector duced the heat rate improvement). But now the development is to ensure that the techndical effi- utility has an incenfive to move to Y: by buying ciency of the systen is at its economically effi- cheaper fuel, with higher sulfur contents, fur- cient value. For example, a T&D systen that ther savings can be achieved without violating has only 2 percent losses might certinly be the standard-with the result that there is no tecnically feasible, but would entail the use of environmental beefi, but only lowe costs- technology (such as superconducting tansmis- From the envionenal perspective, the de- sion) that is prohibitively expensive. On the sired response would be to W: implying that other hand, loss rates of 20 percent and more, the savings from better heat rates be applied to commonplace in some developing countries, the purchase of still lower sulfur coal, say 0.7 are clearly also ecxonmically inefficicnt, and sulfur by weigh investments to reduce losses by rehabilitation In Sri Lanka, where there is no large inmven- of the T&D system are firequently justifiable on tory of old, inefficient thermal plants (as exists, economic grounds alone. But if teconical losses say in India), the most important system effi- are reduced, then to produce the same level of ciency issue is T&D losses. Subsantial benefits to consumers, less fossil fuel will be progress in loss reduction has already been reqired to produce these same benefits, and achieved, but the question might be posed polution emissions wiEl decrease (per unit of whetier the present 12 percent loss reduction benefit to society).' arget of the CEB ought to be firther reduced, However, whether or not efficiency im- and what environmental benefits migbt result provements do, in fiat, lead to environmenal benefits is illustrated in Figure 7-14. Consider Demand-side Management a coal burning power plant at X, which selects A isymatic assssmet of demand-side a fuel with a sulfur content that is such that it management (DSM) options has yet to be con- exacdy mees the ambient standard; for exam- ducted in Sri Lanka indeed, comprehensive ple, in order to meet the maxmmm 24-hour DSM assessments have been conducted for standard, coal with a sulfur content no higher only a very few developing countries to date.27 Nevertheless, fere are indications that there 26 It should be noted that this reaoning applies only to the reduction of technical losses: The enviromental consequences of reduction of nontechnical losses through imnprovemet of collecton produre elimiation of pill:rage, and sD forth, will depcnd upon assumptions made about the resulting financial impact on the utility, and on the price elasticity of dc- mand of those consmers who would now pay the goig tari£ If, for example, improved revenue collection eliminttes the need for a tariff increase in order to meet given balmace sheet ratios, then overall consumption mould tend to in- crease (given nomlal price elastic behavior). On the otter hand, with the same a Dumptions about price elastic behav- ior, consumers who were forced to pay for previously pilfered consuption woild p mbly reduce heir consumption. On balance. one suspects that these effects offset each other, mith the rsult that it is primarly the reduc- tion oftecnimcal, rather tha nontechial losses, that has a direct environmen benefit 2 Such assements have recently been conducted for India (as part of a recent World BankbU.S.AD pow sector re- view) ard Costa Rica (see U.S.AID 1991). 202 Table 7-1: Policy Options Policy emphasis Jifustrative measure System efficiency T&D loss reduction Demand-side management Fluorescent lighting Renewable energy technology Wind power Pollution control policy FCD systems Clean coal technology Pressurized fluidized bed combustion (PFBC) Fuel switching Low-sulfur coal imports Low-sulfur fueloil imports exist some significant opportunities for the in- Wimd e activities in Sri Lanka were troduction of enegy efficient end-use technolo- started on a very modest scale as far back as gies n Sn Lanka. Both the World Bank power 1978. Early activities were tecused on the de- system efficiency study of Sri Lanka (see velopment of small wind pumps mainly for ap- UNDP/World Bank 1983) and te GTZ Mas- plication in the agricultural sector for terplan made some preliminary estimates of the small-scale irrigation during the dry season. At potential for load management and energy con- least one small-scale workshop is now engaged scrvation by the systematic replacement of in- in the commercial-scale production of these candescent lights by fluorescent lights. This machines. measure is, therefore, used in tius study as the Alter a preliminary assessment of Sri Lan- illustative example; the analytical framework ka's wind resources, using past meteorological is capable of analyzing other potential DSM data, the southern coastal region was identified measures in a similar way. as a potential region for harnessing wind ener- gy for large-scale applications such as electric- Renewable Energy Technology ity generation. On te basis of tiat study, a Although there are a number of renewable more detailed study of wind energy resources energy technologies -hat may be considered for in the southem region was initiated in early Sri Lanka, the temhnology wnth the largest po- 1988, funded by the govnment of the Nether- tential impact on the operations of the CEB is lands. The first interim report of this study will wid energy. Utlity-scale solar thermal plants be available shortly. appear to have litle potential, while the use of Based on the analysis completed so far, the photovoltaics probably has greatest potential in southem coastal belt, extending from Hamban- remote rural areas that do not have grid access. tota towards Palatupana on the soudteast, has The extent to which it might actually be possi- been identified as the region that offers highest ble to displace one or more coal fired units with potential for wind power generation. Specific wind plants in Sri Lanka, and at what cost, is electrical outputs of about 800 to 900 kdlowatt- of obvious enviromnenl interest. Figure 7-14: Tbe Impact of Efficiency Improvements F G .| W~~~~~~~~mprove=\ 802 emissions b 80os 2 tnar A B G I ~C1ost 204 hours per square meter and plant factors of fired combined cycle plants, as examined in the about 20 percent have been estimated for this CEB's 1991 generation planning study, have region. The wind power potential within the been found to be uneconomic, this is largely a five-kilometer wide stretch of coastal land be- consequence of high oil costs. But if coal can tween Hambantcota and Palatupana has been be used to drive a combined cycle plant, the in- estimated as 200 megawmatts. The estimate is hernmtly higher efficiency of such units is a ma- based on a power density of 8 megawatts per jor advantage where coal must be imported, square kilometer and a 50 percent land utiliza- and where freight costs may account for a tird tion fictor after excluding forest cover, agricul- ofthe delivered price. tural land, homesteads, and the Bundala Bird For the purpose of exmining the environ- Sanctuary?8 Specifically, our scenario for wind mental impacts of this teclmology, we assume energy developmcat assumes a first the use of 150 megawatts of PFBC units for 5-mecgawatt demoration plant, foflo%xed by the Trincomalee coal plant The assumption three 50-megawatt facilities built over a subse- here is that Trincomalee would be built after quent six-year period. Maswlla Wihth the eariest start-up data for Trincomalee estimated presently at 2002, and a Clean Coal Technaolog six-year lead time, this would imply a techmolo- The shortcomnings of conventional ap- gy commnitment by 1996. That appears to be a proaches to deal with the air pollution impacts reasonable assumption giLven the stte of large of coal plants have long been recognized- rath- scale demonstration projects in the industial- er than fix individual emission problems sepa- ized countries. It would probably be unreason- mrely- SO, control by FGD systems, NOx able for Mawella, which may be needed as control by burner modifications-a much bet- soon as 1998, implying a technology conut- ter approach is to use fimdamentally different mnt by 1993. But certainly by 1996, the corrkustion tchnologiest Over the past decade technology can reasonably be projected as a substantial research effort has therefbre been commerciaBy available. mounted into so-called clean coal tcchnologies, including such technologies as atmospheric and Pollution Control pressuized fluidized bed combustion (AFBC, Fuel gas desulfitzatio (PGD) is taken as PFBC) and inte d gasification combined the representative pollution control measurt cycle (IGCC)Q A number of these are at or very This technology is being widely advocated for close to commercial availability at the utility SO., control in developing oDuntries despite scale in the indusrialized countries, and have high capital costs (that may increase by 15 lately been proposed for application in develop- percent to 20 percent) and difficult operation ing countries as well In Irdia, for example, and mintenance problems (such as sludge IGCC is viewed as an attactive way of dealing disposal). with very high ash coals, and in Palistan, flu- idized bed combustion is under consideration Fue Switching as a way of dealing with the high sulfur con- A variety of fuel switching options are ex- tents of the Lakhra cols. aniined, including (a) 1 percent sulfur coal at In this study we hypothesize the general coal burning plants, 2-5 pecn residual oil for commercial availability of PFBC technology diesels (the base case); (b) 1 percent imported by the end of the 1990s, in combination with residual oil used at diesel plants; l. perent sul- the use of a combined cycle.' Even though oil- fur coal at coal buring plants (identified in the 2 The westen part of his area has long ago been declared a bird smctuary that is considered the habitat ror numerous species of migratoiy birds. Although the danger of wind turbines for bird ire has uot been prord coaclusively, it seems safer to exclude this region for the momnent from wind energy developmcat plans. ' There are four utility-scale demonstration plants presently at or very close to commeial opeation: a 200-megawatt unit at Tidd, Ohio; the 330-megawaut Philip Spam plant in West Vurginia; a 200-megawatt plant at Escovar, Spain, which will bum black Lignite that has 6.8 percent sulfur, 20 perent moisture, and 36 percent asi; and a 135-megawatt plant in Stockholm, which is configured not as a combined cycle but to also produce 225 megawatts (themal) of steam for distric heating- 7. IacorporadarEnvirovnsc aCstts ino PowerDeeopmentPlaanniaA Case Study of SLanka 205 trade-off curves as "low S resid); and (c) 0.7 that may get de&fred; and that in tur fssil percent sulfur coal at coal burning plants (-low plants may be run in those years at a level S coal'). higher than if the hydro plant had been built One might note the distncon betwee fuel earlier-with the result tbat emissions in these switching (as defined above) and envionmetal years are higher tban te would have been if dispatch. In the case of Srin the conven- the hydro plant had bee built earlier. In the tional and environmental dispatch merit orders cae of Sri Lanka, for example, if DSM per- are not likely to be significandy diflrent; both mits, say, the Broadlands hydro plant to be call for hydro to be distched first, followed delayed for several years, thien emissions fiom by the coal plants, as thy become availab!e, Sapu a and Kelanitissa may be higher then diesels, and finally the combustion tur- m those years. Howwevr, DSM also has the ef- bines. Wind enery would also be used when- feet of reducing enegy dmd, which tends to ever available under both cnteria. The offset this efibet relatively remote location of the coal plnts Figure 7-15 illustrs the impact of DSM wiould plae these plants high on the ment or- on SO emissions. In the early ycars until der even under the criterion of mimin im- 2000, the reduction in SO emissions is sub- pacts (rather than emissions). However, in sntiaL However, beond 2001, because other countries there may well be significant Bro2ands and Gig Ganga are defrd, differences in economic and environmental dis- exr4issions are indeed higher (as the exsting patch rules thermal plants operate more hours per year). In fact, over the entire planning horizon, the net Results present value of S4 emissions is 23 percent We have attempted to kep the mmber of lower for DSM (bt 6 percen lower for NOx, abutes small, in the irests of beg able to and 4A perce lowe for CQ emission). fcu on what are important eromentl im- These, to be sme, are modest reductions, a pacts froinm the national viewpoint, and ta consequence of the fact that fluorescen light- propery need to be included at the stategic ing while having a big impact on the evenig planning stage. Purely sit-specific impacts peak demand, has a -eltively smal impa m (such as visual irsion of structures, or tem- terms of energy. Other DSM measures may poay changes to gromdwater tables during wel have higher enery impacts (and hence re- unmnel construction), amenable to mitigation at sult in large reductions in emissions). the project sUta and in any event better ad- Given the conclusion that, i light of its dressed by established EIS procedures, have economic and envimntladvantages, DSM been purposely excluded. Table 7-2 smmna- ought to be part of all policy pacdk for the nzes the attributes used in the case aud. development of the powerf sector,r we therefore redefined the base case as incling DSl: that Dedside Management is, as we examined ombi of programs Some preliminary unms indicated that when and policies, DSM was always a compone of policies wer amined one at a time, DSM oc- ihe combined strateg. For example, wat is curred im every non-inferor and knee set, and identified on Figure 7-16 as 'no coal" not anly lay in the third quadrant in every case. That replces coal plants by a mix of diesels and hy- this should be so is hardly surprising, because, dro, it also includes DSM. The CEB base case i genera, improvig te efficiency of end-use is identified as "no DSM'- generally (but not always) results in across-the- board reductions in emissions. Greenhouse Gas Emissions Ideed, it is cocivable that beuse DSM The reaionship betwe global CO2 con- has the efxct of making it possible to defer ca- cenratons and the actual physical impacts that pacity additions, it is a hydro capacity addition may follow, such as sea level rise or changes in 30 st ssucha rmma onwas adopdby the reptofteOfficialCommiteeto sudy SnLans powerneeds, ap- pointed by the Cabinet Subcmmittee an Forig investents in Apil 1992. maon (3) urges a number of DSM measres for immediate- (including high-efficiaey ligbting) and medium-enm implemetation. 206 Table 7-2: Summary of Euvirerental Attributes Attibute Units Impact Emissions of carbon dioxide [1,000 tons] global warming Population exposure to person-microgram per human health impacts air pollutants square meter per year Biodiversity indx f J &diminution of biodiversity, impact Biodiversity indexon babitat of endemic species Surface temperature> i"C [bectare] ecosstem impacts from thermal plumes Employment discounted incremental employment Emissions of acid rain [1,000 tons per year] potenti for acid rain damages precursors monsoonal rain&ll pattes, ae still poorlya- th e non-inf*rior set plus Kukule run-of-riv. derstood, and in any event are very unlie to Here is an example of where there is no pro- be captured by simple linear correlations. nounced knee-set, typical of siions wher Howe, since Sri Laa's contribution to the non-inferior curve shows no marked worldwide emissions w remain extremely change in gradient. small, the assmpio of linearity is not nmr- If the reduction in other environiental im- sonable. in any event, since the, fous of cur- pacts is valued at zero, then a comparison of rt internaional eforts is primarily on plans in quadrant V-where in ctions in CO2, in the first instance Sri Lan- the environmental objective m3ply a worsenmg ka decision makes will still require inforima- of the cost objctive--indicate that although tion on emissions (espeay in light of the wind energy has an implied cost of $60 per ton possibilities of CO2 taxes or tradable emission of CO2 emons avoided, the no coal option nghts that have been advocated by some). has an implied cost of $44 per ton, ad the coDn- On Figue 7-16 we show the tradef stfuction of the Upper Koumale hydr project curve for CO2 enissions agnt average ie- by 2001 only $6 per ton. mental cost (AIC)- As one might expect, the trde-off curve includes the acceleration and Air Qualty and Health Impacts addition of hydro plats tpper Kotmale and Pror efforts to quantify air qualty impacts Uma Oya), and the elimation of coal from the asocatd with different thermal generation expansion plan (m which a combination of hy- technologies and with different sites in Sri Lan- dro and diesels replace Mawella and Trinco- ka have been very crude "ALr qualitf was one lee). The odter options in quadrant IH include of the attnrbutes used in the Black and Veatch he Kukule run-of-rver project, and PFBC. thermal genrato options stnd alternative If the knee set is defied on the basis of sites were assigned an air quality score in the ymmetrical and equal tolerances of plus or nii- weing summation method that was used to mus 5 percent, then it contins the options on select sites for difierent generating stations. 7. IncouporasiagE,vironmeaial Costs into PowerDedopment Planning:A CeseStudyof Sri Lanka 207 The most serious problen with such expert a dose of two units has the same impact as two judgment scales is not tiat expert judgment individuals experiencing a dose of one unit may be incorrect, but a the relatonship be- each. In fact, it is reasonably well established tween the score and possible mitigation actions that acute episodes of high pollutant concentra- is not denonstrable. In other words, there is no tion, even of short duration, are much more way of subsequently performnmg a sensitivity damaging, particularly to the aged already suf- analysis that might explore the use of altema- fering from respiratory ailments, than chronic tive fuels, alternative pollution control srate- exposure to lower levels. Nevertheless, whatev- gies, or altenative technologies, er the liitations of the simple linear model im- In our study ve made a much more precise plied, the main issue is whether the procedure calculation, by estgiatig the actual population as a whole provides a more objective way of exposure, namely mparing the potential air quality inpacts of alternative sites, technologies, and polution X = Z CJ PJ mitigation strateges than the purely subjective assessnts of "air quality impact" of the type where X is the cumulative population usually made in siting studies exposure to the incremental ambient This model is no substitute for detailed air concentration attributable to the pow- qality simulations that might be conducted at er plant the EIS stage to demonsae compliance wit P is the populatio in the j-th grid specific regulatory rquirements. But for appli- squar 31 cation to system planning and site selection stu- Cjs the increnental average annual dies we believe the model to be a material concentration in the j-th square, at- mprovement over dte sort of purely subjective tributable to the emissions fiom the "expert judgment scale conmody source in question. encountered2 The Cj is esimated by application of the stan- The health impacts trade-off acrve is ilius- dard Gaussian plume model trated on Figure 7-17. The non-inferior curve is defined by the low-sulfir fuel options, PFBC, Qv o (jy -_ and, again, an icreased use of hydro. ljWdCyG exp V2 There are two particularly instructive find- ings here. First is that the no coal option, and wthere u is the mean wind speed (in meters the repl of Mawella by diesels, wors- per second). ens the health impacts unless accompanied by Q is the source ten (in grams per the use of low-sulf fuels. This is simply a second). consequence of the liely locations of the die- H is the stack height (in meters). sels that would replace the coal plants: located It is important to state the imnplied assump- in or close to uban areas, and therefore in tions involved in the definition of such an at- much greater proxinity to population, the asso- trbute as a measure of health damage. First is elated health impact will be large tn that of the assumption that the dose-response function coal plants located far from the population is linear through the origin- neither thresholds c s. nor non-linearities are assumed, this is equiva- The second relates to the effectivenes of lent to saying that one individ experiencing FGD, which is neither in the non-inferior set, 31 Populations were estimated for each 500x500 mete- grid square using census data and land use mps Unfortnmatdy, detailed land use maps wer not available for the Trincomalee area, and therefore the population data was estimated on the basis of very crude approximations that have limited reliability 32 A fi(rther problem concens the extent to which the presce of a myaorpower plant itself acts as a growth pole, and at- tras additional population. For major cities this is not hlly to be an issue, and it can safely be ignored for Colombo. For both the Trincomalee and south coast sites, however, it is conceivable that population may grow faster than would otherwise be the case in the absecc of any major power plant. In light of the difficulty associated with making local area popuation projections, fine tuning the population;estimates would be a somewhat specious exerise, and the pop- ulation estimate for each area was therefore simply taken as constanL Figure 7-15: Impact of DSM on SO, Emissions 3 3% 2 4% -5~~~~~~~~~~~ % 70/ .00,91 9511991203120 C~~~0/O .2 ~~~~2-0% 0,~~~~~10 OV199 1997/ 201 205 20 22O~ (I)~~~~~~~ -3 ~ ~ ~ ~ ~ ~ ~ ~~3 04 j -2- ~ ~ ~ .3% -3 1991 9I 199 I197 9 2003 2005 2007 O9 Figure 7-16: Greenhouse Gas Emissions 5.9 5.8 E +kukuleHD 5.8 -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I 5.7- noDSM 5.6 - 5.5 E 5.4 _ qsnoCoal E\ 0~~~~~~~~~~~~~~ 'S 5.3 - \EC lowScoal 8 ~~5.2 - ael > ci Iresid o T&D>11 5C +wind Trade-ff curve 4.9 - \ +UK2001 o noHydro 4,8 _ QuLiadrant lii ;;pasecase(wIth SM)F +Uo2ol. 4.7 I I I I I I I I1I 1 2 13 14 15 16 17 18 NPV (CO2 emissIons, 2010,10%/*) 210 nor in the knee set On the other hand, PFBC is high unemployment in the construction cost in the knee set (and also lies in quadrant E). esmates. When one also takes into account the CO2 The employment creaton attribute used in emissions trde-off curve, for which FGD is in this study is based on the estimates of local quadrant I (that is, worse than the base case in construction and operat phase employment both cost and CO, emissions), the superiority impacts developed as part of the GTZ Master- of PFBC (and possibly other of the newer plan. The total man-year constuction phase clean-coal technologies not examined here) be- impact is spread uniitrmly over the construc- comes evident non period; all emloyment data are then dis- Indeed, a comparison of the health impact counted to calculate the foOowmg hides: and SO, reduction trade-off curves is useful: if one looks only at emissions, it comes as litle - E s k (Ogj-4Cg surprise that FGD does now lie on the trade-off t (I+ curve (Figure 7-18). However, the low-sulfur whr Ok is the number of opeational staff fuel qption is not sigmnficantly dominated, and ekedinyeartatthek-thfciliy is mcluded in the knee set The point here is C" , is the number of persons that strtegies that may appear efrctive from employed the cons on the standpoint of reducing emisons may or phase at the k-th facility in yeart nay not be effctive to reduce impacts in a r isthe discountrate. cost-effective way. It is implicity assumed that the nominal value These results are a function ofthe assump- of each perso employed is the same across taon made for the sulfr content of domestcally time periods (bef£re discoumtiing This corre- produced residual oil, whose specification for spo to demotatcic political objective of sulfur is no more than 3.5 percent by weight te a tnbutas noted above, wage rate diftr- Because the refinery at Sapugaskanda has been enoes among employees is captured in the eco- runing low sulfur cmudes over the last fe-w nomic attnbute by appropriate wage rate yeas, acta sulfur contents are codrably compaions- Finally, the bulk of the addicto- lower than this: in this case std we have as- al jobs created are for semiskdlled workers and sumed 2.5 percen Unforunately, no reliable new jobs for unemployed According to the data are available on the sulfur content of the analysis conducted by the GTZ Masterplan, to- residual oil actually produced at Sapugaskan- tal employment at existing hydro plants in da. TIS is an ar where I better data needs to 1988 was 988 individuals. The largest single be developed before definitive conclusions are categories were 260 secuity guards, 121 un- derived' skilled laborers, 101 semiskdlled laborers, SO Employment odsw i operators, and 93 other csual and unskilled jobs; these five categories comprise Employment creation is an importn o- 66pet of the total. joctive of national policy, and in Sri LanIa As one might expect, the non-inferior cau there has occurred frequent disussion of te for employmet (Figure 7-19) is defined by the need fOr employme creation in the south, lre hydro projects that requr a relatively where youth unemployment rates are especially large local construction workforce. Note that high. It should be noted at the outset that what since enployment geraon is a benefit from is -aptured in this attribute is the sepazate and the natonal p ve, the non-inferior curve purely political objective of employm cre- ces to the rght; rter than to the leR as in aiion, which is to be dguished fiom stricdy the previously preented cases of costs. economic benefits that would captured by th (Whether this is also true from the local per- use of shadow wage rates appropnate to reflect spective is a matter for the project-level 33 Another issue is the degr to which there is an adequate supply of high-suliur oil available fiom fte Sapgada re- finely: at the moment tis is in excess, but twads the late 1990s it is expected that domestic demands will utilize all residual production. Consequenly, heavy fieloil would need to be imported for those options that require extensive use of diesels Figure 7-17: Health Impact Trade-off Curve 5.9~~~~~~~ 5,8 ~~~ +kukuleHD no DSM 5.5 - IowScoal+Iresld onCa 5,4 - l +FGD nCa 5.3 - m~~~~ lowScoal 5,2 -Iresld Mawella>D wT&D>11 5,1- Ei +w1nd 5 -Trade -off cuirve 4.9 -+ 0 E noHydro ........ 6..U...001..... baseoase(withD M flRf~~~~~~~~M ~ 4,8 Quadrant ill +IP C / 4.8~~~~~~~~~~~~~~~~~~~~~~~~~ 4,7 I k-~~!U02001 4.7~~~~~~~~~~~~~~~~ 2 3 4 5 6 7 NPV (Health Impaots, 2010,10%) 212 envionmental assessment. Large construction of an ecosystn would be unacceptable, and projects in relively remote sites may strain hence assigned an infinite value (even if the the abilty of local communities to handle the habitat involved were of low biodiversity, such influx of temporay workers.) as a sand dune), whereas the loss of one hec- We have assumed no incremental employ- tare if 1,000 hectares remain would be much ment effect for DSM; however, it might be less. noted that a number of studies in the industial- Such an approach to valuation of biodi- ized coutries indicate that the number of new versity is subject to several caveats First, as jobs created by energy efficiency projects is noted, eosyems may require some minim greater than the nimber generated by an area for long-term survival, which implies that equivalent supply-side projec4 the value function would need to tend to mifin- Biodiversity ity' as it approaches that minimum value. Second, and perhaps more iportanly, ihe Given that at the planning level detailed argument is sometimes made that the value to sit-specific infonnation at potential power be ascribed to the loss of habitat assciated plant sites is unlikcly to be available, the only with some regulatory or governmental decision quantification that appears possile is to deve depends upon whether what remains is secure. a probabilistic index that gives the decision- For example, the cost of the loss of 1 hectare of maker information about the likelihood that the a habitat if 1,000 hectares remain might be val- drailed ETS will reveal the presence of an en- ued as negligible, if that remaining habitat is denic species;5 significanty affect ecosystms protected from enachmn On e oer of high biological diversitr, or afifet a habitat hand, if the remaining 1,000 hetares are vil- already in a marginal ondition It should be nerable to encroachment, then the loss ofthat 1 noted that endemicity and biodiversity are not hectare under consideration might be assigned ncesarily corlatect an endenic species may a much higher value. Although not often put be encountered in an area of low biodive*, explicidy in these terms, this line of reasoning and areas of high biodiversity may contain no has in effect been much used by evronmental- endemic species. It is certinly rue, however, ists to oppose power fadilities- For example, fhat extinction of an endemic species would the reasoning goes, because coastal wetands very likely constitte a "show-stopper" at the are ecologically valuable, power plant should project stage; and it is also true that at least as be sited elsewhere. Yet;, a much greater threat far as Sri Lanka is concerned, its endemic spe- to coastal wedands is posed by uncontrolled cies are most likely to be encountered in areas real estate development, but since this is harder of high biodiversity to control, attention is focused on power plants Such an index will have several consituent that make much better targets for attack. elements. First is the nature of the impacted Such reasoning is not logical, because ecosystem itself On Table 7-3 are ranked the there is a confision between costs and benefits. niain ecosystem types encountered in Sri Lan- On the cost side, it really does not maer ka, and assigned to them a value, w,- that cap- whether the loss is attnbutable to a power tures the relative biodiversity value of different plank or to agricutal development, or even habiats. to ilicit felling: the loss is the same in both The second elanent concems the relative cases. On the other hand, the benefits to society valuation, because the value of the area lost is of these two activiies may be quite diierent. a function of the proportion of the habitat that Yet it is only land use planning at the local and is lost. For example, the loss of the last hectare regional level that can address the costs and 34 See, for example, the study by the Council on Economic Priorities, Jobs and Energy (1979). This study. which ex- amied energy options for Long Island, concluded that a consevatn strategy generated more than twice the regponal cmploymen tm what wodd be required by an eqmvalent supply-ide sbategy the ational employment effect was es timated at 1.5 times higher for the consevation stategy. As m Sn LanIka, Long Wsand has no healvy electrical eqip- meat mannuctuing idustry, so the bulk of the cqipment assocated with suply-ide generation projects need to be import (in the case of Long Wsland, from oher areas of the United Sates). '5 A species is said to be endemic to Sri Tlanka if it occurs nowhere else in the wurld. Figure 7-18: Trade-off Curve for SO, Emissions 5,9 5.8l~' +kukuleHD i 5.7 - o noDSM 5.6 5.5- lowScoal + lresid n E 5.4 + FG> S ~~~~~~~~~~noCoal E +FG tv 2 ~I5.3 - a lowscoal ' iresid o M wella>D | 0 512 - T&D>11 5.1 | +wlnd 5 4.9 _ Trade-off cur/e +UK2001 C noHydro .. 4,8 t. .. basecase(with DSM) nt III kuI I !II +WUO2001 4.7 a X XX . C 50 60 70 80 90 100 110 120 130 :140 150 160 NPV (SCQ emissions, 2010, 10%) Figure 7-19: Non-inferior Curve for Employment Impacts 5.9 5.8 k / +kjkuleHD 5.7 m. no DSM 5.6 5.5 noCoal lowScoal+ resid 5.4 ~~~~+FGD E 5.3 - lowScoal Mawella>D 5.2 m I~~~~~~~~resid T&D>11 o 5.1 Trade-off ct rve w +wind 5 4,9 c no-ydro basecase (wlth DSM) +UK2001 El w +Ku eoR l 4.8 L +PFBC +UO2001 4.7 X I I I I 1 1 I I X l.. 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 NPV Oncrementa employmont, 2010, 10%) 7. Incorpornnng Environmental Costs into Powerc bevenePlannl:A Cow StSuy of Si Lanka 21S benefits of akemative uses in a systematic way. another is very complex. Thus, for example, Thus, the biodiversity index associated even if reliable studies were available concemn- with site i B;, is dherefore simply defined as ing t say, the impact of thennal discharges on coral reefs in Ausal or on the Red Sea, it is Bi = >:[ AsiWt unclear what applicability they would have on where A,, is the hectare of ecosystem of type the unique conditions of Sri Lanka. Certain wher At sisthe hectare of ecotstem of type general principles are easy to state; for exam- J as sitel . ple, water temperatures off the coast of Sri w is relatie biodiversity value of La*a are aleady fairly close to klown lethal typej (as defined in Table 7-3). threshold tmpatures for corals. But predict- As it turns out, it appears that the biodiversity i i of a 1,0.1, or 0.001 degee tern- index is strongly correlated witd reservoir w perature rise is essentially impossible. as indicated on Figure 7-20. This is sinply a The is setiof impossare. consequence of the fact that all of the la guidance: hffe present requrement is only ltt projects inundate relatively large amount of no discharge may be of greatertemperature, at natural forest of high biodiversity value: this the discharge point, than 45C; the design ter- would not necessanly be the case elsewhere. pe ofclie onethr45gh desin tem However, when one excludes the five very pseratue of the once-thrgh cooling system discharges for Trinconmalee was set by BVL as large projects with reservoir sizes in excess of 37C, which is well beblw the standard-a fct 15 square-kilometers, the correlation is quite 36 Eee mentioned by the EISY poor, reflecting the diversity of ecosystem However, it is unclear upon whit basis the 45C types encountered at the smaller project s tandard based The biodiversiy impact trade-off analysis Te gmal cs of temal 9charges is depicted on Figu 7-21. Tbe non-inferior to eaerstal eatles of termdischarges curve obviously has as one of its endpoints a charges intO the well-mixed surfae layer "no hWdro" option (in which Broadlands, Ging c o d usuialy have the gendtendenac to re- Ganga, and Umna Oya arc all forced out of the pel fish. On the otr hand, if the discharge is base case): the assumption here is that all of below the thdmocline, themal discharges the thermal projects that would replace the by- would have a gnealy beneficial efect, as tfe dro plants would be at sites of poor biodivers- up woldhng eeager calsed by plume bucyancsy ity value close to -load centers. Certainly theln fec asdb pueboac high dam variant coft Kku.e impliesa ~~tbeWa brings nutrients into the layers nearer the luhi{a variantof Kuue implies a poute - surf=. loss to biodiversity value (B=530) that is eight Howvra er, specific mumical esti- times that of the majority of the policy options mates to the values of this general fimution is (B-50 to 70). imssible One is simply not able Although the wind plants would require a ess ossile One i Im abe rather large land area, the vegetation of the to ake statements of t tnpe that anC aver- areaon te suth oas hasmthe lo bioiv- age surface temuperature increas over an area area on the south coast has rather lowvblodiv- of x square-kilomeers would cause fish- ersit y v alue, and therefre the overall increase catches to deline by y percent Wlat can be in biodiversity iact of this option is judged done as a generic calclation that can be used to compare difient sites is to begin with a Aquatic Ecosystem Impacts definition of what is considered to constitute an The vautinaneacceptable environml risk$ for example, Tihe valuation and asof pote say a temperate increase of no more hn 1C chargesvisr extremely dimagfficulthe l ds at the surfice. One then calculates the surfce chargtn is pxremely difficult The most area over which this criterion is exceeded as a importart problcm stomx frm the fact aquiex- functiLon of the cooling system design proposed. traolations fi-am one aquatc ecosystem tD 36 For example, "... it should be noted tbat the design conditions for themal discharge were based upon the standard seL by CEA that stated axim.um dischage tempem was limited to 45C. As a mitigation measure thl- power plant is designd for a maximnu discharge tempeatur of 35C (with 37C as a worst case situadon), itch allows for a signifi- cant safetr margin" (Trincomalee EIS, Vohmie L p. 8-51). 216 Table 7-3: Relative BiodiversityValue of Main Ecosystem lapes in Sri Lanka Rank Ecosystem Weight (W$ I Lowland wet evergreen forest 0.98 2 Lowland moist evergreen forest 0.98 3 Lower montane forest 0.90 4 Upper montane forest 0.90 5 Riverine forest 0.75 6 Dry mixd evergreen forest 0.5 7 Villus OA 8 Mangroves 0.4 9 Thorn forest 0.3 10 Grasslands 0.3 11 Rubber lands 0.2 12 Home gardens 0.2 13 Salt marshes 0.1 14 Sand dunes 0.1 15 Coconut 0.01 16 Forest plantations 0.01 17 Tea plantations 0.001 18 Rice paddies 0.001 Source This scale was developed by Professor K D. Arudpragasam, Professor of Zoology at the University of Colombo and former Chairman of the Central Enviromnment Authority. There are a number of cooling system op- recommended by BYI, but with dischr to dons that might be considered as altenatis to the much deeper waters of Koddiyar Bay. once-through cooling with discharge directly That such long ocean difser system out- intD shallow bay waters, including long outfalls falls are teelmically feasible is illustrated by a (with or without diffiser system), cooling number of power plans in the United States. ponds, and mechanical or natural draft cooling For example, at the San Onofre power plant in towers. Although the altenatives to te base Calimia7 the intake is about 1 kilometer off- case configurtion all reduce the thermal im- shore, while the diffuser discharg section on pacts and involve higher costs, however, some the outfall starts about 2 kilomets offihore of them also introduce new enviromnental im- and extends for about another kilometer. pacts-such as salt drift and visual itusion To be sure, such a system involves addi- (in the case of nanural draft towers). While a tional costs. However, the alternative of mov- detailed engineering sdy of these alternatives ing the coal plant to the south coast also is beyond the scope of this case study, in our involves vey severe fuel oost penalties. We es- view the potentially most suitable alernative is timate he present value of the incremenl fael- some kind of long outfall with dffuser system, costs for 900 megawatts over a 30-year perhaps involving a different site to that lfetime of the plant at US$142 million,38 3 Whatever may have been the licensing difficulties at this nuclear powe plant, the cooling systdesign was ner an importat issue 3 The annual cost differenitial calculates as follows: 900 [megawatts] x 0.6 x 8760 hours per year] x 1000 iclowaus per megawattl x 2500 [kilocanlories per kilowatt-hourj x 1.20 [dollars per million kilocalories fiu pealty] = Si3 million per year. The fuel cost differential is taken fiom the 1991 CEB Ganation plaming report (p.45) Mwela fbd costs Figure 7-20: Biodiversity Index Correlated with Reservoir Area 1.3 1.2 h KlJKU 022 1.1_.. 1 tGING05 0.9 0.8 0.7 8 0.6 _ GING052 0.5 0.4 _ KELA 071 isi a MAGUO43 0.3 _ 0.2 _ Other small plants |maInty run-of-river) 0.1 D1 0 20 40 60 Reservolr area, W 2 t Figure 7-21: Trade-off Curve for Biodivcrslty Index 5.9- 5,8 w~~~~~~~~~~~ +kukulefiD 5,7 5. no DSM 5.6 5,5 5,4 _ +FGD - ? E 5.3 * lowScoal 5,2 _E Iresid m T&D>11 5.1 m +wind 5- 4H,dr_ Trade-off ourve 4.9 ... .....+ku +UO2001 4.7_ 0 100 200 300 400 500 600 Blodiversrty Index 7. Inrpora6n Ewironmentad Costs ino Power "Dep219nning:4 Case StudyofSnka 219 whereas the incremental costs of a diffuser sys- Demand-side Management tem are likely to be in the range of $25-SO Te fis conclusion, already noted, is the mllionT atactiveness of demand-side maagent On Table 7a4, we ompaee the n ialerenins (DSM) from both the economic and environ- site costs, as indicated in the Trinomalee EIS. mental standpoints. Our repreative DSM The item for "circulating water' includes a meas fluorese-t lightng, es the short capitization of the operating cost diffemntial nies ofla policies for all attributes cx- associated wi different pumping he At th mined; and we therefore redefined the base sdimnt s tal cos t of case to include DSvL A detailed study of a broader set of potential DSM opporunities in the EVI recommended site 2A-2. As alred Sn Lanka ought to be gven high priority. noted in a previous chapter. BVI's incorpora- Because the peak use of lighting ocur in tion of aenr#irfnmlan factors" mnto the site the early evening hours at the time of the daily ranking process seems somewhat unreliable: it M peak, it has an unusaly lag impact is certainly tine that no thermal modelling e- -on the system peak, and hence an unusually sults are presented for site 3A-2. Because this l i capacity requirements. Other site is nearer to the deeper watus of Koddiyar DSM measures would be expted to have a Bay, and because aqaic ecosystem efficts are considerably smaller impact on capacity re- demonstrably the most importat eimen- qu , but may on the other hand also al impact to be considered, we consider this sav larger amounts of enry Inded, our rep- skle to be potellay amatve. s-e toboetaltrcrentative wind energy scenario, which for- With thermal plume modelling studies in sees 155 megawatts of wind capacity on the hand for both these sites and for a variety of - coast by 2001, will prvide significanty oooling wsyem C f losand for compa- rabc sprestemnfatigurlations, and for compa more energy than fluorescent lighting may asbie representative locations on the st save however, as noted previously, wind coast, Tablde-off couldve expndtued. Fon the oplants wll not reduce the peak capacity re- side, and a taeoff curve constructed. For the qieet hsas nesoe h edt moment, we have only one point on this curve: exame atvo indarsyste n- fiom the modeling studies for onctrough text, nir compaisons of teologies in isola- cooling at site 2A (see Figure 7-22), the area of tions of ec cotelxt simply cannot the thenal plume with a IC rise above ambi- captire such interactions. ent, located in well-mixed shallow waters, is 160 hectares (for the fill 900-megawatt devel- Tail Stacks opment at the site)- Tal stacks are expressly excluded as Candidate List and Policy mans ofassung compliance with ambient air Candidat ListandPolicquality standards in the United States. This Implications -was a rional modification of policy, since tall Although the discussion here has thus far stacks were used mainly at Large mme moudt been focused an the two-dimensional trade-off power plants in the ceta United Sttes; the curves, examination of the twodimensional prevailng weather patns with westely plot alone is insufficient to establisih the com- winds for most of the year meant that long- plete trade-off surlc, as noted previously. range tLansport agavated air qualiy prob- Only the application of the set definition rules lems in the northeaster United States and m the n-dimensional attribute space provides a Canada. rgorous definition. On Table 7-5 we suma- The view that smce tail stacks ae map- rize tie set of plans and policies at emerge by propriate for the United States -and Europe, application of each critrionL A discussion of they should also be disallowed in developing the policy imnplications follows. countries as a compliance device is quite wide- ly held. While we certainly are not advocat ___________________ ltail stacks as a compliance meaism, it re estimad at 755 cent per milio kilocaloies, Trinomalee at 635 cents per milion kilocalris. 220 Table 74: DIfferential Site Costs 1 2A-2 3A-2 4 Transmission 5,790 2,670 70 0 Coal handling 0 3,270 1,630 3,840 Circulating water 41,830 0 10,440 14,390 Site clearing 16,380 9,940 0 23,190 Foundations 0 0 0 0 Ash disposal 6,010 8,170 11,950 0 Transportation 3,120 0 0 3,460 Total 73,100 24,050 24,090 89,680 would seem that the option ought not to be dis- Suppose for the moment that the standard missed without a carefl case-by-case ex- for the United States is properly defined at (or amination for tropical islands, given the very near) A, based on the American damage curve high oss of altenative means of SO, reduc- Z-as illustated on Figure 7-23. Assume fur- ton. For example, at the Mawella location, the ther that the cost function for pollution abate- distribution of prevailing winds is strongly bi- ment, say X, is the same in both the United modal, corresponding to the southwest and States and Sri Lanka-a reasonable assump- northeast monsoons Long range trnsport in tion for FGD sstmequip for example. this case disperses pollutants entirey over the But suppose the damage curve in Sri Lanka is ocean during the nofthwest monsoon, and after W, which implies an optimum at D, not A a short overland distance during the southeast Since a large part of the total environmental monsoon Consequently, it in fact, short-range damage relates to the costs of health care and air quality standards might be violated in the societal valuations of morbidity and mortality, vicinity of Mawvella, tall stacks would be a per- there seems litde question that this dmage fctlvy rational solution since any long-range de- function does indeed lie below that incurred the position (that typically occur over distances United States. It follows that use of the Ameri- betwem 200 and 800 lilometers) would occur can standard implies an over-investmnent in almost entirely over the Bay of Bengal or the pollution control by the amount Q. South Indian Ocean. Perhaps the most important point is that even if it is true that environmtal standards are not strictly optimal in economic terms, at Whatever may be the difficulties of explicit least in the industialized countries there are in- valuation of environental exemalities, the ef- stittiional mechanisms that ensure extensive fort ought still t be made provided one is and informed public discussions of proposed mindful of the assumptions and limitations- regulations; the U.S- Environmental Protection While the argument that in the absence of dam- Agency, for cxample, is required to conduct a age valuation studies one ought simply to be detailed economic imrpact analysis of proposed conservative has a certiin politcal appeal, enission standards on the affected industries. adopting the environmental and ambient stan- Such a debate is still almost entirely absent in dards of the industrialized countries may imply most developing counties, and the resource al- a serious misallocation of resources. location dcisions implied by the adoption of standards, or "enviromnental guidelines," 7. IncorporingiwoEironmental CostsintoPowerDevdopMeatPlanning:A CaseStudyofSrLanka 221 Table 7-5: Candidates for Plans and Policies Quadrant tese Non-inferior set Knee set {DSM}z {DSM}2 .DSMh}2 (no coal} (no coall PFEBC} (' EBC0 {PFBC} (Uma Oya} (Uma Oya} (Uma Oya} (imported residl (imported resid} flow S coal+ {low S coal+ imported resid} imported resid} {FGD} {FGD} {Upper Kotmale} {Upper Kotmalel {kukule RoR} {Mawella>D} {T&D losses>11 percentl {no hydro} {kul-'- HD} 1. This is the multidimensional equivalent of the two-dimensional "Quadrant Ilr criterion: these are plans that are better than the base case in at least one attnbute, and no worse in any others. Z. DSM is included in the base case, as discussed in the preceding texL Figure 7-22: Estimated Thermal Plumes at Site 2A Trincomalee Thermal Trin con lee Power Plant Project .. .. v\.,:::t......... . .. ;t tChina o ...........Harbour Estimated Locati6ns of Thermal Plumes for 900 MW Plant ~~~~~~~~~~~.:. .:.. . ........... . :..:.: ii:i:i. i :: :....:;:.-..... .::y 'Cove ,.,., .., ..,, . .,,. . . ... .i.... .............la .. . ...... {.... 1;., \ .::..:;. . ' i'it>' J h\r ,. ...................Trincorn I& Sourcc BV EI;S Figu~xire 864.,r\\ - e wt~~~~~.^..... ............ iX\^ I~~~~~~~~~~~~~~~. ..... . ...............' r> q l . . .;. . . .... . .. .. .. I. +**. v*˘..... 2 @w_ b|A( l\ L ;:. :,.. : ."*.:# .*...... ... .......... .. v\\\" >tR Source: BVI ETS, Figure 8.6-4. Figure 7423: Comparison of Damage Functions and Environmental Standards Sri Lanka and the United States 320 300 | 260 k X< ~~~~~~~~~~~~~~~Total US costs (X+4) I 220 - 200 180L- eoIt' 160 _ Srl Lanka cos C+W) 140) Pollution control cost 120- 100 = 80.U.S.damage co t .~~~~ 0 0.1 0.2 0.3 0.4 0.5 0,6 0.7* 0,B 0.9 - ~~~~~~~~~~~~Degree of pollution abatement 8ri 4 40 - Sri Lanka LdkaaptmmmaUe. opimu 224 ought almost certainly not be made by well- mitigate undesirable impacts of thermal meaning functionaries at the international fi- effluents. nancial institutions responding to pressures The present assumption is that Mawella from American environmental groups. would be developed before Trincomalee. The The argument that health-based standards only justification for developing a higher-cost should be absolute-meaning that the value of site before a lower-cost site would be security a human being is the same wherever he or she considerations that might preclude Trincomalee may be-may have etdical mcnt, but its selec- from earlier development Hlowever, the results tive and isolated application to power sector of this study suggest that if scurity reasons do projects is irrtional. Why should ihe Sri Lau- in fat prwevnt Trmcomalee developm-ent be- ka government be forced to accept an implied fcore 2003 (that is, before a construction start in valuation of $1,500,000 per human life neces- 1997), then it is not clear that one would stil sary to justify an FGD sstem, when expendi- want to develop MaweDla The critical assump- tures on road safety, anti-smoldng, and basic tion here is the location and sulfur content of health case imply human life valuations one if the fuel of the diesels that would be required to not two orders of magitude less. Exprcssed elimate coal at Mawella- diffenly, if the Sri Lanka government can More importantly, our analysis of the justify increased expenditures on human health aquatic ecosystem impacts of a coal plant on care, these resources should be applied where Trwnconalee Bay show that if the reason for they have maxium benefit The ralty in denyig a permit is the poteti most developing countries is that there are risk to the aquatic resources of the Bay, there many more pressing public health problems are more cost-effive ways of dealing with than respiratory diseases induced by incrLmen- that risk than to build the plant at Mlawefl In tal ambient concentrations of sulfates and fine particular, the study indicates that the additon- particulates caused by power plants al costs of $8 per ton of coal inmurred at FGD Systems Mawella (because of the lack of a deepwater Systems harbor that can accommodate vessels of the Our results show that if one wishes to size possible at Trincomalee) are large th spend resources on SO, emission reductions, the incremental costs of a long outfll that the importation of low-sulfur fueloil for use at would discharg heated effluents at greater diesel plants costs less than one third per ton of depth out in the Bay, rte than, close to shore SO removed than by FGD systems. And when as enaged by the present design- We have population-weighted health impacts are consid- not conducted a detailed feasibility study of ered, the value of FGD systems in the relatively such an option, but the evidence is sufficiently remote locations beng considered for coal stong to warrant coser examination. plants is even less, compared to fueloil sulfar CEB Planning Procedures removal at diesel plants located dose to pop- ulated areas. At prsent the CEB does not include the costs of resettlement or the oppotmiy costs of lost producton at hydro plants m its geraton It seems fairly clear that from a cost stand- plnning studies. For some major hydro proj- point, Trincomalee is a superior site for coal ects, this is a significant issue, incrmsing capi- plant development to any south coast location. tal investment by as much as 10 perce MaweLa has specific costs about 20 percent (although at many others, such as at Broad- greaer than Trincomalee. This conclusion re- lads, there are no resettlement costs). While mains unchanged when the cost of environmen- one can reasonably be against the inclusion of tal damages is taken into accout: there is no opportunit costs i decisions, r reason to assume that the environmental dam- settlement costs represent an unambiguous mi- ages of a plant on the south ooast are any cremental cash utay duing the construction smaller hn those of a plant in the Trincoma- phase, and need to be included. In any event, lee Bay area that has been properly designed to since the costs of NOx control are added to the 7. Incorporating Enwironmnudl Costs into PowerDevelopment Planning: A Case Study of Si Lank 225 capital costs of diesel plants, exclusion of di- Although we do not claim to have made rert mitigation costs for hydro plants is any detailed engineering study of ftese is- illogical. sues-and our cost estimates of bringing losses The ability of the WASP-lI[ model to deal down to 11 percent are subject to considerable with the generation planning options in Sri uncertainty-nevertheless the conclusion is an Lanka is subject to some serious question. Not important one because it illustrates once again only are the shortcomings of the model with re- the need for evaluating options in a systems spect to hydro treatment fairly well recognized, context. but this study has raised a number of issues concerning the robustness of the so-caed PBC "optimal" solutions. At the very least, the Pressurized fluidized bed combustion using WASP runs should be verified by runs of a de- th combned cycle appears to be an cxtemely tailed production simulation model, attractive technology for Sri Lanka. Even if the The set of options that are currently ex- installed cost per kilowatt is comparable to that amined in the generation plannng studies needs of a pulverized coal plant with an FGD system, to be expanded. This case sludy suggests that its suitability to be employed in a combined the following options ment serious consider- cycle means that the incrmental costs ationi (a) replacng Mawella with diesels (compared with a coal plant without FGD) are (eaving Trincomalee to be the first coal plant largely offset by coal savings that follow frm to be built, perhaps as a PFBC); (b) replacing imprved heat rates the high dam variant of the Kukule project (as In any event, since PFBC also results in in the 1990 and 1991 studies) by th am-of- substati NOx reductions, and because of river variant heat rate savings lower C02 emissions per net kilowatt-hour from an overall enviromnental T&D Loss Reduction viewpoint PFBC-CC is much more altractive The most important result of our analysis than merely fiting an FGD system to a PC with respect to T&D loss reducton is its rela- plant (which1 all other hings equal increases tionship to the implementaton of DSM mea- CO, emnissions per net kilowat-hour). surest Whether or not DSM is implemented, a The main question concerns the commer- four-year delay in reaching the 12 percent tar- cial availability. However, the prospects for the get (that is, firm 1998 to 2002) increases boti general comnercial availability of 75 to 200 costs and emissions-as illustated on Figure megawat units by the end of the decade seems 7-24 for the cost-CO, emission trade-off. rasonably assure- However, when one examnmes the possibil- ity of reducing T&D losses to 11 rather thn Towards a Robust 12 perent then whether or not DSM is also Energy-environmental implemented makes a significant differece. In Development Strategy both cases, CO2 (and mideed other) emissions will decrease firther. However, in the absence The results of this preliminary investiga- of DSM te fur&er reducton also reduces tion need to be confirmed by more detied pre- costs, whereas if DSM is also impl ed, the feasibility studies that are beyond the scope of cost increases. tthis report- Nevertheless, there are a number of Thbis result is not unexpected: since thie conclusions that may be drawn at this point particular DSM measure considered in this 1. The fluorescent lighting progam to reduce analysis-fluorescent lighting-has a sharp the evening peak load appears to be both impact on the peak demand, and since reduc- economically and environmentally attrac- tion of T&D losses brings a disproportionate tive. However, a more systematic examina- benefit to the peak period (because losses vary tion of other DSM options appears to be with the square of the load), the two measures justified?3 Moreover, we have not ad- interac. through their respective impacts on the dressed here the implementation issues of a load duration curve. fluorescent lighting replacement progra. Table 7-24: Impact of T&D Alternative Loss Reduction Targets 5.9 5.8 12 / by 2002 5.7 No DSM I12%.by 1998 5R6 1%o by 1998 g6. 5.5 'U E 5.4 U) 5.3 : 2 5.2 I 11O% by 1998 5.1 12% by 2002 4,9 12% by 1998 4.8, ' , , 15.8 16 16.2 16.4 16.6 16.8 17 NPM(C02emIssuons, 201 0, 10%) 7ZIncorporatdngEnvironme,alCostsintoPowerDeveJopmeat Planing: A CaseSwudvofSriLanka 227 2. Examination should be made of the possi- Bibliography bilities of reducing T&D losses to an 11 percent rater than to the 12 percent target Amagai, H, and P. Leung. 1991. -The Trade-off presently envisaged by the CEB. As noted Between Ecornmic and Environmental previously, however, the economics of Objectives in Japan's Power Sector,' The such a step i ikely to be strongly denq° p9. Poli- dent an any concurrent implenentation of cies for Sustainable Development DSM mcasures. Manila 3. An initial pre-fsibility study ofPFBC'- May 1991. EnViawmmental Consider- su3. An be u p tedfeasibilimty stuay dc PFBBC ations in Energy Plannitg. Mnila- should be conducted. Prelfininary discus- Black and Veatch Intemational (BVI). March siors might be held with ASEA-Bronm 1985. Trincomalee Fuel Supply and Boveri, the manuficturer of all of the inw- Transportation Study. Report to the portant utility scale PFBC pro.ec to dae, Ceylon Electricity Board, Colonibo, Sri on issues such as commercial availability, Lanka. 1988. Social andEnvironmental _M . arch198Soiladvrnmta suitability to imported Australian coals, lo- Assessment Trincomalee Thennal Power cal availability and cost of sorbets, likely Project cost, and perfornance Brown, AL., R.A. HEndmarsb, and G(T. Macdonald- 1991. "Environmental Asses- 4 Examination should be made of the possi- sment Procedures and Lssues in the Pacif- bilies of replacing the Mawdlla coal uits ic Basin and the Southea Asia Region,7 by diesels Either convenonal oal (o Environmentl Impact Asement Re- PFBC) might then follow in 2001 at Trim- view, I11, pp. 143-156. comalce In any event, unless environmen- ApfroachJorModellingNaonal Enerxy tl studies are begun very soon fir a Sytems. Argonne, Dlinois: Argonne Na- Mawela site, a south coast coal plant can- tional Labomtory, Enviromnental Ases- not be realized by the late 1990s (as re- snt and Information Scinces Division. quied-yhe99CEgneiopln) Cowl on Economic Priodties, Jobs and Energy. quired by the 1991 CEB generation plan). 1979. The Employment and Economic 5. Fmally, it seems clear that the most urgent Impacts of Nuclear Power, Conservaon priority for environmental study is in dte andOterEnerV Optons NewYork area of thermal plume modelling. The sort o E 1989 Inco Risk and Un- of detailed analysis provided in the BVI certinty ui Power System Plannin& En- EIS fir Trincoimalee, for once-through ergy Series Pap #17. Washington, D.C: cooling at the Clappenburg Bay-French idusty and Energy Department, World Pass site (SITE 2A) needs to be extended, CrOUit E., and HF Merrill. 1992. The Trade- at a minimum, to SITE 3A on Koddiyar Strategc Appro Bay, and to one or two representative sites to Power Plarning. Washington, D.C: on the south coast The modlg should Industry and Energy Department, Wold be done not just for once-through coolig, Bak -bwn, J, L Talbot, and G. Le Moigne. 1989. but for a range of options to include ocean Dans and the Environmenr: Consider- outfills, with and without dtffuser systems, aeton, in World Bank Projects. Techncal and for cooling towers. Cost esimates for Pape #10. Washington, D.C.: World the various ns should also be devel- B OFem:ll(;.C. April 1978. Coal, Economics and the oped. Only wiith this fixnformaon in hand Ferrell, Environment: Tradeoffs in the Coal- can the tade-offs across the vanous siting £Jecbic Cycle. hternational Institute for options be realistically assessed. Applied Sysems Analysis (IASA) Re- port PP-78-2. Viennat IASA. 3A good model for whalt seems idicated for Sri Lanka was recetly completed in India, where a comprehnsive evalu- ation of DSM options was imdertaken as part of the World Bank-U.S.AMD Power Sector Revicw Ihe options examined included a divesity of meamr fin the reailitatio of agriculmul pumpsets to high-efficiency indusia motors Clearly the efflicac of differet measmes wil vary coniderably fom county t country, dependig on the devices used in agricultural and in s sdcto See Office of Energy, USATD 1991, for detils of the lndia analysis. 228 Giarnpaoli, D. 1988. Policies and Strategies of the OKO Institute. 1989. 'Umweltanayse von Energie Inter-American Development Bank for Systemen: Gesamt-Emissions-Modell In- the Energy Sector and the Environment. tegrierter Systems (GEMIS)," Darrnsadt, Washington, D.C.: Inter-American De- Germany: Ministry for Economy and velopment Bank. Technology of Hessen. Goldman, C., et al. 1989. Least Cost Planning in Overseas Development Administation (ODA). the Utility Sector: Progress and Chal- February 1990. Social and Environmental lenges, Report LBL 27310. Berkeley, Issues at the Samanalawewa Hydro Elec- California: Lawrence Berkeley tric Project. London: ODA. Laboratory. Pace University Center for Environmental Legal Hobbs, B. August 1979. Analytical Multi-objective Studies. 1990. Environmental Costs of Decision Methodsfor Power Plant Siting Electricity Report. Ncw York: OCeana NUREGtCR-168. Washington, D.C.: Publications. U.S. Nuclear Regulatory Commission. Raskin, P. 1985. -LEAP, A Mficrocomputer Sys International Atomic Energy Agency (AEA). tem for Energy Planning," in AMcrocom- 1991. Senior Expert Symposium on Elec- purer Applications for Energy Planning tricity and the Environment Key Issues in Developing Countries. New York: Papera Vienna: IAEA. United Nations. Keeney, R-L., and KI Nair- 1977. 'Nuclear Siting Stone & Webster Management Consultants, Inc. Using Decision Analysis," Energy Policy 1989. Least Cost Planning Using 5(3): 223-231. EGEAS/COMPASSISTAFF Boston, Mas- Keeney, R.L., and HI Raiffa. 1976. Decisions with sachusetis- EGEAS Capabilities Maual, Multiple Objectives. Preferences and April 1989, also available from Stone and Value Trade-offs. New York: Wiley and Webster. Sons. UNDPIWorld Bank Energy Sector Management Markandya, A. 1990. "Environmental Costs and Program. July 1983. Sri Lanka Power Power Systems Planning," Utilities Poli- System Loss Reduction Study. Report cy, 1, pp. 13-22. 007183. Washington, D.C.: World Bank. Meier, P., and L Ruff 1979. The Spatial Dimen- U.S.AID. Octaber 1988. Energy and Environntent sion of Regulatory Cost Assessment An Appraisal of Energy Assistance Washingon, D.C.: National Commission Strategies for Miniming Environmental on Air Quality. Impacts in Developing Counties. Wash- Meier, P., and M Munasinughe. 1992. Incorporat- inglon, D.C: Office of Energy, United ing Environmental Concerns into Power States Agency for Intational Sector Decision-making: A Case Study Development of Si Lanka. Washington, D.C.: World . Februay 1991. Costa Rica: Power Bank Sector Efficiency Assessmt Washing- Mlntzer, I., et aL July 1990. Environmental Ex- ton, D.C-: Office of Energy, United ternaiiy Data for Energ Technologies. States Agency for International College Parlk, Maryland, Centcr for Development Global Change. University of Mayland. . June 1991. India Power Sector Review: Munasinghe, bi 1990. "Energy-Environmental Is- Demand Forecasting and Investment sues and Policy Options for Developing Planning. Washington, D.C.: Office of Countries,- Proceedings; International Energy, United States Agency for In- Conference on Energy and Environment, ternational DevelopmcnL Bellagio, Italy. Delhi, Idia: Tala Energy Woodward-lyde Consults, Inc. 1991. Decision Research Institute. Framework for Technology Choice: A Nanthakurnar. J. 1990. -Generation Expansion Case Study of One Utility v Coal-Nuclear Planning in a Hydro Dominated System," Choice, Electric Powver Research Institube paper presented at the Fourh RCA Work- (EPRI) Report EA 2153. Palo Alto, CA shop on Energy, Electricity and Nucler EPRI. Power Planning Daejon, Korea. World Bank 1989. Environment and Develop- Natmal Resources, Energy & Science Authority of ment: Implementing the World Bank's Sri Lanka 1991. Natural Resources of New Policies, Development Commitee Sri Lanka. Paper #17. Washigton, D.C 8 National Economic Cost of Soil Erosion in Zimbabwe David Norse and Reshma Saigal Loss in productivity due to soil erosion has an impact at all levels of society. The cen- tral focus of the paper is to provide a systematic methodfor assessing the economic cost of land degradation, and parlicularly soil erosion It descnbes a methodology developed in the mid-1980s to assess Fhe national economic cost of soil erosion in Zimbabwe. Such cosf, it was estimated, could exceed 16 percent of Zimbabwe's agricultural GDP, equivalent to 3 percent of the country's total GDP. These estimates were based on the highly mgnificant relationship between soil loss and losses of nturogen, phosphorous and organic carbon The paper recognizes that the Zimbabwe case stuy suffers from some limitaions. They include whether national economic costs can be reflected by the equivalent cost of mu- trient replacement because local fertilizer prices are only indicative of the actual cost In addition, the link between the loss ofplant minents and loss ofproduction is not established A more complete analysis of the economic cost of soil erosion would need to consider the levels of loss under d[fferent soil conditions, rain regimes and crops grown, as well as natu- ral regeneration of the soil The paper goes on to describe a later study that attempts to fill these gaps from an analysis of soil nutrient depletion in 38 sub-Saharan African countries for 1983 and 2000. While the results of the two studies were broadly similar, many difficul- ties remain in establishing a quantifiable relationship between soilproductivity and erosion. 230 Loss in productivity due to erosion has an Southem Rhodesia. The analysis showed high- impact at all levels of society. At the farn lev- ly significant relationships between soil loss el, declining crop yields lead to failling profits and nitrogen, phosphorus, and organic carbon as a result of lower output. Farnc,s may thus losses from the experimental plots. These find- be forced to make changes in the mix of crops ings were then extrapolated to quantify the na- and the level of input use and in extreme cases tional cost of the erosion-induced nutrient loss even withdraw land from cultivation leading to (Stocking 1986). marginalization and increasing migration to ur- This paper reviews the approach and find- ban areas. Off-site consequences place more ings of the study, broadly following the CIDIE pressure on the local environment in tenns of prescribed format. It is divided into four parts: sedimentation and silt that can clog up irriga- the first discusses the global and the conceptual tion channels and lower the water storage ca- context and sets out the basic problem, the se- pacity of dams, thus incrcasing expenditure to cond summarizes the methodology used; the governments for infrastructure and conserva- tiird describes the physical and country con- tion measures. text and highlights the main findings of thc The central focus of this paper will be to study; and the final part assesses both the lim- provide a systematic method of identifying and itations and the relevance of the methodology, ensuring the economic cost of land degradation suggesting a number of rcfinements. and more specifically that of soil erosion. It e- scribes a methodology developed in the Impact of Soil Erosion on mid-1980s to aszess the national economic cost Productivity of soil erosion in Zimbabwe. Much of the discussion on the land degra- The Food and -Agriculture Organization dation issue centers around te causes and (FAO) has been involved since the early seven- forms of degradation, the extent and rate of ties witl the issues of assessment and monitor- damage resulting terefrom, and the impact of ing of the problem of land degradation such damage on land producvity. There is, (Sanders 1991). According to recen esmates however, only limited quantitative information bsy FAQ, global loss of productive cropland available on rates of erosion in different soils due to soilerosion and degradation is es td and ecosystems, and much of it is largely anec- no be nearly seven million hectares annually dotal. The reason is mainly the complexity of (FAO 1991). the underlying relationships. C.urrent literature, The principal objecfive of the Zimbabwe while replete with broad estmates, does not case study undertaken in 1986 for-FAO's Soil whilt a systematic concaetual framework for Conservation Programme was to collect, col- a coherent analysis of th-e various elements m- late, classifr and anayze data available in or- volved in the physical and ecological processes der to quantify the impact of erosion. The underlyinglanddeion. underlying policy rationale of the study was to The relationship between land degradation bring home to decisionniakers the real but and production is difficult to quantify, because "hidden" cost of not investing in soil conserva- data are Laking on the link betwee ero-ion, tion, which it was estimated could exceed 16 plant production, and economic returns. To es- percent of agricultural gross domestic product tinate the economic sigiificance of soil ero- (GDP) and 3 percent of total GDP. sion, it -.ould be necessary to develop a model The research was carried out by Michael of the physical dimensions of erosion, link Stockg' in collabotion with the Institte of these to changes in crop production and falnm- Agricultural Engineeinsg, Mnstr of Agred- ing systems, and, finally, value these changes. dure, Zimbabwe. It was basdc on data gathered To date such research on the impact of &o- durig the late 1950s and early 1960s in the sion on crop yield has been largely confined to course of a- seres of experiments on soil loss, the United States, Canada, and Australia, runoff and nutrient losses conducted at the Henderson Research Station in the then Dr. Michael Stocling, Schloo of Development Studies, University of East Anglia, UK. & EvwireomentMangement: An EconomicProposolfor Urugmap . - 231 where it tends to focus on field plot trials and There is little agreement, however, on ex- ardficial de-surfaLcing to simulate the removal actly how productivity is reated to erosion, or of successive layers oftopsoil. on the quanttative impact of erosion on yields. Soil erosion is a major form of environ- In part, this arises from the difficulty of defin- mental degradation in the developing world, es- ing frtility as well as the difficulty of conduct- pecially the tropics (due to intense rainfll, and ing controlled experiments to identify and high temnperature, for example). The build-up measure erosion-related yield changes. In this of population pressure and the direct depen- respect the most substantal tropical yield data dence on natural resoures for livelihood fur- is from Nigeria (Lal 1989) where, experiments ther accentuates the natural process of erosion on Alfisols showed that mechanical removal of (through deforestation, overgrazin, and so the top 10 centimeters of the soil resulted in forth). yield declines of 73 percent for maiz. For dcveloping countries there is a virtual Much of the research on assessmeint of absence of tropical experimentl data relabn land degradation concentates on the physical soil erosion to yield losses. Another problem is aspects, such as rates of soil loss tuDgh wind the disparity of exrapolating from the U.S. and water erosion. Hence, the 1986 Zimbabwe temperate experimental data or empirical mod- study represents a significant atmpt in a de- els derived therefrom. Notithstanding these velopmg country to focus on the issue of $utri- liMtations, a stdy conducted by W. Magmth ent depletion from of soil erosion. and P. Arens (1989) in Java established the to- The approach adopted by Stoking is how- tal on-site and off-site costs of deciinig soil ever not the only one available. At te co.n- Productivity at US$340 million to US$406 mil- ceptual level four possible techniques can be lion or 0.5 percent of total GDP. More recent- identified. 4y, for Zimbabwe's neghbor, Malawi, on-site The Productiviy Approach considers the losses from soil ermsion have been esmad to cost of production foregon due to erosioin- be in the range of 0.5 to 3.1 peroen of GDP induced productivity decline, summed for the (Bishop 1990). whole country. If national ctiates of soil ero- Table 8-1 provides an overview of global sion costs are to be mean l, however, sev- soil degradation and excessive soil loss. Of the cral qualifications need to be made What is 4,700 million bectares of agricultural land, production in the erodd case mesred some 900 million (17 percet) are moderatey against? Is it the Eam management sition and 300 million (6 percent) are strongly de- prvailing at the ftime witout declining yields gaded (Oldeman, Hakkeling and Sonbrock (that is, static rather fing production)? 1990). The major man-made causes are defor- Or should the benchmark be an improved man- estation (120 million hectares between 1973 agement system that may include several non- and 1988), folowed by misnagemen of ar- conservation benefits (such as fuelhvood and able land and overgrazing. The most fieqaut crop residues)? The labtt would be appropriate type of degradation is water erosior, followed for a conservation project but not for assessing by wind erosiorn national costs of soil erosion. Erosion lowers soil productivity through a The Land Value Approach has erosion- variety of mechaniSmS, notably the loss of soil iduced losses hereby reducing its value capi- moistare-holding capaci, restriction of root- tlizd into the price of agriculal land. This ing depth, and loss of soil nutrients and organic approac, however, may be more applicable in mater. Ei-osion selectively removes the finer a developed county contet in which rural real and more fertile particles in a soil, thereby re- estae markets are well established. ducing soil fertilty. Eroded soil particles con- The DefensiePreventive Expfetnures tain up to twelve times the concention of technique considers the cost of avoidiig er- nutrients as the original soil. Erosion also af- sion on a national scale by estimating how fiets soils physically m the fbrm of crusting, much it would cost to modify or eliminate te compaction and increased strength with problem. "deletrious" consequc on plant growth- 232 Tablc SI: Soil Degradation by IWe and Cause (classilfcd as moderately to excessively affected) Water erDsion Chemical Physical Total Regions (in mllions of hectares) Africa 170 98 36 17 321 Asia 315 90 41 6 452 South America 77 16 44 1 138 North and Central 90 37 7 5 139 America Europe 93 39 18 8 158 Australasia 3 - 1 2 6 Total 748 280 147 39 1,214 Major causes (in percent) Deforestation 43 8 26 2 384 Overgrazing 29 60 6 16 398 Mismanagement of arable 24 16 58 80 339 land Other 4 16 12 2 93 Total 100 100 100 100 1,214 Source: Adapted from ISRIC7I?JNEP (Oldeman, Hakkeling, and Sombreok 1990). Replacement Cosr is the approach adopted * Compilation and refining of the empirical for the Zinbabwe study. This estimates the data collected from expenmental plot cost of replacing the degraded "productive as- * Analysis of data to determine the quantita- set." considering it as an imdicator of the dam- tive relationships between soil loss and age incurred by degradation. losses of nitrogen, phosphorus and organic Methodology of the Study * Extrapolation of the finding to the domi- nant farming systms of Zimbabwe, quan- to quantfir the impact of erosion by measunng the nutrients lost in the eroded sediments and tfymg soi loss at the national level of - . . ~~~~~three major nutnents and computing the the cost of replacing them. The underlymg polh- thre major nutortient ean comdmto te- cy rationale of the study was to bring home to place t lost oftrte decisiomnnakers the real but hidden cost of not investng in soil conservation. The methodology had three main components: & Environment MangementiAn Economic Proposalfor Urgrnyv 233 Data Base * Applying regression equations to obtain the An intensive erosion rescarch program in loss of nutrients corresponding to the esti- Ziimbabwe was instituted during the period of - meates of erosion the Federation of Rhodesia and Nyasaland be- * Summing up of each farming group to ob- tween the years 1953 and 1964. The data base tain the total loss of nutrients from is drawn from this source and covers 400 plot Zimbabwe years of experiments conducted at the Hender- * Calculating the cost of these nutrients as son Research Station. calculated from current fertilizer prices This study represents a notable attempt at The above sequence for calculating total classifying, documenting, and analyzing the cost to Zimbabwc of the nutrients removed by full record of soil loss and run off, storm by erosion is diagrammaticallv shown in Figure storm and annually, for the many treatments 8-1. and for the fitr soil types used in the experi- ments conducted by Stockdng during 1970-76. Main Findings of the Study A new model of soil loss estimation, SLEMSA The main findings of the studv could be (a modification of the Universal Soil Loss summarized as follows: Equation) was constructed for the specific field 1. Ststiay significant relationships were conditions of Zimbabwe. established between soil loss and the losses The raw data consisted of sludge measure- met taken from the collectin tanks on the o n crosion pIots. This gave records of nutricnt bon from experimental plots. Statistically erosion plots. This gave records of nutrient valid rates of nutrient loss for different 1ev- concentration in percent for nitrogen (N) and orgamc carbon, and parts per million for phos- els of erosion were calculated. phorous (P)_ These three nutrients represented 2. Variations in mean annual losses of soil. the major quantitative impact of erosion on soil nitrogen, organic carbon, and phosphorus chemistry. Furer details of the methodology were dependent on soil type, crop, a-id together wiLth a sumianes arc available in yearz The two most important variables in the FAO report of the study (Stocking 1986). explaining this effect were the rainfall pat- tem and the crop type. Analysis of Quantitative Relationships 3. The erosion process is considerably selec- between Erosion and Nutrient Loss tye i removig nutrents from the soiL Through correlation and regression analy- The ratios wvere highest in arems where mn- sis of the refied annualized data, predictions off was highest, an important finding in were made as to losses of nutrients under given evaluating physical conservation measures levels of erosion. . .that are designed to detain soil but allow The agricultural land resource was run-off Extrapolating the experimental grouped into two main categories-comnercial plot findings to the four main frnning sys- lands (41.4 percent) and commnunl lands 41.6 tems of Znimbabwve. these variations in ero- percent)- Each of these categories was further sion rates were found: sub-divided into gazing lands and arable * Commercial grzing lands with low densi- lands. ties of livestock and relatively complete ve- The extrapolation exercise consisted of the getafion cover had a low rte of erosion. following main steps: - Catem* Commercial arble lands demonstred -* Categorizing of land area into broad firm- similar moderate rates of erosion. This can ing groups and systms by degrees of ero- be attributed to the mainteance of physi- sion hazard cal conservation measures in nearly all * Estimating the rate of erosion on each cases (such as contour bunds, terraces, and system waterways). * Communal arable lands. though adopting similar physical conservation measures, 234 Farming Systems Commercial - Grazing land Communal - Arable land Estimated Rate Erosion Each System Soil loss/nutrient regression Loss N.P,C for each system x Area each system Zimbabwe total losses of *nitrogen *phosphorous *organic carbon x Fertilizer prices Total Financial Cost as Measured by Nutrients in Fertilizers Figure 8-1: Flow Diagram Illustatiug tde Calcldation or Mhe Fancal Cost of Exsin ., _~~~~~-- & Enriroement Mangement: An Economic Proposal for Uruguay 235 Table [-2. LoSses of Nutnents and Organc Carbon from Zimbabwe's Soils (thousands of tons) Farm system Nirogen Orgatnic carbons Phosphorous Commerciail grazing 57 454 5 Commercial arabIc 19 148 2 Communal arable 134 1,324 20 Communal grazing 1,425 13,679 209 Total losses 81,635 15,605 236 had considerably higher rates of erosion year (at 1985 market fertilizer prices and rates compared to commercial areas. This was of exchange). For te arable lands alone, where largely due to their poor state of mainte- there is the greatest investment in tenns of food nance, open access to grazing, and inade- production and fertilis, the financial cost quate vegetation cover. amounted to US$150 nillion. = Communal grazing lands were the most On a per hectare per year basis, the fman- eroded of all. Livestock numbers exceeded cm cost of erosion was found to vary fom - the crying capacity of the land. This US$20 to $50 on amable lands, and US$10 to combined with a succession of droughts $80 on grazing lands, according to the level of denuded the vegetation cover. erosion. On average, 1.6 million tons of nitrogen, If these estimates are correct, they repre- 15.6 million tons of organic matter, and 0.24 sent a substantial percentage of the financial million tons of phosphorus are lost annually by reum per hectare of amable land in Zimbabwe. erosion (Table 8-2 provides details) The ar- This can be illustated by hybrid maize, which able land alone lose 0.15, 1.5, and 0.02 mil- is grown in the relatively better managed corn- lion tons respectively. These nitrogen and mercial lands. The 1984 esimates of variable phosphorus losses friom arable land were about costs and returns on one hectare of maize pro- three times the level of total fertilizer applica- duction (Wmdmill Pvt Ltd. 1984) are illus- tion in Zimbabwe in the 1984-85 season, and tratcd mi Table 8-3. they do not include losses of nutrients dissolved Thus, a US$50 per hectare additional cost in runoff water. of rcplacing nutrient loss would represent be- To monetize the physical losses of nitrogen tween 13 percent to 60 percent of the gross re- and phosphorus, the equivalent cost of fertiliz- tumLs per hectare of amable land under maize ers containing these nutrients was calculated. production. Although these estimates must be For the total nutrient losses fromn the four fnn- considered provisional for reasons considered ing groups, the financial cost based on mineral below, they imply a massive hdden cost on the fertlizer replacement was US$1.5 billion per 236 Table 8-3: Annual Yield and Return of Hybnd Mai7z Land Type Lowyield Average yield High yield Y-eld (kilogram per hectare) 4,000 6,000 8,000 Income (at 180 Z$ per ton) 720 1,080 1,440 Variable costs (Z$ per hectare) 588 700 812 Gross margin (Z$ per hectare) 132 380 628 or (US$ per hectare) 82.5 237.5 3925 economy of Zimbabwe and an undermining of in the original soil, a hypothesis confinned by its resource base for the future. the findings (as to enrichment ratios) of the Physical and Country Context Zimbabwe study) (b) acidification, which in Physcal and Country Context causes aluminum toxicity in soils; (c) de- Results creased water availability due to loss of water- Only part of the physical processes by holding capacity, reduced soil depth, and other which soil crosion afficts soil prductivity can factors; and (d) loss of the soil's structural sta- be captured by the empirical data. The impacts bility through increased bulk density or loss in can occur on and off site, and hy can be organic matter. maniffested in the use ofthe soil, and in changes These changes in soil characteristic have in soil charaaeristics significant implications for crop yields, and the The impact on the use of soil may, for ex- costs of sustaing present yields as more and ample, be a combination of (a) crusted and more fertilizers are requbred to make up for the compacted soils that are too hard to be plowed eroded nutrents. However, resarch efibirs so using animals, resulting in delayed planting and far have not been successful in quantifying fll- lower yields, excess munoff, and associated ly these diverse but interrelated impacts. moistre deficiency; (b) forced changes in land Because the basic data consists of nutrient use and frmng systems occasioned by lowsr and carbon losses from Zimbabwe soils and levels of productivity, and (c) off-site damage measrefments are imrited to assessing the em- such as siltation and sedimentation of drainage st effects an nutrent availability, the study's channels and reservoirs, and inundation of fer- c estimates are incomplete. tile top sodls. Lmttoso h td The impact on soil characteristics and Limitations oftheStudy 1roductive potential commonly takes the form Stocking's study undoubtedly represents an of (a) nutrients consistently lost in the eroded important step in quantification of soil erosion sediments in proportional amounts greatertan impact Te analysis, however, is partial in Environment Mangeament: An Economic Proposal for Uruguay 237 nature and based on a number of assumptions If data were available to assess erosion and generalizations which need refinement. impact on crop-yields, the Stocking study could It is questionable wvhether financial costs have been extended to incorporate cost-benefit can be quantified sufficiently accurately by the analysis. The economic costs of nutrient loss equivalent cost of nutrient replacement. Inor- could be expressed in terms of potential loss of ganic fertilizers supply nutrients in a different production. This could then be compared with chemical form, and eroded soil undergoes the costs of soil conservation or other remedial physical and chcmical changes that make the actions. Such an approach is particularly rel- soil less efficient in supplying nutrients to vant in assessing the economic justification of plants. Fertilizer prices, therfore, can only be environmental protection projects. indicativc of the actual cost. Soil erosion on arable lands is a fimction of Another limitation of the study is that it several variables, such as rainfall, management does not establish a link between the loss of nu- and conservation practices, and land use sys- trients and the loss of production, thus, making tems Other factors ffiat should be distin- it difficult to assess tne profitability of their re- guished from erosion include climate placement. In f-act, the fertilizer cost may not variability, salnty, and str-ctural collapse, Ibr be a good estimte of the economic costs of on- example. Erosion rates alone may also be poor site soil loss. The true measure of is the loss of indicators of productivity loss, particularly net productivity. One would need to take into ac- national losses. Erosion rates may hide the fact count organic matter losses, downstream or that soil is redistributed within a catchment and off-site damages (sedimentation, for instance) is not necessarily lost to agriculture. For some in addition to the on-site productivity losses to soils, even improved management techniques obtain a total estimate of the costs of soil ero- may restore their productivity (generally soils sion. Furthermore, damage firom naturally oc- that have a concentated distrbution of nutri- curring erosion is likely to be much greater ents in the topsoil or shallow rooting depths). than that produced in artificial experimnts, A more complete analysis of soil erosion mainly because natural erosion tends to remove impact would therefore need to consider the prel_entially the most productive constituents natural replenishment of the organic nutrient ofthe soil. content of the soil through manure, deposition, Stocking's analysis also does not take ac- and sedimention, as well as differentiate count of the social cost of worsening food se- losses under different rainfall regimes and cuity and replacement of food losses incurred crops. A number of such possible refinements by erosion with food inports and ad. Such an are considered below. approach could interalize in part the interge- erational extraity incurred by erosion, since The Winand Staring Centre Study it is future generations who would be burdened A later study supported by FAO with debt repayment for food imports. (Stoorvogel and Smoling 1990) and undertiaken The data base appears to be biased to- by the Winand Staring Centre of the Nether- wards tSe better protected soils, and consists land developed a methodology that attempts to maily of low values of soil and nutrient loss. accomplish some of these aspects. It assesses Stocking moreover seems to down-play the sig- the state of agricultural soil nutrient depletion nificance of variations in nutrient losses for dif- in 38 sub-Saharan African countries for 1983 fercrnt crops, in different years and between and the year 2000. The analysis measures the soils, despite extensive evidence of such vaia- net removal of the macronutrients nitrogen, tions. Lal (1989) and other researchers, for in- phosphorous and potaium from the rootable stance, have emphasized Jiffirerences between soil layer using production figures for 1983 soil types. The only excaption in the study is and projected production figures for 2000 pro- tde conclusion that 'the higher rates of erosion vided by FAO by crop and country. These fig- on the nutrient-poor granite sands yield rather ures were disaggregated ito six landhwater lower nutrient concentrations." classes with diffbrernt rainfall and soil moisture regimes: low, uncertain, and good rainfall 238 Table 84: Annual Nutrient [asses fom Aable Land ocking Stud Wand &urinr SLudy Adjusted for Overall counny Erosion Loss output factors, balance N (thousands of tons) (-)153.0 (-)99.5 (-)662 (-)127-5 P 2, 5 (thousands of tons) (-)22-O (-)37-2 (-)32A (-)19_4 Area (millions of hectares) 3.0 - - 4-1 Estimated financial cost (Z$ millions)2 240.0 2632 204A4 (US$ millions) 150.0 164.5 131.9 127.8 1. Adjusted for compensatory nutrient accruals due to manure, deposition, nitrogen fixation, and sedimentation, but not for mineral fertilizer application. 2. Prices as per Stocking study of Z78 per ton for elemental nitrogen and Z$4,700 per ton for elemental phosphorous in 1985. (One US$ Z$16.) areas (rainfed agricultr); problem areas; and The results of this study demonstrated that naturaly flooded and irrigated areas. Data nutrient depletion is already see in sub- from 1983 and projections for 2000 were used Saharan Africa, and is projected to get worse. to estimate frrtilker use by county and by Using this more elaborate methodology, the crop. arable land (4.! 1 million hectares) nutrent bal- The factors deteminng the nutrient flow ance loss for nitrogen and phosphorus for Zim- into and out of the soil were defined and quan- babwe for 1983 were estimated to be 127.5 tified for each land/water class taking into ac- thousand tons and 19 thousand tons count variables such as rainfall, soil fertuity, respectively. croppig pattem, farm management level, fr- In contrast, the Stocking study placed the tlizer and manure application, crop residue average annual loss of nitrogen at 153 thou- management and rate of erosion. These were sand tons and that of phosphorus at 22 thou- incorporated into a land use systems model. It sand tons, since it did not take into account the was assumed that each land use system was output ictors which offset the losses of nutn- homogeneous and thus became the unit for the eats (able 84). calculation of the nutrient balance. The nutrient The Stockng shtdy apparently may have balance of each land use system was assumed overestmated the costs of soil erosion by al- to be govemed by five input and five output most 20 percent due to the neglect of natual factors, nutrient mputs. Although the figures in both The inpu factors were nuneral fertiizers, studies are remarkab!y similar, the Staring manure, nutrient deposition by dust and rain, Centre study incorporates a nmnber of useful nitrogen fixation, and sedimentation. The out- methodological refements that provide an im- put fictors were harvested product, crop resi- proved basis for assessing nutrient losses. The dues, leaching, gaseous losses, and erosion. Stocking study, however, provides more & Environment Mangemneat: An Economic Proposafor Uuy 239 reliable estmates of soil erosion losses per se. hectare per hectare ompared to 15 tons per Both approaches, however, only consider the hectare for commercial arabie lands.) cost of nutrient losses. Many difficulties thus remain in indicatig Secondakry and Off-site Costs a clear-cut relationship betwee soil productiv- ity and erosion. Experments in this area are As pointed out in the previous section, site-specific. General trends in yield loss are there are a number of other costs to be brought difficult to evaluate because experiments rarely into the calculation. Some examples are soil specify initial yield, stage of erosion, and de- crusting and compacion, soil acidification, de- tailed soil ype, especially at the national level clining response to mineral frtilizer inputs; re- In both developed and developing coutries duction of soil watr availabiliy (and the are severemliits to the availability of reli- consequently increased yield variability); great- able data on the rate at which soil erosion takes er competition from aggressive weeds; changes place on diffient soils, the impact of this ero- m land use and farming systems; and off-site sion on crops, fianmer responses, and nonfarm damages factors tha determine the social losses caused Tbese additional costs cannot be deter- by erosion. mined at the nafional level with any precision In conclusion, it is becoming icreasingly because situations and responses are very het- apparent that thre costs of soil erosion in the erogeneous. Nonetheless, they could be very case of Zimbabwe could be well in excess of significant Soil crustng and compaction, for 16 percent of agricultual GDP and 3 percent example, could delay land preparation and of toul GDP since these are costs of nutient pling resuling in crop yield losses of 10 to loss fiom arable land alone. It is citically in- 20 percent If this problem was prevalent on portalt, therefore, to proceed with the refine- just a fifth of the arable land area, the cost ment of the methodologies for castbenefit would be ofthe order ofZ$30 to $60 million. analysis and particularly for soil productivity Similarly, off-site damage resulting in the and yield loss relationships. FAO is, conse- siltation of irrigation channels and reservoirs, quendy, expanding its activities in this area. such that their efifetive woricng lif is reduced by 25 percent, could involve multimilLion dollar Bibliography costs. Empirical Corroboration of Production Buisop, Jbsbma. 1990. Srhe cost of soil erosion in Trends Malawi' chaff report for the World Bank nMawi Country Operations Divisionx Nutrient depletion caused by soil erosion Washington, D.C.: World Bank. no doubt has an adverse impact on levels of November. production and productivity. However, this ad- Food and Agricltual Organizaion of the United verse impact can be offset by countervailing Nations (FAO)NetherIands Confce on Agriculture and the Environment factors, such as introduction of improved tech- f1991. 1Bckgroud document No. 2. April mcal packages mid support services, agricul- 15-19 tural intensification, or simply by favorable LaL 19X9. Effects of Soil Ersion on weather conditions. Consequently it may be CLRp Pr uts in CilC Critical Re- difficult to corroborate the expected loss in views in Plante Sences, 5:4, pp. 303-367. production due to soil erosion with the aid of M , William, and Per Arens. 1989. "The actual empirical data on levels of production. Cost of Soil Erosion on Java: A Naturl In the case of Zimbabwe this can be seen i the Resource Accoundng Approacbf World context of conumnal areas, which according to Bac Eiomnment DepatUWt Woddng testudy were affected quite severely Paper No. 18. WahDington. D.C.: World the Stocking std e ffce ut seceyBn-August. by soil nutrient losses. (Jbe predicted rate O Odeman, LA, ItT.A- Hakkehng. and W.G. Somn- erosion in the case of cormmunal arable lands broek 1990. World map of the stas of was estimated at approxinately 50 tons per hn-anced soil degon: maps and explanatozy note. In cooperaton with 240 Intcrnational Soil Rcfcrencc and Informa- Stocking, Michael. 1986. The Cost of Soil Erosion tion Ccnter, Wageningen. United Nations in Zimbabwve in Terms of the Loss of Environmcnt Programme. Nairobi. Wi- Three Mlajor Nutrients. Rome: Soil Re- nand Staring Ccnter. Food and Agricul- sourees, Management and Conservation turc Organization of the United Nations, Service, Food and Agricultural Organiza- Intcrnational Institute for Acrospacc Sur- tion [of the United Nations] (FAO). vey and Earth Scienecs Stoorvwcl. J.J., and Ema Smaling. 1990. Asses- Sanders. D.W. 1991. 'International activities in sment of Soil Nutrient Depietion in Sub- assessing and monitoring soil degrada- Saharan Africa 1983-2000. Winand Star- tion, paper presenLed at the Intcrnational ing Centre for Integrated Land, Soil and Workshop on Assessmcn and Monitoring Water Rescarch. WageningenlFAO of Soil Quality- Emmaus. Pennsylvania: Romc. Rodalc Research Center. July. Windmill PvL Ltd. 1984. 1984 Zimbabwe Profta Planner. Harare, Zimbabwe. 9 Environmental Management: An Economic Proposal for Uruguay Roberto Aifredo Recalde This paper provides an economic anaIysis of the man national environmental prob- lems of Uruguay andproposes a set of specific economic measures to deal with those prob- lems. The paper (1) shows that the administration of most envirornental resources of the countly have been in the hands of the state and followed a command-and control approach; (2) discusses an economic strategy to environmental management and proposes its use as a complement to the existng methods; (3) reviews the main enviromnental problems of the country, as identified in the National Environmental Study; (4) assesses the importance of these enviromnental problems and identifies their causes through economic analysis; (5) proposes a number of economic-based mechanisms of environmental management for incor- poration in the overall economic policy of the country; and (6) formulates, through an appli- cation of those mechanisms, packages of specific economic measures to deal with those activities which create the most pressing national level environmentalproblems (e.g, erosion caused by intensive agricultural production; indusftial emissions; water resource manage- ment; erratic natural ecosystems management; commercial exploitation ofpublic goods; hu- man occupation of coastal areas; weak management of scenic and cultural urban resources; and uncontrolled disposal of urban garbage). 242 This document provides an economic ignored (even by governmental organizations) analysis of the main problems Uruguay has or cleverly evaded. Second, government usually with the management of its natural and cnvi- lacks funds to adequately to monitor com- ronmental resources, and a proposal of con- pliance with the law, maintain and improve the crete economic measures for incorporation into resources involved, develop suitable technolo- the country's economic policy or environmental gies, and conduct public awareness campaigns, action plan. (It does not involve specific invest- for instance. Third, the technical standards, ment projects or legislative, institutional, and quotas, and bans are economiically inefficicnt educational measures, which have been dealt since they usually bring about underutilization with separately by the National Environmental and deficient allocation of resources (that is, Plan (NEP). the conservation of resources is achieved at the As in most other Latin American countries, expense of unnecessary loss of production or Uruguay's management of natural goodst is well-being). Fourth, the existence of quotas and basically in the hands of the government, which bans, and the concession (generally gratis) of essentially follows a "command and control" licenses, create situations conducive to favorit- approach, as it is called in the lieature. This ism or bribery. Finally, adoption and enforce- approach, also refbrred to as the "regulatory" ment of technical norms demand expensive or "normative" approach, seeks to regulate the information and expertise on ecological, tech- behavior of individuals and institutions utiliz- nological, and sanitary matters that the govern- ig the country's resources. It does so by laws, ment does not have. norms, and rules (that set maximum levels of Although the regulatory approach is valid pollutant emissions, prohibit or restrict use of and valuable in the formulation of environmen- resources, adopt urban and ecological zoning, tal objectives and goals, it provides only par- and so forth), then enforces them.L tally for the insuments needed for their Other government responsibilities include implementation. To overcome this limitation, (a) implementing environmental education and various countries-especially the tralizd public awareness programs; (b) allotting per- ones-have begun to adopt economic istru- wits or licenses for the exploitation or enjoy- ments for the management of environmental re- ment of environmental resources (such as sources. These instruments have shown fishing, consumption of irrigaion water, and themselves to be particularly usefil for (a) as- hunting); (c) directly providing cerain services sessing the importance and priority of environ- that involve nawtral goods (such as potable wa- mental problems on the basis of an economic ter and sewage, maintenance of proteted calculation of their impact; (b) allocating envi- areas, collection and treatment of waste, and ronmental resources more efficiently and ensur- programs of technology development and tech- ing their sustiable use; (c) providing a nical assistance on environmental matters); and system of economic incentives that makes pri- (d) executing projects or works for the mainte- vate individuals and institutions cooperate with nance and rehabilitation of natural goods. the administration and conservation of those re- The experience of otier developing coun- sources; and (d) generating the financial re- trics, and economic analysis, indicate that the sources for meeting the costs of opeation and rcgulatory approach has severe limitations investmt demanded by the management of when applied in isolation from other measures natutral goods. :irj.t, laws, regulations, and bans, are openly Ihis paper was prepared under the auspices of the Department of Regional Development and Environ- ment, of the Organization of American States (OAS), as part of the project 'National Environmental study." under an agreement with the Inter-American Development Bank and the Government of Uru- guay. The author's views do not necessarily represent those of the OAS. In this document, -natural goods" and 'environmental resources" refer to natural resources, ecosystems, natural ys- tans, biological resources, and other products of nature that meet the definition of an economic good: they are scarce 9. Environmental Management: An Economic Proposalfor Uruguav 243 In Uruguay, the application of economic assumed, in this regard, that the government instruments for the management of natural rc- wishes to achieve its objcctives of environmen- sources and ecosystems is practically nonexis- tal management without sacrificing those of tent. As a general rule, economic economic growth and equitable distribution. To considerations are not incorporated into the this end, the suggested approach includes mea- formnulation of environmental policy nor are en- sures for an appropriate valuation of the coun- vironmental factors considered in the fornula- try's natural resources and ccosystems, and an tion of economic policy, suggesting that the efficient allocation of those resources (that is, country could benefit significantly by adopting based on their real value or cost to society as a an economic approach that complements the whole). existing regulatory one. A proposal to achieve that purpose is set forth below. Problems Identified The problems of greatest economic interest Objectives of the Proposal are those of loss of the quantity, quality, or Based upon the above-discussion, the pro- productivity of environmental resources. Quan- posed economic measures should contribute to- tity or quality losses are economic when the ward three objectives: production or consumption activities that gen- First is the incorporation of economic cr.te crate them bring about a present welfare gain nia into the formulation of laws and environ- that is lower than the loss of future well-being mental policies, as well as envirownental from reducing the country's base of productive considerations in the fonnulation of the coun- resources. The losses of productivity are also try's economic policy. The proposed approach relevant from a conservationist standpoint: provides an initial step in that direction by their recuperation pernits not only increasing putting forward specific measures to be incor- economic production but also diminishing the porated into economic and environmental pressure on fragile resources (that is, lower in- policies. puts of them are required to obtain the same Second, environmentai regulations and pol- level of production). icies must be formulated and applied on the ba- sis of feasibiity (not just desirability as has been the case until now). The mere enactment The envirnomental problems identified as of environmental laws and the creation. of a part of the National Environmental Plan are specalized agency are not sufficient for imuple- primarily related to water and land: loss of soil menting environmental goals. Economic crite- through erosion, deterioration of water quality na and instruments are necessary for through contamination, and loss of water pro- formnlating realistic laws and regulations, ductivity through deficient allocation. making their application easible, and cornple- S E menting other instruments (such as institutional Soil Erosion reform, and education and public awareness Serious soil erosion is taking place in farm- programs). The proposal pursues that objective ing areas, especially on the country's richest through (a) the introducdon of explicit mecha- lands-the zone of influence of the city of nisms for financing costs of maintenance, re- Montevideo. The problem causes an annual placenent, and administration of enviromnental gross loss of 31 tons of soil per hectare, which resources; and (b) the incorporation of the pri- compares unfavorably with the 12-ton per hect- vate sector (through economic incentives) in are of the United States; the irreversible (or the conservation and administration of thos. net) loss amounts to 24 tons per hectare, which resources. is equivalent to a total for the country of 14 Lastly, obJectives of economic growth and Million tons per year. Its greatest impact has conservation of environmental resources should been (a) a significant loss of land productivity, be reconciled through the productive and sus- which in some areas has led to the impoverish- tainable utilization of those resources. It is ment and migration of the rural population; and (or Mill be in the fhture) and have alternative uses. 244 (b) a high degree of silting (estimated at 11 Rivcr, polluted with pesticides, fertilizers, million tons per year), which chokes rivers, and biological wastes irrigation canals, and reservoirs, thus raising * Significant loss of fish, in volume and the costs of water treatment, maintenance of number of species, in the Rio Negro, for waterways, and hydroelectric generation. The reasons still not understood (the projected resultant silting has in the past led to the emI- expansion of the thermoelectric plant of bankient and loss of the port of Paysandu on Candiota, Brazil, may aggravate not only the Uruguay River. the pollution of this river but also that of Water Pollution the Laguna Merin) Urban and industrial pollution of water- The country's water pollution has in- courses in certain coastal areas of the At- creased slowly but continuously and is reach- lantic seaboard rcsulting from heavy ing dangerous levels in certain places. Surfice population and tourism pressures, with sig- streams are receiving increasing runoffs of nificant problems in Maldonado and Punta sewage, solid household wastes, and industrial del Este effluents in urban areas, and residues of agri- cultural and livestock production (fertilizers, Inadequate Allotment of Water Resources pesticides, and biological wastes) in rural areas- ~~~~~~~~~~Although no adequate-figures are avail- Oftcn, the very slowness of the polluting able on the productivity of the water resource processn,asks its insidious natuTe. 'Mus ft in its manifold applications, indications are pradual contanuetionosthe Monature.vTh heo bthat its use is seriously inefficient; that is, its gradual cntaination of the Montevideo bay productivity and quality (and, consequently, (itrough the activity of the port and ly reiled s contribution to the growth of production industries) led to the practically irreversible and quality of life) are probably far below loss of what used to be the city's most highly that obtainable if allocations were based on Prized beaches, residential area, and 1O11d5Qt $nometric tech- number will be detnnined on the basis of an niques. In some cases, the appreciation stems stInzate of critical volumes of exploitation. from the elimination or improvement of sources Thus, for fishing, the critical volume will be of noise, objectionable odors, or pollution that whlch assures susminable levels of pro- (such as airports, swamp water, or industal duction; for the discharge of industrl wastes plants). in watrcourses, it will be that which fully uti- In all these cases, at least part of the fi- lizes the watrcourse's capacity for sound nancing of improvements should come firm waste dissolution. Income derived from license taxes based on the enhanced value (caused by sales should finance such managt costs as those improvements) of the adjoining proper- esmating critical volumes and monitormg the ties The advantage of this type of financing is amount and quality of exploited resources. tiat payment is made by the direct bencficiaries Temporary Incentives of the improvements and is not a net burden for thern since it is deducted from actually accrued In order to modify, civic conduct on a basis benefits. of noneconomic factors (moral, cultural, educational, and so forth), die goverment is SpecficMeasures already taking a series of important measures Presented below are the packs of eco- aimed particularly at environmental education nmimc measures that it would be advisable to and public awass. Those measures should adopt in order to solve each of the main envi- be complernented with temporary economic in- romnental problems discussed previously. At centives gered for reinforcing them this stage; however, thy are only Iilustraive of Most of these incentives will be ained at how to apply the six economic instruments dis- engagmg people in new environment-friendly cussed above to specific problems rather than activities (for example; the subsidy suggested being defiitive proposals. The latter would be for establishing associations of beneficiaries). elaborated in consultation with the specialists In general, these incentives should be tepo- and institutions involved and after verification rary; they should be removed as soon as mdi- of the feasibilty of applying each instrument to viduals learn whether the subsidized actvity is each situation. worthy. Consequently, it is important that the subsidy be adopted along with the schedule for 9. Environmenral Manragement. An Economic Proposal for Uruguay 251 Erosion and Silting dissipating resource (water or air) attains a The istig system of taxes on ai.ands, pre-established basic standard of maximum density. Obviously, the higher the concentra- which is based on their potential productivity, tion of industrial plants and the lower the dis- will be complemented by a mechanism of taxa- solving and dissipating capacity of the don or surcharges based on land use. This resource, the higher the tax. mechanism will tax crops and agricultural When the existing overall density is much practices that cause either higher erosion (for higher than the basic standard, a tmporary example, fruit and vegetable cultivation); loss schedule of short-term standards will be estab- of ecosystms (for example, rice cultivation); lshcd acoordingly. The affected industries will or negative effects on other properties (for ex- be consulted (through their associations, inso- ample, eucalyptus cultivation). For teciical as possible) on the design of temporary reasons, the tax will be based only on the stndards in order to identif technological losses occasioned outside the fairm being taxed offs and lower cost altnatives for For example, it will involve the costs of erosion adieving equivalent goals. When the existng and silting outside the farm (such as dredging density is below the basic standard, the tax of ports and reservoirs, and watertreatment for could be zero, in this case, however, it will be eutrophi;cation), but not the loss of soil produc- necessary to define and disseminate, as far in tivity within that fanr. advance as possible, the critical emission levels Temporary centves Will be mtro , beyond wich taxc levels will become positiv. such as free techncal assistance and conces- - -liv y in line wi the argnts sionary crets, for stimulating investment by put forward above, tradeable pollution pennits famers in soil conservation works and svii be issued wich may be frely bought and technologies. To guarantee ticr adopinon, the sold. The total amount of pollution authoridzed activities promoted with technical assistance through the mits (for each major cary of must be earmarked as much for soil conseva- pollutants) will have to be consistent with the tion as for significant production increase. polt ed stalndard (temporany or basic). Loan payments could be tied to actual produc- Urder this arngement, the market determines don m ubshdes d price of the right to pollute (with the tax, All subsidies-direct or auphei-rLated thiat value must be established a prion); the to frtlizers and pescides wil be withdrawn.m en t of pollution levels will be based Additonally, consideration will be given to a onthatprice. selective adoption of taxces on the use of pest- bitiy, basic polution standards will be cides and fertilizers based on an estimate of higher than those of industrialized counties their social impacts on human health, water Sinc eitber the amount of taxation requred to treatment costs, and o lrt reach those standards or the market pnce of The taxes prophosed on land use an on tradeable permits provides a sutable measure pesticides and r shall financ the n of the economic cost of atning them, it is of teporay subsdes and research on te rn- suggested definitive standards be estab- pact oferosion and siltng lisbe only after tha cost is known. Tlis pro- Industrial Wastes and Emissions cedure may requue a change in existing legislation on toxic pollutioa A tax will be levied on industrial dumping of effluents and solid wastes into watercourses Water Pricing and emissions into the atmsphere. Based on A n s thr, amount of emUmmats discbaAgnaitonlev-ysten of water chages will be the amount of contaminants discharged, its yev- established for selling water (or its services) at Of wil depend initally upon the exisng density the opportnity cost it has in di&rent applica- of contaminaion of the resource (air or water) tos h rc fti eorems,tee tions. The pnce of this resource must, there at the plant's location. Subsequently, it will be f a - X ~~~~~fore, adequately ref lect its scrcity and periodically adjusted on the basis of the results i a r c obtained and until the overall pollution of the mcaonate all relevant cit ts, sucr i as dhose oa obSind an urAl th oveall olluon o the mainang its quaIity or increasing the local 252 or national supply. -The design of local rate that region for adjusting irrigation water schedules (of residential and industrial con- charges and land use taxes. sumption in urban centers, and of irrigation in Given that lands with natural ecosystems certain rural areas, for instance) can remain in are generally privately owned and that prescrv- the hands of thc local watcr agency. But the ing those ccosystems is of grcater benefit to so- overall wholcsale" price for each major appli- ciety than to the individual oawner (especially cation must be established at the national level when alternative uses are more profitable to the on the basis of demand. supply, and aggregate owvner), incentives for such use will be pro- water balances. The central purpose of sound vided_ Therefore, the tax on rice cultivation water pricing is the reallocation of water at proposed will be complemented with tax credits both national and local levels, through cost- (or with zero tax) for vetlands use. The rice based charges, that transfer it from lesser uses tax/vetland credit gap can be self-financing (which is presently the case as the cost to con- and broadened until the current process of wet- sumners is low or zero) to more productive uses lands loss is reverted. One important advanage (which is the case when water is rationed of this mechanism is that the mere announce- among users). Local charges must include, in ment of a debate on its legislation will already addition to the basic price, the costs of operat- have the desired effect: the owners of wetlands ing. maintaining, and expanding the installed will tend to hold on to them in expectation of equipment and plants. the outcome. This would not happen, for exam- When a system of quotas is in place, as is ple, with the announcement of a possible land the case with irrigation for rice production, expropriation or a rice cultiwvation ban, since those quotas will be sold by competitive bid- owners would rush to dr'in the wetlands in or- -ding. The price obtained (which is an appropri- der to avoid such expropriation or ban. ate indicator of water productivity in that use) Simultneously with measures for water w.vill be the yardstick- for detenmining whether charges and tax credits, a program of gradual water for that activity should be reduced or government purchase of areas to be protected increased. will be launched. Given that protection of natu- In supplying urban potable water, the ral ecosym is of global (worldwide) inter- structure of consumption that results from the est, the funding for these purhase should be present svstem of consumption-based charges obtained as much as possible from intested will be. continuously monitored. The system outside sources (for example, with debt-for na- provides sufficient incentives (discouragig ex- tre mechanism). cessive or wasteful consumption) but may re- With establishment of the system of pro- quire periodic adjustments based on actual tected areasproposed by the National Environ- consumption pattems. The monitoring of the meal Plan, differential (foreign/national) fees physical quality of water must be accompanied will be charged for access to eco-touristic by the monitoring of economic indicators (such areas. Enitially, the schedule of fees will be as quantity and prices of bottled water con- based on the cost of operating those areas and sumed, and percentage of households with wa- subsequenty (ifjustified by the dmand) on the ter treatnent equipment) in order to facilitate carryig capacity of each (that is, on the maxi- the progamming of water-treatnent mum volume of visitors its fragile nature will investments, allow). Management of Natural Ecosystems Commercial Use of Public Resources Ei correcting market price distortions that In selling public resources, the govemment affect the use of natural goods, priority will be will favor a mechanism of tradeable licenses. given (that is, inunediate application) to remov- This mechanism will be applied mainly to pub- ing distortions that lead to tne destuction of lic natural resources (that is, owned or adinin- natural ecosystems. Consequently, to prevent istered by the government) which are already fiurther losses of wetands in the dqmrment of being commercially exploited, such as beach Rocha, measures must be immediately taken in and river sand, fishing resources, certain 9. Environmental Management- An Economic Proposalfor Uruguay 253 species of native fauna, and navigable waters. coastal lands and of sale of use righs. Taxes (Forest resources-such as firewood, timber, will apply to existing land uses (for example, and seedlings-are excluded because of being the subdivision and sale of fragile and scenic privately owned.) The mechanism could be lands) in conflict with zoning reglaions. Sale based on the one already being applied to min- of use rights will seek to increase the value of erl resources. In all cases, the government will the scenic resource and the cost of its alter- seek to control the total amount of the resource ation, and to fmance works for its improve- exploited through the number of penuits issued ment. These rights will refer to certain uses of and the amount pernitted each license. When private property (for example, construction of there is a licensing system in place (as in the a commercial building) which (a) are restricted case of fishing), the expiration time of out- (but not prohibited) by zoning; or (b) capitalize standing licenses will be shortened and a tax le- on nearby scenic or touristic resources (or on vied on their appreciated value, projects for improving them); or (c) modify An alternative mechanism could be to es- their value. The price of those rights will be tablish a system of taxes or fees for resource based on zoning criteria and on the impact on use. When feasible, however, the system of li- the scenic resource concerned. cense is to be preferred because of the advan- The depreciation of properties stemming tages mentioned before: the govermnent does fim zmning restictions and land use taxes wil not have to calculate the price of the resource be monitored- The tax level can be adjusted to (the market does that); the adminiusaion of the ensure that such depreciation acts as a fine or resource is more easily decentralized; and it is deterrent on speculative practices in fragie or easier for the govermnent to impose a high li- scenic areas but not as a cause of big capital cense price tan a high tax. losses (since many owners have paid property The advice of license holders will be taxes over many years). Consequently, it will sought, so that they may contibute to assuring also be advisable, in selected cases, to grant the sustainable use of the resource and the tax breaks for land depreciation. checkingt of infractions by members and non- A long-term program will be implemented members. However, license holders should play for the gradual buying up of coastal lands in- no part in determining the number of licenses volving a public intest This progrm will to be issued since they will seek to limit that take advantage of the depreciation of pwperties number in order to increase the profits from mentioned above. their activity (as is apparently the case in Mon- Associations of the beneficiaries of toUrism tcvideo winth holders of licenses for taxis and activities (businesses, hotels, tourism, agencies, oommercialfishing)- and so forth) or of scenic/recreatonal re- For public non-natnral resources sources (ndowners of properties adjoining (especially urban transporttion routes) simi- beaches, parks, and scenic spots) will be pro- lar sale mechanisms may be employed (that is, moted as much as possble in order that they tradeable licenses, fees, or taxes), using part of will cooperate with local institutions in the ad- dte income obtained for eliminating or mitigat- mi on ofthese resources. ing the environmental problems stemming from Urban Scenic and Cultural Resouces the involved activity. In the case of buses and btxis, for instance, the sale of licenses (whose In view of the fact that over 80 percent of current market value is, as was seen, approx]- the population of Uruguay lives in urban cen- mately US$190 million) could be employed in ters, the conservation and development of these part to finance the costs of reducing smoke and resources must be given corresponding priority. toxic emissions from their vehicle fleets. Mechanisms similar to those proposed for Haphazard Occupation of Costal Areas coastal scenic resources, including the sale of use rights (based on existing zoning) and subsi- The zoning of coastal areas proposed by dies for the fomation of beneficiary associ- the National Environmental Plan will be con- ations, witl be applied to these resources. Some plemented by a systen of taxation on the use of use rights may be traded between private 254 parties (for example, the right to erect build- users who follow certain procedures for gar- ings of more than a certain height could be sold bage disposal and sorting; those who do not to the owner of an adjoining property who wvill be charged the cost of the service. When wishes to preserve a given view). It should be the collection of fees is feasible, a system of pointed out that the sale of urban use rights sa- zer-average subsidy (that is, a system of high tisfies a demand on the part of owncrs of real charges in high-income sections and low or property willing to reconcile their own interests zero charges in low-income ones, equalling, on between one another and the municipal adiin- average, the cost of the service) is proposed. istration. Properly handlcd, this mechanism Even in this case, the use of zero-base (instead could serve to orient the phvsical and acsthetic of zero) charges in low-income sections is development of the city and generate funds for suggested. advancing the municipality's scenic/cultural Different methods of providing garbage projects. collection services will be tried in different ci- Municipal taxes could be applied on the ties or in different collection zones within each valuation of properties adjacent to sceniclre- major city. This experience will be used to ex- creational resources in order to finance invest- tend the adoption of the most successful meth- ment projects (such as improvement of urban ods to other urban zones or centers. beaches and parks, and elimination of sources A gradual privatization of the service will of noise, objectionable odors, and toxic enis- be sough if the above experience proves its sions) that benefit the ovners of those proper- convenience. Insofar as possible, more than one ties Associations of beneficiaries who would private company will be kept in operation in cooperate m the administration of these re- the largest cities, with limited-term contacts sources could be formed and assignent by zones, in order to maintain a In order to finance the maintenance and certain degree of competion. renovation of privately owned cultural re- The income of scavengers will be increased sources that benefit society as a whole (fir ex- by assigning them ertan remunerated tasks of ample, antique residences of historical value), collection, wuhich may both complemcnt their direct subsidies will be granted to the owners main activity of sorting and put a stop to their (such as technical assistance, subsidized loans, disorderly disposal of wastes all over the city. tax breaks, and even financial contnrbutions). Such methods will be tried out first in selected The administration of commnunity cultural and zones of Montevideo and used also as a means scemc resources (for example, the Old City of for incorporatng the scavengers into pnvate Montevideo) will be subsidized, as will the collection companies. To this end, the formna- formation or consolidation of associations of tion of informal associations of scavengers will inmterestod parties, with technical and financial be encouraged. assistance. it should be noted that while Systms of depositirefund will be adopted subsidies (especially financial) for the forma- for non- biodegradable goods (tires, containers, don of associations of beneficiaries will be certain plastics, batteries, etc-). Under this sys- temporary, those for associations interested in ten, consumers of these goods will pay a high community heritage could be permanent deposit, which will be refunded only after the used good is retuned to certain collection cn- GarbageDisosal r.ters. No receipt stubs will be used, in order that A selective and differential rate system for any person, including the scavenger, will be garbage collection and treatmnent, will be intro- able to make the return. The amount of deposit duced to cover the total cost of the service. The will, at a minimum, cover the cost of collecting systemn will be experimental and its adoption the discarded good in accessible public places. graduated. When the collection of fies is not Ifasible (either in certain cities or certain low- income sections of a given city), a zero-base system could be adopted. Under this systmn, the price of the service will be zero for those 9. Environmental Management: An Economic Prop salfor Uruguay 255 Bibliography .1991. Managing the Region 's Natural Resources: A Development Challenge Anderson, Dcnnis. 1990. An Econoinic Perspec- for the 1990's. (mimeographed). Wash- live on Managemnent in the Public Sector. inglon, D.C.: General Secretariat, OAS. Pads: Organisation for Econoniic Co- 1 Uruy-ANational Enwironmen- operation and Development (OECD). tal Plan. Washington, D.C.: General Sec- Asian Development Bank (ADB). 1990. Economic retariat, OAS. (in print). Policies for Sustainable Development. Organisation for Econonic Co-operation and De- Manila: ADB. velopment (OECD). 1991. Lutz, Ernst. and Michacl Young. 1990. Agricul- "Environmental Management in Develop- tural Policies in Industrial Countries and ing Countries-A Conference Summary." their Environmental Impacts: Applica- Paris Environmental Management Pro- bility to and Comparisons wit1h Develop- gramme, OECD. ing Nations. Washington, D.C: World Panaotou,Theodore. 1990. "The Use of Fiscal In- igNin. centives" Paris: OECD. McGaughey, Stephen. 1988. The Role ofMultilat- P David, and Ail Maa 1989. eral Agencies in Promoting Sustainable ng CunOPportnaitY Cost as a Plan- Development. Washington, D.C.: Inter- eimg Concept m Namtual Resource Man- American Development Bank. agemen" Washington, D.C.: Word Munasinghe, Moban. 1990. -Managing Water Re- sources Lo Avoid ErMronmerRal Degla- rPeatce, David. 1990. -Recen Thindng in OECD dation: Policy Analysis and Countries.- Paris: Organisation for Eco- dqplication: Potiy Washinalysis D.C: d .nomic Co-operation and Development Application." Washington, D.C: World (OECD)- Bank. Munasnghe, Mohan, and Ernst Lutz. 1991. Repetto, RoberL 1988. "Economic Policy Reform "Environmental-Economic Evaluation of for Natal Resource Conservation." Projects and Policies for Sustainable De- Washingtor., D.C.: World Bank ________ Economnic Incenĥivws for Sustaiunable velopment. Washington,- D. C- World ..EukncIcaEafrS l Banlc W Production Washington, D.C.: World Bank.~~~~~~~~~~~ak MYers, Norman. 1990. "The Enviromental Basis R r irk of Environmental Development" Wash- Caa Kir 1990. ESvrotemng Government ington, D.C: World Bank Capacity for Environmental Management Orgnization of American States (OAS). 1984. In- in Latin Americat Washington, D.C.: legratedRegional Planning: Guidelines General Sevretarit QoAS. and Case Studies from 0OAS experienace. UJmted Nations Enivironment Prgamnun (INE andWCashingtudies fiC.: GeneraS Exp ne, 1990. "Action Plan for the Enviromnment Washington, D.C.: General S.crtariat, i Latn America and the Caribbean." -. 1990. EnevironmentalAccounting: n Rgional Office for Latn Ameica and Application to Environmental Soil Loss the Carifbean, UNEP. Washington, D.C.: General Sertariat, Warford, Jereny. 19f9. "Environmental Manage- OASh ment and Economic Policy in Developing - ~~~~~~~~~Count6ies." Washington, D.C.: Woldd BzAL 10 Integrated Environmental and Economic Accounting: A Case Study for Mexico Jan van Tongeren, Stefan Schweinfest, Ernst Lutz, Maria Gomez Luna, and Francisco Guillen Martin 7This paper presents the results of a case study carried out in 1990 and 1991 jointly by UNSO. the World Bank and the National Institute of Statistics, GeograpIy and Informnzacs (INEGID in Mexico, with the objective of integrating and linking enironmental and econom- ic information and to explore whether enviromnentally-adjusted nationalproduct aggregates for Mexico can be derived This work was carried out within the overall analyticalframe- work developed in UNSO 'sDraft HandbookonEnvironmentalAccounting. GDPfor 1985 was 47,391 million pesos. Depreciation ofproduced capital was 5,331 million pesos. Therefore, NDP is 42,060 million pesos. Two EnvironmentallyAdjusted net Domestic Products (EDPs) are calculated EDPI is derived by deducting from NDP the environmental uses related to depletion, deforestation, and kmd use; it is estimated to be 39,662, i.e. about 94 percent of the traitional NDP. EDP2 is obtained byfiarther deducting the cost of degradation, and is estimated to be 36,448 million pesos, i.e. about 87 percent of NDP. In reviewing the results of this pilot study, one needs to keep their tentatve nature in mind The main emphasis in the first stage of developing the SCEEM was on idenifying the realaionship between different economic aggregates and environmentally-adjusted aggre- gates. Tere was less emphasis on the quantitative accuracy of the environmental adjustments. -258 This paper reports the conclusions of a framework for joint economic-environmental case study carried out in 1990-91 jointly by the analysis, called Sistema de Cuenras Economi- United Nations Statistical Office (UNSO), the cas y Ecologicas de Mexico (SCEEM): It World Bank, and the Mexican "Instituto Na- starts with a presentation of the traditional cional de Estadistica, Geografla e Informatica" Mexican national accounts framework for eco- (INEGI), with partial funding from the United nomic analysis, called the Sislema de Cuentas Nations Development Programme (UNDP), to Nacionates de Mexico (SCNM), and extends compile, on an experimental basis, environ- SCNM to SCEEM, successively integrating mentally adjusted economic aggregates for into the analysis the resource balances in physi- Mexico. The aggregates were to be developed cal terms that rcflect oil extraction, deforesta- within an analytical framework of environmen- tion and land use, and degradation concerns, tal accounts linked to the national accoutns of and applying the valuations developed to arrive Mexico, and based on concepts elaborated in at a SCEEM macroanalysis in monetary terms the UNSO's Draft Handbook on Environmen- at pernits a comparison of SCNM and tal Accounting (United Nations 1990). Given SCEEM aggrgates. The mac gat in that environmental and economic information monetary terms are broken down by economic has traditionally been collected by different activities, which pernits comparative evalu- agencies, a special effort was made to establish ation of environmental impacts and use of contacts bewn INEGL the Urban and Envi- naturl resources by different economic activi- ronment Ministry (SEDUE), and other relevant tics in Mexico. agencies with data and expertise in these areas. The analytical data presented in the tables This .paper covers the following topics. throughout the paper refer to 1985. In that First, infonnation is given on the orientation year, a detailed input-output table for Mexico and organzational context of the work on envi- was compiled, ficilitating the adapion of ronmental accounting in Mexico. Next, the pa- SCNM to the integated economic and environ- per describes the environmental concerns, mental analysis of SCEEM. distnguishing beween three concerns-oi ex- traction, dfirestation and land use, and dega- General Orientation of the Pilot dation concems. It summarize for each of the Project concerns (a) the data sources of physical envi- There has been much discussion in Mexico ronmenal indicators that are compiled in order about enviromnental conoens, and a variety of to measure the impact of the concems (and the studies have been carried out to quantify the adjustments that are made to these data in or- impacts on production and the quality of ife?. der to use them for integrated economic- Public opinion and poliicians in Mexico are environmental analysis), and (b) the principles aware that many of the environmental concems of valuation applied in order to arrive at mone- are linked to economic activities and that CInv- tay valuation of the environmental irnpacts. ronental impacts can be mitigated through a Lastly, this repon describes and analyzes the cambination of incentives and regulatons af- efibcts of mcorporg data into an accounting fectig cconomic activities. In order-to assess The author are gratefl to Sweder van Wijubergen and Mohan nasinghe for supporting this wor and to Dale lorgensen and Michael Ward for valuable commen. Also, the authors would like to k Hector de Alzma Romo, Jaime Rodriguez Carmza, Joel Villegas Tovar, J. Dulce Ma. Martinez More- no, Femando Perez Condek Gerardo Calderon Fierros, Alma Morales Lozado, Hilda Martnez Diego, and Silvia Jaime LeUa firom INEGI for valuable assistance. Finally, the authors would like to than the World Bank's Research Committee and UNDP's Meico Progam for providing finding for the sudy. The ancxes to the case study present a detailed account of data sources and valuation methods and include the quanti- tative results ofthe project in considerable detail. Since the anncxes are voluminous, they are not included as a part of this book For eXample. the recent report Mexico in Trausition: Towards a New Rate for the Public Sector (World Bank Report No. 8770-ME, May 22,1991) discusses enviromnnetal policy issues in considerable details 10. Integrated Environmental wad EconomicAccounting: A Case Study for Mexico 259 the economic ramifications of environmental SEDUE and other Mexican institutions work- interventions, there is a necd to link economic ing in the area of natural resources, were sepa- analysis with analysis of cnvironmental im- rately identified. This facilitatcd discussion and pacts. Such linkage makes it possible to deter- fiurther reconciliation of common concepts, val- mine in which economic activities uations, and so forth, and also madc it possible cnvironmental regulations would be most effec- to confront available economic and environ- tive from an enviromnental point of view and at mental data sets, which were used in the past the same time be optimal from an economic for separate analyses, and use them in a joint point of view. The development of such a link- analysis to bring out the interactions bween agc between economic and environmental ac- economic activities and environmental eftets. counting was an important objective of -this Furthermore, within the- resource balances of pilot study. SCEEM, a distinction was made between dif- One of the difficulties encountered in de- ferent resource balances representing differet veloping a joint economic-environmental analy- enviromnental concerns, so that the quantita- sis was the separation between different tive interactions between econormc activities disciplines which deal with economic analysis and each of those environmental concems and with analysis of environmeal impacts. In could be separately assessed. mexico,asinmayothercountries,thedifferent In line with the orientation of SEEA, orientations of economic and environmental SCEEM was compiled in an integrted man- analysis not only are reflected in different disci- ner, with much emphasis on identifyng qua- plines and experts supportig those disciplines, tfring, and establishing the relation between but also are represented by different institu- the environental effects on different macroe- ions. In preparing government policies in conomic aggregates and less emphasis on fields related to the present study, the two dis- quantitative accuracy of the enviromental ad- cipiines are mainly represented by INEGt justments. The rcason for this emphasis is our which has developed macro economic accounts belief that first the orientation of the integrated that are used by the government in economic environmental-economic analysis should be es- policy making, and SEDUE, which is the focal tablished, thus, integatig expertise, conct, point in the government for environmental poli- and data, and only at a later stage when there is cy making. The first step in implmentg the common understanding of concepts and format project in Mexico was to establish the neces- of such integrated analysis, data improvements sary contacts between experts in econonic ac- can be attempted, which would not only lead to countming in INEGI and resource accountig in improved accuracy of isolated data, but would SEDUE and other Mexican isttuion4 dealing also lead to improvements in the intgrated with natural resources. analysis. An accounting framework was developed For instance, after completion of the envi- that could be used as the quantitative instm- ronmental project in Mexico, detailed stadies ment dincgh which envionmeal data ana- may be caried out with regard to soil erosion lyzed by SEDUE and other Mexcan caused by agricultural production or with re- institutiLons, and nacroeconomic data elabo- gard to air and water pollution caused by co- rated by INEGI, could be linked. The joint nomic activities in the Federal District of economic-environmental accounting fiamework Mexico. In such studies, economic and envi- SCEEM is based on the System of Economic ronmental data would have to be compiled in Evronmental Accounts (SEEA).3 SEEA was the same classification detil and on the basis adapted to the Mexican c ces and re- of compilaton metiods that are compatible in quirements, such that economic accounts in- terms of valuation, esfimation procedures, and cluded in SCNM and compiled by INEGI, and so on. resource balanc known by experts frm 3 SEEA is the accounting framewor that is included in te Draft HandbR k on Enviroment Accountin at is being developed by a consultant to the Unitd Nations Statistcl Office. In the remaining tex, refece will be made eithe to SEEA or thc HandbookL 260 Improvement of data for integrated envi- for the purpose of urbanization. The transfer of ronmental and economic analysis should not be land also affects the practice of cultivation, the only objective of fiuture activities in envi- which is sometimes carried out by burning fo- ronmental accounting, though. There should be rests for temporary cultivation and abandoning also much emphasis on further developing the the land after a few harvests. Both the logging rudimentary analyses of environmental impacts of timber and the tmnsfer of forest areas to on growth and performancc, which are pre- economic uses generally involve not only the sented later in the paper. Only through such exhaustion of depletable resources-timber and improvements of the analytical tools would the forest land-but also the destruction of these improved data have an impact on policy mak- natural resources as ecosystems. While the lat- ing through the replacement of traditional eco- ter loss of the forest as an ecosysten is an envi- nomic analyses by integrated analyses that are ronmentally critical concern, it is not dealt with based on environmentally adjusted aggregates. under the deforestation and land use concem as defined here, which solely deals with the quan- Environmental Concerns titative reduction of two natural resources, that Addressed and Measured is timber and forest land. The environmental concems exanuined as The degradation concems do not deal wffi part of the project activities deal wiffi quantita- the quantitatve exhaustion of natural re- tive and qualitative efflcts on natural re- sources, but rather with the qualitative degra- sources. They are grouped together in the dation of the ecosystem. They include in the present report in three main groups: oil extrac- present study the contamination of air and wa- tion concerns, deforesttion and land use con- ter through the generation and deposit of resid- cern, and degradation concen. The first two uals, and also the environmcntal impact of groups anr also called te drepletion and land leavmg garbage and solid wastes behind as a use concerns. result of production, including domestic pro- The oil depletion concern is dealing wr duction activities of households. Also covered the quantitative exhaustion of a natural re- i grdation conces a land e source that is an important source of revenues sion and ground water loss. for the Mexican economy. It is reflected in oil th i ither analysis of the impacts of the extraction and findings of new oil reserves. The three of environmental concerns, it has traditional economiic aggegates do not take ibeen assumed that the quantitative depletion into account any allowance for depletion, and land use cncerns of oil extraction, defor- which constitutes a loss of fulture revenues; in estation, and land use presented above have an ocher words, the oil resources are considered as imeate impact on the productimv and free goods in traditional economic analysis. mcoe-generatng capacity of the Mexican The oil depletion aspect is analyzed here sepa- economy. The depletion of oil and the loss of rztely from environmental impacts, that is, the tres aindiatly affect the future income- air and water pollution caused by oil extracton generating capacity of the oil extraction and and the subsequent oil refining activities; these logging activities. Also, land t ed friom environmental impacts, which afibct the qualitr the euroirnhent to economic activities n agrs - of natural resources (air, water), are dealt with culture, livestock hoein anid for purposes of under tLe uegradation concern described below. uasization will have an itmediate (in rthis The deforestation and land use conces positive) impact n the incomegeneratig deal with two aspects of deforestation. These capacty of economic activities. The qualitative include the loss of timber as a result of con- degdation concerns, on the other hand, have mercial and noncomnmercial logging at a speed apUost inmediate effects on he welfre of te beyond the capacity of nature to replace it population and long-run eflbs on the produc- through natural growth, and the transftr of un- tive and income-generatiag capacity. The ef- exploited forest areas to uses in economic acti- fects are on the quality of life through efifets vities, eithier in the form of agricultural la, of air, water, and solid waste pollution on land used for holding cattle, or land a healt through the effects of land erosion on land used for holding caffle, or. Imid umnsferred~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 10. Integrated Environmental and EconomicAccounting: A Case Study for Mexico 261 the quality of recrcational arcas, and through A variety of data sources were used for ground water loss which immediately affects compiling physical indicators that reflect the households in their consumptive activities. Ulti- three tvpes of cnvironmental concerns men- mately, the degradation concerns will have an tioned above. They arc summarized below and impact on production, either through health ef- described in detail in the annexes to the case facts of workers in the production processes or study.4 through secondary effects on growth of prod- Information on proven reserves of oil was ucts in agriculture, fishing, and forestrv which, obtained from published information by the in turn, affects the generation of net product in publicly controlled Mexican Petrolcum Com- ftese economic activities. These secondary ef- pany PEMEX. Annual extraction data were fects on production are, howevcr, more diffi- also obtained from this source. New discover- cult to identify and analyze because many ies are not published but obtained as a residual effects-and not only environmental from the published data on oil deposits and oil ones-may operate together, and time lags of extraction. the effects on production may be very Information on the (opening and closing) considerable. stock of forest resources was obtair.- ih'm the The distinction between immediate effects National Forest Inventory, a report prepared by on production and the long-run effects on pro- the Secretaria de Agricultura y Recursos Hi- duction through primary effects on well-being, draulicos. The report included information on does not entirely coincide with the distinction land surface covered by forests expressed in between the two groups of concerns Oil deple- hectares and also data on the volume oftrees in tion may also have long-run effects through the cubic meters. There was no detailed infonna- air and water pollution caused by oil extraction tion on the type of trees. The same Secretaxia activities. Deforestation may have severe long- de Agriculura and Recursos Hidraulicos also run effects: the loss of ecosystetns may affict compiled a Regional Inventory of the Use of productivity, particularly in agriculture, and Land, which provided information on land use through health effects on the productivity of in general and not only for forest areas. In- wvorkers in other industries, and in general it fonration on changes in the use of land was may have a deteriorating effect on the quality obtained on the basis of assumptions that were of life. On the other hand, some of the degrada- based on studies about the relation between in- tion effets such as land erosion and ground creases in the use of land and the correspond- water loss may have irnmediate effects on pro- ing growth of agricultural production, gowth ductivity. particularly in agriculture. of production in cattle farming, and growth of urban centers. As a result of these assump- tions, decreases in the forest area over time The compilation of additional data on could be calculaed. physical indicators with rgard to balances of Data regarding the degradation concerns natural assets was accompanied by an exten- were compiled with regard to land erosion, air sion of the present SCNM to include asset bal- and water pollution, ground water use, and the ances of produced assets as well. The generation of solid wastes. information used to extend the SCNM in this Land erosion was esfimated on the basis of manner took as a point of departure the SCNM the General Ecology Report elaborated by the data regarding gross capital formation. Addi- Comision Nacional de Ecologia The report tional infomation was obtained from the Eco- identfies areas with very severe, severe and nomic Census 1986, supplemented by data moderate land erosion and presents an average from the Census of Commerce and Services erosion per hectare (27.54 tons per hectare) at 1980, and the Survey of Capital Formation in the national level. The infonnation on erosion the Enterprise Sector, both carried out in 1980 per hectare was applied to the areas uwed for by the Banco de Mexico. agriculture, cattle farming, and forestry and thus national totals in tons were derived and 4 See footnote 1. 262 allocated to the three economic activities men- infornation was the General Ecology Rcport tioned. Reconditioning of soil, resulting in the prepared by the Comision Nacional do Ecolo- elimination of some of the soil erosion, was de- gia, which provided infonnation on the daily ducted from soil erosion in agriculture. average generation of solid wastes at the na- The main source of information on air tional level (.693 kilogram) and at the Federal pollution were 1989 reports resulting from a District level (.987 kilogram). The report dis- long-term program carried out by te Federal tinguishes between different solid residues, District of Mexico, called the -Programa Inte- which are grouped together in the preparation gral Contra la Contaminacion Atmosferica" of data between biodegradable and nonbiode- Based on the results of this project, contamnina- gradable. This information was used together tion coefficients could be calculated for five with population data to arrive at a national lev- types of contaminants-sulfur dioxide, nitro- el of solid waste generation by households. gen oxide, hydrocarbons, carbon monoxide, and fuli suspended particles. The contamina- Valuation tio coefficients related the emission of these As will be shown below environmental ac- substances to the number of motor vehicles, the counts in physical terms can give indications of owtput of the oil refining industry, the output of the direction in which environmental costs and electricity plants, and the output of other man- capital would affect the traditional economic ufacuing -industries. The coefficients were aggregates. However, for an integrated analy- used to calculate national totals for each of the sis, it is essential that environmental variables contaminants, based on the total number of ve- be expressed in the same monetary units as the hicles in Mexdco and the output of a variety of economic variables. industries causing air polution. Two types of criteria are used in valuing The extent of water pollution was mea- the natral assets and changes therein. The first sured in terms of biochemical demand of oxy- type calculates the value of stocks of assets as gen (DBO). This is an effective measure the sum of discounted values of future income because additional oxygen is needed in order to streams; the value of change in the stock of break down excess of organic substances found natural assets is based on the changes in future in the water. The sources of infibnnaion were income streams as a consequence of additions private studies. Contamination coefficients ex- to the reserves of natural assets, or depletions. pressed in DBOs were estmated for a variety The second type of valu3aonl, which is only of industrial sectors and also per capita to cov- used for valuing quality changes in the natnual er water poilution by households. Application assets, is based on the cost of avoiding such of the coefficients to output in each industry changes These general criteria translate into and totl populaion provided national totals on more specific methods for each of the natural DBOs. resources and environmental concerns covered Information on ground water resources in SCEEM. was obtained from a variety of reports, but the For exploitaton of oil and timber reserves, main one was the General Ecology Report pT- two alternative valuation methods are used. the pared by the Comision Nacional de Ecologia. first one is the net rent method as developed in The use of ground water was estimated on the projects carried out by the World Resources basis of water used per capita and water used Institute projects (Repetto 1989), and the se- per Unit of industial production in a number of cond one is the method developed by El Serafy industries including agriculture, cattle firming (1989), which is based on the calculation of a and electricity production. The reduction in the depletion allowance. The net rent metod cal- ground water resources would be the difference culates te value of natural resources in the between this use and what is annually replen- ground as the difference between the market ished through pr itation. value of lumber and oil and the cost it takes The generaion of solid wastes was only (including a normal profit) to extract this ri- calculated for the household sector; industrial source for commercial exploitation. The net wastes were not covered The main source of rent value is calculated separately for opening 10. IntegratEnvironmental and Economic Accounting: A Case StudyforMexico 263 and closing stocks as well as for changes in assumed to be equal to what it would cost to thosc stocks as a result of dcpletions and re-inject water into underground water reser- additions. voirs. The cost of water and air pollution was The method dcvds!oped by El Serafy values estimated on the basis of the cost it would take the natural resource as a finction of the sum of to reduce such pollution to acceptable levels. the discounted values of future- income strams that are generated by it. [t assumes that a part Accounting for Environmental of operating surplus of mining and forestry Impacts on Economic Activities would have to be reserved for re-investment In order for the integrated envirownental (depletion allowance) in order to assure that the and economic analysis to have maximum ef- sum of the discounted income streams fect, the existing analysis taken as a point of (covering net operating surplus and compensa- depare should be the one that is fiequently tion of employees) over the limited life of the used as a basis for government policy decisions natural resources would be equal to the sum of and that furthermore includes a maxinum discounted income streams over an infinite pF- number of variables in which environmental riod obtained if net product were reduced by impacts could be incorporated. The analysis the depletion allowance and the depletion al- tiat best fits these conditions in the Mexican lowance were re-invested. The amount of the deplfionallwanc pe uni of rodctio is case is input-output analysis that is based mn epletion allowane per ul nit of production i s Mekico on the highly developed data base of oabtined as a result of these assumptions. The SCNM. It is convenient to take SCNM as point values based on the net rent and El SWtafy of departure when developing SCEEM, as the methods have their own economic meaning and SCNM data base and its analytical features are are, therefore, used in combination in the inte- easiy understood by both users and producers grated analysis of this paper. of such statistics in Mexico, where in the past The value of land is generally based on the enive use has been made of input-output sum of the discounted value o f fiitrX, inconC mdata and analyses for policy purposes. By streams in tie difirent economic acties in sdtrcturnng SCEEM in a manner similar to which land is used. It is calculated separately SCNM, boti producers and users of SCNM for forest land, land used in agiculture, land cSn easily compare the daa base of SCNM used for holding of livestock, and land used for ith the panded data set of SCEEM and can urban needs. In each type of economic activity, widi - exprate how ag ates of SCNaM the value per hectare of land is based on the thusge wpen redefined in SCEEM. Also, for sum of the discounted values of net value add- anelytical usersit would be easyCto see ho the ed per year per hectare of land. In the case of SCEMnlyi analysis would difer from the anwdy- forest land, the value of the land is based on net sis that is tradtionally carred out in the con- value added in forestry. Only forest land as an sis of SCi M. environmnental asset has been given an aconom- The present section starts with a macrna- ic value; other land not used for economic pur- nalysis of the economic and environmental poses other an furest land has not been data, followed bv a similar analysis in which valued. Revaluaton of land withi the same ac- separe economic activities are identified. tivity between the beginning and end of the pe- riod is not taken into account, only changes in Macroanalysis the value of land when transferred from one ac- tivity to another. Features of "Sistema de Cuentas All type of degradation have been valued Nacionales de Mexico " (SCMAf on the basis of avoidance cost per unit of con- . Behind the present national accounts of taniinant or other unit of degradation. In the Mexico, that is, SCNM, is a very simple ana- case of land erosion, this cost was valued as the cost of fertlizer to maintain d productv- lytical model, winch serves a vanety of analy- ity of the land as it was before erosion took use The sianple model includes a supipy and place For ground water loss, the cost use identity and production functions With help of Table 10-1, some basic fiatuues of 264 Table 10-1: Mexdco InputfOutput Schcmc (SCNM) (billions or Mccian pesos) Economic activities Final Produced Producfion Rest of world consumption capil Economic supply P M Total 75,706,918 4,897,328 Economic uses Ci Ex C I Total 28,315,216 7,305,293 34,948,897 10,034,840 Gross product Y 47,3191,7027 Table 10-2: Enlarged InpuiJOutpot Schemc with Produced Asset Balance (billions of MCican pCsos) Economic activities Rest Final Produced Producion of world Consumption Capital Opening assets Ko. p. ec produced assets 111,162,310 Economic supply P M Total 75,706,918 4,897,328 Economic uses Ci Ex C I Total 28,315,216 7,305,293 34,948,897 10,034,840 Depreciation Depr. -Depr. 5,331,186 (5,331,186) Net product Yn In 42,060,516 4,703,654 Revaluation Rev. p. ec produced assets Closing assets 11K1. p. ec produced assets 115,865,965 10. Iniegrated Environmental and Economic Accounting: A Case S:ulyfor Mexico 265 SCNM can be described, or natural resources play an important role, is The supply and use identity can be written to expand SCNM with the corresponding asset as balances of produced assets, which arc not regularly compiled in Mexico, but which form P+M=Ci+C+I+Ex (i) an integral part of the system of national ac- counts. This expansion of SCNM. which is in which P is production, represented in Table 10-2, changes the input- M is imports, output (ilo) scheme from one that can onlv be Ci is intermediate consumption, used for static analysis of the input-output type C is final consumption, to a dynamic input-output model in which capi- I is gross capital formation (or invest- tal output ratios can be identified and uscd in ment), and analysis of growth. The extended scheme of Ex is exports. Table 10-2 includes three additional elements A second identity defines gross product or as compared to the scheme of Table 10-1 that value added (Y) as the difference between pro- is: KO0) is opening stock of produced assets: duction and intermediate consumption: Kl(4J is closing stock of produced assets; and Depr is depreciation. Y =P - Ci (ii) The lree eklments are used to define the additional asset balance identitv, which eLx- When substituting this income definition plains the relation between the opening and into the first identity, a third identity is derived, closing stocks of produced assets -(Kl , ) which links gross product and expenditures: and KO(p04 on the basis of changes in pro- duced capital. Further changes are reflected in Y = C + I + (Ex - M) (iii) (net) product and (net) capital formation. Both The supply and use identity (i) is repre- were defined gross in Table 10-1 and are re- sented in Table 10-1 as the identity betwen the placed by net versions after deduction of de- totals of the second and third row, tiat is, preciation in Table 10-2. T'he asset balance presented in the last col- 75,707(P) + 4,897(M) 28,3 15(Ci) + umn ofthe table has the following format: 34,949(C) + 10,035W + 7,305(Ex).Y KIP . = KOP, + (I-Depr) + Revp. The product identity (ii) is shown in the first column as the difference between total One element, Rev ,˘ (standing for revalua- economic supply or output (P) minus total eco- tion of produced economic assets), is added for nomic uses or intermediate consumption (Ci); the purpose of completeness. In Table 10-2 no that is, gross product (Y) is value has been entered for this element howev- er, as revaluation of produced assets is not tak- 47,392(Y) = 75,706(P) - 28,315(Ci). en into account in the additional data that were compiled. The asset balance in monetary units The product-expenditure identity (iii) is not as presented in Table 10-2 therefore takles the immediately obvious from the presentation of following quantitative fbnnat: the table, it is reflected in the identity between Y on the one hand and the sum of C, I, and 1152866(Klpa) = It l1162(KOpj + (Ex-M) on the other, that is, [10,035(0) - 5,33 1(Depr)] 47,391(Y) = 34,949(C) + 10,035(1) + Because one of the elements that explains [7,305(Ex) - 4,897(M)]. the change in the value of produced assets is depreciatiorL incorporation of depreciation is The firs. step in derving SCEEM, in an essential element of changing the static i/o which asset balances of cnvironmentl assets model of SCNM to a dvnamic one. The 5 For ease of presentation, the last three digits of the figures presented in the tables are omitted in the tcxt 266 incorporation of asset balances into the cx- of the produced assets that were included in panded version of SCNM is therefore accom- SCNM. The first additional column refers to panied by a corrcsponding change from using nonproduced assets that are directly "used' in gross product (Y) in Table 1O-1 to net product economic activities together with produced as- (Yn) in Table 10-2-and gross capital forma- sets; both groups are labeled economic assets. tion (1) to net (In)-and this changes the na- The second additional column refers to assets tional accounts identity prcsented above as that are only "affected" by economic follows): activities-so-called environmental assets. Economic assets are used as production factors Yn = C + In + (Ex - M) in the generation of output, and production analysis requires that full balances including in which Yn and In are net product and capital stocks of economic assets are available. Envi- formnon concepts from which depreciation ronmental assets are not considered as produc, has been deducted. tion fihtors in this sense; their contribution to Conceptual Framewmork of "SisIema de the generation of output is not filly understood ConceptuasEcolF camew cofSisemias de or perceived in existing analysis and this is Cents Ecologcas e Ecoomas de generally reflected in the nonavailability of in- Mexico" (SCEEMd) - fonnation on asset balances, includmg data on SCEEM takes SCNM as the point of de- stocIs ofthose assets. parture in ims design, while intmoducing a num- in the case of Mexico, oil resaves and land ber of modifications. It includes a wider asset uased in agriclture, livestock masnnamt and boundary, covering not only produced assets, niation are treated as nonproduced geou but also nonproduced natural or enviromnental note assets, while water (including ground assets. Thus, SCEEM includes imputations for wat s), air, soil (lost through erosion),ran also additional expenditur tms that are related to all Forests arc deal itu as environental as- depletion and degradation of nonproduced as- sets. Forests are included as environmental as- sets. Furthermore, taling into account the im- sets buse it was not possible to distinguish puted items, SCEEM incorporates modified m the data between virgin forests and forests ooncepts of net product or value added, which at are used for commercial logging and that are derived by deducting not only the tradition- should have been treated as an economic asset al cost items, but also imputed items that corre- in the same mainer as oil. Both ground water spond to envirounental cost of depletion and and soil are treated as nonproduced environ- degradation. FiWally, SCEEM changes the con- mental assets either beatuse their contribution cept of capital formation as used in the tradi- as a production factor to the generation of out- tional analysis of SCNM and introduces a new put is not sufficiently clear (ground water) or concept of capital accumulation, which takes becase no asset balances could be obtained into account not only changes in produced as- from available data (soil). sets as a result of production and depreciation Correspondlng to the incorporation of a of such assets, but also changes in the stock of wider asset boundary in SCEEM, additional nonproduced assets resulting from new finds of costs are incorporated that reflect the use or de- nonproducul assets and deteriorations of non- terioiation of nonproduced assets as a result of produced assets as a consequence of economic economic activities. In Table 10-3, two types activities. of imputed costs are represented. The first type The modified features of SCEEM can be (Cid&,p) axe imputed costs related to depletion easily appreciated from the presentation of Ta- and losses reflecting deteroration of land that ble 10-3. They can be compared with the fea- is flisferred from the environment to econom- uires of the traditional SCNM, which are ic activities, and the second type of imputed identified by the shaded areas in the table. cost (Ci,,) covers the deterioration of the &v- SCEEM in Table 10-3 includes to di- ronmental assets as a conscqaunce of economic tional columns for the incorporation of asset activities. For purposes of this pilot study, the balances of nonproduced assets alongside those cost of depletion includes the cost of depletng 10. Integrated Environmenial and Economic Accoundng: A Case Stuvfor Medaco 267 Tablc 10-3: System of Ecmomic Aomunts of Mcwo (SisUm de Cuts de Mero (SCEEM)): Basic Sucturc Economic actvities Environment Economic assets Nonprodred Final environ- Con- mental Rest of sump- Poduced Norpro- ,ssets Producton world don assets duced assets Opening assets Konp.ec .o.np.e.v Economic supply --- Economic uses conomic ur~~~~~~~~~~C . . ;. . ~.. .. .. : -t-Ci E.; K-, C ,--'-I- =t Depreciation D---epr-f ---;- Depr - Net domestic Product (NDP) Yn In._ _ _ _ _ Environment -Lnp.env uses: Lnp.ec- Depletion and land use concerns Cidepl -DpLnp-ec -DpLnp.env Environmentally -IAnAenv adjusted net Ynl lAn.ecl product Degradation Ci.degr Degr.npenv concerns -IAn.2env Environmental- Yn2 IAn-2ec ly adjusted net (=IAn-lec) product Revaluation R e Rev.npec Rev.np.env Cosing assets __ __ pe. K1.np.ec KI.np.env 268 oil, timber, and ground water reserves. Thc im- assets (LAnj) and the other to net accumulation puted cost of land use refers to the of environmental assets (IAn,J). trees-representing the ecosystem-that are Net accumulation of economic assets lost as a consequence of transfer of forcst land is defined as the change in the produc- to agricultural land, land used for holding of five capacity, that is. capital used in cattle, and land used for the purposes of urban- production, including not only pro- ization. The imputed cost refcrred to as cost of duced assets, but also nonproduced degradation includes the cost of air and water economic assets. Net accumulation of pollution, the cost associated with solid waste environmental assets is the net change materials, and also the cost of land erosion and in the quantity and quality of environ- ground water loss. menta assets as a result of economic Following the introduction of imputed cost activities itemns in SCEEM, two modified net product The enlarged concept of net accumulation concepts are introduced, called enviromnentally of economic assets includes net capital fonna- adjusted net domestic product I and 2 (EDP I tion and two additional elements related to non- and EDP2). EDPI-or what is represented in produced economic assets, including an the table as Ynl-is derived by deducting from element representing "investments" in nonpro- net domestic product (NDP) in SCNM the en- duced assets, that result from the transfer of vironmental uses related to depletion and land enviromnental assets to economic activities use, that is (Ij,,p, and another element representing the depletion of nonproduced economic assets Ynl =Yn-Cidpl (Deplj,j In the case of Mexico, the "investment" element includes the transfer of and EDP2-or Yn2-is obtained by firther land and mineral reserves to use in economic deducting the cost of degradation, that is, activities, while the depletion element refers to the depletion of oil. Depletion of timber and ground water is not included in net accumula- Y- 2YnZCi~+Ci~) -ition of economic assets, as timber and ground There are two reasons why a distinction is water are not treated as economic assets but made between EDP1 and EDP2. The first is rather as environmental assets. Also excluded that the valuation of depletion cost and the cost from net accumulation of economic assets is of land use as summarzed above is directly revaluation of produced assets (Revp,), non- linked to the market value of the assets that are produced economic assets (Rev.11j, and non- depleted or transferred to econonic use. On the produced enviromnental assets (Revp.). odher hand, imputations for the cost of degra- T1he table disinguishes between net accu- dation are much less close to market valuations mukltion of economic and enviromnmental assets and therefore much more controversial cle- related to Ynl and Yn2, which are defined as ments in the analysis. A related reason is that follows: the cost of depletion and land use taken into ac- count in the calculation of EDPI relates to the IAni0 = IAn,1 = hi + (oL. - Depli,j use of economic assets, wheras in the deriva- IAn, = -(Tp.env + Depl,,~p, ) tion of the EDP2 not only the cost of depletion TAIA -ff np.env + Dept,,p + Degr,,p) of nonproduced ecomic assets is uken into account, but also the cost of affecting nonpro- When approaching Ynl and Yn2 from the duced environental assets such as air, water expenditure side, while using the above defini- (including ground water loss), and soil (soil tions of LAn,, and IAnn,, the following identi- erosion). tics holdc Another feature of SCEEM that is different from SCNM is the introduction of two new Ynl = C + (QAn,l - IAP-n1) + (Ex-M) concepts of net capital accumulation. One such Yn2 = C + (lAn2e, - LkAn) + (Ex-M) concept refers to net accumulation of economic 10. Integrated Environmental and EconamicAccouning: A Case Study for Me-ico 269 Both identities show clearly the change in relation between closing and opening as- the traditional national accounts identity, after sets of oil in terms of million of barrels: incorporation of environmental assets. Net ac- KI,p,-= KO,pe- (Depl,,p - cumulation of economic assets (IAMJ.) is only 70,900 = 71,750 - (.1,265 - 415 ) partly reflected in Yn. An important component 2. Defonstation and Land Use. The concern of net accumulation of economic assets is di- of deforestation and land use is represented rectly based on the transfer of environmental in Table 10-5. It includes three separate assets to economic activities; and this is re- elements of enviromnental uses-the log- flected in a negative entry for LAn,. ging of trees as part of the forestry activity SCEEAI Applied to Environmental (7,626 thousand cubic meters); the transfer Concerns of forest land to economic activities in agriculture, livestock holding, and urban- The conceptual scheme ex7plained above is ization (277,589 hectares), and finally the applied below to the three cnvironmental con- losses in terms of trees resulting from the cerns separately. transfer of land from the environment to 1 Oil Exitracfion. The incorporation of the oil economic uses (35,474 thousand cubic extraction concern in Table 104 is very meters). simple. It includes only two elements of en- The cost of logging of trees is presented as vironmental uses, that is, oil depletion and intermediate cost of production (Cidp) and as new finds of oil, both of which are ex- forests are treated as environmental assets, a pressed in physical terms. Oil depletion counterpart reduction in nonproduced environ- (1,265 million barreis) is shown as an ex- mental capital (Depl,). The trsfer of for- tra environmental cost (CLd,1) and a reduc- est land to economic uses is reflected in an tion in the value of nonproduced economic incrse in nonproduced economnic capital assets (Dcplpcj. New finds of oil (415 ([,) and a decrease in nonproduced environ- million barrels) is presented as an addition mental capital (-I1P,J. The transfr losses re- to nonproduced economic asset (,,,J and sulting from this transfer of land are shown as as a reduction in the quantity of environ- intermediate cost of the activity (an example mental assets (.-L,.,,) bfor such an activity might be the construction As a result of the incorporation of these iusty, which is responsible for the output two elements, there are changes in EDP1 called "improvements to land), which carries as compared to NDP, and in net accumula- out the preparation of land for economic use tion of economic assets as conpared to net (CdpJ, and a reduction in enviromnental capi- capital fornation of SCNM in Tables 10-1 tal (-Depk.pa). and 10-2. Even though the envirmmental EDP1 (Y,1) is reduced as a result of log- uses are in physical terms, it is easy to see ging costs and losses of tres due to the trans- the direction in which these macroaggre- fer of land to economic uses. Net accumulation gates would change. EDP1 (Ynl) wouid of economic assets (LAn1J as compared to net decrease with the amount of oil extracted capital fornation, though, is increased with the (Cj&pj) and net accumulation would change amount of forest land transferred to economic as a result of the difference between what uses. The difirence betweem the reduction in is extracted (Depl,p_j and the new finds of Y., and the change in net accumulation of eco- oil reserves (I,j. The diflrence between nomic assets as compared with net capital the decrease of Ynl and net accumulation fornauon, is entirely reflected in losses of ean- of econonic assets would be the decrease ronmental capital as presented in tbe column of in nonproduced environmentl assets (oil thee rvi(onient(-An,vi. reserves) that are transerred from the envi- The asset balances ta are affected by the ronment to economic uses (-I,p). The as- deforestation and land use concerns are those set balances of oil are presented in the for land in econonic uses, which are presented column for nonproduced economic assel, in the column for nonproduced economic as- which shows the following quantitative sets; forest land, which is included in the Table 10-4: Enlarged Input/Output Scheme with A5set Balances (bllions of Mcxican pesos) and Adjustments for Net Changes in Ol Rserves (physical units) Economic activhle Envison- Ecw;omic asseu ment Nonpro Nonpro. duced Pizysical duced envtron- unit of Final Produced economic mental measure- ProducIoni Rest of world corutmpdion asus easses asseLs ment Opening assets Ko.p.ec Ko.np.ec )So.np.env Produced asset 111,162,310 Pro - 71,750 millions of barrels Economic supply P M Total 75,706,918 4,897,328 Economic uses Ci EX C I Total 28,315,216 7,305,293 34,948,897 10,034,840 Depreciation Depr -Depr 5,331,186 (5,331,186) Environmental uses: Oil extraction concerns Ci.depl -Depl.np.ec Oil depletion 1,265 (1,265) millions of I.np.ec -I.np.env barrels Ncw finds or oil 41S (415) millions of barrels Nct product: EDPI Ynl . IAnl.ec lAnL.cnv Closing assets KI.p.ec Kl.np.ec Ko.np.env Produced assels 115,865,965 Oil 70,900 millions of barrels Table 10-5: Enlarged Input/Output Scheme with Asset Balances (billions of Mcxican pcsos) and Adjustments for Changes In Land Use and Deforestation (physical units) Economic acdvildes Economic Assets Environment Nonpro- Monpro- PlytycaI7: duced dRwced unit of *rdu n Pest of Fon. l Prtuced economic cnviro,uncnn mcasurec Prodwcdlon world couLwmpti'on aossa assegs al! asseus mend Opening assets Ko.p.ec Ko,np.cc Ko.np.cnv Produced asset 111,162,310 Timber 3,125,268 thousand 139,741,568 cubic melets Land S6,078,532 hectares a. Economic supply P M total 75,706,918 4,897,328 Economic uses Cl C I total 28,315,216 7,305,293 34,948,897 10,034,840 Dt preciatlon Depr -Depr 5,331,186 (5,331,186) Environmental uses- Deforeslation concerns Ci.depl -Depl.np.env thousand (7,626) cubic meters Logging 7,626 e Forest land Iransfer to 277,589 (27,589) heclares economic uses -De~.n4~env thousand 74) cubic meters Transfer losses 35,474 Net product! EDPI Ynl LAnl.ec -lAnl.env Closing assets KS.p.ec Kl.np.ec Ko.np.env Produced assets 115,865,965 Timber 3,082,168 thousand cubic meters Land 140,019,157 5S,800,943 hectares 272 column for environmental assets, and timber, Comparison between SCNM and SCEFEM which is also presented in this columnu. The Aggregates anadAnalyses, in Monetary three asset balances in quantitative form are as Terms follows: KI KO + I In order to overcome some of the limita- 140,019,157= 139,741,568 + 277,589 tions of the above analysis in physical terms, a land used for economic purposes (hectares) parallel analysis in monetary terms is presented below. The monetary analysis is presented in K 1, - Kn two tables. Table 10-7 includes the monetary 55,800,943 = 56,078,532 - 277,589 valuation of the depletion and land use con- forest land (hectares) cemns, including oil extraction, deforestation, and land use concerns. The table arrives at a K1 =K0, - Depl,, concept of EDPT (Ynl) as defined earlier in 3,082,168 = 3,125,268 - (7,626 + 55,474) Table 10-3. Table 10-8 presents a similar trees lost (thousand cubic meters) analysis in monetary terms for the degradation efficts that are added in this table to the mone- 3- Degradation. Table 10-6 incorporates he tary valuation of the depletion effects. After in- 3 egactsof dadation. Theseinclues th corporation of both eficts into the latter table, effmUof demdatin. lese icludesoil the concept of EDP2 (Yn2) is obtained. erosion in terms of tons of soi lost, sold waste n.atenas resulting fromhouseho1. EDPJ. Table 10-7 reflects in monetary waste mat.rWs resultng from houschold terms the same environmental effects as activities-also in tons-ground wae presented in Tables 10-4 and 10-5. By ap- used in terms of thousand cubic meters, waterm- polltione inmteris of theusandcc b plying valuations to the elements of the depletion concerns related to oil, deforesLa- chemical demand for oxygen (DBO) usd tion, and land use presented in those tables by nature to destroy the foreign substances i physical terms, estimates are obtained in the water, and finally air pollution in for EDPI and also for net accumulation of tens of tons of various chemicals that are economic assets ([AnJ and enviromnml emitted by industrial production processes. assets (-LAni - The net product and cor- The cost of degradation is presented as en- oems - ~~~~~~~~~~responding expenditur items are presne vironmental cost of producdon (Cd 1 - vromnta coto rdon the line for EDP1. Because difierent with counterpart entes m the column for vues have been used for o and the environmt, representi the deteriora- stocs and flow Items, revaluaton tion of envirownental capital (-Dcegr,J. The effect of including these enviromnental elements (Rev. and Rcv1c4 have been uses Ls to lower EDP2 (Yn2) in comparison incorporated additionally in the table. with NDP (Y)nl2. The national accounts identity without en- withNDP YA) cludd i thetra"nal ironmental adjustments, whinch was in Ta- SCNPL There is no effect on net accumulation, 1 atents as all degradation effects are recorded as af- fecting nonproduced emvironmental capital. The difference between the negatve effect on Yn2 and no effect on net accumulation of eco- * dS as c * f - ~~~~~~~in quantitative temi s: nonnic assets as compared to net capital forma- ' uniaietns tion is entirely reflected in the degadation 42,060,516= 34,948,397 + 4,703,654 + effects reducing the quality of enviromnental (7,305,293-4,897,328) capital (-Degrq,,,,,) presented in the colurrm of changes in Table 10-7 to: the cmnvronment Ynl = C + (An,,. - 1An,,) + (Ex-M) or, 39,662,772 = 34,948,897 + (24,245,455 - 21,939,545) + (7,305,293 - 4,897,328) Table 10-6: Eolarged Inputl/Oupu Scheme with Asset Balances (billions of Mcxican pcsos) and Adjustmeints for Degadation of Air, Water, and Land (physical unils) Economic acdividar .~~~~~~~~~~~~~~~~~~~~~~~~~~~~~w,n . -. EeanomZe essets Znviron - Ecwioinlc assets mait Nonpro- Phsical duced ,Vonproduced unit of fRcst of FY:at Produced economic eovlronmrn- meAsure- Production world conAmplon assets asse. far aosses mnntr Opening assets lCo.p.ec Ko.np.ec Ko.np.env Produced asset 111,162,310 Land Water Air . . ~~~~~~~~~~a Economic supply P M . Total 75,706,918 4,897,328 Economic uses Ci Ex C Total 28,315,216 7,305,293 34,948,897 10,034,840 Depreciation Depr -Depr 5,331,186 (5,331,186) Emvironmental uses: Degradatlon concerns Ci.depl -Degr.npeav - Land Soil erosion 420,992,059 -420,992,059 tons Solid wastes 18,228,157 -18,228,157 tons Water thousand a .ibic Ground water use 2,456 -26meters Waler pollution 2,359,275 -359,273 tonsDSO . Air Sulfur dioxide 6,646,070 -6,646,070 tons Nitrogen oxides 1,804,408 -1,804,408 tons Hydrocarbons 2,383,030 -t,83,08 lon5s Carbon monoxide 17,967,872 -17,67,872 tons Suspended particles 477,529 -477,S29 tons Net Product: EDP2 Yn2 lAn2.ec (=lAnl.cc) -IAnJ.env Closing assets .Kl.p.ec K1.np.ec Ko.np.env Produced assets L l5,86,965 Land Water Air 274 The elements in the above expression that market value minus cost including a nor- have changed are net product and net accu- mal profit. An alternative value proposed mulation. All other elements are by El Serafy (1989) suggests that both oil unaffected. and timber be valued as the sum of dis- The table includes revaluations for oil, tim- counted values of depletion allowances that ber, and land. The revaluations for oil and would be needed to secure a continuous timber are residuals obtained as the differ- income stream (after deduction of the ence between the value of closing minus depletion allowance) even after the natural opening stocks and the net changes due to resource has been depleted. Both for oil new finds and depletions, which are all val- and timber the depletion allowance is much ued separately. The revaluation of oil and lower than the net rent. For oil the net rent timber therefore includes two elements: the is 1,162 Mexican pesos per barrel and the first one refrrs to the change in the value of depletion allowance is only 160 Mexican resourees remaining between the opening pesos per barrel; for timber the net rent is and closing balance sheets, and the second 21.527 Mexican - pesos per -cubic meter, reflects the revaluation of the extracted oil while the depletion allowance is only 1.46 or timber between the opening balance pesos per cubic meter. If these much lower sheet and the moment of exaction. valuations would be applied to the extrac- The revaluation for land used in economic tion of oil and timber, the cost of depletion activities is negative. The reason is that would be nuch lower and thus EDP1 land is transfred at its value as forest would be higher. The quantitative result of land, which is derived as the sum of dis- these altenative valuations as presented in counted revemnes accruing if all timber Table 10-7a, supplament, shows that would be harvested for purposes of lum- EDP1 would increase from 39,662,772 to bcring this value was estimated to be 41,795,147, and te element of revalution 38.15 million pesos per hectare. Once in- due to extaction would be correspondingly corpoated into the column for nonpr- decreased with the same amount from duced economic assets at this value, the 84,820,456 to 82,688,086 land is then revalued to the value per hec- 2. EDP2. Table 10-8 presents the same deg- tare of its use in economic activities. In iadation elements as were included in Ta- most instances this value was much lower ble 10-6 in physical terms. All than the value per hectare of forest land counterpzrts of the degradation cost pre- (agriculture: 2.64 million pesos per hect- sented as Cldegr in the column for pro- are; livestock holding: 1.99 million pcsos duction are included in the column for the per hectare); forest land that was changed environment None of the degradation to waste land after use in shifting agricul- therefore affects net accumulation of eco- tural cultivation was assumed to have no nomic assets: the only effects are on net value after this use. In the case of urban- accumulation of environmental assets ized land, ihe value was higher, that is, (IAn2env). 75.50 million pesos per hectare. As a result The revised national accounts identity is ofthese gerally lower land values in eco- presented in the row of net product, that nomic uses, the revaluation elemen is, (-18,290) is negative in Table 10-7. The valuation used ffor oil and timber in Yn2 = C + (An2a - IAU2ar) + (Ex-M Table 10-7 is the mc. rent value, that is, the or m quantitative terms: ' The large differences between the two valuations are due to the long peiod in which oil and timber would be available. Seval quions may be asked with regard to the alterative valuation. Would it indeed be feasible to find in a country like Mexico an alternative investment potential that would be able to absorb the depletion allowances of oil and timber production. If tbht were the case, why would oil and tinber not be exploited more rapidly? However, if resources were exploited more rapidly, prices of the products (oil and timber) may drop as a result of increased supply, and altenative investment possibilities nay be reduced, which would result in lower interest rates. This would mean, that in the long run, the net rnmt method and the El serat method may result in similar valuations. 10. Iseegraaed Eviromaeat:a and EconomicAccosdutg A Case Sthdyfor Mexico 275 able 10-7: Enlargd Input/Output Scheme wnth Asl Balancs and AdjiustmctS for Oil Dcpiction, Land Use, and Dclfration (billios of Mexican pcsos) fcronea tdvtida Environ- Economic sse mat Nonproduced Nonprduced SAm of FI Produced economsc avironmntal roducgon - wd coneampo ass - s are as= Opening assets lCop.ec Konpec Kop.env Psuded asset 111,162,310 Tunber 46C,88,404 Oil 66,84,000 Land 339g259,491 1,483,758,406 Economic supply P M Total 75,706,918 4,897,328 Economic uss c Ex C otalO 28,315,216 7,305,293 34,948897 10,034,840 Depreiation Depr --Depr 5,331,186 (5,33t,186) EnvironmentS USe= Oil atraction concerns CLdepl -DCpL.npeC Oil depletion . 1,469,930 -1,469,930 New firds of oil Lnp.ee -Lnpenv 48Z.320 -"2230 Deforation concerns Logng 164,165 -DepLnapeur -164,165 Forest land Transfernrd to economic uses Lnpec -Lnp.env 20,529,501 -20,529,501 Transfer lossts 763,649 -DepLnp.cov -76S,649 Nct product: EDPI Ynl Fx - M C lAn2.ec 24,245,455 -IAnlcnv 39,662772 2,407,965 34,948,897 -213939,445 Revaluation Ret.p.cc Rcvnnp.ec Rev.nps.av. Produced asse2s rumber 32,368,256 Oil -S,452,200 Land -18,290,522 Cloing assets KLp.ec Klnp.ec Kolp.cc Produced ae115 55 Timber 78,428,847 Oil 118,048,500 Land 341,498,470 1,463,228,905 276 36,448,314 = 34,890,558 + (24,245,455 - 25,095,664) + (7,305,293 - 4,897,328) which is different in a number of respects from the traditional identity in SCNM defined in Table 10-2 as Table 1-7a: Supplemnct EDP and AsetBaboasAlteatively Yn=C+In +(EX-)or Valued on the Dask of Nct Rat and a Depletion Alowance in quantitative terms: (bills of Maiwn pesos) 42,060,516 = 34,948,897 + 4,703,654 + (7,305,293 - 4,897,328) Net capital accumulation of economic assets minus that of environmental assets is negative as compared to a positive val- Openings stock ue for capital formation in economic ac- Oil 66;84 000 counting, and this results in a much lower value for Yn2 as compared to Yn. It fin should furthermore be noted that in the New finds of oil 482,230 derivation of Yn2, an additional deduc- Newletion oil30 -tion has been made for environmental ser- VDepleton vices produced by the government in the let rent (1,469,930) form of sanitation services. They are Depletion allowance (202400) treated as intermediate conswuption of Depletion allowance{(202,400) (domestic) household production activi- 'runber, net reduction ties. Because these expenditures (58,339) Valued on basis oft are dealt with as final expenditures in Net rent (9Z7,814) SCNM, this treatment lowers final con- Depletion allowance (6x,969) sumption (C) from 34,948,897 (in Table Depletion allowance (62,969) 10-2) to 34,890,558 and correspondingly Adjustment to EDP1 2,132,375 reduces Yn2 fiurther as compared with its EDP1 39,662,772 value in previous tables. EDP1.ADJ. 41,795,147 The incorporation of values for output of enviromnental protection services into the - Revaluation, based on -etable permits comparison with the core- depletion allowance 82688,082 sponding degradation effects. From the Net rent 84,840,456 macro presentation in Table 10-8 it can be Adjustment to revaluation (2,132,375) observed that the total value of environ- mental protection services (151,194= Closing stocks 92,855+58,339) is only 5 percent of the TIm ber '78428,847 total value of environmenlal degadation Oil 118,048,500 cost (3,156,119). It is questionable, how- ever, how to interpret these figures, as it is not certain whether the degradation ef- fects measured are gross, that is, before the protection services were carried out, or are aet values after incorporation of the effects of these services. Given the man- ner in which the degradation effects are estimated, it is likely that at least some of the effects measured are gross effects and 10. Integrated Environental and Economic Accounting: A Case Studyfor Meico 277 Table 10-8: Enlarged Input/Output Schemc with Asset Balances induding Adjustments for Dcgmdation and Environmental PrOttion Ecpcnditurti (billions of Medcan pesos) Economic acahida Envia-onem Economic asse Nonproduced Fma conymp- Nonproduced eaviranmwloe P- ducton. Rest of wold lon Produced capil cconomicOc anew Opening asscts Ko.np.cc Konp.ec Ko.p.enav Produced assets 111,162,310 46,988,404 Timber Oil 66,584,00 Land 339,259,491 1,483,758,406 Water Air Economic supply P M Total 75,706,918 4,897,32B Economic uses ci E c I I I Total 28,315,216 7,305,293 34,94897 10,034,840 of whidLr Environmental P.enwp P.cnvp Protection services Industty 92,855 291,385 Households 58,339 Depreciation Depr -Depr 5,331,186 Environmental uses CLdcpl Lnp.ce -Dcpl.np.cnv. Oil extraction, -Depl.np-ec deforestation, and land use concerns 19,541,801 -21,939,545 (tot2l) 2,397,744 Deygadation concerns CLde.-- -Degr.np.env. Land Soil erosion 448,880 -448 - 00 Solid wasies 197,269 -197,26 Water -191,568 Ground water use 191,568 -66',456 Water polution 662,456 Air -234,792 Sulfur dioxide 234,792 -137,442 Nitrogen oxides 137,442 -127,409 Hydrocarbons 127,409 -83,477 Carbon monoxide L072,826 Suspcndcd partlides 83,4Z7 -3,156,119 Subtotal degradation 3,156,119 Nct product EDP 2 Yn2 Ec - M C iAn.ec2 r=iAn.ecl) -lAn.erw2 36,44B,314 2,407,965 34,890,556 24,245,455 -25,095,664 Degradation Rev.p.ec Revwnp.ec Revnp.env Pmduced assets Timber 32,388,256 Oil 52,452,200 Land -18,290,522 Cosing assets KI.pAec Kl.pcc Klpcc Produced assets 115,865,965 Trimber 78,428,547 Oil 118,048,500 Land 341,498,470 1,463,228,905 Water Air 278 that therefore there is some double count- for operational govemment policies. Therefore, ing, because the cost of environmental the analysis has been extended below to identi- protection services and the degadation fy the depletion and degrdation effects by eco- they try to eliminate are deducted at the nomic activities and determine the sectors that same time to arrive at EDP2. are using the economic assets in their produc- 3. Comparative Analysis of EDP] and tion processes. The sectoral analysis focuses, EDP2. The analysis presented above is among others, on three elements: value added, summarzed in Table 10-9. The table balances of econonic assets including pro- shows in dramatic form how changes in net duced as well as nonproduced assets, and also product from NDP to EDP1 and EDP2 the environmental protection expenses made by would have consequences for analysis. The different sectors. The quantitative results of table shows that final consumption is 83 this study are reflected in Tables 10-10 and percent of NDP and nct capital formation 10-11. is 11 percent. When changing to EDPF, fi- Value aded. Table 10-10 presents a nal consumption increases to 88 percent of breakdown of the production data by economic EDP I and net capital accumulation is less activities, including not only the traditional out- than 6 percent. It is true that net accumula- put and intrmediate consumption components, tion of economic assets would be nearly 12 but also the values of depletion, degradation, percent of EDPI, but the efet of this in- and land use efficts Envimental protection crease is eliminated because a large part of expendies are also identified in the table for the net accumulation in economic assets is each industry and (domestic) household pro- directly taken from the environment (the duction activities, and presented as "of which" enviromental capital is reduced by 6 per- items that ae reflected m the output and inter- cent of EDPL). When extending the analy- mediate consumption figures of each industry. sis to EDP2, final consumption is farther The net product concepts-NDP, EDP1, and increased to nearly 96 percent and net EDP2-have been identified in each of the capital accumulatimo becomes a negative three sections of the table: NDP is calculated -2 percent, which is the net result of an in- first fllowed by EDP1 after incorporation of crease in net accumulation of econonic as- the depletion and land use effects, and dten sets to nearly 13 percent of EDP2 and a EDP2 is calculated after incorporation of the decrease of envn al pital which degradation effects. The environmental uses amounts to -15 percent of EDP2. presented are the same as in Tables 10-7 and 10-8; the totals fir the national economy be- twee Table 10-10 and those two tables Comparison of SCNM and SCEEhM coincide. Aggrega*esbyEoj"CN ndmicActv SCAn additional economic activity, called Aggregates by Economic Activities household producton activites, is introduced The analysis above of environmental im- in Table 10-10 in order to allocate the environ- pacts on economic aggregates has been carried mental impacts of household consmption. out in macro format, showing how main aggre- This column also includes the environmental gates of net product are affected by the incor- expenses made by households. The environ- poraion of depletion and degradaton effects; menal proteton expenses made by govern- how net capital formation changes into a con- ment on behalf of households, which are cept of net accumulafion that refers to all eco- teted in SCNM as final consumption and nomic assets, produced as well as thus added to NDP, are deducted in the table nonproduced, and finally how final consump- for the calculation of EDP2 in the same manner tion and net product are affected by a diffeent as this was done in Table 104. tatmn of environmenl expenditures that The oil concern presented in the table only are included in final consumption of SCNM. covers oil depletion that is recorded as environ- While such macroanalysis is useful, it does not mentl cost of the oil industry. New finds of oil prmvide the inforation hat would be needed are not dealk with in this part of the table, 19. Inregrated Environmental and Economic Accowundg: A Case-Studyfor Mexico 279 Table 10-9: Comparative Analys of Expenditure Distribution of NDP, EDPl, and EDP2 Paean of Pacer of Pacen of MIP NDP EDP] EDPZ ZDLP EDP2 Net productfcxpenditurc 42,060516 39,662,772 36,44%14 Final consumplion 34,94.897 83109 34,948,897 85.12 34,890,558 95.73 Capital accumulation, net 4,70,654 11.18 2,305.910 5.81 -850,209 -2.33 Economic asscs 4,703064 11.18 4,703,64 11.86 4,703,654 12I0 Environmental asseLs -2,397,744 -6-05 5S53,86; -1524 Expons-imports 2,407,965 5.73 2,407,96S 6.07 2,407965 6.61 because they arc treated as net acc on of and household prduction activifies. Solid economic assets with a counterpart negative wastes are assumed to be only generted by entry for net accumulaton of environmental as- households in their capacity as consumers. sets. The oil industry includes only extracion Therefore all envirnmental impacts of solid and not refining, wastes have been allocated to the column for The deforestation concern includes two ele- household production activities. ments: logging and transfer losses due to the Capital Stock and Capital Accumulation. tansfer of environmental land to economic The balances of ecenomic assets by economic uses. The depletion cost of logging is allocated activities are presented in Table 10-11. The to the forestry industry. Reforestation is as- rows of the table include, for each sector, asset sumed to be the result of forestry activities and balances for produced assets and relevant non- therefore deducted from trees lost through log- produced assets. The table distinguishes be- ging of timber in forestry. The transfer losses tween fixed assets and stocks. For the fixed are allocated to agriculture and animal fming assets, the asset balances include as columns and breeding, insofar as they concern the losses the opening and closing stocks, gross fixed due to transfer of enviromnental land to these capital fornation, consumption of fixed capital economic activities. Losses allocated to agri- and net accumulation of economic, nonpro- culture also include losses due to forest fires duced fixed assets. The closing balance of and losses of trees due to the conversion of for- fixed assets is equal to the opening balance est land into waste land; allocation of these plus gross fixed capital fonnation, minus con- losses to agrculture is based on the assumption sumption of fixed capital plus net accumulation that these losses