&DPw /7/ Public Disclosure Authorized

1 71 1z1 Discussion Papers Issues for Public Disclosure Authorized Infrastructure Management in the 1990s Public Disclosure Authorized

Arturo Public Disclosure Authorized Recent World Bank Discussion Papers

No. 112 StrengtheningProtection of IntelleaualProperty in DevelopingCountries: A Surveyof the Literature.Wolfgang Siebeck, editor,with Robert E. Evenson,William Lesser, and CaslosA. Primo Braga

No. 113 World Bank Lending for Small and Medium Enterprises. Leila Webster

No. 114 UsingKnowledgefrom Social Science in DevelopmentProjects. Michael M. Cemea

No. 115 DesigningMajor Policy Reforn: Lessonsfromthe TransportSector. Ian G. Heggie

No. 116 Women's Work, Education,and Family We!farein Peru.Barbara K. Herz and Shahidur R. Khandker, editors

No. 117 DevelopingFinancial Institutionsfor the Poorand ReducingBarriers to AccessforWomen. Sharon L. Holt and Helena Ribe

No. 118 Improvingthe Pe!formanceof Soviet Enterprises.John Nellis

No. 119 PublicEnterprise Reform: Lessonsfrom the Pastand Issuesforthe Future.Ahmed Galal

No. 120 The InformationTechnology Revolution and EconomicDevelopment. Nagy K. Hanna

No. 121 PromotingRural Cooperativesin DevelopingCountries: The Case of Sub-SaharanAfrica. Avishay Braverman, J. Luis Guasch, Monika Huppi, and Lorenz Pohlrneier

No. 122 Pe!formanceEvaluationfor Public Enterprises. Leroy P. Jones

No. 123 UrbanHousing Reform in China:An EconomicAnalysis. George S. Tolley

No. 124 The New FiscalFederalism in Brazil. Anwar Shah

No. 125 HousingReform in SocialistEconomies. Bertrand Renaud

No. 126 AgriculturalTechnology in Sub-SaharanAfrica: A Workshopon ResearchIssues. Suzanne Gnaegy andJock R. Anderson, editors

No. 127 UsingIndigenous Knowledge in AgriculturalDevelopment. D. Michael Warren

No. 128 Researchon Irigationand DrainageTechnologies: Fifteen Years of WorldBank Experience.Raed Safadi and Herve Plusquellec

No. 129 Rent Controlin DevelopingCountries. Stephen Malpezzi and Gwendolyn Ball

No. 130 Patters of DirectForeign Investment in China. Zafar Shah Khan

No. 131 A New View ofEconomicGrowth: Four Lectures. Maurice FG. Scott

No. 132 Adjusting EducationalPolicies: Conserving Resources While RaisingSchool Quality. Bruce Fuller and AkhliuHabte, editors

No. 133 Letting GirlsLeam: PromisingApproaches in Primaryand SecondaryEducation. Barbara Herz, K. Subbarao, Masooma Habib, and Laura Raney

No. 134 ForestEconomics and PolicyAnalysis: An Overview.William F. Hyde and David H. Newman, with a contribution by Roger A. Sedjo No. 135 A StrategyforFisheries Development. Eduardo Loayza, in collaboration with Lucian M. Sprague

No. 136 StrengtheningPublic Service Accountability: A ConceptualFramework. Samuel Paul

No. 137 DeferredCost Recoveryfor Higher Education: Student Loan Programsin DevelopingCountries. Douglas Albrecht and Adrian Zidermnan

No. 138 Coal Pricingin China:Issues and ReformStrategy. Yves Albouy

No. 139 PortfolioPerformance of SelectedSocial Security Institutes in LatinAmerica. Carmelo Mesa-Lago

(Continued on the inside back cover.) 1 71~u World Baik Discussion Papers Issues for Infrastructure Management in the 1990s

Arturo Israel

The World Bank Washington, D.C. Copyright C 1992 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A.

All rights reserved Manufactured in the United States of America First printing August 1992 Second printing February 1994

Discussion Papers present results of country analysisor research that are circulated to encourage discussion and comment within the development community. To present these results with the least possible delay, the typescript of this paper has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibilityfor errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressedin this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliatedorganizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibilitywhatsoever for any consequence of their use. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates,or its Board or member countries concerning the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitation of its boundaries or its national affiliation. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts01923, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabeticaltitle list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is availablefree of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'1ena, 75116 Paris, France.

ISSN: 0259-210X

Arturo Israel is seriior adviser in the World Bank's Infrastructure and Urban Development Department..

Library of Congress Cataloging-in-Publication Data

Israel, Arturo. Issues for infrastructure management in the 1990s / Arturo Israel. p. cm.-(World Bank discussionpapers; 171) Includes bibliographical references. ISBN 0-8213-2193-5 1. Infrastructure (Economnics)-Developing countries. I. Tide. II. Series. HC59.72.C3187 1992 363'.09172'4-dc2O 92-24118 CIP - iii -

Foreword

Infrastructure- broadlydefined as electricpower, irrigation,transport, teleconications, water supply, and sanitation- will play a key role in facilitatingthe resumptionand acceleration of growth in developingcountries. Well-performing infrastructure services are crucial for enabling countriesto participatein and benefitfrom such developmentsas the globalizadonof trade. They are equallyimportant to effectiveimpl6mentation of povertyreducdon programs. The closelinks between infrastructureservices and the environmentare also well recognized.

Expectationsabout the role of infrastructurein developmenthave changedover dme. While infrastructure'simportance to developmenthas long been understood,its impacthas been blunted by other constraintsthat blocked or deflectedits effects throughoutthe economy. Over the last decade,policy reform, especially in the contextof structuraladjustment, has broughtmany countries to the point where they are now able to use the resourcesdevoted to infrastructureservices more effectivelyfor development.

The studyanalyzes a nmber of issuesthat are commonto all infrastructuresectors. Some, such as the reliabilityof infrastructureservices, emerge from the need to adaptthose servicesto the stricter requirementsof an open economyand to improvetheir performance.But most of the issues emergefrom exploringthe specificimplications for infrastructureof the changingrole of the private sectorand competitivemarkets in the provisionand regulationof infrastructureservices. A dominant theme arising fromthis analysisis the need to reorientthinking about infrastructureaway from the traditional emphasison supply and toward demand considerationsand the quantitativebut also qualitativeneeds of users, clients, and beneficiaries.

The studysummarizes much of the backgroundwork undertakenfor a sectoralpolicy review of infrastructure.It makes several operationalrecommendation and proposesresearch priorities. Follow up work to this study constitutesan importantcomponent of the Deparment's researchand policy activities. V. (LN' Louis Y. Pouliquen Director Inrastuctu and UrbanDevelopment Department - iv -

Acknowledgements

A numberof Bank staffand consultantscontributed to the preparationof this report. In particular,Hans Adler contributedsignificantly to the preparationof the chapterson the statusof infrastructureand reliability;Frannie Humplick(MMI) conducted the major backgroundwork on reliability;Elliott Berg and AnnMcDermott are the principalauthors of the chapteron privatesector fmancing;Richard Moore (AmericanUniversity) contributed to the analysisof institutionalissues; Alan Krupnickand WinstonHarington (Resourcesfor the Future)advised on environmentalissues and WilliamFox (Universityof Tennessee)on economicaspects; Jennie Wishart (from the Hiclding Corporation)helped prepare the statisticalbasis for the analysisof the conditionof infrastructure; and BavaniKrishnamurthi conducted research in severalareas. The paper has also benefitedfrom the commentsof a large numberof Bank staff and outsiders,of whichwe wish to highlightWarren Baum,Philip Blackshaw, Ralph Gakenheimer (MIT), AliceGalenson, Jeffrey Gutman, Ian Heggie, Rita Hilton, ChristineKessides, Ayse Kudat, Louis Pouliquen,Louis Thompsonand Herman van der Tak. BarbaraGregory, Gaby Menassie and Rita Vartaniandid an excellentjob in producingthe successivedrafts and preparingthe studyfor publication. The uual caveatsapply. v -

Table of Contents

-Page

1. Conclusions and Recommendations ...... 1 Overall Conclusions...... 1 OperationalRecommendations ...... 2 Research Priorities ...... 4

2. The Current Status of Infrastructure.. 6 Introduction.. 6 InfrastructureCharacteristics .. 7 The Bank's Experience in Infrastructure .. 9 Infrastructureat the End of the 1980s. 11 Problems with the TraditionalAnalysis .. 21

Table 2.1: Bank/IDA Lending for Infrastructure.10 Table 2.2: Bank/IDA Projects PerformanceEvaluations, 1974-90.10 Table 2.3 Achievementof InstitutionalDevelopment Objectives: OperationsEvaluation Department, 1978-87.11 Table 2.4 Central GovernmentCapital Expendituresin Infrastructure During 1970s and 1980s .13 Table 2.5 Access to Infrastructure, 1980 and 1990: Rural and Urban PopulationCovered .14 Box 2.1 Conditionof Roads .16 Box 2.2 India: Irrigation Sector - Condition,Operation, and Finance .18

3. Infrastructure's Unkages with the Rest of the Economy ...... 23 MacroeconomicLinkages ...... 23 Pricing and Cost Recovery ...... 26 Effects on Poverty ...... 27 Effects on the Environment ...... 28 The Need for Planning ...... 29

Annex 3.1: InfrastructureSector AdjustmentLoans/Credits ...... 31 - vi -

Page

4. The Reliabilityof InfrastructureServices ...... 33 Definition...... 34 Demand for Reliability ...... 34 Effects of Unreliability ...... 35 The Growing Importanceof Reliability ...... 40 How to Build Reliabilityinto Operations ...... 45

Box 4. 1: Women and the Reliabilityof InfrastructureServices .39 Box 4.2: Global Market Connections: DevelopingCountries are GettingInvolved ...... 41 Box 4.3: Effect of InfrastructureDeficiencies on India's InternationalTrade Chains ...... 44 Figure 4.1: InternationalTrade Cycle ...... 36 Figure 4.2: Holistic Representationof InfrastructureReliability ...... 46 Figure 4.3: Railroad Lines Serving Industry ...... 47

5. ReorientingInfrastructure Toward Demand: Using Marketsand the Private Sector ...... 49 The Role of CompetitiveMarkets ...... 49 The Private Sector Role ...... 53 The Central Role of Regulation ...... 56

Table 5.1: InfrastructurePotential for Competition ...... 52

6. AlternativeMechanisms for PrivateSector Flnancingof Infrastructure...... 59 Types of Financing ...... 59 Costs of Private Finacing ...... 65 Inventory of Privately Financed Project ...... 65 Potential Costs and Benefits ...... 66 Potential for Use ...... 68

Table 6.1: Comparisonsof Terms of Three BOT Projects ...... 61 Table 6.2: BOT Variants ...... 62

Annex 6.1 Privately Financed Infrastructureby Country ...... 70 Privately Financed Infrastructureby InfrastructureType ...... 73 - vii -

Page

7. DemandOrientation through Competition Surrogates: Decentralizationand Voice ...... 76 Decentralization ...... 76 Voice ...... 81 The Bias against Maintenance...... 86

Table 7.1: InfrastructureActivities of Potential for Exit and Voice ...... 83 Box 7.1: InstitutionalPerformance and Irrigation Development the Need for BeneficiaryParticipation ...... 85

Appendix...... 88 Table A. 1: World Bank Lending for Infrastructure(Projects & Loan Amount) ...... 88 Table A.2: World Bank Lending for Infrastructureby Region & Subsector ...... 89 Table A.3: World Bank Ex Post Rates of Return of InfrastructureInvestments U-nweighted-AverageRe-estimated ERR ...... 90 Table A.4: World Bank Ex Post Rates of Return of InfrastructureInvestments Weighted-AverageRe-estimated ERR ...... 90 Table A.5: Staff AppraisalReport Ratings for InstitutionalDevelopment Programs ...... 91 Table A.6a: InfrastructureSupply Performance ...... 92 Table A.6b: Conditionof Infrastructure ...... 93 Table A.6c: Infrastructure- Operating Performance ...... 94 Table A.6d: Infrastructure- Financial Performance ...... 95 Table A.7: TelecommunicationsSector: UnsatisfiedDemand, 1979 & 1988 ...... 96

References ...... 97

Chapter1

Condusionsand Recommendations

OverallCondusions 1.1 This chapter summarizesthe main conclusionsof the study and discusses their operationaland researchimplications. The first, generalfinding is that the conditionof infrastructure (electricpower, irrigation,telecommunications, transport, water supplyand sanitation)at end of the 1980swas broadly similarto that of a decadeearlier. Most governmentscontinued relatively high levelsof expenditure,helping to increasecoverage-especially in powerand water and to reducethe urban-ruralgap. However,the generalcondition of infrastructureis poor, performancestandards are low, and the financialsituation in most casesis still weak. Generalizationsare difficultto reach becausethere are majordifferences among subsectorsand countries.

1.2 Second,there is remarkablylittle clarity about the nature and intensityof the real, financial,and fiscallinkages of infrastructurewith the rest of the economy,but the negativeeffects of distorted policies on returns from infrastructureinvestments have been reasonably well documented. These macrolinkageshave been historicallytackled through rigid planning-for example, in input-outputmodels-and insufficientknowledge is availableas a basis for effective macroeconomicand sectoral policy management. The interactions among the infrastructure sectors-their complementarityand substitutability-alsoappear central.

1.3 Third, the challenges faced by infrastructure in developing countries-trade globalizationand liberalization,the increasingreliance on competitivemarkets, sharperproduction and marketingtechniques, faster changesin demandpatterns, increased urbanization, and expected shortagesin investmentcapital-suggest that the way to enhanceinfrastructure's role in growth is mainlyby increasingproductivity, particularly in maintenance. To investmore to compensatefor operationalinefficiencies, as in the past, is unrealistic. Also, demandchanges suggest that there will be increasingemphasis on the qualityof servicesand not only on their availability.

1.4 Fourth, the measurestraditionally pursued to enhanceproductivity have not been successful. Recent developmentssuch as policy reforms and a changingrole for the public and private sectorsare steps in the right direction,but not sufficientto achievethe requiredincreases in productivity. A new approach is needed that focuses on enhancingthe demandorientation and consequentlythe accountabilityof services. By this we meana closer attentionby operatorsto the nature and patternof demand-to markets-and an increasedrole for clients,users, or beneficiaries in the planning and operationof the services (voice). This contrastswith the traditionalsupply orientationof infrastructure,largely determined by the predominanceof technicalelements and the fact that many serviceshave been monopolies,public-sector operated, or both. This alternative approachshould be complementedwith a more explicitattention to the real, financial, and fiscal linkagesof infrastructurewith other economicactivities.

1.5 The specificcomponents of this proposeddemand orientation are: (i) a focuson the reliability of services, as a way of providinga more demand-orientedperspective to issues of -2 - operational performance; reliability is proposed as a practical focus to deal with the issue of quality of services; (ii) an extension of competitive markets beyond what is now the case in most countries-profiting from recent institutionaland technological innovations-as the simplestand most effective way of achieving a demand orientation; (iii) an expanded role for the private sector in the operation and financing of services, for a similar reason; and (iv) an expanded role for competition surrogates-particularly decentralization and voice-in the many infrastructure activities that have to remain as monopolies and/or in the public sector.

1.6 Fifth, private sector financing of infrastructure has considerable potential in many countries. However, some mechanisms such as build-operate-transfer (BOT) schemes are of limited applicability because of the absence of suitable regulatory frameworks and an unwillingness by private sector operators to share risks. Traditional and widely used methods such as concessions, self-help, or even bond financing, where feasible, are more likely to succeed and their impact can be improved substantially. Turning the focus toward these traditional methods will help the developing countries for which BOTs are not a viable alternative.

1.7 Sixth, although the link between infrastructure and the environment was not explored in detail for this paper, the study highlighted the negative environmental impact of distorted prices and badly maintained equipment and fixed facilities. Similarly, the paper did not explicitly explore the links between poverty and infrastructure services, but the scattered available evidence confirms the conclusion that inadequate and unreliable services affect the poor much more intensely- exponentially-than they affect higher-income groups. In fact, a working infrastructure might be one of the most effective poverty reduction activities.

1.8 Seventh, the concept of infrastructure is being constantly expanded, to include not only fixed facilities but also equipment and "social' infrastructure. This approach tries to present as unified a large set of activities that are quite heterogeneous. This trend ought to be reversed, to focus on more homogeneous groups. Activities that can operate in the private sector in competitive or contestable markets should be dealt with as standard industrial activities sometimes requiring regulation. Public sector attention on infrastructure should be concentrated on those public goods that need to be operated as monopolies or quasi monopolies. These public goods face difficult problems: their performance can be enhanced only through competition surrogates and, since these surrogates are less effective than competitive markets, the performance of these services is likely to always be weaker, thus requiring strong political support.

1.9 Finally, the study has identified a number of areas (listed below) about which applied research and experimentationwould be desirable. Defining a research program aimed at dealing with these issues should be given high priority. Research on infrastructure in developing countries has been scarce in recent years.

Operational Recommendations

1.10 The findings point to alternative approaches that can help guide developing countries and the World Bank as they tackle infrastructure development. The set of alternative approaches should include: -3 -

1. Emphasizing improvements in operational performance, quality and reliability of services, and maintenance and rehabilitation, instead of increases in capacity through new investments.

2. Promoting competitive markets in infrastructure services, whenever possible, making sure that the institutional and technological innovations that allow competitive markets to function have been fully used.

3. Ensuring that a country has the institutional capacity to manage sectoral policies and to regulate infrastructure services, as needed. Sectoral policy management most likely will require some form of sectoral or subsectoral planning.

4. Promoting, whenever feasible, private sector participation in the operation and financingof infrastructure services; for most countries, complexmechanisms such as (BOT) schemes are unlikely to succeed, but more traditional approaches such as concessions, self-help, or infrastructure funds or banks hold promise.

5. Focusing attention and political support in the public sector on infrastructure services that need to remain public monopolies-to enhance their performance with measures such as the setting of clear rules, decentralization, and increased voice for users.

6. Including as a key ingredient of infrastructure policy management the promotion of stronger participation of users in the decisions of public sector infrastructure suppliers- particularly of those not operating in competitive markets-to give greater weight to demand.

1.11 At the Bank, these approaches could underlie the dialogue with member countries. The approaches should be reflected in the policy dialogue, but also in the preparation of country- strategies and country economic and sector work. The country strategies should include an assessment of the expected role of infrastructure in the countries' development strategies. The strategies should be based on multisectoral infrastructure studies, with a perspective similar to that of this study, to assess common problems across infrastructure actors. The strategies also should be based on traditional sectoral studies and on studies of specific issues such as reliability, the role of competitive markets, regulation, and voice.

1.12 Individual operations could also reflect these alternative approaches. In addition, a multisectoral operation focused on one or more of the special issues discussed here-for example, reliability in the context of an export promotion cycle, the promotion of competitive markets, or decentralization-can be valuable.

1.13 Two other points are worth highlighting. First, telecommunications is emerging as a key sector for most developing countries. The Bank should reevaluate its relatively modest role in this sector, particularly in the context of adjustment or other operations that include privatization. Second, regulation is essential to assure the success of the new sector policies-and this probably is one of the most worrisome aspects of infrastructure. Regulatory problems are evident even in the best of circumstances in developed countries. The difficulties in developing countries are worse: the -4 - conceptualbasis for regulationis not well-developed,the neededinstitutional capacity is consistently underestimated,and politicalsupport is seldomavailable. Improving the qualityof regulationmight be the most importantchallenge for developingcountries trying to improveinfrastructure.

Research Priorities

1.14 For severalreasons, such as other prioritiesin developmentstrategies and a narrow focuson projectanalysis or sectoralplanning, research on infrastructurein developingcountries has been scarce. Surprisinglylitte is known about the specific effects of infrastructure;about the effects,nature, and extentof externalities,and aboutthe natureof macrolinkages.Comparative data amongcountries are rare, inconsistent,and unreliable. The Bank shouldtake the initiativeto help set a researchagenda as soon as possible. Severalpriorities emergefrom this study.

1.15 First, more needsto be knownabout short-and long-termeffects. One aspectwould be the linkagesof infrastructurewith the rest of the economy. The scarceresearch that has been conductedhas little-if any-policy application.A large research agenda can be defined around effects-for example,variations in effectsof infrastructureaccording to regions'economic structure, population, and economicdensities; effects of infrastructureat different stages of development; effects on different units-households, productiveunits, communities;effects of changes in the structure of demand-such as trade liberalizationor elinination of locational subsidies-on infrastructureservices; dynanics of interactionand sequencingamong infrastructure investments; and so on. Other considerationsare the externalitiesof infrastructureon the environnmentand on poverty reduction-and long-termaspects such as the shape and densityof urban areas. Researchalso could help separatethe differingeffects on growth of the quantity,diversity, and reliabilityof services. For individualprojects or groups of projects,the methodologyof cost-benefitanalysis also needsto be improved through the inclusion of as many factors as possible-for example, the environment-andof distributionaland fiscaleffects. Decisions are constantlybeing made about these issueswithout adequateknowledge.

1.16 A second,new area of research should cover technical,institutional, and political ways of extendingcompetitive markets and alsothe possibilitiesof private sectorparticipation. For exnample,consideration should be given to redirectingtechnical research to these two objectives instead of treating technologicalprogress as exogenous. More systematicresearch is needed to assess the potentialof specificinstitutional solutions under differentcircumstances. For example, the low potentialof BOT schemessuggests the need for more research on them, in light of their apparentpopularity. Also, for the remaninigpublic monopolies, more researchis neededabout the potentialof decentralizationand voice.

1.17 A third research area should be the design and implementationof regulation for infrastructure-beginningwith a detailed assessmentof recent experience in developed and developingcountries. A conceptualframework needs to be developedfor the managementof sectoralpolicies including, for exaample,sets of performanceindicators. In fact, policymanagement and researchare seriouslyhampered by unreliabledata. The Bank shouldconsider whether to help produce an infrastructuredata bank or develop other solutions to assure reliable intercountry comparisons. At a minimum,the Bank shouldrecord comparableinformaton on its operations. -5 -

1.18 Last, more reech i neededon the reliabilityand qualityof inatucture sevices, particularlyin the contextof intersectoralinfrastructure operations, such as trade logistics. - 6 -

Chapter2

The CurrentStatus of Infrastructure

Introduction

2.1 Infrastructure(defined in the next section)faces a number of challengesduring the 1990s. The first is to support the resumptionof growth after a decade of stagnationin many developing countries. Many of these have gone through adjustmentprocesses resulting in considerablyrestructured economies, more consistenteconomic policies, less distortedprices, and an ongoingprocess of public sectorrestructuring. However,the supplyresponse to adjustmenthas not been as strong as expected,in part becauseof inadequateinfrastructure services. Second,the productivityof infrastructureis key in determiningthe way in which countrieswill deal with global trends such as the increasingpace of urbanization,the openingup of the economies,the globalization of trade, and technicalinnovations in manufacturingand logistics. Third, growingconcerns with the environmentand poverty reductionalso are heavilylinked to the productivityof infrastructure. Thesechallenges look particularly daunting because infrastructurein developingcountries is alleged to be in a "crisis," in the sense that a shortageand deteriorationof services and investmentsare seriouslyinterfering with the achievementof developmentobjectives. 2.2 Infrastructurehas had a varying role in developmentstrategies. In the 1950s infrastructure,transport in particular, was given a central role: if governmentswere to provide infrastructure, economic growth would greatly accelerate, a view heavily influencedby the experienceof developedcountries. But this assumptionbecame discredited by experience: in most cases infrastructurewas not sufficientto ensure growth. Other measureswere necessary,such as investmentsin directly productiveactivities or in human resource development. Infrastructure becamesecondary or subsumedin other factors(although infrastructure investments have continued at fairly high levels). The relativelylower priority assignedto infrastructurein development strategiesreached its nadir with the switch to a focus on macroeconomicand sectoral policies. Inevitably,this emphasison the macro-variablespushed the priority givento infrastructurecloser to the bottom;this is where it seemsto be now.

2.3 Policy reform programs during the 1980shave resulted in infrastructureservices facinga more consistentset of macroeconomicand sectoralpolicies and a reductionin their expected direct distributionalrole. Servicesare beginningto operatein more competitiveand less regulated markets;many are being privatizedor operatedby the privatesector. Most of these developments are recent and their impact still hard to assess, but they are potentiallyimportant developments, becausedistorted policies and pricesmake most attemptsat operationalimprovements ineffective or counterproductive.The stage is now set for achievingsubstantial improvements in the productivity of infrastructureservices. 7 -

2.4 The purposeof this paper is to analyzea selectednumber of issuesconsidered key to enhancingthe productivityof infrastructureservices in the 1990s. The paper attemptsto look at infrastructureas a whole, reviews recent developments,highlights new insightsemerging from subsectoralor country comparisons,illustrates successful approaches, and proposes a number of measures aimed at increasingthe productivityof infrastructureservices that could be of use to developingcountries, the Bank, and the developmentcommunity.

2.5 This chapterdefines infrastructure, reviews its current status, and containsa brief overview of the Bank's role in infrastructure. Chapter 3 discusses what is known about infrastructure'seffects on the economy-includingeffects on povertyand the environment.The issues surroundingreliability and how it may be used to improveproductivity are coveredin Chapter4. Chapters 5 and 6 review approachesfor reorientingthe managementof infrastructuretoward demand-using markets and the private sector where possible-including some private financial mechanisms. Competitionsurrogates are discussedin Chapter7.

Infrastructure Characteristics

2.6 All definitionsare conveL..ions.The more literaldefinition of infrastructure,which refers to fixed facilitiesor installations,does not provide a sufficientbasis for decidingfrom an operationalpoint of view which sectorsto include. The Bank's administrativeorganization into infrastructuredivisions or departmentsreflects historical convenienceor accident and does not provide a workable approach for deciding on this matter. Few, if any governmentsuse "infrastructure' as a way of organizingtheir administrativeservices. However, the test of the usefulnessof a definitionis whetherthere is a groupof sectorswhich share commoncharacteristics and problemsto a sufficientdegree to justify their being lumpedtogether for analyticalpurposes. Not all the sectors chosenneed to share all the characteristicsor problems,but there should be significantelements of commonality. In this paper, the services providedin the followingfive sectorshave been included:(i) electricpower, (ii) irrigation,(iii) telecommunications, (iv) transport, and (v) water supplyand sanitation,both in the urban and rural dimensions.

2.7 Infrastructureservices result from combiningphysical facilities, hardware- vehicles and equipment-and other inputs. The focusof this paper is on those servicesand not only on the physicalfacilities.

2.8 Traditionally,infrastructure has been providedmainly by the public sector and has the followingother characteristics:considerable lumpiness in its investments;economies of scale, resulting in monopolies;high level of externalities (positive or negative); intermediate input characteristics;important network effects; and difficultiesin recoveringcosts. This characterization is still generallyaccurate but distinctionsmust be made, for example,between urban, rural, and inter-urbaninfrastructure; trunk and feeders;fixed facilities and movingones; and facilitiesoperated by enterprises-publicor private-or by central governmentagencies.

2.9 Infrastructuremay be examinedfrom several perspectives. For example,from a maintenance perspective, there is a spectrum of fixed facilities according to their usage. At the simpler extreme, fixed facilities combinedwith fixed machinery-a thermal power plant or a -8 -

telecommunications center-are the easiest because their maintenance is related mainly to normal wear and tear. At the other extreme, highways require the most complex maintenance because the wear and tear is related not only to factors such as time, geographical dispersion, and climatic conditions, but also to usage by a heterogenous fleet of vehicles with different axle-loads and with potential overloading. Maintenance of other fixed facilities is simpler. A railway track carries more homogeneous vehicles-and, institutionally, the provider of the fixed facilities has owned the vehicles, at least until now. Fixed facilities that transport water also have relatively simple maintenance needs. In fact, from this-and other-points of view these subsectors can be grouped into the "water cluster," the "machine cluster," and the "vehicle cluster."

2.10 From the point of view of demand, highway infrastructure is the most complex: road transport interacts with a large number of sectors, has high network effects, and is a widespread intermediate input. Other complications are the lack of direct contacts between suppliers (highway departments) and users, and the difficulty of operating highways-as distinct from road transport-competitively. On the other hand, in activities-such as telecommunications or railways-run by enterprises in which suppliers are in direct contact with users, the potential for operating in competitive markets is higher and the activities are in a much better position to respond to demand. Some subsectors have a locational monopoly that enables them to largely disregard demand, especially in thin markets-for example, airports, ports, and water supply.

2.11 In terms of competitive markets, relevant factors are the lumpiness in investments, economies of scale, the demand structure, the linkages between users and suppliers, and the complexity of the required institutional and managerial structures. At one extreme are activities that clearly could operate competitively and in the private sector-such as urban transport. An example at the other extreme is large water supply systems, for which, with few exceptions, a competitive private-sector environment can't be created. Other subsectors in a similar, but not as extreme, situation are waterways and major water storage facilities. The other subsectors fall in-between. A corollary to this analysis has to do with the degree of government intervention. Taking only the potential for monopolistic conditions (encompassing lumpiness, long asset life, and economies of scale), externalities, and potential for cost recovery, highways again appear extreme, needing a high level of government intervention-as do other activities such as major irrigation systems and dams. At the other end, telecommunications, road transport, or thermal power generation seem to require a minimum of intervention relative to other infrastructure subsectors.

2.12 This preliminary analysis suggests a spectrum of infrastructure subsectors ranging from those in which policy or institutional solutions are extremely difficult, to those in which the technical aspects are relatively simple and lend themselves to efficient operation in a competitive market-in public or private hands-with minimal regulation. Highways-especially urban and interurban-require the most difficult solutions. Other fixed infrastructure, such as major irrigation or water storage facilities, has less acute but similar characteristics. And although the potential for operating railways competitively is high, they have complex managerial needs. Maintenance is a key issue for all infrastructure because institutional,financial, and political factors conspire to consistently underestimate and underfund maintenance. Thus, a complex maintenance function implies an exponentially more complex maintenance operation. 2.13 From an operationalperspective, a clearerdistinction ought to be made amongthese different groups of infrastructureactivities. Activities that have few traditional infrastructure characteristicsought to be treated like commercialor industrialactivities. Conversely, for activities that have many of the complexcharacteristics of infrastructure,countries will have to look for the best possiblepublic sectorrole-and one that needs to be strengthened.

The Bank's Experience in Infrastructure

2.14 A comprehensivereview of the Bank's experiencein infrastructurewill be provided in a forthcomingpolicy paper. But three key trends are germaneto this analysis.

2.15 First, in spiteof the well-publicizedchanges in developmentstrategies, infrastructure has continuedto represent between one-third and one-halfof BankJ/ lending since fiscal 1975 (Table 2.1 and AppendixTables A.1 and A.2). In real terms, infrastructurecontinues to be most important,but its compositionhas changed. The importanceof transporthas declinedwhile that of electricpower has increasedconsiderably. In fiscal 1970-79,transport lending was 38 percent of infrastructurelending, electric power, 28 percent; and water and sanitation,10 percent. In fiscal 1980-89,the percentageswere 27, 34 and 10 percent,respectively.

2.16 Second,infrastructure operations have performed better than average,with 85 percent of all infrastructureprojects analyzed by the OperationsEvaluation Department judged to be satisfactory,compared with a Bank averageof 77 percent(Table 2.2). Telecommunicationswas the best sector-with 96 percent satisfactory-and irrigationthe least favorable,with 25 percent. The ex-postrates of return for these sectors have been consistentlyhigher than the Batik averages, particularlyfor telecommunicationsand highways,often by significantamounts (Appendix Tables A.3 andA.4). Theoutcome ofthe institutionaldevelopment objectives inthose operations,however, has been less favorable. Of the 1,250 projectswith insdtuionaldevelopment components reviewed by OED in 1978-87,only 36 percentwere consideredto have achievedsubstantial success, although infrastructurehad higher successthan most projectsin other sectors. For example,utility projects achievedsubstantial success in 42 percent of the cases (Table 2.3). Even in the sector with the lowest level of substantial and partial achievement-agriculture-the success was greatest in irrigation. A more detailedreview of 40 infrastructureprojects reports that in more than 52 percent of the projects, the treatmentof institutionaldevelopment was rated good, comparedwith projects in other sectorsfor which the figure was 38 percent (AppendixTable A.5). Becausesustainbility requires the achievementof institutionalobjectives, it is not surprising that the outlook for sustainabilityis also relativelylow-although againbetter for infrastructurethan for other sectors. Of 320 infrastructureprojects evaluated by OED in fiscal 1986-90,only 59 percentwere considered likelyto be sustainable,largely because of insdtutionalproblems. That figure was still better than the 51 percentaverage Bank-wide.

I/ Bank includes IBRD and IDA. - 10-

Table 2.1 World Bank Lending for Infrastructurel, (in millions of 1987 U.S. dollars)

1970-79 1980-89 % Share % Share % Share of infra- % Share of Infra- of Total structure of Total structure Sector Amount Lending Lending Amount Lending Lending Irrigation 8,087 7.9 15.0 11,079 6.6 14.8 Electric Power 14,909 14.5 27.7 25,553 15.2 34.2 Telecommunications 3,079 3.0 5.7 2,348 1.4 3.1 Transport 20,225 19.6 37.5 20,001 11.9 26.7 Urban Transport 2,259 2.2 4.2 8,497 5.1 11.4 Water Supply & Sanitation 5,341 5.2 9.9 7,331 4.4 9.8 Total Infrastructure 53,900 100.0 74,809 100.0

Total Bank Lending 103,015 167,625 l Infrastructure Lending as Proportion of Total Lending 52.3 44.6

Note: Some figures do not add up to totals because of rounding. Source: World Bank data. l! Bank includes IBRD and IDA.

Table 2.2 Bank/IDA Projects Performance Evaluations, 1974-90

No. of Projects PercentageRated Sector Evaluated Satisfactory Irrigation 163 75 Electric Power 220 92 Telecommunications 72 96 Transportation 463 83 Water and Sewerage 132 86 Total Infrastructure 1,050 85 All Bank Projects 2,587 77

Source: Operations Evaluation Department. - 11 -

Table 2.3 Achievement of Institutional Development Objectives: Operations Evaluation Department 197W87 (percentages)

Sector Substantial PartLal Negligible Utilities 42 42 15 Transport and Urban 34 45 20 Agriculture 33 42 25 Industry 48 32 20 DevelopmentFinance Companies 40 45 15 Education 31 56 18 Human Resources 23 54 23 Total 36 45 19

Source: OperationsEvaluation Department.

2.17 Third, the Bankapproach to infrastructureoperations has evolvedlike that for other sectors. Initially, infrastructureoperations were purely investmentoriented-with only modest institutionaldevelopment components, focused mainly on the implementingagency. Projectsused conditionality-oftenunsuccessfully-to impose financialdiscipline and also generally supported public ownership and operation of public services. This initial focus has been broadened to encompasssectoral operationsgeared to policy reform and sector-or subsector-wideinstitutional reform programs. These broader operationshave tried to address the real, fiscal, and financial linkagesof infrastructurewith other economicactivities. There are, of course,important variations amonginfrastructure sectors.

Infrastructure at the End of the 1980s

2.18 Thisstudy assessed infrastructure in the late 1980sin developingcountries using three major sources:the Bank's sectoraldata banks, Bank sectoralreviews, and detailedinformation on nine countries-Bangladesh,Brazil, Colombia,Hungary, India, Kenya, Nigeria, , and Thailand(see AppendixTable A.6 for a full treatment). Thesecountries illustrate widely varying infrastructurestatus in all major regions, at different income levels, and of differing sizes and populationdensities. The data have serious limitations,however. First, for some countries and sectorsthe most recent data are generallyfor 1985-87-althoughit is doubtfulthat the situationhas changeddramatically since then. Second,the data are not consistentenough to facilitatecomparison of the 1980ssituation with that of the 1970s-thus, the generalconclusion that progressin improving infrastructurein the 1980swas slowerthan in the 1970scannot be statisticallyverified. In addition, the margin of error is large; the conceptsof infrastructure,operational efficiency, and financial viabilityare not alwaysconsistent among countries or sectors;and monetarydata are rarelyadjusted - 12 - for inflation. Becauseof these shortcomings,the reviewwas supplementedby extensiveinterviews with Bank staff.

2.19 In mostrespects, the reviewfound the statusof infrastructurewas not greatlychanged in the late 1980s from a decade earlier: poor conditions,low performancestandards, and weak financeswere still the norm. Although,even during adjustment,many governmentscontinued high levelsof spendingfor infrastructure-usuallynot as high as before-others did not. Spendingranged from substantialinvestment in East to reductionsin some countries in and Latin America. Partly because of differencesin spendingvolume, servicesvary among countries and subsectors. Some countries were able to expand coverage and reduce the urban-rural gap. And supply exceeded demand where growth was slower than expected. Yet despite increased spending, infrastructurein East Asia generally has not been able to keep up with demand. The type of spendingalso varied. Some countries maintainedservice levels by increasedinvestment in new capacity alongside a deteriorating stock. And, whatever the reason, infrastructure generally has responded slowly to new demand. Adjustmentand policy reforms-countrywide or sectoral-have not yet greatly affected services, prices, or financial condition.

2.20 Whether this scenario constitutes an infrastructure crisis is largely a matter of perception. Infrastructure was not a major barrier to growth in the 1980s-particularly in stagnantor decliningeconomies. But what could indeed qualify as a crisis is that infrastructure in its present conditionwill prove a major barrier to future growth, when other barriers have been eliminated. This crisis differs from that in several developed countries, in which a previously adequate infrastructurehas been allowed to deteriorate due mainly to lack of funds instead of serious operationalproblems.

2.21 Expendfture and coverage. This study reached four conclusions about expenditure and coverage. First, public sector expenditure in infrastructure in the 1980s continued in real tenns at levels higher than in the 1970s. Second, as a consequence, the increase in infrastructureservices during this period was generallysufficient to meet the increase in demand from populationgrowth and from higher income-when such an increase occurred. Third, investment also allowed infrastructureto reach a larger share of the population. But, fourth, available capacity was seldom well utilized.

2.22 A study of 53 countries indicates that governmentinvestment for infrastructure was generallyhigher in real terms in the 1980sthan in the 1970s(Table 2.4), and even adjusting for populationincreases. The nine-countrystudy and sectoralinformation generally supportthis conclusion. In a study of road investments in 35 countries, central government capital expendituresin real terms increased in 23 of them in the 1980scompared with the 1970s, and the country studies do not indicateany broad slowdownin road investmentsby the mid-1980s. In the 1980s, the made a specialeffort, the WaterSupply and SanitationDecade, to expand these services. Annual investmentsincreased 50 percent in real terms from about $6 billion in the early 1980s to $9 billion by the end of the decade. The capacity of teleconmnunicatonssystems also was substantallyexpanded. A study of 18 developingcountries - 13 - indicates that investmentsincreased from $4.6 billion in 1979 to $5.5 billion in 1988 (in 1992 prices). Investmentswere higher in 11 of the 18 countries. However, a study of ports in 16 countriesshows no significantdifference between investment levels in the 1980sand 1970s. The country studiesalso found that road networks were expandedand improved in all countries, that traffic generally increasedat 5 to 10 percent annually, and the number of vehicles continuedto grow rapidly. Network expansionsin the first half of the 1980sgenerally ranged between5 and 20 percent for most countries. In five of the seven countries for which comparativedata are available, the network increases were greater or similar to those in 1975-79.

Table 2.4 Central Government Capital Expenditures in Infrastructure During 1970s and 1980s (in real terms)

Number of Countries

Total RAil Capital Pipe- Te_- Sector Expenditure Roads lines Ports Energy Water comm.

Total No. of Countries 53 35 24 16 18 17 20 No. of CountriesWhere iCapitalExpenditues Were Higher in the 1980s 32 23 11 7 12 14 14 Lower in the 1980s 21 12 13 9 6 3 6 ComparedWith 1970s IIIIII

Note: Average annual expendituresin 1987 prices in 1975-79and 198089 are used for comparisonof expendituresin the 1970s and 1980s. Source: IMF GovernmentFinance StatisticsYearbooks

2.23 The supplyof power increasedconsiderably in the 1980s. A studyof 70 developing countriesindicates that it reachedabout 2,000 TWh (Terrawat-hours)in 1989,more than doublethe level of 1980 (Mooreand Smith1990). Annualelectricity consumption increased between 1980 and 1987 from 470 Kwh to 650 Kwh per person in 41 developingcountries. In the nine countries studied,power generationand consumptionincreased at annualrates rangingfrom 5 to 13 percent, lower than the increases of the 1970s, reflecting mainly lower economic growth. Several countries-particularly Bangladesh,Brazil, and India-faced power shortagesin spite of large expenditures, partly because of investmentsthat were not balanced between generation and transmission. In telecommunications,in spite of the higher investments,capacity in most countries is insufficient-causingcongestion, reduced service quality, and annual losses to the economy estinated in some countriesat more than the annualinvestments in telecommunications.Unmet demandfor telephoneconnections in the developingcountries might exceed 20 millioncompared with about43 millionservice connections (Wellenius 1990). A studyof 28 developingcountries indicates - 14 - that between 1977 and 1987 waiting lists have increased, or remained essentially unchanged in 22 of them (Appendix Table A.7). In 1988, telephone waiting periods in a study of 18 countries averaged about 10 years.2/ Finally, port traffic increased in all nine countries studied in detail. Capacity has been adequate to handle general cargo traffic, but ports have not adjusted fast enough to the rapid increase in container traffic.

2.24 As a result of investments and increases in supply during the 1980s, access to infrastructure services has continued to increase, with a growing proportion of people in most countries having access to electricity, safe water, telephones, and transport. The gap between urban and rural populations remains large, but has declined somewhat. The population having access to electricity increased between 1980 and 1990 from 35 to 43 percent, but grew faster in rural areas (from 27 to 33 percent) than in urban ones (from 70 to 74 percent), thus closing the gap slightly (Table 2.5). In water supply, by 1990 about 69 percent of the population had access to safe water, about 63 percent in rural areas and 82 percent in urban. This was a significant improvement over 1980, when rural access amounted to only 30 percent and urban access 77 percent. As a result, almost one billion more people in rural areas had access to safe water.3/ Access to water increased in seven of the nine countries-but not in Bangladesh and Colombia-but many cities faced water shortages severe enough at times to require rationing. Service deficits tended to be concentrated among the poor. More generally, in an increasing number of countries, especially in East Asia and Latin America, nearly all people in urban areas have some access to electricity and water supply. Access to telecommunications, however, is still relatively low in spite of increased capacity. A study of 13 countries shows that in 1989 the number of telephones ranged from 0.5 telephones per 100 people (India) to 18 (Portugal), compared with 40 to 60 in developed countries overall.

Table 2.5 Access to Infrastructure, 1980 and 1990: Rural and Urban Population Covered (in percentages)

Urban Rural Sector 1980 1990 1980 1990 Power 70 74* 27 33* |Water 77 82 30 63 Sanitation 69 72 37 49

Source: World Bank Databases; and U.N. Report of the Economic and SocialCouncil, July, 1990.

21 InternationalTelecommunications Union, Pyramid Research,Inc.

3/ U.N., Report of the Economicand Social Council on Achievementof the InternationalDrinking Water Supplyand SanitationDecade, 1981-1990,July 1990. The effectivenessof this "access' is hard to assess. Many of thewenew systems become quickly non-operationaland in need of rehabilitation,or access is hampered, for example,by deficienciesinside dwellings. These figures are likely to be overestimates. - 15 -

2.25 Although expansion of infrastructure capacity has continued, it has not been equally well used. The existing capacity is generally adequate-or even exceeds demand-for roads in Africa, railways in Africa and in Latin America, for port capacity to handle general cargo, for power in many countries, and for water supply in some countries. The major reasons for excess capacity include the unanticipated slow increase or decline in demand because of stagnating economies irnthe 1980s and-for a few sectors-the long-term decline in demand that has not been matched by reduced capacity (for example, in railways and port capacity for general cargo traffic). Excess capacity is sometimes the result of investment imbalances, as between power generation and transmission and distribution, or when irrigation capacity is wasted because of inadequate investment in canals. On the other hand, the existing capacity is fully used or fails to meet effective demand for roads in many countries in Asia and in some Latin American, European, Middle Eastern, and North African countries, and at peak hours in most large urban areas; for railways in most Asian and Eastern European countries; for power and water supply in some countries, particularly China and India; for telecommunications;and for port capacity to handle container and bulk traffic in almost all countries.

2.26 Railways continue to be a special case. Most have lost traffic in the 1980s although China and India are important exceptions, for well-known reasons. The decline has been sharpest for passenger traffic because of the continuing expansion of roads. In most of the nine countries studied, the railways could carry more traffic with relatively small additional investments. Investment data are not readily available but the level of investments by central governments in railways, pipelines, and air transport combined did not differ much in the 1980s from the 1970s, suggesting a decline in railway investments.

2.27 Infrastructure condition and performance. The condition of infrastructure continues to be poor and to deteriorate in a number of cases. For example, the vast expansion and improvement of roads in developing countries in the 1960s and 1970s "created an infrastructure that has been crumbling in the 1980s and threatens to collapse in the 1990s if not quickly strengthened and protected" (World Bank 1988, p. 5). In 85 countries, 26 percent of the paved network is in poor condition and requires rehabilitation and reconstruction and 42 percent is in fair condition, requiring resurfacing if these roads are not to decline to poor condition (Box 2.1). The condition of unpaved roads is even worse, with 33 percent in poor condition. Regional differences are not very large, except that roads are in particularly poor condition in South Asia because of a rapid growth of traffic on roads built to obsolete standards. In many countries extensive new construction offset the deterioration that resulted from insufficientmaintenance. Subsectors with a rapidly changing role seem to be in the most precarious condition. Railways, for example, are in generally poor condition in Africa, , and Latin America-and somewhat better in Asia.

2.28 Operational standards have remained largely unchanged. The infrastructure sectors suffer from serious operational inefficiencies, including excessive labor, high losses, low reliability, and inadequate maintenance-and there is no evidence of significant improvement in the 1980s. In developing countries,ports averaged about 13 tons of break-bulk cargo per ship-hour, compared with 23 tons in industrial countries. For containerized cargo the difference was even greater-36 tons compared with 100 tons (Hickling 1990, Table 17). Telecommunicationssystems in many countries are frequently congested and a large proportion of attempted calls is not completed. In Morocco, only 20 to 30 percent of peak intercity and international calls were completed. Annual faults averaged 0.9 per line in service, nearly double an acceptable level......

...... ~I 1111 i ill..

~~~~8 1~~~~~~~~~~~i~~~~~~~~~ .. h2~~~~~~~~..

* .~.. ~ ~~~...... - 17 -

2.29 The efficiency of many electric power companies did not improve in the 1980s. A study of about 100 countries indicates that power losses averaged 18 percent of net generation in 1980 and 1987 and the number of consumers per power employee remained at about 70. In Nigeria, for example, system losses were a very high 35 percent, of which 15 percent were technical losses and 20 percent due to poorly maintained meters, theft, and underbilling. The quality of service, including reliability and losses, of many railways is frequently low and is not adequately tailored to shipper needs. A study of 26 railways indicates that they were not able to increase their efficiency in the 1980s.4/ But there are wide differences. For example, the turn-around time of freight cars-the time from one load to the next-averages only three days in China, but averages 25 to 30 days in many African countries.

2.30 In water supply, losses are generally high and Bank efforts to help reduce them have not succeeded. In Brazil, water losses averaged 29 percent in 1987. Losses were also high in India, quality was unsatisfactory, and the supply often was interrupted. In Nigeria, most water-treatment works were unreliable because of frequent equipment failures-so that water quality was poor. In irrigation, water distribution and application are frequently inefficient, unreliable, expensive, and inequitable-as found in five of the country studies. As a result, cropping intensity and yields are well below their potential. There are also extensive losses of production from waterlogging and salinity. The operation of the irrigation network in India has deteriorated considerably in recent years (see Box 2.2). In Morocco, the water efficiency in the Bank-assisted Doukhala project was about 50 percent, compared with an appraisal estimate of about 65 percent.

2.31 The issue that continues to emerge as central is inadequate maintenance, although the degree varies widely among sectors and countries. Current studies give several reasons to explain this shortcoming: insufficient funds and import restrictions in railways; inappropriate maintenance facilities, a shortage of spare parts, an inconsistent policy on equipment maintenance in ports; and insufficient funds and shortages of skilled personnel in water supply and sewerage (see Chapter 7). Inadequate equipment is a common problem. For example, rudimentary signaling and telecommunications systems in railways; inappropriate and excessively costly technologies in water supply; and the low penetration of modem technology in telecommunications. Bad initial construction is believed to have compounded the problem in irrigation, water supply, and ports.

2.32 Weak effects of refonms. Of the policy and institutionalreforms that are of particular interest to infrastructure, the elimination of price distortions, elimination or reduction of subsidies and subsidized services, and a concern for the financial health of enterprises and centralized agencies are among the most important. (Other objectives, such as a larger role for the private sector or deregulation are discussed in Chapter 5). Improved finances of infrastructure operators and a less- distorted price structure are essential but-unfortunately-there is little evidence that such changes have occurred (Appendix Table A.6). Instead, first, some subsectors that have maintained their monopolistic position-telecommunications and most ports-continue to be financially sound. But other monopolistic subsectors such as electric power have seen their financial position eroded largely

i/ Hickling 1990, Table 16 and World Bank RailwayDatabase. tz~~~~~~~~~~~~~~-

AaL 1W C

fio'.r.t ~ t. k'b...... Wfb~~ ~ ~ ~ ~ ~~~~ ...... O...... ~ ~ ~ ~ ~ ~ ~ :X::i ~ ~ ~ ' 0~~: on~~ ~~~~~~~~~~~~~~~g ow ...... ~

...... fA ...... ~~ ...... I~~ ~ ~~ ~...... ~~~~0......

Z C ...... ~ ~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ - ~. f ...... wwx ...... ~ ...... I- . .. . if0 1 1 X' .. ...d .: I~~~~~~~~~~~~~~~ - ppq:: ; ~~~~~~~~~~~~~~~~~~. .. . . b~~~~~~~~~~~~~'~~~~~~~~~~~~~~~~~jIIIi~~~~~~~~~~~~~~~~~~~~~~a o:...... ~~~~ .....~ ~ .. ~~~~~~4~~~~.. ... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... - 19 - because of more intensivepolitical control of tariffs than in previous periods. Second, railways continueto sufferfrom weak financeslargely because they are exposedto competitionbut don't have enoughoperational independence. Third, chargingfor water supplyand irrigationcontinues to be an operationaland sectoral problem. Fourth, countrieshave generallymanaged to collect user chargesto cover road operatingcosts, but trucks do not cover the costs they incur.

2.33 Ports have generally managed to maintain a sound financial position through monopolisticpricing and other measures. But competitionis increasingfor many ports. Tariffs often don't reflect the costs of various services-and there is a danger of financialdifficulties if tariffs are not restructuredto reflectthe growth of containerand bulk traffic. In the nine country studies, port revenues coveredoperating costs, the only exceptionbeing Nigeria. The financial situation of most telecommnunicationsagencies is sound, given their monopoly position and demand in excessof supply. Financialrates of return tend to approach20 percentand the annualreturn on net revaluedassets averaged14 percentin a studyof ten countries. In a studyof 37 countries,cost recovery averaged 172 percent.S/ In fact, telecommunicationsagencies often are an important source of public funds. In Morocco, the system was expectedto transfer $2.1 billion to the government-60 percentof gross revenue-in 1987-94(Wellenius 1990, p. 3). Financingproblems can arise, however,from the high foreignexchange costs of these investments.

2.34 A study of electricpower utilities who were Bank borrowers in 29 countries showed that in 1987-89, only six of them earned at least 8 percent on their investmentsand-for 13 borrowers-rates of return had declinedsince the 1970s(World Bank 1989b, para 28. Thus, the proportion of investmentrequirements that could be financed internallydeclined from 25 to 17 percent (WDR 1988, p.142). The companiesin only two of the nine samplecountries were in a strong financialcondition. Most raitwayscontinue to run heavy financiallosses which imposea heavyburden on governmentbudgets. A reviewof 21 railwaysshowed operating ratios generally exceeding130 percent and only railwayswhich were still the dominantfreight mode-because of the country's policy decisionto protect railways-had significantlylower operatingratios, which still exceeded 100 percent.61 The losses tended to be particularlyhigh for passenger services. The practical result was the well-knownvicious cycle of railways: delayedmodernization, inadequate maintenance,poor services,and traffic lost to roads.

2.35 In irrigation, water charges often do not cover the costs of operations and maintenanceand are not relatedto the valueof water used. Low collectionrates encouragewastage of water, causingsalinization, high water tables, and erosion. Low chargesalso lead to inadequate funds for maintenanceand increasethe burden on governmentbudgets. In India, for example, recoveryrates rarely exceed70 percentof operatingcosts. The tariff structurefor watersupply and sewerageservices in most developingcountries is not relatedto costs-and revenuesrarely finance a reasonableproportion of new investments. Moreover, there are extensivelosses from theft, inadequatemetering, and a failureto bill and collectpromptly. Cost recoveryin urban areas is much

I/ IHickling1990. Table3 andWorld Bank database.

§/ lHickling1990. Table 18 andWorld Bank Railway database. - 20 - higher than in rural ones. By and large, the poor financialcondition in water supplyreflects the view of many governmentsand societiesthat water shouldbe free of charge.

2.36 For roads, the overalllevel of user chargesis adequateto cover actualexpenditures. A studyof 14 countriesindicates that in the mid-1980sroad user revenuesexceeded expenditures in ten of them by marginsranging from about 20 to 35 percent. If maintenanceexpenditures had been increasedto adequatelevels, revenueswould still have exceededexpenditures (Heggie and Quick 1990). Heavytrucks, however,rarely pay fully for the additionalcosts they imposeon road constructionand maintenance. Thus, it is likely that other vehiclesare payingtoo much. What is not consideredin these calculationsis the proportionof chargesthat shouldbe defined as general revenuesand as such deductedfrom the road revenues. If this adjustmentis made, the financial situationof roads is less favorable.

2.37 Variationamong countriesand subsectors. Clearly, there is a wide variety in infrastructureamong sectors, subsectors,and countries. This sectionhas generalizedabout trends as much as possible,but a few additionalpoints canbe made about sectoraland countrygroupings.

2.38 First, a numberof subsectorshave had dynamicgrowth in most countries, often in spiteof negativegeneral economic conditions: telecommunications,road transport,container traffic in ports, and airports. Demandfor these serviceshas grownfast and often countrieshave not been able to supply the necessary services. All indicationsare that demand for these services will continueto increaserapidly, especiallyif countriesresume fast growth. Conversely,railways and generalfreight traffic in ports have continuedto decline,reflecting structural changes in their roles. Water supply, sanitation,and irrigationhave had a mixed record and are more difficultto classify. Water supplyhas been a dynamicsector but the actualdemand is difficultto gaugebecause of price distortions.

2.39 Second,despite the wide variationsamong countries-depending on factorssuch as incomelevel and stage of development,size, populationdensity, economicand political structure, and urbanization-five countrygroupings can be identified:

* High-growthcountries in Asia and now in Latin America-Chile, Indonesia,Korea, and Thailand. These countries will continue to experience high demand for infrastructureand a key issue is how much infrastructureshortages will limit their growth.

* Middle-incomecountries in Latin America,Europe, the , and North Africa-some highly indebted-which are implementing important adjustment programs. Here, as indicated,infrastructure deficiencies were probablynot decisive in explainingthese countries' poor performance in the 1980s,but substantialincreases in productivityare neededif infrastructureis going to contributeto the resumption of growth.

* China and India, which comprisealmost half of the developing-countrypopulation and do not conformto the patterns coveringmost countries. Factors such as the crucial role of railways, power shortagesin India, and the slow developmentin - 21 - telecommunicationssets them apart. The two countries are likely to experience importantstructural changes in infrastructurein the 1990s-particularlya largerrole for road transport and a fast expansionof subsectorsneeded for a more open economy,such as telecommunications,container ports, and electricpower.

* EasternEuropean countries, and those in the formerSoviet Union, which have some of the characteristicsof China and India and will experience several structural changesin infrastructurelike those describedfor China-but also will have changes in the spatialdistribution of productionthat will greatlymodify the level and structure of the demandfor infrastructure.

* Most countriesin Sub-SaharanAfrica, which are undergoingadjustment but stillhave stagnanteconomies, which also requiredrastic increases in infrastructureproductivity despite extremelyweak institutions.

2.40 Third, the analysis in this chapter also suggests that the interaction among infrastructuresubsectors is crucial. An extreme exampleis that of India, where at one point insufficientrailway capacityto carry coal resultedin underuseof thermal power plants which, in turn, affected telecommunications,water supply plants, and water pumps for irrigation. These interactionsneed to be exploredmore systematicallyto better assess investmentpriorities.

Problems with the Traditional Analysis

2.41 One constanttheme that emerges from an extensivereview of the literature is that sectoralanalyses and actionsin the 1980sto improveinfrastructure performance were traditional- similar to past efforts. No breakthroughoccurred, except perhaps in policy or institutional reforms-which have yet to showsignificant results. Most of the poor performancehas beenblamed on technical,financial, or politicalproblems-the politicalproblems being inconsistentpolicies and inefficientpolicy implementation. Inadequate maintenance often is raisedas a centralfactor, but so are obsoleteequipment and poor construction. Institutionaland managerialfactors-including lack of skilledstaff-are certainlybehind many of the reasonsdiscussed, for example,technical indicators such as poor billing and collection,poor maintenance,low wagonturn-around time, power outages, and so on.

2.42 Financialproblems derive from having many services providedby enterprisesor agencies running a deficit and dependingon resource transfers from the central government. Foreign-exchangescarcities indirectly affect services in most countries, even if the agencies are running a surplus. For highwaydepartments, having adequateroad-user charges has not assured financialsolvency, because central governments usually have not transferredall collectedor needed funds to the highway agency. Overregulationand excessivegovernment interference on prices, staffing,investments, and servicesto be providedcontinue to be citedto explainpoor performance.

2.43 The close interactionsamong fmancial,political, and operationalissues have been analyzed, but the extent of those linkagesis still not clearly understood. For example, public enterprisesforced to operateat a deficitand pursuingmultiple and contradictoryobjectives defined - 22 - for politicalor socialreasons cannot achieve a reasonablelevel of operationalperformance. Under such circumstances,it becomesvery difficultto providethe incentivesand motivationto staff.

2.44 Theseundoubtedly are valid reasonsfor poor performanceand need to be addressed, particularlythrough policy and institutionalreform programs. But, as indicated,many of these reasonsare too supply-orientedand insufficientto achievethe performanceimprovements required in the 1990s. A demandorientation is also needed. - 23 -

Chapter 3

Infrastructure's Linkages with the Rest of the Economy

3.1 Since the mid-1970s, infrastructure's economic role in developing countries has been largely overlooked. Instead, such concerns as investments in directly productive sectors, human resource development, and consistency of economic policies have held center stage. Infrastructure is attracting renewed attention, however, because of the developingeconomies' poor supplyresponse to adjustment programs. The need to reassess-and reach a consensus on-infrastructure's proper role in development is not a matter of theory, but of fiscal urgency. No matter how much an economy is opened to the private sector, governments will still have to provide a substantial share of infrastructure-and thus plan how to allocate resources. Such decisions often err-partly because of political or social considerations, but also because of an incomplete understanding of infrastructure's interactions with the rest of the economy. Even well-developed project appraisals fail to account for the fiscal implications.

3.2 The Bank, like other development institutions, has paid scant attention to the economic effects of infrastructure. A review of a sample of public expenditure reviews and sectoral studies reveals that the links between infrastructure and the overall economy are only discussed in general terms-and until recently never systematically (Annex 3.1). Although a few sectoral studies deal with country-specific issues-such as the linkages between trade and transport in India-these studies show that the Bank has no systematic opinion about the main operational issues-such as the proportion of public investments that should be devoted to infrastructure. Yet the new challenges-the globalization of trade, increasing market orientation, and efforts to reduce poverty and protect the environment-beg for an understanding of the linkages.

3.3 This chapter briefly considers current knowledge about the linkages between infrastructure and other economic activities-the need for planning, the role of pricing, and infrastructure's ties to poverty and the environment. The intent is only to highlight issues needing research.

Macroeconomic Linkages

3.4 The evidence on macroeconomic linkages is sketchy-largely confined to developed countries, and highly suspect in its results. For example, a study by Aschauer on the United States and other developed countries concludes that the return on investments in infrastructure is about 40 percent, but the analysis has been found to be incorrect (Aschauer 1989). A particularly difficult problem has been trying to determine whether infrastructure investments have promoted growth or whether expansion of the economy has led to greater demand for infrastructure. It is likely that the causality goes both ways, but research has not separated the different effects in a way that provides precise answers. Also, the notion of diminishingreturns from infrastructure investments is supported by several studies. Research also suggests that public infrastructure investments first crowd out private investment in the year they are made, but lead to greater private investment in later years. - 24 -

Andthe WorldDevelopment Report 1991concluded that a distortedpolicy environment substantially reducedthe return on the investments.

3.5 One shortcomingof mostof thesemacroeconomic explorations is that they do not take into accountthe efficiencywith which increasesin capacityare used. In this respectthey are supply oriented and assume operationalperformance is adequateand implicitlyunchanged in the time covered. In fact, whether an investmentin infrastructurecapacity achieves its full potential contributionto economicoutput depends not onlyon operationalperformance but also on the demand for the service: if effectivedemand is lackingor other servicesare desired, consumptionwill be less than expected.

3.6 This dearth of knowledgeis quite remarkable. The Infrastructureand Urban DevelopmentDepartment is undertakingresearch to assessthe marginalproductivity of infrastructure investmentsin a sampleof developingcountries. The approachis to collectevidence at the macro and micro levels and explore ways of reaching more meaningful conclusions through a complementaryanalysis of both levels. The available information hampers cross-country comparisons:data from differentcountries are oftencollected using different definitions. Evenbasic data-such as the proportionsof infrastructuresectors in GDP-are hardly credible. For example, in 1987 infrastructurewas 16.8 percentof GDP in Bangladeshand 14.1 percent in Morocco,but only 8.1 percentin Brazil or 8.9 percentin India. Most analysisof effects and linkageshas been based on secondary-usuallyaggregated-data. Apparently,systematic primary data on the effects of infrastructurehave not been reportedanywhere.

3.7 Influence on output. Infrastructurecan influence output through supply and demand. It can affectthe supplyof total outputby being a direct input in production;making private capital or labor more productive,or enhancingan area's abilityto produceby attractingworkers or private capitalfrom other regions. Infrastructurecan affectthe demandfor total outputby creating income for workers and entrepreneursthrough the constructionof facilities,crowding in private investment by encouragingit throughcomplementarity, or crowdingout privateinvestment when this is replaced by infrastructureinvestments-and by generatinga greaterdemand for infrastructureservices as the economygrows. The supplyeffects are more longterm than the demandeffects.

3.8 Real,finanda4 andflscal dbiensions. Infrastructure'slinkages-real, financial,and fiscal-with the rest of the economyhave importantpolicy implications.In additionto the linkages to output, the employmenteffects and the importanceof infrastructureinvestments and servicesin the total economyare equallyrelevant. A similaranalysis can be made of financiallinkages-for example,the effectsof borrowingby publicinfrastructure enterprises on the availabilityof creditfor private infrastructure. The fiscal linkagesinclude deficits of public enterprises, subsidies,user charges, and cost recovery. In addition, sectors can be substitutes,as between transport and telecommunications,but moreoften they will be complements,as the linksbetween transport, power, and telecommunications.

3.9 Unfortunately,most evidenceabout the effectsof infrastructureinvolves individual investmentsor services. The most readily availabledata show that Bank-fmancedinfrastructure projects-particularlythose in transport-have higherinternal rates of returnthan the Bank's average - 25 -

(AppendixTable A.3). Althoughthese figureshave seriousproblems of interpretation-sectorsuse differentappraisal methodologies, and ex-postevaluations have incompletestreams of benefits-the figures indicatethat infrastructureinvestments might be amongthe most productive,and suggesta shortageof infrastructurefacilities.

3.10 Long-term effects. The long-term effects of infrastructuremay be the most important-for example, effects on land use, the location of economic activities, the pace of urbanization,and the shapeand densityof cities. Infrastructurecould be key in determiningwhether a developingcountry becomes, for example, an urban societybased on high density and public transportor one with low densityand privatetransport-or, worse, the former for the poor and the latter for the rich. In fact, urbanizationwith its underlyingeconomies of agglomerationmight be one of the main factors explainingeconomic growth in a number of countries. Once urbanization begins to level off, as in Latin America,growth rates begin to slow. This hypothesisneeds to be explored,because conventional wisdom has causalitygoing in the oppositedirection.

3.11 A relatedissue is how the cost functionsof differenttypes of infrastructurebehave under different assumeddensities of economicactivities and population. There is no reason to supposethat the cost functionswill vary followingsimilar patterns. Telephoneswill be different from water, which will be differentfrom the varioustransport modes. Similarly,one wouldexpect differentrelationships for the sectors among countries,because of varying relative factor prices, amongother things. Thesecost relationshipsare the buildingblocks for understandingthe long-term effects of infrastructure. Locationis one of the key organizingideas about interactionsamong differentinfrastructure elements, an importantreason for dealingwitAh infrastructure services as a whole.

3.12 Knowledgeabout long-term effects varies considerably among sectors. For example, for all the bias on short-term effects, they have been considerably researched for transport-particularly highways-and reasonably well for electric power. The big gap is telecommunications-anodd gap in light of telecommunications'obvious major influence on foreign trade, services, and productionmore generally, the links with informatics,and the huge unmet demandin most developingcountries.

3.13 Deteriorationof services. Anotherarea needingresearch is the differencein effects between the addition of infrastructureservices and the deteriorationor suspensionof services. Availableevidence suggeststhat the deteriorationor suspensionmight have more effect. (See Chapter 4 for a discussionof the Nigerian experiencein this respect.) The topic is especially relevant in light of the continuingdifficulties with maintenance. Deterioratinginfrastructure is known to cause increasedcosts for users and suppliers,unreliable services, congestion,private generationof servicesat much higher costs, and-in extremecases-a breakdownof economicand other activities. But the pattern and intensityof those effects for each subsectorare not known.

3.14 Sequenc{ng. Little is known about the proper sequencing of infrastructure investmentscompared with other changesthat canaffect development, and aboutthe thresholdpoint wheninfrastructure becomes a real limitationto development.In recentyears, a numberof countries in Asiahave managedto grow fastwithout their infrastructuregrowing commensurately. Congestion - 26 - has ensued,sometimes dramatically, but the jury is stillout on whetherinadequate infrastructure was a major barrier to growth. The effects of inadequateinfrastructure, and whether countriescan withstandconsiderable shortages and continuegrowing at high rates, may dependon whetherother factors are in place-for example, a dynamic private sector, an open economy, and consistent policies. But if there are no other dynamicfactors, adequateinfrastructure can act as a central stimulus. For example,the "incubator"theory postulates that the confluenceof infrastructureand other services in central urban areas induces the establishmentof new small enterprises, which otherwisewould not be profitable.

3.15 On the other hand, somedeveloping countries have too much infrastructureat certain sites-which can tie up neededresources. The explanationsinclude speculative investments intended to stimulatedevelopment-such as new industrialcities in -tariffs set too low to ration demand, incorrect decisions on where infrastructurewill be needed, and outmigration after infrastructure is built. As indicated, a corollary is the need to know more about the complementaritiesamong sectors-for example,between transport and telecommunicationsregarding industrialdevelopment.

3.16 Anotherbroader way to look at sequencingis from the perspectiveof countriesat different stages of development. A systematicanalysis-now unavailable-of how infrastructure needs evolve during developmentwould help with judgmentsabout the adequacyof infrastructure at particular stages.

Pricing and Cost Recovery 3.17 Pricing and cost recoveryissues are well-treatedelsewhere, particularly regarding roads, the most complexcase (Heggie1991). Here onlyselected issues directly relevant to the other topicsin the paper are highlighted.A key issue is howpricing policiesin the differentsectors reflect their institutionalstructures. Other issuesare that pricing and cost recoverypolicies have not taken intoaccount the fiscaleffects and the cost of public funds,and the majornegative effects of distorted pricing on resource allocation,on operationaland managerialstandards of infrastructureservices, and on the environmentalimpact of infrastructure.

3.18 Pricing and the instutional context. Pricing policies reflect the institutional structure of the sectors and subsectors. For infrastructure that is organized in public enterprises-power, telecommunications,railways, and urbanwater and sanitation-efficientpricing is based on long-run marginalcost. But in sectorsorganized as governmentdepartments-roads, irrigation, and rural water and sanitation-the efficiencyprice is defined in terms of short-run marginalcost. This pricing divergenceis unjustified-reflectingonly the institutionaldifferences. The divergenceexists becauseenterprises rely mainly on the direct pricing of their services to recover costs and financeadditional capacity-making long-runmarginal cost usually significantly higher than short-run marginal costs. By contrast, governmentdepartments do not depend on internally generated funds to cover costs-instead using direct budgetary allocations. These differencesin treatment create distortionsin the public sector-and particularly for competition amongdifferent types of transport. - 27 -

3.19 Pricing, cost recovery, andfiscal effects. A pervasive problem with infrastructure pricingpolicies is the neglect in many casesof the budgetaryeffects of those policies. Becauseof a strict applicationof short-runmarginal cost principlesor of an inability-or unwillingness-to increaseprices, governmentsoften have been forced to cover costs by subsidization,writing off debts, lendingat preferentialrates, or allowingassets or servicesto deteriorate. However,in recent years the cost of public funds has been as much as twice that of private funds, most governments have limitedtax capacity-and extensivegovernment borrowing may no longerbe feasible. Thus, an analysis of the fiscal effects of pricing policies must be incorporated in the policies' implementation. Such a framework, being developed for roads-in the forthcomingpolicy paper-overturns traditionalthinking by suggestingthat prices be based on the most economically efficientway to recover total costs.

3.20 The negativeeffects of settingprices too highor too low are beginningto be analyzed more systematically.Charging too little and/or under-allocatingfimds has led to a viciouscycle of inefficientoperations, unreliable service, and deteriorationof capitalassets. All these developments implypoor service,which in turn has adverselyaffected consumers' willingness to pay and thus the abilityof the supplyinginstitutions to cover costs. In irrigation,for example,the proportionof costs recoveredhas been between15 and 45 percent. Institutionalissues firther exacerbatethe negative effects-as in caseswhere chargesare not collectedby the suppliersand need to be allocatedthrough a politicalprocess. Allowingprices that are too high in the current climateof deregulationcould induce the entry of private sector operatorswho can take the best clients away from the public suppliers. Theexistence of considerablecross-subsidization, as happensextensively in infrastructure, exposespublic operatorsto selectiveprivate competitionwhich will prevent the continuationof the cross-subsidies-usuallypolitically imposed-and greatlydeteriorate the suppliers'financial situation. The problem is that differentiatedprices are justified on economicefficiency grounds: "Ramsey' pricing-charging higherprices to users with low elasticityof demandand lowerprices to users with higher elasticities-minimizesthe welfare loss from chargingmore than marginalcosts.

3.21 Entry of alternative suppliers can hurt economicefficiency or cost recovery by underminingthe economiesof scale of the public-sectoroperators-thus increasingoperating costs and triggeringa viciouscycle. This situationmay be exacerbatedby overconsumptionand possible congestionwhere the prices are lower.

Effects on Poverty

3.22 The effectson povertyneed to be exploredfurther. Infrastructureproblems can affect the poor disproportionately-suchas forcingthe poor more often to obtainmore-expensive or less- safe alternatives(Whittington and others 1990). Higher-incomegroups usuallymanage to capture the best services available. The differencein the quantity and quality of infrastructureservices between the rich and poor sections of cities is the most glaring illustration. The negative environmentaleffects of inadequateinfrastructure-such as air andwater pollution-are mostevident in denselypopulated inner cities, where the poor are concentrated. Inadequateservices also affect the poor's labor productivityand access to employment-for example, by harming their health. Inadequateaccess to services-such as transport-also can reduce the time availableto the poor to - 28 - engage in income-earningor domesticactivities. Activitiesthat have a high potentialto generate employmentfor the poor-small enterprises-are particularlyhurt by infrastructuredeficiencies.

3.23 Generally,approaches to infrastructureservices need to be developedthat willmost affectpoverty reduction. And more knowledgeis neededabout the effectsof infrastructureat the householdlevel. Programsto enhancethe poor's accessto infrastructurethrough subsidieshave provedgenerally unsustainable because they are extremelydifficult to target.

Effects on the Environment

3.24 Evidenceof seriousand worsening environmental degradation in developingcountries and recognitionby donors and countriesof the complementaryrole that the environmentcan play in developmenthave raisedquestions about the linkagesbetween infrastructure and the environment. It is important,however, that scrutiny of these linkagesnot be based on assumptionsdrawn only from the experienceof developedcountries, where expansionsin infrastructureare usually seen as entailing some environmentalsacrifice. The World DevelopmentReport 1992 has focused on environmentalissues and proposesa comprehensiveaction program for developingcountries (World Bank 1992). Here, only a few issuesparticularly relevant to infrastructureare highlighted.

3.25 'The environmentalproblems related to infrastructurein developingcountries are a strange mix. On the one hand, rudimentarycooking and home heating appliances-in poorly ventilatedareas using dirty fuels-may exposemany peopleto harmfullevels of indoor air pollution. Millionsmore must copewith a high incidenceof waterborneand foodbornedisease. Bothproblems were more or less eliminatedin developedcountries a centuryago. On the otherhand, as economies develop,countries are increasinglysubject to the ills of industrialization-suchas hazardouswaste dumpsand urban photochemicalsmog. As a generalrule, infrastructuredevelopment helps correct the problemsin the first categorywhile exacerbatingthose in the second.

3.26 Thus, althoughthe negativeeffects of infrastructureexpansion should be considered, the improveddelivery of infrastructureservices can also help maintainand improveenvironmental quality. Indeed, the benefits to commerce,tourism, and work force productivitythat may be associatedwith infrastructuremake a compellingeconomic case for its improvement,even if its direct externalitiesare ignored.

3.27 It is still not too late for most developing countries to prevent the kind of environmentaldamage experiencedby developedcountries. The environmentaleffects of many infrastructurechanges depend on the detailsof their use-for example,where a highwayis built or the type of fuel used for power production. Issues to be studiedinclude, for example, the linkages betweenlong-term effects, urban infrastructureinvestments, and the shapeand densityof large urban areas.

3.28 Pricing. Pricing that is too high or too low can indirectlyharm the environment. Excessive charges for services that help protect the environment-such as sanitation-may lead consumersto find alternative,less-safe solutions. Thus, local governmentsspend 30 to 50 percent of their operatingbudgets on the collectionof solid wastes (WorldBank 1989c). Most problems, - 29 - though, result from failure to charge users the marginalcost of a service. Subsidiescan lead to excessiveconsumption, inefficient production, and insufficientfunds for proper maintenanceor expansion. Poor maintenancemay causesewer backups, for example. Of 200 wastetreatment plans in Mexico, 80 percenthave been unusablebecause of poor maintenance(Bartone 1989a). Or, in Lima, Peru, less than 60 percentof the refuse collectionfleet is availableat any one time-which causes "the almostuniversal transformation of plannedsanitary landfills into open dumps"(Bartone 1989).

3.29 Subsidiesalso may have environmentalconsequences by favoring one mode-of transport or power production, for example-over another. Low taxes on heavy vehicles may encouragetheir use at the expenseof railwayswhich may be less polluting. Cheaperprices for coal-comparedwith, say, naturalgas-may increaseair pollution. Proposalsto includeexternalities in pricing-for electricity,for example-range from economicincentives, such as emissionfees and tradable permits, to laws requiringthat a certain percentageof emissionsfrom a vehicleor power plant be treated. Emissionsfrom output are relativelylarger in developingcountries, because the qualityof fuel burnedis lower, the technologyis more primitive, and the operatingpersonnel are unlikelyto be well-trained.For instance,particulate emissions from powerplants in Chinaare very high because,in part, unwashedand unsizedcoal with an ash contentaveraging 27 percentis burned, as opposedto the ash content of 11 percent burned in the United States. Althoughper-capita electricityconsumption is still low, the high income-elasticityfor electricpower impliesthat this sectorwill grow rapidly with development.

3.30 Ibfrastucture defciencies. Many of the negativeenvironmental consequences of infrastructure in developing countries are directly correlated with operational deficiencies-particularlylimited access, unreliableequipment, and distributionlosses. Limited access and unreliable equipmentproduce some of the same problems associatedwith inadequate pricing. In addition,unreliable equipment may increasepollution. Start-upof stem-electricboilers whenmaintenance is poor causesa large emissions"pulse" of hydrocarbonsand flyash. Or, backup plants used to make up for power shortages-as in Nigeria-are likely to be more polluting. Reducingpower distributionlosses-caused by line losses,for example-can limitthe needto expand capacity.

The Need for Planning

3.31 The widespreaddisenchantment with planning and with the public sector-and the emphasison the private sector-have led to the virtual abandonmentof infrastructureplanning. Some formof planningis needed,however, because of the publicsector's continuingimportant role in providinginfrastructure. Whether governments are laissezfaire or interventionist,they still need to decide how to allocatepublic sector investmentsto infrastructure. In addition,environmental, political, and social considerationsspecific to each country require an overall perspective. Governmentsthus need informationabout the interactionsbetween infrastructure services and key macroeconomicand sectoralvariables, the costs and benefitsof non-economicobjectives pursued throughinfrastructure services, the effectsof currentinfrastructure policies on services,the expected rates of return on current public sectorinvestments, and about other economicobjectives-such as - 30 - exportsor regionaldevelopment-that may be influencedby infrastrucure or influencedemand for infrastructure.

3.32 As infrastructurebecomes privatized, and marketsbegin to function,the distinction between planning and sectoral policy managementmay be increasinglyblurred. As the focus changesto policy management,the first priority should be to identifythe main ingredientsneeded for that task-a conceptual framework, information and monitoring requirements, supporting economicand social analysis,and institutionalcapacity. However,although a better understanding of these issues can greatly improve policy managementand investmentdecisions, cost-benefit analysiswill need to remaincentral to investmentdecisionmaking, but improvedthrough the explicit inclusionof externalities(e.g., the environment)and distributionaland fiscal impacts. Combining macro- and microanalysisto improvethe qualityof decisionswill continueto be an art. Infiastnt Sector Adjus_ment Loans/Credits

OPERATION INFRASTRUCTURE LOAN/ INFORMATIONAVAILABLE ON OBJECTIVES OF THE OPERATION SECTOR/S CRED1T SUPPLY CONDITION, INVOLVED IN OPERATING AND FINANCIAL $ PERFORMANCE. MILLION 1. Water Supply and SanitationSector Water Supply & Sanitation 80 72% in urban & 80% in rual areas haveaccess; Pdivatization,efficiency pricing, and reduction of Adjustmantprog, Cote d'lvoire, pod operatingperformance in the urban areas, dependenceon public budget support. Action includes: (Repont No. P-5089-AVC,June 1990) inadequatemaintenance in the ural areas; coat radonalizationof tariffs,transfer O&M to users in rural recovery poor in rural areas; net internalcash areas, financial restructuringand regulatoryreformn. generationfinanced only 7% of capital inveatmente. 2. Public F.nterpdieSector Adjustment Power, Water, 40 Water and power utilityperformance improving To extendand deepenongoing reformn efforts, facilitate Program, Mauritania.Report No. P- Communications.(Port Sector as a result of an earliertechnical assidance privatization,financial restructuring, and reduce state's 5293-MAlU,May 1990) marginaly involved.) projectsupporting reforms. Port setor financial role. performancenot satisfactoryin 1989;post & teleconmmunications'earlier reform measures modest,and progressingvery dowly. 3. Road Transport and Road, Teleommunications. 380 Restritive regulatoryenvironrnent for road Private setor development:deregulation in the trucking TelecommunicationsSector transport,inefficient operations and poor sector,privatization rAd regulatory reform in the Adjustment BProject, Maxico.(Report mu.tmodal service; overallroad cost are telecommunicationssector. Prdcingand taxation No. P-5254-ME,May 1990) recovered,but there are cross subsidiesbetween distortionsin the two sectorsare being eliminated. automobilesand truch.Trucks pay oniy 50% of Measuresto imiprovehighway safety and to monitor road costs, evenafter increasinAgdiesel prices in complianc of pollution sandards. Dec. 1989 (increasd 15%before); rail tariffs below cost; about 20% of Mexico's federal highway networikis badly deterioratedbecause of inadequatemnaintenace over severalyears. Telecommunicationssysem is a major constraint to economicgrowth: long waiting for connectiona,unreliable, and heavily conpsted service. 4. EnergySector Adjustment Power 100 Suboptimalgeneration, underdeveloped Implementatonof comprehensivemedium and long term Prtoam, Cote d'Ivoire (ReportNo. distribution.Inefficient public enterpises: high polcy famework for energy sector. Irove P-5165-RC, December 1989) electricityprices. Operatig performance ivestment planning,increase operational efficiency and deterioratedsince 1982-maintenanceinadequate; srengtheningfinancial management of the electrc utiity, transmissionhas been affectedby frequent includingintroducton of more efficient flexiblepriing interruptions;distributon has sufferedfrom mechanisms. equipmentdamtages-weak financial managenent- overly ambitiousrural electrifcationprogram contributingto poor financialperformance. 5. Transport Sector Transport 60 Highway network has deterorated after years of To reduce cost and the price of transport is the principal Adjustumnt eatwnnt Ptoject, Chad inadequatemaintenance; monopoly-induced objective.The credit would support:promotion of (ReportNo. P4977-CD, March 1989) inefficienciesin truckingand high transpOrt competitionand a greater role for pnvate sector in road coWst. constructionindustqy, improving rsowuree mobilization and allocationin the sector, abolitionof existingmonopoly and administred tariffs, road rehabilitation,adequate 0 maintenance.

k,4 6. EnergySector Adjustent Credit, 175 Transmissionsystem development does notmatch t Bangladesh (ReportNo. P-4549-BD, Power generatingcapacity. Distribution loss high, Z March 1989). systemneeds renovation.Unreliable and poor qualityof supply,very high systemlosaes (33- o W 3Xof generation),lowmanpowerproductivity, high accountsreceivables, tariffs below costs; t internalfunds financed 10% of investmentsby the electricityutility in 19808. SupportGovt.'s reform program:pricing and cost recovery;tariff restructuringand manimizationof subsidies;balanced investment in tranamission,distribudon, and generation; improvementin institutionalperformance. 7. Public EnterpriseSector Adjustment Power and Water 40 Managementand financial difficulties,because of Improvepublic enterpriseperformance, reduce Loan, Mali (ReportNo. P-4734-MLL institutionalstructure in the power and water dependenceon public finances Actionprogram includesa Jun.e1988) utility, organizationalresPcturing of 2EPM(electricity and water company),and OPT (postsand teleconmmunications company)(e.g., creation of the new pots and telecommunicationacompanies). S. Power SectorAdjustment Loan, Power 300 Access to 58%; 95% in urban areas; supplywell Improveef6ciency of power generation,transmisson, and Colombia(Report No. P-4676-CO, in excess of demand; inadequatemaintenance of distribution;rationalize investments, bring electricity November1987) distributionfaeilities, resulting in their pricingin line with long run marginalcost, financelrge deterioration,and overloading;high system share of investmentfrom internalcash generation losses; serious financialdifficulties (because of improvethe regulatory framework nd utility high debt service requirements,heavy ara, managenent, and financialrestructuring ofthe sctor. emp.from the public sactor, and high losses, and lack of reglaktoryframework and inadequate supervision);Exceptions arePM (En*presad PflblicasdeMedelln and, CYC(Corporaciortde AutonornaRegiona l del C;uca). Tarif stracturas are distoeted:resdnil turiffsare lowvrelative to LRMC; subsidizedconumption blcks are large; idusal taiffs in Bogotaare high. 9. Enegy Seotor AdjusmentLoan, Power 325 High openrti cos (attributedto absenceof Hybrid-SupportGovt.'s reform program(inmcese Turkey (tepot No. P-4513-TU,June sytematic maintenance,and low levelsof prvate 1987) utilizaon); tariffs increased,still below LRMC. setor participadon,grater financialautonomy, and focus on high-priorityinvestments), and providepart financeto investmentsin energy sector. 10. PublicEerpie Sector Power,Teecommunication 80 Policy reforms:tariff ettingfor water, power, and AdjustmentLoan, NaNer(Report No. and Water telecommunicationsto cover operatin costs, debt service, Pi4522-NIR,June 1917) nd a reasonablesre towardnew investmentcosts; privatesector development;and legal and institutional reform (accountability,supervision, financial autonomay

II. PublicEnterpris Sector Transport,Power, Water, and 20 OperadtnglosECS in the railway ector, requiring Reductionin the siZeof public ector, eliminationof Adjustmet Loan, Jica (Report Telcommunication central govt. budgetarysuppOt; port authority intr-governmental arras; improvedoperating babnes No. P-4579-JM,May 1987) has generatedoperating supluss ince 1980; of public enterprise targeted; nd monltoringand control Power srvice quality nd acial permance of public enterprise. Actionprograms concernmg umprovua 'TheNatonal Water Commission infrasutucr include:privatization and/or linuidationof sffers frnm managameitproblema, shorag of loss-generting railway services; periodictarff funds (iavy arrears from govt. agencies,failure adjustmentsin the water sectorto cover Cost inceass, of tariffs to keep pace with inflation);Steady and eliminationof arrears; and privatizadonof the improvementin the quality, reliability,and telecommuicationssystem. acce sibilityof telecoAmunicatonsinin recent ye r; however, stdl-extensivedelays in rectificationof technicalfaults. Scienc: WodldBan Apprissl documents - 33 -

Chapter4

The Reliabilityof InfrastructureServices

4.1 The centraltheme of this paper is the need for enhancingproductivity of infrastructure through a stronger demand/clientorientation. One key ingredientis an increasedfocus on the reliabilityof services. Reliabilityis importantfor two reasons. First, sincereliability is essentially in the eyes of the user or client, it helps to inducea more demand-client-oriented perspectiveon traditionalissues of operationalperformance. Since operationsalways have been, by definition, supply-oriented-mostlyin the handsof engineersand accountants-a differentfocus linking demand issues more closely with operationsseems a more effectiveapproach for increasingproductivity. Second, reliabilityis an increasinglyimportant ingredient of infrastructurequality. In the early stagesof development-especiallyin growthstrategies oriented toward self sufficiencyin the lastfour decades-the first priority has been access to services. But access is no longer enough. The increasingproportion of internationaltrade in most developingeconomies, the globalizationof the world economy, the acceleratingpace of industrializationresulting in more high-value-added products, and the major changes in manufacturingand logistictechnologies-for example,just-in- time inventories-increasinglyrequire services that are reliable. For some activities, such as an artisanshop or even irrigation,frequent electricity cuts are expensive.For high-technologycomputer work or the handling of frozen materials, however, frequent interruptionscould bring ruin. A country cannot begin to participate in the globalizationof financial markets without good telecommunications.The essenceof just-in-timeinventories is absolutereliability in the distribution and communicationssystems.

4.2 The concept of reliability is important not only for relatively high-technology activities,however. If developingeconomies become truly integratedin the internationaleconomy they will not be able to keep a dual economy-one integratedand another isolated-for long. Even relativelylower-technology activities, such as the export of perishables,have requireda high level of infrastructurereliability. Also, there is evidencethat unreliable services affect the poor much more negatively. A key ingredientof successfulpoverty reduction strategieswill be reliable infrastructureservices.

4.3 Of course, the importanceof reliability varies significantlyamong activities and subsectors-and countries. At one extreme,for developing economies that are pursuingan export-led growth strategy-such as Chile, Korea, Malaysia, and Thailand-the reliabilityof infrastructureis alreadya serious issue that would apply to most aspectsof their economies.At the other extreme, in still-closedand undeveloped economies in Sub-SaharanAfrica, the issue of accessibilityof services is still central. The paradox, however, is that the longer countriesdelay emphasizingquality or reliability,the harder it will be to accelerategrowth-particularly to participatefully in international trade. And although reliability is more important, for example, for the transport and telecommunicationssubsectors in foreigntrade, it can be a useful focus in all sectors. - 34 -

4.4 The importanceof the conceptof reliabilityhas emergedfrom the need, due to the developmentsdescribed above, to focuson the qualityof infrastructureservices rather than only on the quantityor accessibilityof services. Not only that, but for a numberof economicactivities very specificlevels of qualityare required. It is no longerpossible to think of infrastructureservices as having one, uniform level of quality, as is usually the case in discussionsabout the role of infrastructure. The notionof "service"needs to be disaggregatedto be able to explorethose levels of qualityrequired by differenttypes of activities. After all, the centralobjective is to providethe qualityof servicesin the most effectiveway. In this context, reliabilityis just one approachthat promisesto be operationallyuseful.

Definition

4.5 For users, reliabilitymeans that a servicehas a high probabilityof being available in the quantity,quality, and at the time required. This conceptof reliabilitycan be directly linked to a specific use.7/ Users indirectly decide about production because of the level of quality-reliability-of infrastructureservices they expect. Substantialchanges in that reliabilitywill induce changes in the users' behavior. For example,a brokerage firm might expect almost 100 percentcompleted connections in telephoneservice; a reductionto 70 percentwill greatlydisrupt the operation. Or a farmer mightcount on weeklytruck transport;if the truck does not appear in a key period, the farm's output might be lost. In most cases-and especially in developing countries-unreliable service is the result not only of technical aspectsbut also of managerialand financialfailures.

4.6 Reliabilityis measuredin differentways for different infrastructureservices. For electricity,it is measuredby the number, length, and time of power outages, and by changes in voltageand frequencylevels; for freighttransport, by the meetingof deliveryschedules and by loss or damage to goods being transported;for passenger transport, by delays in travel time; for telecommunications,by delays in dial tone, uncompletedtelephone connections, and interferences with the sound; and for water supply, by the timelydelivery of promisedamounts of water and its qualityfor drinking,irrigation, or other uses.

Demandfor Reliability

4.7 The demandfor service reliabilityvaries considerablyamong economicactivities, dependingon the level of technical sophistication. Thus, the importanceof reliabilityvaries at different stages of economicdevelopment. At one extreme,there is little demand, or expectation, for reliability in subsistenceagriculture. The demand begins to appear in the early stages of commercialagriculture or industrialproduction. In general,manufacturing requires relatively more reliable power, transport,or communications.A shift from manufacturingto a service economy requires even better communicationsfacilities. At the other extreme are the activities-such as exports of perishables or modem technologiesusing sophisticated computer work-that are

71 The definitionmight also be extendedto includeprice reliability,but thatdimension has been left out of this analysis. - 35 -

impossiblewithout reliable power, water, transport,and communications.Although the demandfor reliability will vary significantlyamong countries, in all countries, foreign trade-particularly exports-will typicallydemand higher reliability.

4.8 Suppliersof infrastructureface a varietyof demandsfor reliability. One distinction is in demandsfor infrastructureas an intermediateinput, for whichthe effectsof unreliabilitycould be serious-for examplein terms of lost production,idle capital or labor, or spoilage. Demand could comealso for infrastructureas a final or consumergood, in which case unreliabilityis more in the form of an inconvenienceor lost time, althoughmany other negativeeffects could ensue, on health or time availablefor directlyproductive activities. Thus, suppliersmust adapttheir services to those varyingdemands for reliability. For example,in transport,perishable goods are sensitive to time delays as well as damage during processing and handling. Precision goods demand smoothnessof processingto avoid damageand mightnot survive transportover roads or railway tracks that are in poor condition. In water supply, the demandsfor water quality differ for households,industrial users, fire departments,or road cleaners. Similarexamples can be given for the other types of infrastructure. Thus, the natureof the demandmakes it difficultto evaluatethe reliabilityof a whole infrastructurenetwork, since such networksare needed for competinguses, each of which has a different reliabilityrequirement. This is why a clear user's perspectiveis operationallyso important.

4.9 The effects and perceptionsof reliabilitymay vary accordingto time and location. The reliabilityof infrastructurecomponents may vary from link to link-as in a systemsupporting an internationaltrade chain (Fig. 4.1). For example,in India the current objectiveis to improve competitiveprices and deliveryschedules-and thus exportperformance. Thisobjective differs from past goals to consolidatethe economyand achieveself-sufficiency through import substitution. The past objectivesdepended on infrastructuredifferently than they do now. Systemswhose reliability may have been adequatefor past objectivesdo not live up to current demands.

Effects of Unreliability

4.10 Infrastructurereliability is generallypoor in developingcountries, although it differs widely among countriesand sectors. Manyof the nine countriesanalyzed in depth for this study suffer from frequentpower outages and voltage drops. In northernIndia, electricityshQrtages in 1989 led to powerrationing for industry,agriculture, and domesticuse; the grid collapsednine times in six months, interruptingabout one-third of the systemand requiringthree to five hours each time for restoration. Voltagedrops of 25 percentare common,often causingrural water pumpsto burn out. Failure rates of local telephonecalls ranged from 8 percent in Bombayto 43 percent in Calcutta,compared with 2 to 3 percent for a well-operatedexchange in a developedcountry.

4.11 Inadequateroad maintenance also canmake roads impassable, as in Ghanain the early 1980s, when access to interior areas was curtailed. In Nigeria the rural road network had deterioratedso badly that produceoften had to be movedby head-loading,which costs about forty times more than transporton well-maintainedroads. Poor rural roads also made water from tank trucksmore costlyand unreliable. Water is periodicallyrationed and-because of equipmentfailures -water treatmentworks are not reliable, threateningwater quality. In India, 20 to 25 percentof Figure4.1 InternationalTrade Cycle

Infastructure Trade

accessibility accessibility

Ports d o Inbnd Transporton 4 Warehoues Dlttbutlon Airport Networks Points

deliveryspeed

orderprocessing orderprocessing inventory order routing routing management processing

Telecommunications Source: Humplick1991 - 37 - rural water supplyschemes are out of operationat any time. One cause of unreliablerural water is the fluctuatingelectricity supply. An unreliable water supply for many irrigation systems contributesto farmers' reluctanceto maintainirrigation structures and pay for water.

4.12 Infrastructureunreliabilt In Nigeria. Lee and Anashave done the mostthorough studyof the effectsof infrastructureunreliability, focusing on Nigeria(1990). The investigationwas intendedto detail the effects on manufacturingfirms of well-documentedfailures in infrastructure services-electric power, water, urban transport, and telecommunications. Some of those deficiencieswere significantin Nigeria. For example, in 1988, more than two-thirds of the manufacturingestablishments suffered more than five power outages per week and 17 percent experiencedmore than ten-with almost20 percentof them havingto shut down operationsduring those power failures. 4.13 This researchshows that large and smallfirms are affecteddifferently. Large firms spend significantresources investing in their own facilitiesand servicesbut small firms are more seriously affected by infrastructuredeficiencies. Because of the lumpinessof the investments requiredand the high operatingcosts, smallerenterprises can rarely affordto get their own standby generators,boreholes, radio transmitters,or messengermotorcycles. Since manynew jobs in urban areas comefrom new smallenterprises-generally 60 to 80 percentof newjobs in major citie-the negativeeffect of infrastructuredeficiencies is substantial. And the higher costs to Nigerianfirms affect their abilityto competeinternationally.

4.14 Althoughall 179 firms studiedare connectedto the powergrid, every firm withmore than 20 employeeshas its own standbygenerator-and has invested,on average,$130,000 on power facilities. But, only one-thirdof the smallfirms-fewer than 20 employees-can generatetheir own power. Similarvariation by size of firm is found in the private provisionof water supply. No sampiledsmall enterprisehas its own water supply; about 14 percent of the firms with 20 to 50 employees have their own supply; and two-thirds to three-fourths of medium-to-large enterprises-with 100 or more employees-have private wells. No small firm has a private communicationssystem, such as a messengermotorcycle or radio transmitter,but becausepublic telecommunicationsservices are inadequate,70 to 90 percentof the firms with 500 employeesor more have their own radio transmitters.

4.15 The capitalvalue of these privateinfrastructure facilities averages 25 percentof the total valueof machineryand equipmentfor firms having fewer than fifty employeesand about 10 percentfor firmshaving fifty or more. The percentageof total capitalcosts varies accordingto the type of service. The averagecapital cost for electricitygeneration is almostfour times larger than the shareof capitalfor wells and treatmentfacilities. The averagecost of generatingelectric power privatelywas about nine timeshigher than the averagecost in advanceddeveloping and developed countries. This illustratesthe firms' willingnessto pay for reliableservices. The samplefirms, as a whole,used only25 percentof their privatelyinstalled generating capacity. The presenceof these largely idle reserves of generatingcapacity contributes to a high averagefixed cost.

4.16 Costs. The costs of unreliabilitycan accrue not only to users but also to suppliers, who can lose their originalinvestment if they don't act to improvereliability; operators facing future - 38 - difficultiesin managingdemand, such as dealingwith congestedservices; and non-usersas a result of negativeeffects such as pollution,noise, environmentaldegradation, and lower land values. The typesof costs resultingfrom unreliability vary, dependingon usage. For example,in power supply each of the users has a differentcost, and there are well-acceptedapproaches for calculatingthem (Munasinghe1979; Lee and Anas 1990). For other forms infrastructure,such as transport, water supply, and waste disposal, there is not one well-acceptedapproach. An exceptionis railways, where there are well-documented-althoughpartial-approaches (Martland 1974).

4.17 Also, there are direct and indirect costs of the results of unreliableservice, such as power outages. The direct costs occur during an outageand the indirect costs arise if users adapt their activitiesto become less susceptibleto outages-such as paying for other energy sources. Industrialoutage costs includedamaged materials and lost production,minus any recoveryduring regularworking hours or on overtime. Changesin voltageor frequencycan damageelectric motors. Commercialusers incur losses in sales and services from outages if computersand other office equipmentcannot be used, when there are no lights or elevators,when cash registersor gasoline pumps stop working, or when refrigerationstops in stores, restaurants,and hotels. The costs of power outagesdepend not only on their lengthand frequencybut also on such factors as the time of day or the day of the week and whetherthe outagewas forewarned.

4.18 Effect on cost recovery. The potentialfor cost recoveryseems to be directly linked to service reliability. Studies have shown that unreliabilitycontributes substantially to users' unwillingnessto pay, even when they can afford to do so. Farmers are unwillingto pay water charges not only becauseof a long traditionthat water shouldbe free, but also becausethe supply of irrigationwater is frequentlyerratic and unavailablewhen it is most needed. One study found that householdsin a large Nigeriancity-Onitsha-were paying private water vendors more than twicethe operationand maintenancecosts of the pipedpublic distribution system because the quality and reliabilityof the private service was better (Whittingtonand others 1989). Another study in Nigeriafound that "householdsdo not trustthe governmentto providea reliablepublic water supply" and cited tiiat as a key reason for the low willingnessto pay water charges(Whittington and others 1990,para. 106). Anecdotalevidence suggests that unreliabilityis also importantto cost recovery in electricpower, telecommunications,and water supply,thus contributingto the pervasivefinancial weaknessof public agencies.

4.19 Distbutional effects. Availableevidence indicates that the distributionaleffects of unreliabilityare extremelyregressive. First, the Nigerianstudy showsthat small firms have much less possibility of generating their own supplies-or protecting themselves from unreliable infrastructure.Even when such firms do attemptto providetheir own infrastructure,economies of scale make the costs higher for the smallerfirms. Second, the cost paid by the poor for reliable suppliesis often severaltimes the publicprice. In Nigeria, in some instances,the poor pay nine or ten times the price of public water service or-worse-they have to devote much their time to obtaining water. In Istanbul, recurrent costs assumed by householdsto cope with unreliable infrastructurerange from 1 to 5 percent of annualincome, with the pooresthouseholds facing the higher proportion. And, in general,women bear muchof the burdenof these extra costs (Box4.1). Third, as indicatedthe rich often manage-as SamuelPaul (1991)puts it-to "capture"the few good servicesthat are available. Developingcountries are full of examplesof this phenomenon-fromthe - 39 - perfectlymaintained road and telephoneline to a politician'sfarm, to better servicesto the richest sectionsof a city, to preferencegiven to particularshipments by road or rail. The poor have little voice and are not in a positionto pressure suppliersto improvetheir reliability.

,settlemeus she s:f wssrl ntvibl, wmnal ogdstSgancsbrnig ,hm. Whentrp se. e y~~~~~~~~~~~~~~~~~~>'M~~~ ~~~~~~~~~~~~~~~~. . , i.'..S.i:" >" ......

thanhoursmen-up morO to fitneit a weak in ruralIndia and twelve hoursmore 'anrura N.pa...... 1990)... he.

tolaehmetohu.beoeteslr*.th....h SS8SSSi 8T: SS B, 8 '~~~~~~~~~~~~~~.hI wr.wkngfrbssan ; '''...... ufrn hevytraff"ic...... Manyh 2B'''S'h spensixy toeighy mintae hors.weekto thir jbu~someget iscorage andd outso ofawrk Kdat direct to.fo0 t Inethenhoral einunieaioe~. unrelabIlityinfepeffe Te to thef unsprelid upplyabryleo infrasructureservice-inoludngifatreuctouretion the averagsmoe tranport.wegh eergynadlarriedb t the wa*-icreacath.inoo~atlbtit3a b*twn th housholdworkand4 s . se icomegenertingactivties .--- 8B .. =~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. Poorwome...... - hav househ level,thd hoheeeolnreliabilitfewmecanimshic thy hroghcn canrduc pomnseresignritcand suh rpatbilty unerataucte hos woen labords. n Angh extenehom huseod exampleheu l i watersupplyin Iostanol, wher no householdwkStnudrinkho tap wateprandy horsehold prethlogshav orswic

lau2 y.Womn whoserl theusbaworkeds unreli apse inr oastedctrel on thehubpaoor ispmoredlrning wateth.

aouseollevel MnpmsadwB uneibiiycn Eswhere pose igndaificant haltsi hazardssua.o obseremed.I As ruramleP is

water supply in Istanbul, where no household can drink tap water and households thus have to switch - 40 - to alternativesources, reduce consumption,install redundantcapacity, or have their own treatment facilities.

The Growing Importance of Reliability

4.21 The globalizationof the world economyis a momentousdevelopment that will have important positive effects on developingcountries (Peters 1991 and 1992). Just exposing an economyto foreign competitionimposes stricter demandson the quality and reliabilityof most activities,including infrastructure services, than demandsprevailing in a closed economy. If a countryalso wantsto join the more advancedproduction processes, such as just-in-timeinventories, the demandsare even stronger. But stricterdemands apply to all activitiesin internationaltrade, not only the advancedones, such as the productionsof automobileor electronicsparts. For example, the strict inventorycontrols kept by retailersin developedcountries impose stricter delivery schedules for importsof artisan work. Initially,the pressureexercised by internationalcompetition will have a salutaryeffect only on tradablesbut, over tirne, suchpositive effects will be spreadthroughout the economy,especially if the exportand importbase is broad. The analysisbelow also illustrateshow institutionalweaknesses affect the provisionof infrastructureservices and how administrativebarriers are primarilyresponsible for delaysin distributionsystems.

4.22 The worldeconomy is facinga newkind of industrialcompetition (Alavi 1990). New internationalrivals have emergedwith advantagesin costs, productivity,or governmentsupport. New productsor processeshave createdunusual possibilities in marketingand production. Rapidly changingtechnology in productionand communicationshas transformedthe competitivedynamics of the world economy. The adaptabilityof supplyto day-to-dayvariations in demandis a central characteristicof the new world competitiveenvironment. Althoughin this new environmentprice competitionremains important, quality, delivery time, marketing,services and the ability to adapt rapidly to changes in demand are becoming crucial to capturing and maintaining markets. The globalizationof worldmarkets and newproduction processes-accelerating in the 1990s-will require major improvementsin infrastructurequality if developingcountries are to participatesuccessfully. These improvementsare important because more than 80 percent of exports from developing countriesgo to developedones. Alavi's studyregards physical infrastructure as "the backboneof competitiveness"and findsthat "a sophisticatedphysical infrastructure-roads, rail, air, and marine transport, telecommunications,power, and so on-is a preconditionfor productivitygrowth and successfulinternational competition" (Alavi, para 4.32).

4.23 The increasingglobalization of marketspermits producers to seekcost advantagesfor suppliesand other inputsworldwide. At the sametime, consumerpreferences in many marketsare changingrapidly as informationabout new productsspreads quicklyand increasingincomes make consumersmore selectivebut also more capricious. These changes require quick responses in supplies, production, and marketing. Flexibilityand prompt global delivery frequentlybecome as-sometimes more-important as cost and quality. An exampleis a textile companyin Taiwan which is a key supplierof finishedgarments to a buyer in (Box 4.2). Fashiondesigns are handled in Taiwan,but tailoringtakes place at an offshoremanufacturing base in Thailand. The factory in Thailandimports textiles from a weaver in India, cotton from the United States, and syntheticyarn from Indonesia. The finalproduct is shippedfrom Thailandto Germany. In all this, - 41 - the supplierhas to be very responsiveto the buyer's strategyof reactingquickly to frequentchanges in fashion.

}'-~~~~~~~o 42=

~~re_~~~~~~Goa ~ietos

'~~ :..-..--...:~~~~~Dvlpn..''. Comtre a.. ,,.,ttil uueTtie t.o apii Frd:Craetclyitgae etl n ama tnnfcue and rter h opnsaky upiro iihdgaret to MRode aioa Theslor Om.y) Wiefs deinin Re ade nteTie edue r, alrn pa-take:S *. a::.. enl :::horesta.:..he.::fcwin Ras Rnhnui .us RfRag~k(hd not h ..'.'' fatr _,pot _-ult deinrtXtie rmawae nRjsa Ida n otnfo 'e-" th UsiieeSats). Slhtcymi upidfo ceia dar tpe n (ldos4Tefnlpoucsaearsipdfo Bnkkt usedrtTnalti,OetFtr hasto be ver re.nsv to ModemFashion.. :R.:.:.-::: ons strgy .,astk ~ .,:,. Infntei Stores.Tn.o i.nai(JiedSae)prhsslre ut fpatty

4.24 lad uy FoThe slasticn plb tofsmarkets hasngmenaccompranies foy tnspw throdughtionproes. ta rel donmin aciztninvenhomies basuing thioejutrin-atiom e donventorysytm Just-in-time bsistesfoc - 42 - closerrelationships between buyers and sellersbecause buyers expectperfect quality and absolutely reliabledelivery. An exampleis a steel processingplant in Californiaowned jointly by a Brazilian, a Japanese, and an American company. The plant is suppliedwith steel plates produced in Brazil-and plates madein Mexicousing mineralore fromBrazil. Becausethe plant providessteel products to several car and equipmentmanufacturers, most of whom have adoptedjust-in-time production arrangements,the supply lines for ore and steel plates have to be extremely well- coordinated.

4.25 The effectivemanagement of thisprocess requires up-to-date information on supplies, production,and sales-and an extensiveand reliablecommunications system. Information,in effect, becomesa substitutefor inventoryand other resources-and improvesservice while keepingcosts under control. Althoughthe costs of majorproduction inputs have increasedin the last fifteenyears, the costs of informationhave decreased. The quantity, quality, and complexityof information exchangedbetween buyers, sellers,and third partieshave increasedconsiderably in importancewith wider acceptanceof just-in-timeand related processes. The timeliness,accuracy, and availability of informationare the singlemost importantservice dimension. A study of 1,450 businessesin Canada,Japan, the UnitedStates, and six WesternEuropean countries found that their managerswill attach increasing weight to the availabilityof reliable telecommunicationsin the 1990s before deciding to enter new markets or selecting new business partners (Peters 1992). But good communicationsmust be backed by reliable delivery of supplies, which involves not only internationaltransport but also efficientproduction processes. These, in turn, require reliablepower and domestictransport.

4.26 The new productionprocesses require smaller consignmentsand more frequent shipments. The neededmultimodal, worldwide perspective has directlyaffected transport methods and arrangements. In manyregions, ocean transport is now providedwith increasingfrequency and speed-especially for high-valuecontainerized goods-and with a degreeof reliabilitythat enables shippersto adhereto their just-in-timearrangements. Two-thirds of the servicesinto and out of key Pacific ports are providedon a fixed day of the week. Similarly,the long-termshift of domestic freight traffic from railwaysto truckinghas been reinforcedbecause of trucking'sgreater flexibility in respondingquickly to changesin deliveryrequirements.

4.27 A studyof manufacturingprocesses in WesternEurope indicates that of the timetaken from the supply of raw materials to final retailing, 2 percent is used in production, 5 percent in transport, and 93 percent in storage at variousstages of processing(UN 1987). In some developed countries,the use of logisticsto improvethis processhas resultedin reductionsin order cycletime of up to 80 percent, inventoryreductions of 30 to 70 percent, and reductionsin costs of the final product of up to 30 percent. An importantcontribution to this has been the liberalizationof govermnentregulations preventing competition in transportand telecommunications.

4.28 Studiesof specificexport commoditiesconfirm the new developments. A study of bicycle and footwearimports into the UnitedStates found that the three importantproduct criteria are price, quality,and on-timedelivery: "Reliable,complete deliveries are essentialfor retailersto take advantageof seasonal sales peaks and fashion cycles" (Egan and Mody 1990, para 2.10). Becausethese peaks and cycles are short, even deliverydelays of only a few days can mean lost - 43 - salesand substantialmarkdowns. Similarly,a studyof the electronicsindustry found that competing effectivelyin world marketsrequires not only low productioncosts, but also high qualityand rapid deliverytimes (Dahlman1990). Theserequire a strongerinfrastructure than mostnew entrantshave.

4.29 The new productionprocesses present developing countries with a difficultchallenge. Althoughmany have succeededin diversifyingexports and increasingthe volumeof manufactured exports-though from a low base-they may now be facinga slowdown,in part becauseof serious logisticalconstraints. For example,the costs of containerizedexports from Indiahave risento one- third more than the averagefor competingAsian countries-and the time requiredto get products to marketsis also longer (Box 4.3). The causes includepower shortages,capacity constraints in ports, roads, railways,and warehousing,and poor telecommunications.Other key impedimentshave been outmoded trade legislation and customs procedures, uncoordinated management of infrastructure,and limitedavailability of informationservices. As a result, exportsand importsface long delaysand high costs. Container-handlingcosts in India are 80 percenthigher than in Japan and the United States.

4.30 The situation is similar in many Latin Americanports. Accordingto the UN EconomicCommission for Latin Americaand the Caribbean,the cost for handlingone ton of steel in Santos, Brazil, was $22 in 1988, comparedwith $3.18 in Kaohiung,Taiwan, and $3.80 in Rotterdam. The major reasons for the high costs include insufficientinvestment, government regulations,and union work rules that require overstaffmg,prevent work at night, and penalize privateterminal operators who handle only their own cargo. Brazilwas expectedto lose $5 billion in potentialexports in 1991 becauseof high port charges (The New Yorklimes 1991).8/

4.31 India's effortsto export computersoftware have failedto meet expectations,in part becausethe stateof Karnatakahas been ableto supplyonly half the contracteddemand of electricity, forcing operatorsto buy their own continuouspower supplies. One of the largest international companieshas brought in its own satellite dish to assure adequate communications. As The Economistreports, this has discouragedexports of computersoftware and encouraged"exporting highlyskilled Indians" (May 4, 1991,p. 70). The few successcases-like horticulturalexports from Kenyan harvest to European marketsin less than 18 hours and large exports of fresh fruit from Chile-were made possiblebecause of the eliminationof bureaucraticred tape that had prevented privatecompanies from establishing effective logistics. Similarly,Idonesia, Malaysia,and Thailand have achievedsubstantial growth in trade, essentiallyby establishingliberal regulatory environments.

4.32 Thus, infrastructure services have become crucial for succeeding in industrial dispersal and export diversification. But even if these services are physicallyavailable, their effective use is often hindered by governmentpolicies involving, cumbersomeexport license proceduresthat delay shipments; transport regulations and fragmentedand obsolete document requirementsthat prevent efficientmultimodal transport; insuranceand liabilityrequirements that don't coverthe costs of tradedcommodities and thus inadvertentlyincrease risks; bankingprocedures that prevent speedytransactions, such as the issuanceof letters of credit;and classificationcodes for

Ji/ A billion is 1,000 million. 44 - local goods and services that don't match international codes. According to Peters, 23 documents are needed to clear imnportsin India, 118 for exports. Singapore processes 25 percent of its trade and the related transport tranactions without documentation. - 45 -

How to Build Reliabilityinto Operations

4.33 The conceptof reliabilityhas two direct operationaldimensions. One is the need to determinethe optimum-or most adequate-level of reliabilityfor a groupof users. Reliabilityhas to be consideredas an economicconcept: suppliersand users are only interestedin optimizing reliabilitybecause of cost. Thus, a cost-benefitanalysis of reliability-linking supplyand demand in individualmarkets-is needed. The seconddimension is to assessin a particularcase whetheran action program focused only on enhancingreliability would differ from a conventionalprogram aimedat enhancingthe productivityof a whole infrastructureservice. This is the crux of the matter: a focus on reliabilityshould also mean lookingat productivityfrom a different-and, it is hoped, more effective-perspective.

4.34 Cost-BenefitAnalysis. A full analyticframework for a cost-benefitanalysis of reliabilityhas been developedas backgroundfor this paper (Humplick1991) (Fig. 4.2). The main points are highlightedhere.

4.35 To conducta cost-benefitanalysis of reliability,a whole activitycycle-subdivided into smaller cycles-should be the unit of analysis. Whole cycles can be, for example, a multisectoralone focusedon a productfor export,or any farm-to-marketor factory-to-marketcycle. The perspectiveof the whole cycle is in the hands of the user-or a planningagency-but each step mightbe in the handsof a differentsupplier. A railwaysystem, for example,is just one subsystem of an internationaltrade chain (Fig. 4.3).

4.36 The next componentof a cost-benefitanalysis is an analysisof the suppliers'options for enhancingreliability. Beyond the obviouspoint that one way of increasingreliability is to enhanceoperational efficiency, suppliers can provide redundancy at the designand investmentstages. This is a complex technical concept, but a key to assuring reliability of several infrastructure components.2/ The conceptualframework then deals with the user's options for reacting to unreliability. Alternativesinclude continuing to demandthe unreliableservice, for lack of options; providingalternate facilities, when feasible; exerting pressure, individuallyor collectively(voice); or reducingor stoppingdemand for the serviceby reducingor discontinuingits own productionor changinglocation. The frameworkthen includesthe estimationof the costs-private and social-of unreliabilityand the cost-benefitcalculations.

4.37 Focusingon Reliabliy to IntproveProducivity. An emphasison reliability-instead of on general agencyperformance provides a sharper focuson productivity,and, in many cases, points up a need for a multisectoral,multiagency approach. The essence of a sharper focus on reliabilityis the concentrationon the productivityof one or a few activitiesperformed by a supplier, presumablythose on which the potentialpay-off of higher-qualityservice is greater-such as containertraffic in a port; one sectionof a railway,one city in a telecommunicationsnetwork, or one generatingplant in a power system. The alternativeis the way most analysisof performance is done, taking a whole agency-a port, a telecommunicationscompany, or a railway-as the unit

2/ For example,redundant capacity could be activewhen all capacityis used ali thetime; passive, when redundantcapacity is dormantand activatedonly when needed; or standby, an intermdiatesolution. Figure4.2 HolisticRepresentation of InfrastructureReliability

Usage Activity ..'.'fwrn.. Production Process v

Q.: t:- t.:i, .. ?xe ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...¢ Z...... :

Dynamic Response dleliabil,t _ Usage of Userto Infrastructure P&o,tY _ ~~~Reliability

SupplyResponse Informationto Suppliers PerceivedFailure

Infrastructure SystemSupply -a' Industriesproducing goods for trade depend on intrastructure Systems Y as an input. The reliability,capacity, and cost of infrastructure Process servicesat any point dependson the managementand main- tenancepractices of infrastructureservice suppliers and the dynamicresponse of all usersto infrastructureunreliability. Usersdemand different levels of reliabilitydepending on their PerceivedFailure particularusage actvity, and adopt reliability-enhancingstrate- gieswhen the levelsprovided do not meettheir needs. The extentto which usersadopt such strategies can be usedas informationto infrastructureproviders when selectingoptimal investmentdecisions to improveeither the reliabilityor capac- Source: Humplick1991 ity of infrastructuresystems. - 47 -

Figure4.3 RailroadUnes Serving Industry

Pick-upand DeliveryPoints

INDUSTRY1 INDUSTRY2

IntermediateYards

RYARD AJ

YARDF

Tracklin nnecting ick-upand delivery points

TO RAILROADX TO RAILROADY

Source: Humplick1991 - 48 - of analysis. The hypothesis favoring the first method is that by concentrating on a more limited activity requiring high standards, the nature of the operational improvements required will become clearer and the potential for improvement greater. The expectation is that the higher operational standards achieved through this method will spill over to the rest of the entity and the pace and nature of progress will be faster and more pointed-that users really needing better services will get them faster.

4.38 This approach has risks. One is that resources could be taken from other services, perhaps reducing their quality and productivity, or the spillover effect might never materialize, especially if the supplier faces many exogenous negative factors. These possibilities must be weighed in each case but the approach looks promising. The alternative, of focusing on an entire agency, has been tried often and has not worked.

4.39 The other way to use reliability to enhance performance is to look at whole multisectoral and multimodal cycles, such as for key exports or for one industry. Interagency coordination would be needed to improve operations over a whole cycle. This full-cycle perspective is particularly useful for planning by a public or private entity-to understand the full implications of an activity and to help identify any bottlenecks in the cycle. A similar analysis can be used for sectoral policy management-for industrial, infrastructure, or foreign-trade policies. Such analysis makes individual suppliers aware of being part of the cycle, allowing for pressure to be put by users or other suppliers in the cycle. The total costs of unreliability become more obvious and quantifiable. More generally, a focus on service reliability gives a clear and quantifiable objective to programs aimed at increasingthe productivity of infrastructure suppliers and the quality of service. It helps to make those programs tangible and useful for the clients that need the better services. - 49 -

Chapter5

ReorientingInfrastructure Toward Demand: Using Marketsand the PrivateSector

5.1 As indicated,one of the main causesof the low productivityof infrastructureservices in developing countries appears to be their supply orientation-focusing on providing services-largelydue to the services'monopolistic position and the tendencyof managementto focus on the technicalaspects of the operation. Thus, this report is mainlyabout the need to increasethe productivityof infrastructureservices through a strongerdemand orientation. A key ingredientfor achieving such an orientation is institutionaldevelopment. This paper concentrates on two institutionalapproaches for enhancingthat demandorientation. First, it assumesthat competitive marketsand broaderparticipation by the privatesector are the mosteffective avenues for achieving a demandorientation. Such an approach is discussedin this and the next chapter. Second, if competitionand private sector participationare not feasible, their effects need to be simulated through competition surrogates. Two surrogates particularly relevant to infrastructure are administrativedecentralization and increased participationby clients and beneficiaries. These surrogates are reviewed in Chapter 7. The analysis assumes that countrywide policies and institutions-such as trade liberalization-are designedto inducecompetition.

5.2 Of course, other institutionalissues also are importantfor enhancinginfrastructure productivity. Traditionalinstitutional strengthening and sectoralpolicy reform will continueto be centralto the Bank's strategy. After all, infrastructureprojects have had greatersuccess in achieving institutionaldevelopment objectives than have most projectsin other sectors.

The Role of CompetitiveMarkets

5.3 Much has been said about the role of competitivemarkets in achievingan efficient allocationof resources. But only recentlyhas the focusbeen extendedto seriouslyanalyze markets' effects on the use of resources. A fundamentalcharacteristic of competitivemarkets is that they provide incentivesand disincentivesfor effectiveinstitutional performance. Competitioncauses pressure-reducing discretionarybehavior in an organization,and imposinga disciplinethat leads to improvedperformance. For these reasons,the simplest,most effectiveway to achievea demand orientationis to expandthe realm of competitivemarkets. The largerthe proportionof infrastructure services that can be operatedin reasonablycompetitive environments, the better. The analysisfor this paper concludedthat the scopefor competitivemarkets in infrastructurecan be expandedin most developingcountries, but that the potentialfor these marketsvaries considerablyamong sectors, some of which should remainpublic entities. In addition, the government'sregulatory capacity is a key componentin the role that competitivemarkets can play.

5.4 Expansionof compeive markets. Thereare three main waysto expandcompetitive marketsin infrastructure.One is to adapttechnological innovation to reduceeconomies of scale and allow the participationof a largernumber of suppliers,or permitpricing of serviceswhere that was not possible before. The second, often closely related to the first, is institutional-including - 50 - legal-innovationsthat restructureor divide marketsin a way that allows participationby a larger numberof suppliers. The third optionis to dismantlethe disablingenvironment, eliminate the web of regulations,restrictions, and governmentinterventions that prevent the functioningof markets. Regulationlikely will still be neededto establisha level playing field. 5.5 The conventionalwisdom about the monopolisticnature of many infrastructure activities stems from technologiesavailable decades ago. Those technologieshave evolved considerablyin recent years. For example,new technologyin thermal power generationmakes it economicalto generatepower on a muchsmaller scale than before. In addition,institutional changes canseparate power generation from transmission and distribution-withthe distributingentity having more of a monopolisticposition. Technologicalinnovations in telecommunicationsalso allow more competitionby making possible the provision of different types of services by a number of companies.The tremendousprogress in the last two decadesin developingdecisionmaking tools and databasesystems has helpedinfrastructure services that are complexand management-intensive,such as railways, to enhance their capacity to operate competitively. Other developments-such as metering for water and other utilities or mechanismsfor road pricing-have also allowed more market-likeoperations.

5.6 However,technological change is still treated as an exogenousdevelopment. What has not been done is to turn this perspectivearound: becausecompetitive markets are now seen as one of the most effective ways of increasingthe productivityof infrastructure, technological innovationshould be directed towardthat goal. This issue needs to be exploredfurther.

5.7 The numerousinstitutional innovations that allowthe expansionof competitionhave occurred mostly in developedcountries. Most importanthave been the attempts at breaking up previouslylarge infrastructureentities to allow competitionamong the parts or with other entities. Typical amongthese are the subdivisionof nationalpower, telecommunications,or water-supply companiesinto regional or functionalunits; the breakupof railwaysinto companies handling different types of services-freight, passenger, and suburban-or companiesfor maintainingthe tracks and others for running services(Moyer and Thompson1992); and the breakup of ports into companies running individualberths or services and competingwith each other. Among these changes,it is necessaryto distinguishthose that genuinelyresult in more suppliersserving the same marketfrom measuresthat are forms of decentralization-suchas regionalizinga nationalwater company.

5.8 Competition can also be enhanced through contestability-competitionfor the market-a particularlyimportant avenue for infrastructure(see Baumol, Panzar and Willig 1982). The scopeand techniquesfor makingmarkets contestable have been developedconsiderably in recent years. The potentialis high for a numberof activitiesthat wouldcontinue to operateas monopolies or quasi monopoliesin the public sector: water supply, ports, airports, toll roads, and power transmissionand distribution.How to improvethe effectivenessof contestabilityis an importanttask in this respect. This subjectis furtherdiscussed in the next chapter,around the issue of concessions.

5.9 One form of extensionof competitivemarkets that is becomingprevalent could be called 'competitionwithin the supply,"in whichpublic entities providing infrastructure sub-contract through competitivebidding to other public or privateentities to undertakea number of activities: maintenance,repairs, fee collection,or the running of specific services. Insofar as there is no - 51 - collusionin the subcontracting,this has provedto be an effectiveway of introducingcompetition and enhancingoperational productivity.

5.10 Sectoraldifferences. Someinfrastructure subsectors lend themselves more easilyto competitive markets than do others. The potential depends on how much each subsector approximatesa private or a public good. Criteria for distinguishingbetween public and private goods are excludabilityand subtractability,or the degreeofjointness of conswnption. The more a good exhibitsthese traits, the greater the chancethat it can be operated competitively. Another importantcharacteristic is that somenonpublic goods may also be naturalmonopolies (Triche 1990; Savas 1982).

5.11 Excludabilityis the degree to which a potentialuser of a good or service can be excludedif the user does not meet conditionsset by the supplier. Althoughexcludability is easyfor many goods-for example,goods with a unit price in a store-excludabilityvaries, dependingon the cost-effectivenessof allocatingthe goods. The impracticalityof excludingindividuals raises issues of free-riders-who can consumea good or servicewithout paying. For example,one problemin allocatingroad maintenancecosts is identifyingwho shouldpay, becauseexcluding free-riders is impractical. Similarly, difficultiesin supervisingelectricity consumption in squatter settlements meansthat cost allocationamong users requiresincreased monitoring. How to assurethat those who consume share in the cost of productionis not simply a choice of market productionversus governmentproduction. And, even when govermnentdoes not producethe good, it still can help organizeproduction.

5.12 Subtractabilityrefers to how muchone user's consumptionof a servicesubtracts from the abilityof others to consumewithout raising production costs. Differenttypes of infrastructure serviceshave differentdegrees of subtractability.For example,highway use by one vehicleusually does not affect use by others. Yet, at certainthresholds of use a highway may be congestedor deteriorated.Upstream pollution in water systemsor irrigationsubtracts from the abilityto consume downstream.Often, therehas been a tendencyto createexcess capacityin powergeneration, so that consumptionby one does not affectthe availabilityfor others.

5.13 Thesetwo characteristicscan offer a fourfoldclassification of goods (and services): public goods, with low subtractabilityand low excludability;private goods, high subtractabilityand high excludability;toU goods, low subtractabilityand high excludability;common property goods, high subtractabilityand low excludability(Table 5.1). This tentativeclassification offers a first cut at the determinationof appropriateinstitutional arrangements. The two extremessuggest in principle government(public goods) or competitive-market(private goods) ways to provide infrastructure services, although the institutionalalternatives for toll and common-propertygoods are less straightforward. The practical issue is how far toll and commonproperty goods can, through technologicaland institutionaldevelopments, approach the conditionsof private goods and be operatedunder competition. The exactpotential in each casewill dependon factorssuch as market size and the macroeconomicand institutionalframeworks.

5.14 For more privategoods, relianceon externalcompetitive forces-where contestable markets apply-becomes more appropriate. But for more public goods, the use of competition surrogatesbecomes essential. One clear exampleof the difficultyof applyingcompetitive market - 52 -

Table 5.1 Infrastructure Potential for Competition

Excludabiliq

Low High Subtractabiliq Public eoods Toll goods urban roads toil highways Low local rral roads telecommunications waterways ports highways airports wastewatertreatment urban water systems power transmission power distribution COmmoA:2ro29eq9ods Private eoods irrigationsystems road transport High rural water systems urban transport hydro power generation rail transport thermalpower generation water storage storm drainage on-farm irrigation

criteria is in roads-it is in everyone'sinterest to use the service,but in no one's interestto pay for it. Becauseit is difficult to excludeindividuals, roads are overused.However, this is one of the cases in which technologicaldevelopments, such as electronicpricing, might changethe economic nature of the activityby making cost-recoveryfeasible, at least for some roads. Some forms of infrastructurelend themselvesto combinationsof alternatives-such as when generation and distributionof electricitycan be separated. Althoughthere are few economiesof scale in power generation,a fairly competitivemarket might be possible, but transmissionand distributionmight have to operateas regulated monopolies.

5.15 The potential of competitivemarkets for infrastructure services in "average" circumstancesis much higher than was presumedonly a few years ago. But, equallyimportant, some services likely will never be suitable for competitivemarkets-such as most interurban highways and urban roads, large water supply and main irrigationnetworks, and hydro power generation. Thesemight be closerto defenseand the administrationof justice than to road transport or telecommunications. Governmentsmust acceptthat these activitiesare likely to remain in the public domain as monopolies. For highways, the Infrastructure and Urban Development Departnent's paper on user chargesand accountabilitysuggests several ways to simulatecompetitive - 53 - pressuresthrough administrative and other means (Heggie1991). It is unfortunate,however, that the emphasison the importanceof competitivemarkets and privatizationhas deflcted attentionfrom the need to continue exploring ways of enhancingthe performance of monopoliesand quasi monopolies(Chapter 7.)

5.16 Toll and commonproperty goods also have serious and difficult problems. For example,with commonproperty goods such as somewater systems,one of the difficultiesis how to assurethat individualshave incentivesto contributeto productionand conservation.If individuals cannot be excludedor if the group is too large to hold individualsaccountable, free-riders may obtain.

The Private Sector Role

5.17 There has been increasingpressure to fmd alternatives to public provision of infrastructureservices. Motivatingfactors include heavy claims on current revenuesand a lack of borrowingcapacity that limitgovernments' abilities to financemajor investments-while demand for many types of infrastructureis rising faster than populationor income. And the opportunitycost of donor fundingof investmentis rising as donors have expandedinto policy-basedlending. Odtr factorsencouraging privatization are that publicsector management capacity is stretchedthin and at some governmentshave been convincedby efficiencyarguments that private investmentis more efficientthan public investment(Bennet 1981; Butler 1985;Hatry 1983;Roth 1987). Althoughthe most powerfulway to achieve a demandorientation is throughthe private sector in a competitive market, the current push to privatize infrastructuremay prove excessive. The potendal for privatizationvaries amongsubsectors. Even for subsectorsthat have good potential,other factors must be consideredbefore designing privatization policies for infrastructure(World Bank 1992).

5.18 Privatlaton as a condnuum. Private sectorparticipation can take differentforms. The easiest and most obvious one is a service providedby a fully private operator, but other alternativesinclude joint private-publicventures, contractingout, franchisesand concessions,a varietyof build-operate-transfer(BOl) schemes,and voluntarismand self-help.BOT and concession schemesare analyzedin Chapter6. Aspectsof voluntarismand self-helpare reviewedin Chapter 7.

5.19 Preconditons for private secdorpartcpaton. Aside from a suitable legal and institutionalframework at the countrylevel, preconditionsfor successfulprivate sector participation are a competitiveenvironment or, alternatively,an effectiveregulatory capacity; suitableplayers; and carefulplanning of the privatization. 5.20 Many privatizationefforts have not paid enoughattention to the competitivenessof the relevant markets. Collusionand oligopolisdcbehavior may require careful regulation-for example, in the Belize Second Road Maintenanceand RehabilitationProject and the Guatemala SecondaryRoad Maintenance Project. Bothprojects call for contractingout maintenance,yet neiher says muchabout the need for a viableand competitiveprivate construction industry which is virtually absent from both countries. Similar conclusionsare reached by Moore and others (1987), who studied contractingout in Honduras, stressingalso the need to examinethe political, legal, and regulatory environment. They conclude that allowing the private sector to operate without a - 54 - competitiveenvironment may lead to inefficiency,collusion, and the skewingof supply. Part of these negativeeffects can be mitigatedby allowingforeign frmnsto participatealongside domestic firms, as is done in Guatemalaand Honduras. Where naturalmonopoly conditions prevail-as in water supply-the negativeeffects can be worse (Triche1990).

5.21 A good exampleof an effort to build private sector capacitywhile encouraging competitivenessis the ZimbabweSecond Highway Project. The Bank and the governmentagreed to use a mix of contractorsand public constructionunits, someof which were to be force account (work done by the staff of the public agency). Theseforce accountunits were to receivepayment based on the averagerate for competitivebids for similarworks-the units' rental of equipmentand vehiclesat lower-than-marketrates was a partialsubsidy. The rationalefor subsidiesduring an effort to encouragecompetition is that they assure somecontinuity of delivery. The issue of continuityis not limitedto developingcountries. A commonconcern in the United Statesin contractedservice delivery is the potentialfor service disruptionbecause of strikes,bankruptcies, or slowdownsby privatebusiness. If part of a service continuesto be conductedby government,the potentialfor such occurrencesis reduced. New York City contractsout garbagecollection but retains a small fleet to assure a minimal level of continuity,to prevent oligopolisticbehavior, and to assure competitivepricing from the private sector (Stevens 1987). And, to prevent successfuloriginal biddersfrom takingunfair advantagein subsequentbidding rounds, an "independent"-inthis case, government-standard for measurementhas proved useful. This practice might be even more importantwhere preexistinglevels of competitionare low or where local monopoliesexist.

5.22 The constructionindustry is another example. Kirmani (1987)has identifiedkey policybottlenecks in the developmentof domestic-privatesector-construction industries. Despite vast activity involvingconstruction, the Bank's record in the developmentof domesticcontractors is grim-partly because of weak governmentcommitment to develop the industry outside the government'simmediate purview. In part, the problemis an inappropriatepolicy and institutional environment. Some of the more serious bottlenecksinclude a lack of governmentinterest in the problemsof the constructionindustry; poorly administeredprocurement procedures; difficulties in importingequipment, parts, and materials;inadequate institutional arrangements for credit, bonding, and insurance;and delaysin paymentsto contractors. Thus, specialallowances often are neededfor the developmentof privatecontractors. For example,in the MadagascarSeventh Highway Project, all rehabilitationand periodicmaintenance work is to be contractedout. To help in the development of the industry, a comprehensivetraining schemehas been prepared-for which contractorswill compete to enter-and field works will be conducted with tutorial managementassistance, coordinatedthrough a trainingarm. Aftertraining, contractors who have performed successfully will be approvedfor bidding. The projectalso envisagesproviding payments up front to ease liquidity problemsfor local small bidders.

5.23 The Malawi Northern TransportCorridor Project has made a consciouseffort to ensure incentivesfor privatesector development. Malawi Cargo Centers are to be establishedto be responsiblefor cargo operations. Three entitiesare to be involvedin ownershipand management: the government;a parastatal and parent companyof Malawi Railways; and a mixed enterprise comprising Malawi private sector investors, Malawi Railways, and the Petroleum Control Comnmission.The governmentwill transfer assets to the parastatal. These assets will in turn be leasedto MalawiCargo Centerscompanies-which will manage,promote, and marketthe use of fte - 55 - facilitiesand the NorthernTransport Corridor. The intent is to reducerisk and to encourageprivate sector involvement. These arrangementsprovide incentivesfor all parties to perform well. The parastatalresponsible for supervisingthe use of assetsand facilitieshas a direct financialinterest in the successof the Malawi Cargo Centers. And the expectationis that the financialperformance of other transportservices will be improvedwith the successof NorthernTransport Corridors. Private investors-although not bearing the entire risk of the venture-will share in gains and losses. The ultimateguarantor is the government.

5.24 In C6te d'Ivoire another form of private sector participationis the use of leasing -affermage-in which the municipalityconstructs a facility and contractswith a private firm to operateand maintainit. In this case, the governmentcontracts out for Abidjan'swater supplyto the Societede Distributiond'Eau de la C6te d'Ivoire (SODECI),a private entity. The govermnent retains control of price setting and the municipalityreceives a fee for the amortizationof the constructioncosts (Lewisand Miller 1986). Severalinconsistencies have hamperedSODECI. Until 1987, SODECIwas compensatedfor shortfalls. And becauseinvestment decisions resided in the (centralized) Water Directorate-with no risk to SODECI-optimistic estimates of potential consumptionled to overinvestment.This was particularlytrue in rural areas, where overdesignand overconstruction,weak monitoring by SODECI,and the failureto incorporatecommunity initiatives led to a cross-subsidyof rural water by high tariffs on urban users. (C6te d'Ivoire Water and SanitationProject 1990; Triche 1990).

5.25 In fact, in many contractualarrangements the major issuesare costs and risks-and ways can be foundto reducethese burdens on the privatesector. In the UnitedStates local and state governmentsregularly form public-privatepartnerships to provide many infrastructuralservices. Leasebuyback schemes have been widelyused. The public sector's capacityto financelarge-scale constructionunderlies the initial investments. These are then leased to the private sector for operationand maintenanceand a fee is paid to the municipalityfor the amortizationof the initialdebt burden. This practiceis increasinglycommon with airportsand other facilities,where the up-front costs and risks to the privatesector are high (Paganoand Moore 1985). Maintenancecosts become a responsibilityof the lessee. The applicabilityof these relationshipsto developingcountries is large where the burdens of the lumpinessof initial investmentand the issue of risk are yet more problematic,but where the need for ensuring accountabilityand operationalefficiency is higher.

5.26 A final importantprecondition in many developingcountries is the public sector capacityto prepare the units for privatization. Althoughin most casespublic enterprisesshould be privatizedwithout attempting to restructurethem, restructuringor downsizing,if it is necessary,is a highlytechnical undertaking for which few governmentshave the capacityin the right places. To prepare and negotiate a deal-for example, the price and methodsof financing-is also highly technical. Suchexpertise can be boughtfrom investmentbanks or consultants-athigh cost-but the public sector still needs the capacityto negotiatewith the bankersor consultants,to evaluatetheir work, and to ensurethat the consultants'recommendations and actionsare in the publicinterest. To have this capacitydoes not require a big organization,but a few well-qualifiedprofessionals with technicaland politicalindependence (Israel 1990). A recent studyof municipalsolid waste services in Latin Americaoffers an importantword of cautionin this respect. - 56 -

The existence of public agencies that are capable of negotiatingand monitoring contractsefficiently without unnecessary burdens on private operatorsis important. Unfortunately,[public] municipal waste managersin developingcountries too often have no professionalpreparation in the field and turnover is high, preventing continuityand coherencein dealingswith the privatesector. Trainingand technical assistance in this area can be provided through municipal strengtheningprojects financedby the Bank and other[s]..., (Bartoneand others 1990,p. 9).

The CentralRole of Regulation

5.27 Perhapsthe m6st important-and most neglected-aspect of the institutionalreforms under way in most developing countries is the need to regulate public-and private-sector infrastructureservices. This is a key counterpartto the larger role of competitivemarkets and the private sector. Liberalizationpolicies mean that govermnentswill no longer supply the services directly or regulate prices and quantitiesproduced, but still need to regulate the functioningof competitiveand noncompetitivemarkets. Infrastructureservices require three types of regulation: for competitivemarkets, for monopoliesor oligopolisticmarkets, and for the subcontractorsof public enterprisesor agencies.

5.28 First, even in fullycompetitive markets-rare in infrastructure,except perhapsroad transportor coastalshipping (if it is opento internationalcompetition)-the government must regulate to ensure that the rules of the game are clear and equallyapplicable to all participants. Regulations set technical and safety standards,labor laws, and financialand enviromnentalstandards. The environmentalstandards may have importanteconomic trade-offs that usually require extensive negotiations. And even if a market is competitive,governments need minimalmonitoring-not intervention-to ensure that the market remainscompetitive, that entry is unrestrictedand that no collusionamong producers is occurring.

5.29 Second,for monopoliesand oligopoliesthe governmenttask is much more complex. Governmentmust oversee the monopoliesand be able to intervenewith regard either to prices, investmentlevels, rates of return on assets, or other measuresof performance-if such indicators suggestthe monopolisticpower is being misused. Much has alreadybeen written about this topic. Third, an often-overlookedaspect of regulationis internalregulation-when public enterprisesor central agenciescontract out or provide licensesto privatesector operators to perform a numberof tasks. This is in contrastto externalregulation, performed by governmentagencies. The public agencyneeds to have the internalcapacity to monitorthe contractorsand ensure that the work meets specifications. In other words, the agencybecomes the regulatorfor these contractors.

5.30 DismantlUngthe disablingenvironment. An effectiveregulatory strategy also needs to dismantlethe disablingenvironment by eliminatingunnecessary regulation (and the attendant bureaucraticred tape). Much of the decline in the productivityof infrastructureservices in recent times has been attributedto the uncontrolledand contradictoryexpansion of the regulatoryand legal frameworkassociated with "doingbusiness." Where private entrepreneursmight provide specific services-for example, local small-scaletransport, rural road maintenanceprograms, rural road constructionprograms-the burdensof compliancewith overly sophisticatedstandards discourage responsivebehavior. Entrepreneursfacing severely circumscribedopportunities to compete-and - 57 - unable to change governmentpolicy through the use of voice-walk away from ("exit") formal markets(Hirschman 1970).

5.31 In the PakistanPower Project,despite efforts to createincentives for private sector participation,private sector interestwas declining. In 1985,the governmentannounced incentive measures to encourage private sector participation-for example, special fiscal incentives, an extensionof incometax holidays, and exemptionson import duties on equipment. Yet, response remainedweak. Studiesconducted before appraisal suggested that a principalimpediment to private sector participationwas the evaluationand review process. Cumbersomeprocedures involving undertrainedstaff became sourcesof incredibledelay. Pressureto speedup reviewsled to ad hoc evaluations,uncertainty, and inconsistency-anddisinterest by investors. In the ColombianPower SectorAdjustment Loan (12/87)it was noted that for greaterefficiency and effectivenessin energy production,there was a needto addressthe "highlylegalistic regulatory enviromnent." Lengthy and time-consuminglocal regulationsand proceduresaffect every aspect of projectimplementation.

5.32 The need for proceduralreform is also an issue for donors. Carapetisand others (1984)point to the consequencesof rigid standardsof road constructionrequiring expensive capital- intensive methodswhen local community-basedalternatives often work better. Also, excessive regulationcan have other negativeeffects. In the Philippinesofficially sanctioned fare structures often presumeoperation on all-weatherpaved roads. Thus, privateoperators rarely seekpermits on unpavedroads. As a result informaltransporters operate on these sectionsand charge as much as two-to-fourtimes the rate officiallyset for "good' roads. The consequenceis a disenfranchising effect and few pressuresfrom operatorsto improvethose roads. 5.33 In other cases, the issue is one of inadequateenforcement of regulations. In the ColombiaSeventh Highway Project, one major issuewas the government'swillingness and interest in enforcingaxle-load and dimensionregulations. Failure to enforcethese regulationsmeans rapid deterioration of road systems. Another common example is the inability-often the legal incapacity-of suppliersto prevent the unauthorizeduse of a service,resulting in less than full cost recovery. This is especiallyrelevant in water and power. The result is reduced incentivesfor efficientproduction-and inefficienciesin allocationand distribution. In the SecondManila Water Supply Project, high levels of unaccounted-forwater were recognizedby a reasonablycompetent water authority. Whenthat authorityattempted to expelillegal users, however,much time and effort were neededto obtain legaljudgments because of the bureaucraticnature of the legal process, and the authorityrecovered only part of the value of the illegallyconsumed water. Existingpenalties appear to be inadequateto deter illegal use. And numerousstudies of rural water supply have pointedto failuresof projectdesign to considerthe potentialdiversion of water for nonhouseholduse in agriculture. Churchilland others note:

Experiencehas shown that rural villagershave often diverted water intended for drinking, cooking,and washingto irrigatingcrops.. .This is particularlytrue where water is made available in large quantitiesthrough the use of piped distribution systems. Villagersin Senegaland Nigeria, for example,were foundto be tapping a large percentageof their piped water for agriculture...(1987, p. 53). - 58 - 5.34 In Burma's SecondTelecommunications Project a similar problem arose with the inabilityto enforcepricing policiesand collection. The governmentprohibits the disconnectionof servicesfor nonpayingcustomers for a full year. The result is weak internal incentivesto collect and a loss of operationalefficiency. A similar,yet more dramatic, case can be foundin Indonesia, where governmentis the delinquent. The majorreason for low levelsof collectionwas the arrears of other state enterprises. Bill collectionby the NationalElectricity Authority for nongovernment customerswas generallysatisfactory: the accountsreceivable fell below two months' billing. In contrast, outstandinggovernment accounts receivable represented more than one year's billing in 1981-82(World Bank 1988a,p. 47).

5.35 Improving the regulators. To ensure that infrastructureoperators meet safety, financial,economic, and other standards,governments need an institutionalcapacity of higherquality than is traditionallyfound in regulatoryagencies in developingcountries. A regulatoryagency needs to be able to acquire and process all relevant informationfrom the regulatedunits and elsewhere; conduct technical, financial, and economic analyses of the regulated units and the subsector; coordinatewith other public and private entities;and interact with and prevail over the regulated units. These functions are difficult to perform well in any circumstance,but particularly in developingcountries. They requirehigh-caliber staff with enough prestige, statusand independence from the regulatedunits to deal with other governmententities on an equalbasis.

5.36 For example,a regulatoryagency for water supply, overseeinga few private water companies operating as regional monopolies, would need several economists and financial analysts-to monitoroperating costs and the financialevolution of the water companies,and be able to link the water sectorwith the rest of the regionaland nationaleconomy. The agencyalso would need staff to cover the technical aspectsof regulation,and perhapssocial scienceprofessionals to deal with related social issues. In an averagecountry the numbersof required staff would not be large, but would have to be preferablyof a higher professionallevel than those working in the regulatedunits.

5.37 Few countriesappear to have thoughtthrough the governmentcapacity needed to regulatetheir privatesector strategies. The new regulatoryagencies are more like think-tanksthan traditionalbureaucracies and mostpublic sectorsin developingcountries are not able to attract and retain the needed high-caliberstaff. Other areas also are competingfor such staff (Israel 1990). Efforts are beginningin developingcountries to tackle regulationsystematically. For example,the main objectiveof a recent technicalassistance project for telecommunicationsin Mexico has been to help developa regulatorycapacity. If past experienceis any guide,however, the privatesector- orientedstrategy will not succeedunless the regulatorycapacity proves to be first class. Witnessthe difficulties developed countries are having effectively regulating infrastructure. Water, telecommunications,electric power in the UnitedKingdom, and the airline industryin the United Statesare cases in point. - 59 -

Chapter 6

Alternative Mechanismsfor Private Sector Financing of Infrastructure

6.1 Infrastructure investnent is still a public sector responsibility in most of the developingworld. But governmentshave begun exploring or encouragingprivate financingof infrastructureinvestments and services. This chapter reviewssome of the newer or less familiar options available to govermments.Three main types of financingtechniques are analyzed: (i) investmentsto build, operate, and then transfer a systemto the government-usuallycalled BOTs but withmany variants,including concessions, which usually entail rehabilitating an existingsystem, anddevelopment gain, which capturessome of the positiveexternalities of infrastructuredevelopment by giving the private investor the right to develop adjoiningproperty; (ii) self-help, including everythingfrom cooperativesdelivering telephone services to power cogenerationto feeder road construction;and (iii) bond financing. The chapterbegins with definitionsand illustrationsof these techniques and presents an inventory of existing projects, showingthe extent of their use in developingcountries; reviews the cost and benefitsof private infrastructureinvestment, using the BOT approachas example;and drawsconclusions on privateinvestment potential in infrastructure.

Types of Financing

6.2 Build-operate-transferand variants. BOT projects involve a private company- usually a consortiumled by an internationalconstruction company-that finances, builds, and operatesan infrastructuresystem for a fixedtime duringwhich the governmenthas a regulatoryand oversight role. BOT projects are designedto generate enough revenues to cover the project company'sinvestment and operatingcosts plus an acceptablerate of return on capital-usually 15 to 20 percent. At the end of the project,usually in 15 to 25 years, the systemis transferredback to the government. BOT projects were originallyconceived to transfer commercialrisks to the privatesector and thus free governmentfunds for otheruses. To date, lendersand projectcompanies have been reluctantto committo a projectunder nonrecourseconditions, so a governmentguarantee is usually negotiated. BOT arrangementsare deal-specific,with the allocationof risk among creditors,the sponsoringgovermnent, and the projectcompany varying significantly by project. The actualallocation of risk is perhapsthe key to a project's attractivenessto the participants.

6.3 One of the most successfulBOT projectsto date has been Hong Kong's East Harbor Crossing, a combinedroad and rail tunnel under Victoria Harbor, which was financedentirely through the private sector using long-termfranchises offered by the government. The project is managedby a privateconsortium, which put up equityworth 29 percentof projectcosts. Observers attributeHong Kong's successin organizingand executingthis and other BOT projectsto careful project appraisal, competitiveproject bidding, and the government'swillingness to bail out the project in case of financialdifficulties. The govermmenthas also been willing to supportprivate investorsby rearrangingits own prioritiesto providenecessary ancillary investments, such as access roads. In return, the projectcompany has assumedall commercialrisks. - 60 - 6.4 The Bankis helpingto setup Pakistan'sfirst BOTproject-a 1,292-megawattoil-fired powerplant to be locatednear the mouthof the Hab River. Totalproject costs are estimatedat $1.1 billion to $1.3 billion with 20 percent equityfinancing from a combinationof project sponsorsand convertiblebond issues; 30 percentof fundingfrom Pakistan'sPrivate SectorEnergy Development Fund, createdpartly with Bank funds; and 50 percent from commerciallenders and export credit agencies. After more than four years of preparationtime, constructionis expectedto begin in mid- 1992. The projecthas been slowedby the need to establishlaws and regulationsconcurrently with projectnegotiations.

6.5 A projectfrequently held up as an exampleof everythingthat is wrongwith the BOT approachis the Gazi powerplant in Turkey. After six years and millionsof dollars in development and negotiationfees, this BOT contract for a 1,000-megawattplant has still not been awarded. Problemswith this projecthave includedthe inexperienceof the governmentand the sponsorswith this form of contractingand their subsequentshifting demands on such issuesas risk allocationand tariff levels. An additionalproblem has been poor organizationof the competitionfor contractors and designs. The InternationalFinance Corporation, among other creditors,has pulled out of this project and its fate is uncertain(Table 6.1 comparesthe terms of these three projects).

6.6 In the decadesince the term BOT was coined,several variations have been spawned (Table 6.2). For a refurbish-operate-transfer(ROT) projectin the Philippines,a project company shippedin an old power plant, refurbishedit, and broughtit on line in much less time than it would have takenfor a green field investment.Develop-built-operate (DBO) is a new idea being proposed for at leasttwo urbantransport projects (Ankara and Karachi),in whichthe projectcompany assumes no commercialrisk initiallybut is financiallyaccountable for buildingand operatinga systemwithin performance specifications. The company then assumes commercialrisks incrementallyand conditionallyas the governmentsets up appropriateregulations and capitalmarkets.

6.7 Concessions. Concessionsare an older form of the BOT idea in which a private companycontracts with a governmentto build or expandand operatean infrastructuresystem for an agreed-uponperiod of time during whichit bears full commercialrisk. The companypays a fee to the governmentand is responsiblefor all investmentand maintenanceduring the concession period. Compensationis throughtariff revenues;contracts are designedto be long enoughto allow concessionairesto recouptheir investment.Concessions are typicallyin water and sewerageand are most commonin France, Spain,the former Frenchcolonies, and some countriesin Latin America. 6.8 In CMted'Ivoire, the operationof Abidjan'swater supply systembegan as a lease contract awardedfollowing a competitivebid in 1960. As indicated(para 5.24), the contractwas revised in 1987 to give SODECI,a privateentity, a full-fledgedconcession. The water supply systemis one of the best-operatedbut most expensivein Sub-SaharanAfrica. The new contract reduces SODECI's compensationbut increasesits autonomyin investmentplanning and execution in the urban water sector. This adjustmentwas a response to criticisms that fragmented responsibilitieshad led to a lack of accountabilityand efficiency. In the past thirtyyears, Ivoirians have taken an increasingrole in the managementof SODECIand most of the company'sequity is now held by privateIvoirian stockholders. Table 6.1 Comparisons of Terms of Three BOT Projects

Totl Prepration Return Concession Government ie LEtCost Equity Debt Thue on Equity Bid? Period Support Fincing Structure

agKong $500 mion 29% 71% 2byearsfrom Unknown Yes 30 year road Govt. franchise Equi: sat Harbor bid tender to franchise;22 allowsfiul access to -100% from sponsors rosxaig conitact; year rail road and rail est. 4 yn fianchise revenuesalong route; Debt: constuction comniitmentto - 80% revolvingcredits construct of - 20% _instllmentsale supporting facies

istan $1.1 billion - 20% 80% Unsolicited 18 - 23% No 23 years Securitypackage Equity: ab River $1.3 billion bid received includes: - 60% from sponsors wer Plat 12.87; subordinationof - 40% convertible constuction governmentdebt; bonds not yet begun force majeure; inflation; forin . debt: exch; fuel supply; tax - 65% export exemptions; credit protectionagainst agencies change in tax or duty - 35% Private levels; predetermined Sector pric set to guarantee Energy DevelopmentFund real return on invt key $1.3 billion 20% 80% 40 months 16% Yes 26 years Take or ay; tax Equity: azi Power since bid waiver, forex guars; - 67% sponsors hat submitted; force majeure; - 33% Turish status subord'd govt debt to government uncertain be issued to cover operationalshortfalls Debt: - 100%export credit agencies - 62 -

Table 6.2 BOT Variants

BOO (build-own-operate)in which there is no transfer back to the government

BOOT (build-own-operate-transfer)with a special "T" for transfer and training

BROT (build-rent-operate-transfer)

DBO (develop-build-operate).

ROT (refurbish-operate-transfer)

6.9 AlthoughBOTs and concessionsare similar contractually,they have carved out distinct market niches. BOTs-the emphasistends to be on the "B," the building phase-and consortiumsare usually headed by an internationalconstruction contractor. Concessions,on the other hand, focusmore on the managementof a system,usually a utility, and their consortiaare usually led by a Europeanutility company. AlthoughBOTs tend to be stand-aloneprojects from scratch in the power generationor transport sectors, concessionsare usually awarded for the expansionand operationof existingwater or power utilities.

6.10 The differencebetween concessions, lease contracts,and management contracts needs to be kept clear-although the three may overlap or be combined. Concessionsare a contractual arrangementin which a privatecompany invests in new infrastructure. Lease contracts,as defined by Vuylsteke(1988) share some of the characteristicsof concessions-a contractorpays a fee to govermnentin exchangefor the right to operate and maintaina facilityaccording to commercial criteria for a fixed time. But the contractor is generallynot responsiblefor new investment. Managementcontracts are a step furtherremoved from concessions. Here, a companyis hired to operatean infrastructuresystem for a fixed-feewith no financialexposure. All risk rests with the government.

6.11 Developmentgain. A financingtechnique that has becomepopular in the countries of the OECD is knownas capturingdevelopment gain. Usingthis technique,some of the positive externalitiesof new infrastructureinvestment are internalizedby givingthe privateinvestor the right to developadjoining property and thus enjoy some of the benefitsthe investnent creates-notably - 63 - higher property values. This technique,which requires large land holdings and a sophisticated propertydevelopment sector, has been used successfullyin HongKong and is proposedin Thailand. HopewellHoldings, one of the major players in BOTshas gained cabinet-levelapproval on a $2.8 billionbid to span Bangkokwith 60 kilometersof mostlyelevated rail and 48 kilometersof highway. The proposalis for a BOT-likecontract in which the governmentwould provide neitherequity nor financialguarantees. Hopewellwould put in equity worth 10 to 15 percent of project costs and would earn much of its return throughbuilding and leasingretail space in the rail stations. Some incomefrom the real estate developmentswould also subsidizerail fares. Hopewellwould pay the state rail system,which ownsthe properties,a fee for accessto these propertiesover the anticipated thirty-yearlife of the contract.

6.12 Self-help. Mobilizinglabor and capitalto constructsmall-scale, local infrastructure is one of the most commontechniques for supplementingpublic resources-with varyingdegrees of organizationand varying amountsof state supportor opposition.What the projectsin this category shareis the smaller-scale,self-help element of the initialinvestment-even if the eventualsize of the enterpriseis substantial-for example,an electricalcooperative in Santa Cruz, Bolivia,had grown to 50,000members by 1981and was providingpower to that city and more than 80 nearbyvillages. In some casesthe net effectis large; Chinais estimatedto have some 100,00030-kilowatt minihydro projects. Water,power, and telecommunicationscooperatives prevail throughout Latin America and Asia-and feederroad constructionand maintenance are managedlocally in many Africancountries. 6.13 Oftenlabor contributions reduce the financialcost of theseinvestments. For example, in the BanglungDistrict of Nepal, local communitiessuccessfully constructed sixty-two suspended bridges through a combinationof local inputs and governmentfunds. Althoughthe national governmentprovided some special materials, the local communitiesprovided all of the locally availablematerial and labor. Householdsnot able to participatedirectly in the constructionwere asked to contributefood or money. Costs to the governmenttotaled only about $50,000-and the amountsmobilized locally were substantiallygreater.

6.14 At the high end in terms of local financial input, cooperativesthroughout Latin Americaprovide telephoneservices in townsor villages. Typically,cooperative members finance the purchaseof a smallexchange, cables, telephonesets, and other necessaryequipment for a local telephone network, which may or may not be hooked up to the national system. Such rural telephonecooperatives have operatedfor decadesin Brazil,where the cooperativescan connecttheir subsystemsto the nationalsystem, subject to technicalstandards and usage charges. Experiencein the Sao Paulo area showsthat the rural telephonecooperative system is particularlyeffective in mobilizinglocal participationand private investmentfunds. Initialand recurrent costs are said to be lower than those of the national utility because of the lower costs of local labor, some self- construction,low overhead,and the use of simpletechnologies.

6.15 Perhapsthe most dramaticform of self-helpis the substantialinvestment by private enterprises in generatingcapacity for their own use in response to frequent power outages or unsteadyvoltages. This power is essentialto the functioningof many developingeconomies; in Indonesia,captive capacitytotals an estimated4,600 megawattscompared with 6,420 megawatts installedby the public power corporation. Anotheroutstanding example is private investmentthat has replaced public invesunentin tubewells,as in Pakistan, or public investmentin large-scale - 64 - irrigationfacilities-as in Nigeria where between 1984 and 1989 almost 9,000 northernNigerian farmers investedin shallowtubewells and pumps. The investmentshouseholds make to deal with infrastructureunavailability or unreliability-especiallyin water and sanitationand power-also are a form of self-help. Theseinvestments can be substantialin someeconomically powerful households but, as indicated,are more importantrelative to incomein poorer households.

6.16 Bonds and infrastructurebanks. In many Asian and Latin Americancountries, publicutilities issue government-backedbonds to raise private capital. Municipalbonds are rarer becausemost municipalitiesare not creditworthyand can get money more cheaplythrough some other vehicle such as central governmentallocations, local taxes, or donor funding. A common constraint to bond financing in developing countries is weak local demand for investment instruments. Public financialmanagers have been able to circumventthis low demandin several ways. In Pakistan, securitieshave been issued denominatedin foreign exchangeto encourage overseassavings repatriation. Othercountries have madebond purchasescompulsory. For example, the municipalityof Taegu City, Korea, is financingroads through a bond issue. All residents wishingto registertheir vehiclesmust purchase these bonds. Thus, this transportinfrastructure will be financedby potentialusers. In Brazil, ELECTROBRAS,a holding companywith equity in roughlyninety utilities, gets its revenuefrom loan repaymentsand what are essentiallycompulsory loans from the larger electricityconsumers who mustpay a billing surchargeeach month. At year's end, the customersreceive long-term bonds as compensationfor thesepayments. In addition,utility or road bond issuescarrying government guarantees are often boughtup by nationalpension funds, providinga valuablestimulus to local capitalmarket development. A good exampleof this practice is the bonds that have been issuedto fiance Indonesiantoll roads.

6.17 Infrastructurebanks are a public targeted-lendingfacility, typically found in Latin Americaand financedthrough a combinationof bond issues,government funds, and externaldonor support. The banksonlend their capitalto local governmentsfor infrastructureinvestments. Loans are often guaranteedby the nationalgovernment. Infrastructurebanks mobilize domestic funds for investmentand offer an attractivevehicle for donorfunding; the WorldBank has helpedfund twenty- six such banks. Most of these eventuallycollapse under the weightof bad debts. Their financial difficultiesusually stem from lending at below-marketrates and havingthe government-guarantee facilitiesbear most or all of the risk on the loans.

6.18 An exceptionis Mexico'sBanco Nacionalde Obras y ServiciosPublicos, which is financedthrough a combinationof bonds, governmentsources, loan repayments,and multilateral agencies such as the IBRD and Inter-AmericanDevelopment Bank. The Mexican bank makes advancesto federalagencies and to states, municipalities,and their agenciesthrough several lines of credit-for example, a water supply and sewerage investment fund lending to states and municipalities;a regionalurban investmentfund; a nationalprogram for urban developmenttargeted to small municipalities;and a higher-interestprogram for urban developmentused mainlyby the largest cities. Interest rates are positivein real terms and above the cost of capital, but loans to smaller municipalitieshave lower rates than loans to large municipalities. The Mexicanbank's overalf financial situation is strong; it is responsiblefor loan appraisal and supervisionand has authorityto issue bondsearmarked for individualsectors. -65 -

Costs of Private Financing

6.19 Costs associatedwith private financinginclude the diversionof domesticresources to a lower-priorityactivity-to the extentthat projectselection criteria are distortedby the quest for leverage-to bring in the foreign investmentand make the project work. There are other costs. Necessaryancillary investments by the governmentcan often be significant.Private investors prefer to investin existinghigh-demand areas insteadof extendingservices to new, unservedareas-because revenue is surer-even thoughdoing so may be againstthe best interestsof the city or region. In addition, it is not clear that privately funded constructioncosts less than public alternatives, especiallyfor projectsthat are not bid competitively.Foreign owners and creditorsmay also seek compensationfor the inherentriskiness of these projects and for the high transactioncosts in their preparation and negotiation. Given these considerations,investors in BOTs and BOT-like projects-concessions or developmentgains-might be expectedto demandhigh rates of return. Althoughthe sparse informationat hand suggeststhat financialrates of return are not higher than in the OECD countries, rates are certainlyhigher than those soughtby public sectormanagers.I1/ The important point is the rate of return on equity-which is usually much higher because of leverage. But there is little informationabout this aspect.

Inventory of Privately Financed Projects

6.20 An analysis of private financingof infrastructurein the developingworld (Tables A6.1 and A6.2)-by countryand by infrastructuretype-points up that by early 1991 fewerthan ten BOTs had begun operating,although more than twice that number had been initiatedor were in advanced planning. Power generation and transport are by far the most common types of infrastructurefinanced through BOTs. Most projects are concentratedin a few Asian countries. Concessions,on the other hand, are establishedand more widespread. They are foundtypically in French-speakingAfrica and in the water sector, but in Latin Americathey are all power utilities. Developmentgain is just being introducedin developingcountries-and has been proposedonly for mass transit.

6.21 Self-helpis a commonway of financingsmall-scale infrastructure projects-in all sectors-throughout the developingworld, often using cooperatives,which are especiallycommon in Latin America. Bonds-in Asia and Latin America-are used to financemajor investments in all infrastructuretypes. Infrastructurebanks are foundprimarily in Latin America,where they typically lend for urban infrastructureinvestment, such as roads and water and sewerage.

IQ/ Financialintern rates of return vary for transportprojects in the OECDcountres. The Cheapeake Bridgein theUnited States yielded about 5 pect, theMont Blanc Tunnel 17 percet, the ChannelTunnel (beforerecent cost overrns) 17 to 19 perent, and the Paris-Colognefast train (TGV)6 to 11 percent. The Hong Kong Metro hadnegative retus, whoile e lasttwo Hong Kong tunnels were expectedto yield 14 to 16 percent(Havlicek 1990). Projectedrates of retum of BOTsin developingcountries appear to be higher. IheHab River Powerproject in Pakistanis aning 18 to 23 percent;the GaziPower Plant in Turkey,16 percent;Malaysia's Labuan Water Supply Project 18 to 20 percent;and Bangkok'sSecond StageExpressway, 21 percent(Augenblick and Custer1989). - 66 -

Potential Costs and Benerits

6.22 Private investmentin infrastructuregenerally offers four potentialbenefits: reduced risk for the publicsector; greater efficiency and innovationin constructionand operationand reduced burdens on public sector management(when investmentis linked to private ownership and operation);additional funding for investment;and positive externalities-tangibleand intangible. Having a greater private sector presence in the economy leads to a better image in the world investmentcommunity, more responsive and appropriatepublic policy, improved management attitudesand practices, and greater exposure to outside ideas and market developments. Bond financing,with no direct influenceon systemdesign or operation, could achieve the latter two benefits. The characterand magnitudeof a self-helpproject's benefitsmay be somewhatdifferent.

6.23 Additionalk. Additionalityis usually defined as the net increase in investment resources made available to the economy as a result of private investmentin infrastructure. Resourcesfor investmentcome from domestic savings-private and governmental-and foreign capitalinflows. A projectmust increasethe total of these resourcesto be additional. The question of the existenceand level of additionalityis contentiousand complicated. Skepticismabout these benefitsappears widespread among Bank staff, althoughno systematicanalysis has been done.

6.24 The extent of additionalitydepends on the patternof financing-whetherforeign or domesticfinancing, using debt or equity. Foreign equity componentsare clearly additional.The foreign equity providers engaged in a BOT or BOT-likeproject are far more likely to invest elsewherethan in the host country in the absenceof the project. The inflow is a net incrementto nationalsavings-and there is no reason why other savings, foreign or domestic,should fall as a result. The value of the equityinflow might be reducedif, as some practitionerssuggest, project proposersraise the prices of project inputsto offset the cost of their 'forced" equity participation. But exactly how this would happen in a competitivesituation is not clear and, in any case, such considerationsshould be reflectedin the project'srate of return. The "deduction"from equity is thus not likely to be significant.l/

6.25 Foreign debt financingis probablylargely additionalin most countries. The two credit sources are commercialbanks and internationalor bilateral aid or export credit guarantee agencies. One basic questionis whetherthese institutionshave fixedcredit ceilings-risk limits-for the borrowingcountry and whether,in the absenceof the projectin question,they would lendto that limit anyway. Neither is likely. Unless countriesare heavilyindebted, commercial banks would probablybe ready to lendto a BOT or BOT-likeproject with a specialsecurity package, in countries where they wouldnot otherwiselend at all. Even if individualcommercial banks have risk limits, the BOT or BOT-likeproject can engagea larger numberof banksin the financing-thus expanding total credit availability. The aid and exportcredit agenciesare likelyto find private-publicventures attractive lending opportunitiesand thus expand their country lending ceilings. The other big questionabout additionalityof foreigndebt financingis whethera BOT or BOT-likeproject crowds out other investmentsbecause of the local resources it absorbs. Whether crowding out occurs

11/ In principle, it is necessaryalso to look at this issue of additionalityin the long term, when the net effect will depend on future repartitionof equity, dividends, tiedcontracts, and so on. - 67 - dependson the project's fmancingpattern and the economy'sflexibility. The need for govemment guaranteesmight reduce the extentof any additionalityby subtractingfrom the govemment'sability to borrow for other purposes.

6.26 A projectdrawing on private domesticequity is additionalonly if it attracts flight capitalor raises interestrates and inducesnew savings. Fiancing usingdomestic debt is generally nonadditionaland probablydistorts investmentallocation-as governmentdiverts resourcesfrom possiblyhigher-productivity uses to benefitfrom the leverageof the BOT or BOT-likeproject. For bondsdomestically issued and held, the attributesof both domesticdebt and equityhold. There is additionalityto the extentthat flightcapital is repatriatedor interestrates and savingsrise, and there are allocativedistortions.

6.27 Because of the sparsity of experienceand evidence, it is hard to see a typical financingpackage. Foreignequity investnent, however,is presentin all BOT or BOT-likeprojects, generallyat 10 to 30 percentof total investment. Foreign commercialbank credits are invariably a major element,usually carrying an exportcredit guarantee. Host governmentequity participation does not seem to be a major element comparedwith contributionsusing subordinateddebt and guarantees. Thus, the additionalityof mostBOTs and BOT-likeprojects is probablyrarely less than 25 percent of the total projectcost and is likely to be substantiallyhigher.

6.28 Reducedgovernment risk. A centralfeature of BOTsor BOT-likeprojects is limited- recoursefmancing, in which owners and creditorsreceive no governmentguarantees-but can lay claim only on the projectcompany's assets and revenues. To the extentthat projectrisks are borne by privatefinanciers, government risk is reduced. But governmentrisks are usuallyreduced by no more thanone-third. In the typicalcase, there are guaranteesof minimumtraffic volumeor take-or- pay arrangements,as well as governmentguarantees on projectdebt. The Hong Kong East Harbor Crossingproject, discussedabove, is rare in that the ownersbear substantialrisk.

6.29 Effidency. Efficiencybenefits from privateinfrastructure investnent can come in either lower-cost,more innovativedesign and constructionof the physicalplant or in more cost- conscious,market-sensitive, and dynamicmanagement. These benefits,however, might be partly offset by the preclusion of constructionin stages and the need to introducetolls that add to constructionand operatLigcosts.

6.30 Doubtsexist aboutthe magnitudeof bothefficiency benefits. The first benefitis most likely to be significantfor a BOT or BOT-like project for which bidding is competitiveand specificationsare general. But, the highnegotiating costs, the endemicuncertainties, and the nature of BOTs-with their unsolicitedproposals and frequentinformal commitments at early stages-will always make competitivebidding difficult or ineffective. So efficiencygains at the design and constructionstage may not be substantial.

6.31 The gains from more effectivemanagement could be greater, because operating inefficienciesare usually a greater source of resource waste than are bad design or excessive constructioncosts. But inefficientoperations often result less from bad managementthan from inappropriate macroeconomicpolicies, weak regulatory systems, or crippling institutional arrangements. And at least some of the efficiencygains from private ownershipand management - 68 - could be obtained by contractingor leasing to private management. In principle, these two possibilitieslower the potentialnet benefitsfrom reducedburdens on management.

6.32 Positive externaies. Successful adoption of a program drawing on private investmentfor infrastructurewill almostsurely bring many positiveexternal effects, althoughtheir magnitudeis uncertain. The reputationof a countryand the worldwideperception of its investment climatewill almostcertainly improve; openness to such investmentis a strong signalof commitment to privateinvestment in general. Local capitalmarkets may be strengthened. Even the processof negotiationcan have positiveeffects by makingexplicit the nature and magnitudeof the risks and uncertaintiesinvolved in the project and by making clearer where the regulatory systemblocks investmentand frustratesentrepreneurship.

6.33 The process can leave governmentnegotiators and the economybetter prepared for dealingwith private-especially foreign-investors. A facility'smanagers can be expectedto become articulateadvocates of prudentmacroeconomic management, regulatory reform, full-costpricing of services,and bettermaintenance of ancillaryinfrastructure. The project thus willhave demonstration effects for managementof other firms-state-ownedor other-and provide a new kind of window to the world. And, finally, such a project will bring a new form of liberalizingyeast to the economy. Thesepositive external efforts may be the most importantof all the potentialbenefits.

Potential for Use

6.34 Some typesof privateinvestment have more promisethan others and the potentialis greater in somedeveloping countries than in others. For many reasons,the classicBOT-a major green field investmentin roads, tunnels, or power stations-is unlikelyto spread much beyondthe few faster-growing,highly market-orientedeconomies in Asia where it now can be found. Concessions,self-help investments, and bond flotationare more widely applicable.

6.35 The slow implementationof BOTs partly reflects their newnessbut also indicates more fundamentalobstacles. First, few countrieshave regulatorysystems well-developed enough that definitionsof rights and obligationsof private investorsand the state can be straightforward. Contract negotiationsgenerally occur in a black box with an abundanceof gaps and ambiguities. And, the BOT arrangementis still so rare that replicablemodels do not exist. Each countryhas a uniqueenvironment, partly because of the uniquecharacter of each deal, and partly becausedetails are often secret. Changesof governmentcan further complicatematters. So transactioncosts in these projects are relativelylarge.

6.36 And-because infrastructure remains highly vulnerable to government interference-few privateinvestors in infrastructureappear willing to bear the kindsof risk that they acceptin other investments. For BOTs, private investorstend to require a higher return on their capitalthan most governmentsfind acceptable,and the implicitcosts to governmentcan mountas ancillary investmentsare required-access roads, new transmissionlines, and so on. Thus the poorest countriesare not likelyto be able to attract BOT investments.

6.37 Concessionshave a better track recordthan classicBOTs and are found throughout developingcountries. Unlike BOTs, concessionshave had well-wornmodels for many years. - 69 - Regulationsto make them work are alreadyin place in at least a few countries, makingthe path clearer for interestedinvestors and governments. In addition,unlike BOTs, concessionsdo not usually involve starting from scratch but, instead,involve private takeover of an existingpublic utility. New investmentis incremental. Comparedwith BOTs, then, the concessionsare inherently simpleand less risky. The ROT projectin the Philippinesis a variationon this approach.

6.38 Self-helpprojects, because of their smallerscale and independencefrom government oversightwill remainthe most commontype of privatelybuilt and managedinfrastructure. Their smaller scale typically limits their effects, but-as seen in private power generation in Indonesia-there can be notableexceptions. Often the most positivesupport governmentcan offer is the removalof regulationsthat oppose establishmentor expansionof systems. Bans on the sale of self-generatedpower to utilitiesare one of the more commonobstacles to expansion.

6.39 Bond financingof utilitiesor roads has successfullymobilized available savings in many countries, althoughnot in Africa. External bond financingrequires a functioningcapital marketand an attractiveinvestment. In countrieswith limitedindividual savings and a limitedcapital market, pensionfunds can provide a valuablesource of domesticcapital.

6.40 For the last severalyears, privateinvestment in infrastructurehas been identified-in the investmentcommunity, the Bankand the popularpress-with BOTs, but the BOT in its classic form is perhapsthe least promisinginstrument reviewed here, at least for the developingcountries that most needprivate investment in infrastructure.And while analyticand policyattention-in and out of the Bank-has been heavilyfocused on BOTs, the specialneeds of the poorest countriesin this area are inadequatelyserved.

. - 70 -

Annex 6.1 Page 1 of 6 Privately Fianced Infrastrucure by Country

Develop- NlGIONlCountryllnfttru oeType BOT Concession mentGain Self Help Bonds

AFCA Chad Utlity p CMteD'Ivoire AbidjanWater Systm 0 Ghana Rural Roads 0 Miadaguca Water Utility 0 Power Utlity 0 Telecommuliications 0 Malawi Rural Roads 0 Nigeria Sef :,generaon 0 Tub Wells0 Sao Tome and Prinipe Sao Tome Utility P ASiA U.-U ElectricityCooperatives 0 Water Cooperativw 0 China 600 MW Coal-firedPower Plant P 700 MW Coal-firedPower Plant 0 MinihydroProjects 0 Water Utility 0 Hong Kong Cros Harbor Tumel 0 B. Harbor Croi Tate's Cairn Crossmng lanaul FxexdCross P W. Harbor Crosng P Mas Transit Railway 0 Power Utility Water Utility 0 India Telecommunications 0 Railways 0 BombayPower 0 Power P ElectricityCooperative 0 Indones North-SouthLink 0 Other Toll Roads p Other Toll Roads P CogeneratedPower 0 ElectricityCooperatives 0

Key: 0 - Opfadonal I - Inklated P = Plamed - 71 -

Annex 6.1 Page 2 of 6 Privately Financed Infrastructure by Country

Develop- REGlON/Country/lnastructure Type BOT Concession ment Gain Self Help Bonds

ASA (condaued) Korea 0 Taegu City Roads Macau Water Supply 0 Malaysia Labuan Water 0 Labuan Power Transmission I N. Kelang Straits Road 0 KL Kepong Interchange 0 North-SouthExpressway P Kuala LumpurInterchange P Nepal SuspendedBridges 0 rnilippIe MaRL ElectricLight Co. 0 Private Power Dist. Cos. 0 ManilaWaterworks & Sewage 200 MW Gas Turbine Plant O 700 MW Coal-firedPower I 300 MW Coal-firedPlant I ContainerPort I 117 ElectricityCooperatives 0 TelecommunicatiodsCooperatives 0 Water Cooperatives 0

2nd Stage Expressway I BangkokMass Transit P P Other Road Systems P Cogeneration 0

ElectricityCooperatives 0 EUROPE THE MID-EAST, AND NOMTHAFRICA 0 Urban Infratructure Morocco Water Supply 0 Oman 100 MW Power I Pakistan Hab River Power I KarachiMass Transit P Water and Power 0 Tube Wells 0 Turkey 1000MW Aliaga Plant I 200 MW KaractepePlant I Gazi Power Plant P Cogeneradon 0

Key: 0 = Operaonal I -nitated P = Planned - 72 - Annex 6.1 Page 3 of 6 Privately Financed Infrastructure by Country

Devlop- REGION/Countryllnfrtructure Type BOT Concesion mot Gain Self Help Bonds

EUROPE, THE MID-EAST, AND NORTHAFICA (cotinued) Yemen O Rural Roads LATIN AMERICA AND THE CARIBBEAN Ar entina ETectricPower 0 Water Cooperatives 0 Barbados Light and Power Company 0 Bolivia LaPaz Power 0 Water Cooperatives 0 Electricity Cooperatives 0 TelecommunicationsCooperatives 0 Brazil ELECTROBRASUtility 0 Electricity Cooperatives O Telecommunic. Cooperatives 0 Small Local Infras. 0 Chile Santiago Water 0 Sewerage P Electricity Cooperatives 0 Water Cooperatives 0 Colombia Sewerage P Urban infrastructure 0 0 Urban Infrastructure 0 Electric Power 0 Electricity Cooperatives Costa Rica Cogeneration I Electricity Cooperatives 0 DominicanRepublic 40 MW Power Generator 0 Ecuador Guayaquil Private Power Co. 0 Electricity Cooperatives 0 Mexico Small Urban Infras. 0 Nicarapua Electricity Cooperatives 0

Key: 0 = Operadonal I - Iniated P = Planed - 73 -

Annex 6.1 Page 4 of 6 Privately Financed Infrastructure by Infrastructure Type

INFRASTRUCTURETYPEI Develop- Region/Country BOT Concession ment Gain Self Help Bonds

POWER - UTILTY Africa Madagascar 0 Asia Bangladesh 0 Hong Kong 0 lndia 0 Indonesia 0 Philippines 0 0 0 Vietnam 0 Europe, Mid-East, N. Africa Jordan 0 Pakistan 0 Latin Amerca & Caribbean Argentina 0 Barbados 0 Bolivia 0 0 Brazil 0 0 Chile 0 Colombia 0 0 Co5ta Rica 0 Ecuador 0 0 Mexico 0 Nicarapua 0 POWER - GENERATION Africa 0 Nigeria Asia China P 0 lndia p 0 Indonesia Malaysia I Philippines o

Thailand 0 Europe, Mid-East, N. Africa Oman I Pakistan I Turkey I p 0

Key:, 0 = Operational I = Iitated P -Planmed - 74 - Annex 6.1 Page 5 of 6 Privately Financed Infrastructure by Infrastructure Type INFRASTRUCTURE TYPEI Develop- Region/Country BOT Concession ment gain Self Help Bonds

POWER - GENERATION (conftnued) 0 Latin America & Caribbean Colombia Costa Rica I Dominican Republic 0 TRANSPORT Africa Ghana 0 Malawi 0 Asia Hong Kong 0 0 I p P India 0 Indonesia 0 P 0 Korea 0 Malaysia 0 0 P P 0 Nepal I ilgpapines I Thailind ~~P P p Europe, Mid-East, N. Africa Jordan 0 Pakistan P Yemen 0 Latin America & Caribbean Argentina 0 Brazil 0 Colombia 0 0 Mexico 0

WATER & SEWERAGE Africa Cote d'Ivoire 0 Madagascar 0 Asia Bangladesh 0 China 0 Hong Kong 0 Macau 0 Malaysia 0 Philippines I

KPv? 0= fn.rtfinviml T = T" dP Planned - 75 - Annex 6.1 Page 6 of 6 Privately Financed Infrastructure by Infrastructure Type INFSTRUCTURE T7PEI Develop- Region/Country BOT Concession mentGain Self Help Bonds

WATER & SEWERAGE (continued) Europe, Mid-East, N. Africa Jordan 0 Morocco 0 Pakisan 0 Latin America& Caribbean Argentina 0 Brazil 0 Bolivia 0 Chile 0 p 0 Colombia P - ~0 0 TELECOMMUNICATIONS Africa Maagsca 0 Asia India 0 Philippines 0 Latin America & Caribbean Bolivia 0 Brazil 0 IRRIGATION Africa Nigeria 0 Europe, Mid-East, N. Africa Pakistan 0

Key: 0 = Operational I = ad P = Planned - 76 -

Chapter 7

DemandOrientation through CompetitionSurrogates: Decentralizationand Voice

7.1 The previoustwo chaptersconcluded that competitivemarkets and the privatesector have a larger role to play in enhancingthe productivityand demandorientation of infrastructure activitiesthan is usuallyacknowledged. However, the intrinsicnature of someinfrastructure-many highways,urban roads, major water works, and power distribution-requiresthat they continueto be providedas publicgoods by the public sector,usually as monopolies. The extra accountability and demandorientation that automaticallyare part of a competitivemarket are unavailablefor such activities. Those effectswill have to be simulatedthrough competitionsurrogates, a second-best approach.

7.2 The mostpromising surrogates for infrastructureservices appear to be administrative decentralizationand voice-active participationof clients,users, or beneficiariesin the running or managementof a service. There is no reason, of course, why marketsand competitionsurrogates cannotbe used simultaneously. Becausemany forms of infrastructurewill continueto operate as regulatedmonopolies-or under monopolisticconditions-these competition surrogates can be used to achieve stronger market disciplineand demandorientation. Clients' voice regarding private monopolies,for example, could ensure better serviceand lower cost-with clients reinforcingthe role of regulators. Markets can be decentralizedto achievecloser localparticipation.12/

7.3 Competitionsurrogates are less powerful than competitivemarkets in enhancing productivityand a demandorientation. As a consequence,those services that must remain as monopolies-public or private-can probably never be expected to achieve the operational performanceattained in competitivemarkets. There is no completesolution to this problemand countriesshould learn to face that fact and act accordingly.

Decentralization

7.4 The option of decentralizationoffers in principlean effectiveway to achievehigher productivityand a demandorientation. This avenue, however, should be approachedonly when preconditions for success are in place. Local administrativeunits providing infrastructure traditionallyhave been weak fieldunits of centralagencies, dependent on centraladministrations for decisionmakingand financial resources. Increasing dissatisfactionwith the inefficiencyand unresponsivenessof central governmentshas fueleda search-often half-hearted-for decentralized alternatives. One result has been a tendencyto transfer responsibilitywithout careful consideration

J2/ Severalaspects of decentrlizaion and voice awenot reviewed in this paper-for example-the role of NGOs. See Paul and Israel 1991. - 77 -

of localor regionalauthority and absorptivecapacity. But becausepublic-sector decentralization is becomingthe norm-particularly in East Asia, EasternEurope, and Latin America-a clear position is needed on the subject.

7.5 Central governmentshave been criticizedfor inefficienciesresulting from reduced opportunitiesfor local initiatives;inaccurate information flows to decisionmakers;the reductionof civil servant responsivenessto citizens' demands, because of limited incentives for central governmentministries to considercitizens their clientele;and imperfectcontrols over civil service personnel. Other criticismscover resource allocationsthat favor central governmentsover field units; confusionof responsibilitiesand accountability,resulting in weak incentivesat all levels; tendencies to be more concerned with glamorous and overly elaborate-often capital- intensive-constructionand infrastructureprojects at the expense of routine maintenance;and citizens'perception of infrastructureprovided by the centralgovernment as free becauiseof the lack of connectednessbetween provision and consumption(Thompson, Connerley, and Wunsch 1986). Other studiespoint to additionalproblems of overcentralization:incomplete financial independence, weak local trainingand professionalcompetence, and inadequatetechnical support for operations.

7.6 AdmInistradvedecentralzaion. Administrativedecentralization involves the transfer of responsibilitiesfor financing, allocating, planning, and managing resources from central governmentto subordinateunits., There are three types of administrativedecentralization, which could coexistin one organizationfor differentfunctions:

3 Deconcentrationinvolves the shifting of functionsto local agents while retaining decisionmakingand financial control centrally. Authority, responsibility,and accountabilityare retained centrally.

* Delegation,a more comprehensiveform of decentralization,involves the transferof responsibilityfor most decisionmaking,yet the retention of some authority and accountabilityin the centralgovernment.

* DevoMioninvolves the fullertransfer of authorityfor decisionmaking,responsibility for activities,and accountabilitymechanisms to subnationalunits, usually involving the grantingof corporatestatus to those units.

7.7 Decentralizationis one of the institutionalreforms that mayhave the highestpotential for failure and distortion-for example, through conflicts between levels of responsibilityand authority,or responsibilityand control. The weaker a country'sinstitutional structures, the wider the differencein institutionalstrength between central and local agenciesand the greater the risks

I3/ Decentralizationcan take threeother forms. Political dectizaion involves the expansionof ciizen involvementin decisio , usually throughlarger representationor beneficiaryparticipation. Spatial dcentralizationinvolves the diffusion of activities across regions to build local or regional capacity. Market decentralizationinvolves theuse of market m anismsto provideinfrastructure by relyingon the revealed preferences of individual conunmers. See Rondinelli 1990; Rondinelli and Nellis 1986; and Silverman1990. - 78 - of decentralization.Paradoxically, the countrieswith agenciesmost in need of decentralizationare the least likely to successfullydecentralize. Decidingabout the most reasonabledegree and form of decentralizationrequires estimatingthe likely negativeeffects in a specific institutionaland politicalcontext.

7.8 In cases of infrastructureactivities where there is considerablehomogeneity, capital intensity,need for central politicalcommitment, or limitedlocal capacity,central control may be more appropriate. In some countries,power generationor water resourcedevelopment requires a national strategyfor coordinatingchoices and providingfor initial capacity. But, where tasks are highlyspecific and the principalconcern is with technicalefficiency or feedbackfrom clients,some form of decentralizationmay be more appropriate. Specificityin tasks and clear, fair rules and rights may not only call for decentralizedsolutions, but also offer more sustainableoutcomes. Providing clear, fair rules, rights and priorities in irrigation projects-for example, rules for allocatingwater amongfarmers, the responsibilitiesof beneficiariesfor operationand maintenance, penaltiesfor noncompliance-duringdesign can facilitatesustainable program development. At the same time, functional specializationat the different levels-for example, on planning, design, construction, operations and maintenance, and regulatory activities-is facilitated by a clear organizationof rules of operationand staffing.

7.9 Conflicts of responsibilit and authoriy. Decentralizationand public enterprise reform are closely parallel: key to both are the interactions between local or decentralized agencies-or public enterprises-and the central governmentagencies that define and monitor the decentralizedagencies' operating rules. Experience shows that a large share of the factors determiningthe performanceof these local and decentralizedagencies are controlledby the central government. One severe limitationof infrastructuredecentralization has been the gap between responsibilityand authorityor capacity.Transfers of responsibilitywithout a correspondingtransfer of both authorityand resourceshave often spelleddisaster, generating distrust at the local levelsthat receive added responsibilities. Central governments-often uncommitted to granting autonomy-maintaincontrol over key appointments,the budget,procurement, revenue generated and labor regulations,thus seriouslyhindering the capacitiesof local agenciesto operateeffectively.

7.10 Effortsto clarifylevels of responsibilityand authorityare evidentin the SenegalDebi- LampsarIrrigation project. SAED, a state corporation,was createdto plan and developirrigation in the river basin. The centralgovernment, however, required compliance with complicatedfinancial and administrativeprocedures-for example,on biddingand expenditures-andkept central accounts. Delays, inefficiency,and the deteriorationof SAED's financesresulted. To avoid these problems during implementation,the govermnentagreed to regular budget transfers to SAED and eased requirementson prior approvalof expenditures.Mutual financial and operationalresponsibilities of SAEDand the governmentwere definedmore clearly. In addition,efforts were madeto link SAED activitiesto beneficiaries:decentralization was intendedto give farmers greater autonomythrough separateproject management teams with technicaland administrativeindependence.

7.11 Anotherserious concern in a decentralizationprogram is the coordinationof legaland administrativeprocedures affecting different units. In the BrazilianSouthern States Water Supply and SewerageProject, the borrowerwas the NationalHousing Bank, the centralinstitution and main source of investmentfunds-allocating resourcesamong states, setting goals and priorities, and - 79 - supervising and assisting as necessary. Start-up delays resulted from confusion and poor coordinationin the effort to shift responsibilitiesto the state level. At all state water companies, procurementwas a problembecause of doubtsabout the applicabilityof WorldBank guidelines. The National HousingBank had not issued guidelinesbecause several state and national regulations conflictedwith the WorldBank's directives. Furthermorethe WorldBank's subprojectselection and approvalcriteria did not conformto NationalHousing Bank project preparation guidelines which the state water companieshad to follow. In fact, subprojectselection criteria varied with each of the fiveexisting World Bank loans to The NationalHousing Bank. By 1986,the NationalHousing Bank ceased to exist-its functions were subsumed in yet another central governmentalbody-and confusionabout guidelinesand authoritypersisted. 7.12 Dramaticand suddendecentralization programs have often createdmore coordination problems than they have solved. In the Upper Volta Rural Roads project in Ghana rapid decentralizationcompounded serious problems of local capacityand inefficiency.The initialintent was to strengthen the central Secondary Roads Maintenance Service, but during project implementationresponsibility for maintenanceoperations was shiftedto regionaldirectorates which couldnot accommodatethe increasedtasks. A seriesof projectsgradually shifting responsibility may often be more appropriate.

7.13 One way to analyze the potential for decentralizationis to use the criteria of excludabilityand subtractabilitynoted in Chapter5. Where both criteria are high, the logic of decentralizationobtains. Where one or both of these do not apply, alternatives to full decentralizationmay be more appropriate. For example,in water supplyand sanitationmost studies point to the efficiency of a three-tiered combination of centralizationand decentralization. Excludabilityapplies to this sector-especiallyin the urbancontext-making decentralization feasible. But the naturalmonopoly of water supplyand sanitationsuggests a need to allocateresponsibilities amongdifferent levels. At the nationallevel, an agencyshould be responsiblefor finance,long-term planning, standard-setting,coordination of training, advice and support, procurementof imported parts and technical assistance. At the regional level-for example, state or provincial governments-an agency should be responsiblefor the implementationof nationally approved standardsand regulations,supervision and support of local systems,and planningand training for local managementand technical staff. And entities at the local level-for example, municipal agenciesor communityorganizations-should be responsiblefor management,collection of fees, monitoringof use, operationand maintenanceof the local system, and local budgetaryplanning (UnitedNations 1990).

7.14 A similardivision of responsibilityand accountabilityseems appropriatefor other infrastructureactivities where naturalmonopoly conditions exist, reasonablelevels of excludability and subtractabilityare evident,responsiveness to localclients is needed,and it is possibleto enforce linkagesbetween use and payment-for example, for power transmissionand distribution.14/ In other cases where monopoly conditions are not as prevalent, but where excludability and

14/ Thisdecentralization would create difficulties for ruralelectnfication, however, which generally operates at a loss, becausemass-subsidization would not be possiblein a decentrlizedscheme. - 80 - subtractabilitydo exist, an emphasison lower levels of administrationmay be more appropriate. Such wouldbe the case for most transport, irrigation,and rural water systems.

7.15 As indicated, the discussionof these alternativeshas to be in the context of the relative strengthsof institutionsat differentlevels. Local institutionsoften are weaker than central ones, and successfuldecentralization requires a major local institutionalbuildup that might not be feasiblein the short term. The politicalcontext is also crucial: a weak local institutionis more likelyto fall under the influenceof localpressure groups than becomeresponsive to legitimatelocal needs.

7.16 Conflcts of responsibiiyand control. Failure to link responsibilityand effective controllimits the abilityof organizationsto operateon a commercialbasis, providingfew incentives for cost recovery. In Benin, the telecommunicationspublic enterprise,the OPTB,has since 1984 been subject to direct control by the Ministry of Informationand Communications.Financial transfers fromthe centralgovernment have occurredwithout effective autonomy and accountability at the enterpriselevel. The result has been limitedincentive to improveoperational efficiency, low collectionrates, and insufficientcost recovery. Insteadof being competitiveand self-financing,the OPTB has drawn on the central treasury to cover deficits. The recent introductionof greater operationalautonomy and a contractplan that establishesclear rules for dialogueand negotiationmay provide a soundbasis for improvingperformance-although problems of full financialcontrol still remain.

7.17 A key aspectof decentralizationis financialmanagement and control. Althoughfull cost recoveryis alwaysadvocated, there are practicaldifficulties. An exampleis the XiamenPort Project in China. In the 1984 port reform, the Ministry of Communicationsbegan to encourage autonomyand decentralization.Reforms included self-financing of port operations,investments to be funded by internal cash generation,and increasingcompetition among ports. Before these reforms, the Ministryof Communicationsset tariffs, allocatedcargo, providedgrant financingfor investments,covered any deficits, and collectedmost surpluses. As a result, there were few incentivesfor any port to becomeefficient. Nowonly a portionof the cargo is allocated,tariffs for unallocatedcargo are allowedto be set competitivelywithin a narrowrange, and profits are retained by ports and subjectto taxation. Decentralizationhas becomeeffective.

7.18 The failureto providefinancial and decisionmakingautonomy is particularlycritical in sectorshaving reasonablyspecific, highly technical, and self-containedinfrastructure activities. In these cases, there is little need for central govermmentalcontrol. Such is the case in telecommunications.In the JordanTelecommunications Project problems of autonomytroubled the agencyand the projectfrom the outset. Authorityfor publictelecommunications and postalservices is vestedir the Ministryof Communications.TCC is a semiautonomouspublic agency in chargeof providing and operating domestic and internationaltelecommunications services. TCC has a monopoly,but its institutionalcapacity has not kept up with the expansionand improvementof the network-and, for all practical purposes,it is operated as a central governmentagency, without autonomy. In such a case, the centralgovernment motivation might be direct appropriationof the surplusesgenerated by an agency. - 81 -

7.19 Another approach for determining appropriate levels of decentralizationin infrastructurerelates to specificfunctional activities (Silverman 1990). Retainingcentral roles for policy formulationand overallplanning, but decentralizingoperations and maintenancein specific and technicallyroutine activities,may prove mosteffective in weak states. For example,the design of the SeventhMadagascar Highway Project demonstratesconcern for the appropriatedivision of responsibilities. Formerly, all activities were centralizedin the Ministry of Transportation's headquartersin the capital-includingthe designingof road maintenanceoperations, the dispatching of brigades,and the instructingof the regions-which led to inefficient,inflexible road maintenance operationswith a lack of adaptabilityand continuity. Decentralizationof road maintenancewhile retainingcentral administrative control over policyplanning and implementation,and the monitoring of regionaiprograms through performance accounting systems, allowed for a reasonabledistribution of responsibilityand accountability.

7.20 The World Bank has focused attentionmore recently on the use of contracal arrangementsto introduce competitionsurrogates-in particular contract plans or negotiated performance agreements between owners-the central government-and managers of public enterprises(Nellis 1989). Althoughcontract plans vary greatly, somegeneralizations are apparent. First, contract plans may have been oversold as a means for resolvingfimancial and efficiency dilemmasof state-ownedenterprises. Althoughthere are some clear potentialbenefits-such as improved accountability, consistency in supplies, better personnel management and closer relationshipsbetween rate structuresand costs-performancehas not improvedsignificantly. Many of the dilemmasare becausecontract plans are not true contractualrelationships with autonomyof financialdecisionmaking and enforceabilityfor noncompliance-particularlynoncompliance by the centralgovernment. Fairness and clarityin rules and relationshipshave not been achieved-perhaps as a result of limited governmentcommitment to the transfer of power implicit in the contract. Politicalwill and the institutionalcapacity in centralgovernment agencies to superviseand monitor behaviorare needed. Still, someof thesegeneralizations might be prematureand the positiveeffects might take longer to materialize. A more systematicdialogue than in the past between public enterprisesand central governmentagencies over a number of years mightbear fruit.

Voice

7.21 Market-orientedand administrativechanges are necessary to enhance a demand orientationin infrastructureservices-but not always sufficient. Voice-participation-of clients, users, or beneficiariesis particularlyrelevant for infrastructure,given the relativelylimited range of competitivemarkets for many subsectorsand the often problematiceffects of administrative solutions. And sustainableprograms require a sense of ownership,with services responsiveto beneficiaries'or clients' felt needs. In extremecases, such as highways,in whichthere is no direct contactbetween users and suppliersand governmentsoperate the infrastructureas a localmonopoly, how to bring voice as an additionalform of demandorientation or accountabilityis a major issue. The closerthe connectednessbetween client and service,the greaterthe probabilitythat concernfor sustainableservice will be expressed. Issues of appropriatetechnological level, of community participation,of willingnessto pay all revolve aroundthis sense of ownershipand voice. In the absence of voice, beneficiaries-such as farmers in irrigationprojects-may choose to exit. In project after project, the role of clients in infrastructureoperation and maintenancehas been a difficult issue. Yet the establishmentof mechanismsfor voice has been the key to success for - 82 - several projects.l5/ The higher the level of market failure, the greater the importance of developingvoice. 7.22 Voice, Samuel Paul says, is "the degree to which [users] can influencethe final outcomeof a servicethrough some form of participationor articulationof protest/feedback"(Paul 1991a,p. 13). Voice can be expressedin a variety of forms, dependingon the country and the circumstances-for example, in user or client associations, nongovernmentalorganizations, cooperatives,political groups, or directintervention in the managementof services. As distinctfrom "exit"-an economicaction that can be takenindependently by an individual-voiceusually requires someform of collectiveaction. Individualpressure is seldomeffective, particularly for the poorest segmentsof the population. Voicehas a cost, and could be assessedby the participantsas a cost- benefitcalculation. In principle,the more competitivea market,the lessdesirable is the voiceoption relativeto the purelyeconomic one of exitingthe market. However,voice is effectiveonly if people feel that they will be heard, that they do "own" or substantiallyinfluence the operation.

7.23 The potentialfor voicedepends on severalfactors. Some are fairly obvious,such as the relative importanceof a serviceto a client, and the public's income, education, and cultural characteristics.But there are other factors, such as legal and institutionalbarriers, closelyrelated to the political environment; cultural mores (for example, availability of public hearings); informationalasymmetries, in the sensethat suppliersoften have more informationthan the clients; and the level of differentiationin the service. The less differentiateda service,the strongeris the voice option,because all users-weak and strong-have a stake in it.

7.24 Paul's analysishelps to tentativelyclassify infrastructure subsectors according to their potentialfor exit and voice(Table 7.1). Such a schemedoes not reflectthe actualcontinuum, but is operationallyuseful. Most subsectorsoffer a relativelyhigh potentialfor the use of voice. The assumptionis that in cases where the potentialfor exit also is high-dissatisfied users will exit (quadrant4). The interestingones are those sectorsin whichexit is less feasiblebut the voiceoption has a high potential (quadrant2). The implicitassumption is that urban environmentsare more suitable for exercisingvoice than are rural ones, because of the generallybetter economicand educationalconditions of the population. The mechanismthat is preferablefor exercisingthe voice optiondepends on each subsector.

7.25 The difficultcases are those for whichthe potentialsfor exit and voiceare both low (quadrant1). This difficultyis reflectedin the manyunsuccessful efforts to build up a voiceoption in interurbanhighways, waste treatment,or major irrigationworks. Paul's analysisconfirms the importanceof findingworking alternatives for developingvoice and participationin those subsectors. The developmentof voiceand exit appearsto be closelylinked to the extensionof user chargesfor these services. The more developedthese charge systems,the higher the chancesof buildingup effectivevoice and participationoptions. It is interestingthat quadrant3, the high exit and low voice alternative,is empty. If exit is an option, a weak voice is probablyirrelevant.

15/ This is not the place to review the extensiveliteraure on communityparticipation, beneficiary roles, and "listening to the people." For further information,, see Paul 1991; Korten 1980; Salmen 1987; Uphoff 1985; and Cemea 1991. - 83 -

Table 7.1 InfrastructureActivities, Potential for Exit and Voice

VOICE

Low High

rural roads urban roads rural water supply hydro power generation major water storage large-scaleirrigation waste treatment airports Low major irrigationsystems publicports hydro power generation telecommunications interurbanroads small irrigationsystems waterways power distribution thermalpower generation

EXIT 1 2

3 4

HIigh urban transport road transport rail passenger rail freight

Source: Based on analysesby SamuelPaul (1991a).

7.26 The relevanceof clientparticipation also canbe consideredin terms of the distinctions betweenpublic and privategoods. With publicgoods, the needto enhancea demandorientation by including clients, users, or beneficiariesbecomes paramount. In these cases, fear of lack of accountabilityand the need for explicitincentives for suppliersbecome most evident. In rural roads, the need for involvement,for voice to be institutionalized,is critical at all phases of planning. Plannig inputstend to comefrom severalagencies and fromthe communityserved; in construction, there is a need to keep costs down, puttng labor-intensivetechniques at a premiumin most cases; during operationsand maintenance,there is a needto developcommunity capabilities with effective technicalassistance and financialsupport from the central government(Cook 1985). One crucial - 84 - area for encouragingvoice and localparticipation is in maintenance.Many projects-especially rural roads and irrigation-are built on the assumptionthat communitymaintenance will occur, yet it rarely does. The track record of authentic communityparticipation in design, planning, and constructionappears weak (Box 7.1). Failure to incorporatebeneficiaries affects performance in several ways. For example, there is a top-down mentalitythat pervades most public works departmentswhich may require bureaucraticreorientation. Over-standardizationand program complexityin administrativeprocedures and standards,bidding procedures, or disbursementsmay discourageparticipation.

7.27 A wider use of voice is encouragedby treating project design as a learningprocess insteadof a blueprint. The learning-processapproach helps reducediscontinuities between planning and implementation,because implementation often requires creativeadaptation. For example,the Nepal SecondMahakali Irrigation Project includescontinuing support for pilot farms established under the first project, which provide training for extension workers and farmers. A project coordinationcommittee is to be formedto assistimplementation and operation. The committeewill consist of the project manager,representatives from the district level (Panchayat),and a farmer representativefrom secondarycanal operations. The major functionat this levelis communication among farmers and project staff. Farmers would participatethrough the committeein project decisions-which should help them identifywith the project, developa sense of ownership,and commit to operationsand maintenancetasks. This committeeis to continue to operate after construction. The project shouldtransform informal farmer groups into formal farmer irrigation associationsdiscussing project details, negotiating participation and cost-sharingarrangements with farmer groups, and training and preparing groups for operationsand maintenanceof completed subprojects.

7.28 Other optionsfor ensuringvoice and participationare lengtunan systemsin which, for example,each beneficiary is assignedpart of a roadnetwork; developing local contractor capabilities; self-helporganized by local leaders; and use of nongovernmentalorganizations as implementors. In Zaire's Third Water SupplyProject, the lead implementingagencies have involvedlocal villagers and nongovernmentalorganizations in the operationand maintenanceof the rural water supply. Attemptsto shift these responsibilitiesto local communitiesare to be accompaniedby trainingand technicalassistance. On-the-spot repairs and periodicmaintenance are organizedto help incorporate villagersand local mechanicsin these activities. The expectationis that this participationand sense of ownershipwill lead to higher collectionrates. In the SenegalDebi-Lampsar Irrigation Project, great care was takento incorporatefarmer groups at variousstages of projectdevelopment. SAED, the state water corporation,provided literacy and administrativetraining. Farmer groups are responsiblefor coordinatingactivities such as water delivery, crop spraying,and the maintenance of secondary and tertiary canals and dikes in their irrigation quarter. To increase farmer participationin perimeter management,a joint SAED-farmercommittee was established,with authorityon specificquestions of implementationof the agricultureprogram. The mechanismis provinginvaluable for giving farmersa voice in operationsand in helping SAED to becomemore responsive. For delicate problems such as establishingpenalties for debt delinquencies,the committee'sdecisions have been more acceptableto farmersthan in the past becausethese decisions are made by their representatives. - 85 -

0;~lQ 41-2isz__z=i;.

;~~~TltiftOR ?uqruiF =d1r~to Kvipe

A rbecetsuyb h prts Evalatio rtiu tur io~eXn Eriatopca ,a 11E Thlpbe _n _hilnan _~nta i -sinini. En ita 1 2 ectpt p.%.rwEI

.Th~rjcswr our h Thipnepram iga id ?eaiaarito r~e aud,i

'Taln,th Notb1 Itiato Xs.jt rI SadI.

~~~~~~~~~~beecaisi l tg fp*"'"'dlipRma. mhl

7.29 p Toeteffiioetivencessandiotn ofta vae~'wgouicte responsible enhancperasionga mecainismsto insr the opentrflotofinfomtion yse adfromae cliet orueto theniimpwalementingw ageny. In th Piipariueas Third iorsdrjet,scon tersumig shli$etiochre atmedforte preownetb helpedawithpterotin ma vrketing,an feedback. In the MexicaP orsRlblttooe permmaonroject, b c port~s ervice ompne allowedportusarstinithe pob rtodeelopet comhittees ~pto ~ ircourageti#stivnessson ctoa demapndtves. invrstmentprgrmand daevlpd major inetmendecisiaions ofThe cotemitteeyis4 comptose ofmbersa

seethe algovenmlcts tatgfen cuies andrgepresentativesofthepivhatpene saector aindthaacademiicsan resea commugites. Anustetoli decsonaing p rocs*w.~hnttes ainvling fey~nsra kesapovde4shasbe - 86 - establishedfor infrastructureinvestments in the region. In each case, mechanismsto enhancethe opportunityof voice have made-or are expectedto make-the differencebetween success and failure. In addition,ownership, feedback mechanisms, links betweenlocal and apex organizations, reduced project complexity,clarity about responsibilities,and accountabilityhave helped create incentivesfor providersand users.

The Bias against Maintenance

7.30 Most infrastructuresectors exhibit a systematicbias in favor of new constructionat the expenseof maintenance-and even at the expenseof efficientoperations. In highways,railways, ports, and electric power there has been surprising consistency in the creation of excess capacity-and the problems associated with continued emphasis on construction instead of rehabilitationand maintenance. There are regionaldifferences-with East Asia needingexpansion of the systems-but this bias is quite widespread.

7.31 Key factorsin the neglectof maintenanceinclude the inabilityto excludeusers who feel little responsibility,biases in donor incentivesagainst supervision and postprojectaccountability, the failure to includebeneficiaries in projectdesign, and top-downapproaches that emphasizethe use of complextechnologies and have little concern for long-termmaintenance. Other key factors include the use of govermmentforce-account agencies with little incentiveto provide day-to-day continuityin operationsand with distortedincentives for governmentstaffs, who perceive their long- term interestsand politicalgain in the more visible side of infrastructure-construction. A recent studyin LatinAmerica found an enormousbacklog of roadmaintenance and rehabilitation.Although the overall backlogwas about 45 percent in nine of twenty-twocountries or states there was a backlog of from 50 to 100 percent (Mason and Miquel 1988). The main reasons for this poor performancewere age of network, low productivityof maintenancestaff, inadequateallocation of maintenancefunds, misallocationof availablefunds, and lack of monitoringand enforcement(for example,of weight limitations).

7.32 In each case, however, what is identified are mostly proximate causes of the maintenancedilemma. I has been arguedthat one reason for this predilectionfor new construction over more labor-intensiveoperations and maintenanceis the result of the specificityand technical rationalityinherent in capital-intensiveand standardizedconstruction activities (Israel 1987; Moore 1989). In part, this focus on short-termtechnical rationality-constructionactivities in this case-versus the longer-termprocess of institutionaldevelopment (recurrent maintenance activities) is endemic. Providingshort-term solutions can conflictwith long-terminstitutional development needs. As indicated, in roads these underlyingdifficulties are exacerbatedby the institutional separationbetween the collectionof revenues-the ministryof finance-and the expenditures-the road agency.

7.33 Part of the bias towardnew constructionstems from the natureof donor fmancing. A recent study of the Philippines showed that Bank efforts continued to concentrate on road construction-even while recognizing the deplorable condition of rural road maintenance. Maintenancewas treated as a subordinatedactivity; its design was simplisticand a residualeffort; casual monitoringand supervisionprevailed. Agenciesresponsible for district road maintenance performedalternatingly among the national,provincial, and local levels, without much planning. In addition, decentrafizatonefforts occurred in a planningvacuum, resulting in obscure guiding - 87 - principles,confusing interagency arrangements, and poorly sorted-outoperational responsibilities. This bias toward constructionis being reducedat the Bank, which has already financed several highwaymaintenance operations.

7.34 There is a tendencyto view externalfinancing for new investmentsas free moneyand a resultant tendencyto allow deteriorationto occur-including in the United States. Grants for expandingurban water and sewer systemsencouraged localities in the 1970sto overbuildsystems that later requiredextensive maintenance. Maintenance,however, depended largely on the local capacityto raise revenuesthrough fees and taxes. New systemsput in place became long-term burdens for recurrent budgets, which do not competeeffectively during fiscal crises. Scarce resources,low political visibilityof maintenanceoperations, and short-termhorizons for political leadershipand bureaucratsmeant that maintenanceactivities became low priorities (Paganoand Moore 1985).

7.35 The lack of attentionto maintenancepartly reflects a lack of politicalcommitment. Not only does the high visibilityof constructionactivities give them high politicalpriority, but the weaknessof low-incomelocal-especially rural-constituencies in makingtheir demandsheard means that in the absenceof voice, consdtuentsoften exercisetheir exit option-nonmaintenance,minimal commitmentto communityoperating activities, and nonpaymentfor services. In a sense, fee paymentshould be viewednot onlyas a rationalizationof cost recoverybut as a politicalstatement.

7.36 The failure to involve local communitiesin decisionmakingcan often lead to the applicadonof inappropriatetechnologies which favor constructionover maintenance. In Sierra Leone, CAREhas had principalresponsibility for rural road programs. The originalintent was to build low-costlabor-intensive roads. Instead, capital-intensiveand relativelyexpensive construction took place. The main reason was CARE's decisionthat higher-costfeeder roads were neededto "build in" maintenancebecause of governmentweaknesses in this. At the same time, however, CARE unwittinglyundermined existing self-help traditions by decidingthat speciallocal groups should maintainthe roads. In addition, there is no local identificationwith the roads because beneficiarieswere not involvedin the construction(Carapetis and others 1984).

7.37 Thus, the managerialand fmancialproblems used to explainthe neglectof maintenance are importantbut do not go to the heart of the matter. Both the present incentivesystems in the countriesand the donor agenciesmake such a bias almostinevitable. In manycountries, government agenciesgive externalassistance little value, yet valueinternal resources highly. Donorsstill give lendinga high priorityand give localrecurrent expenditures in the countriesa low value. Until this asymmetryis solved,the bias will persist. Table A. 1 World Bank Lending for Infrastructure Number of Projects and Loan Amount (millions of U.S. dollars)

1915 1916 1971 1913 1919 19W 196 196 1963 196 19IS 196 196 1966 1969

NojzAmt lb. IaAmt 1b.aAmI WIhIAmt1b.IaAatlAmt Pb Now.IaAnt !hI&AM. M8.IautNoAmtM.LAW aAzt liolnAMn No.IzAzt!lAmtPbIo IzAAmd tion ~13473.7 12 594.0 72 769.0 17 759.9 19 7673 19 13193 24 1353.8 11 8260 19 1029.9 13 902.1 14 1081.2 17 1405.1 8 418.4 13 941.7 5 580.4 iecbwar 10 503.7 20 949.3 17 9515 19 11462 18 1354.9 25 23923 17 13230 22 2131.2 16 17682 25 2651.3 17 2250-3 26 2786.9 24 3016.9 17 2006.9 17 32829

MMMamiatm 8 199.05 642 2140.0 5221.1 2110.0 4131.033292 9395.8 257.03 1665 5121.6 3 OA 7 682.3 136 7 161.0

,utation 27 968 35 1370.9 34 10474 33 10922 38 1904A4 28 14445 28 10628 30 1614.2 30 1923.6 29 2596.9 27 2138.7 24 14982 25 1745.9 34 26425 22 1830.8 k& ~~142952 20 738.1 50 6505 21 655.7 28 1264.9 18 7953 17 5703 20 1055.3 19 1007.7 20 1526 13 8981 12 781.7 16 1217.7 19 1313.6 8 774.2 WVy 5 437,053225.3 4125354 259.2 S36.74 3370 3271.02 1030 2 45D0 5676.7 4755.042329.9 4 300 7 856 4 3320D ts ~ 5 1640 82295 9 246.6 7163 38SU86 3120 258U73330.9 7 2Y7.7 3333.86 381.7 7 34.6 5148.2 5259.6 3175.4 or ~3925 2 780 125.000.0 2 68.0 0AISU00 0.00 0.00 0.00 0.00 0.00 0.00 0 0 0-0 co IormmLs 0 0D00 000 0 01 I1500 00U0 004103 1 1250 2250821 3.84 103.9 1 2.0 0 0.03 2133 7549.2

7 SUP*A 10 145.1 11 334.6 13 300.7 16 3752 22 1018. 16 6310 11 534.6 10 441.2 19 810.9 13 640.8 13 7'M. 14 604.8 16 9695 6 5353 10 7912

a8 230 83 333 88 W9 90 3595 99 M6 92 5913862 4665 62 5406 86 559 83 6958 76 637 84 6345 80 6833 71 6162 61 6646

Bmr&Lmnvkg ~5896 663 ~ 7 8411 10011 11482 1229 13016 14477 15524 14386 16319 17674 19221 21367

*t i.ms 392% SW0% 45A96 42.7% 503% 513% 375%6 41.6% 38.% 44.8% 443% 38.9% 382% 321%6 311% Pzupztkmo

a. %MBaAdq PotV=

N WomdhaDhUIe9tSVU31WK Table A.2 World Bank Lending for Infrastructure by Region and Subsector (million US dollars)

Africa Asia EAsen LAC Total for AR 4 Regm

Sector 1979 1984 1989 1979 1984 1989 1979 1984 1989 1979 1984 1989 1979 1984 1989

Inigation 15.0 0.0 0.0 489.5 752.1 325.0 138.0 142.3 184.4 125.0 7.7 71.0 767.5 902.1 580.4

ElectricPower 10.1 219.6 138.4 722.8 1280.0 1993.5 276 208.7 415.0 346.0 843.0 736.0 1354.9 2651.3 3282.9

Teleconmmnication 20.0 0.0 103.3 90.0 8.5 12.7 0.0 128.0 0.0 0.0 30.0 45.0 110.0 166.5 161.1

Trnsporation 433.1 463.7 248.7 596.2 1322.7 1136.8 406.6 309.4 296 468.5 501.1 149.3 1904.4 2596.9 1830.8 Roads 332.3 390.2 34.7 289.5 572.0 708.0 331.6 199.4 0.0 311.5 421.0 31.5 1264.9 1582.6 774.2 Railways 80.0 66.0 7.0 206.7 500.7 150.0 0.0 110.0 138.0 96.0 0.0 37.0 382.7 676.7 332.0 Pots 113.8 7.5 0.0 0.0 250.0 112.4 75.0 0.0 63.0 0.0 76.3 0.0 88.8 333.8 175.4 Other 7.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 61.0 0.0 0.0 68.0 0.0 0.0 Sector Loans 0.0 0.0 207.0 0.0 0.0 166.4 0.0 0.0 95.0 0.0 3.8 80.8 0.0 3.8 549.2

Water & Sanitation 141.0 32.9 238.2 591 151.5 0.0 117.0 427.8 233 169.8 28.6 320.0 1018.0 640.8 791.2

Total Infiaatmucture 619.2 716.2 728.6 2389.5 2514.8 3468.0 937.6 1316.2 1128.4 1109.3 1410.4 1321.3 5055.6 6957.6 6646.3

TotadBank/IDA 1202.1 2368.3 3924.7 4207 7002.6 7832.5 2336.6 3125.8 3767.5 2264.8 3027.5 5842.1 10010.5 15524.2 21366.8 Lending

Infrastructure as 51.5% 30.2% 18.6% 56.8% 50.2% 44.3% 40.1% 42.1% 30.0% 49.0% 46.6% 22.6% 50.5% 44.8% 31.1% Poon of Total

Source: World Bank AnnualReports. *Urban Loa Not Included.

21, Appendix - 90- Page 3 of 9

Table A.3 World Bank: Ex Post Rates of Return of Infrastructure Investments

Unweighted-AverageRe-estimated Economic Rates of Return Sector 1980 1982 1984 1986 1988

Irrigation 30 22 20 14 11

Electric Power 8 13 11 11 11

Telecommunications 21 18 24 13 20

Transport 23 20 Highways 23 17 22 Railways 20 13 22 Ports 16 24 17

Water Supply & 7 7 11 8 12 Sanitation

Average of All Bank 17 14 12 16 17 Operations

3 Urban Loans Not Included. Source: World Bank, OED Annual Review of Project Performance Results.

Table A.4 World Bank: Ex Post Rates of Return of Infrastructure Investments

Weighted-Average Re-estimated Economic Rates of Return

Sector 1980 1982 1984 1986 1988

Irrigation 30 24 17 13 10 Electric Power 6 14 10 12 10 Telecommunications 21 18 24 13 19 Transport 25 18 23 19 27 Highways 26 17 24 Railways 21 13 22 Ports 16 29 17 Water Supply & 6 6 12 8 10 Sanitation l l l_l WeightedAvg. ERR 17 21 16 16 15 All Sectors l l l_l

* Urban Loans Not Included. Source: World Bank, OED Annual Review of Project Performance Results. - 91 - Appendix Page 4 of 9

Table A.5 Staff Appraisal Report Ratings for Institutional Development Programs (percentages)

infrastructure (40) Weak Adequate Good

Roads (7) 12 25 63

Ports (2) 100

Water/Sewerage(8) 25 25 50

Power/Energy(7) 14 28 57

Telecommunications(1) 100

Irrigation (7) 43 28 28

Urban/Municipal(7) 14 28 57

Total 20 28 52

Non-Infrastructure(92) 38 24 38

Source: Moore (December 1989); Gray and others (June 1990). Table A.6a

Infastucture-Supply ftrfonnance

couNm3Y kohU RALWAW3 llS Pow IOOW WA= KLWY OMA3.uGMaK

BANGLADESH Networkexpanded; and p ger ArightAcce only to 5% of Epnded. Supply inreaed; acce taffic icesand. trafficdeclined. population. declined,becaus of poplation iCrease; usban access lightly improve. BRAZIL Networkexpanded; reight taffic increased; Containertraffic Power consumption Expnded. Graly expanded;access 90% of irigation taffic inreaed;vehicle longdisae pasuenger increaed. grew; Acces 67%; urban 84%; mral acce 71%; devloped by private fleet grewcongestionin trafficdeclined. acs 89%; inadequate and urban access89%. setor. urbanare. transmissioncapacity for generaton capaity. COLOMBIA Networkexpaded; Freight*an pasenger Trafficdeclined; Acces 95% in large Acoes to water 57%-,to More than half of priority to rural roads; tffic declined. expectedto increae in cities; consumptiongrew; ewrage 41%; lower irrigationin private trffic increaed. 1985-90. exes reservecopaity. access in rural areas. sctor. Poor on-farm developmentin public irrigation. HUNGARY Networkexnded Preightand panger Accessimproved. Expated; capacitytoo Accessimproved in rural slightly;trafficincrased; trffic declined. unall for the economy; and urban aras. vehicleRed grew. large unsatisfieddemand & long waiting. INDIA Nctworkexpanded: Freight a paenger Trafficincread; Access 14%; shortaes Expnded; long waiting Supplyincreased; access One-thirdof eultivable emphasison runl roas; traffic increaed. containertaffic persist;unbalanced periods for connections; improved.Coverage in area irrigated;irrigation trffic andvehicle fleet increased.Trffic close investmentsin svcr underinvcetmcnt. rural area increaed potentialhas risen to grew. to capai in someports genration, transmision, more rapidlythan in 61.5 nmillionha. (1984), andexces capacityin and disttibution. urban ars. Supply with an cstimated93% of smallerpotu. intermittentand potatial ued. unreliabl in manycities; acut shorage in Madras.

KENYA Nework expaded; Freighttrffic declined; Ample cpacity; traffic Accessonly 5%. Access 100%in urban taffic increased. pasenger traffic generallygrw. and 15% in rural area. wmRO(CO Networkexpaded; Freight nd pasenger Traffic increased. LImitedaccs; inWlled Long waiting(up to 6 Access improvedin Expanded. traffic ad vehicle et trafficinreasd. capacityconsant after years); spply urban, declinedin rural grew. 1985. insufficiceL area. NfGERIA Ntwork epanded Freightst passegr Port capcity adequat Access 16%; many smal Acce limitedto 20% of Supplyinadequac; access aubsetisly; traffic trafficdeclined. for currenttrffic lves. urban center not yet countryares; long to watr 20% rural, increased. electrified.Longwaiting waiting(up to 10 years) 100% urban; acut for connection. spply shortage in

THALAND Netwok cxpanded; Freighttraffic increased; Traffic increased; Pr capita consumption Supplyepanded; but Accessimproved in rural Irrigaion cxpanded; trafficrmad vhc Red p nrtfic containertaffic gowth higher than veragefor demandgrowth fase areas; declinedin urbsn weak distribution incrased. decined. faaer thnconvenional countriesat similar are". systems. cargo trffic growth. incomelevels. Table A.6b Condition of Infrastrctre

COUNIY aBD RAJLWAYS 0OWrIS IEfwlCU WAIU SUPWLY llIAThI

BANGLADESH Mains edenwok Aged; ailing. Defets in distribution Generall good; mostly sysvtem;poor maintenance buitt in thc early 1960s; 45X/40%715X- of3sk= ptants. older facilitieam poor main unpavednetk: condition. 50% in poor condition (1984d) Inadtquate mansenance. BRAZIL ai p ved etwork: Tracks in poor eondition, Large part of sytem poor/fair/ood: locomotivesold; poor built in Iat 20 years, 24%I34%42% in 1987- _aintenanc. in satisfactorycondition. 88; Indequate maintenace, ubstantial nvw contuction. COLOMBIA Main pavednewk Tracks and equipmentin 3 out of 4 ports Distribution ystem has Good conditionin large Good for private Poorlfairlgood: poor condition.Poor deteriorated,beaw of deterioratedbecause of Cities,but inadequate irrigation sdsnm, 21%137%/42%il984; maintane, poortenntenance,old inadequaemintenane, maitenaceotherwie; but poor for public networkdeteriorated; ae, andsomeae andivetment. nappropriatetchDology. sytema; levelof Mint exp poor constuction. inadequate adequa now. mainteac and rehabilitation. HUNGARY Part of tbe sytem needs Track and fleet in good Modernsytem in good Technicallyobsolete; ralbiltatio; condition;track condition. 28S of exchages need faintenancnadea inocl shortage replcemeat in 986. for the increasedtraflic, of maintenance

capacity. INDIA Mi ad n k: Traks in poor condition; Obsolee system;poor Plants, transmissionand Obsoletetechnology; Water supplysystems in Maintenance obqoletcmotive power, maintenance;scere distributon y te in substandardequipment. poor sate m many cities; adeguatein 3 %tminal,ad foreign exdhug poor coniion fctvc equipment. Punjab. maugenanccinadequate. Iocomotives, constraint. iadequat mainc shortageof funds for pre parts needed for KENYAn ~~~~~~~~~~~~~~~~~~~~~~~~~~repair. KCENYA Mainpaved aetwonk Locomotivesin poor Baes, oil terminas, Rural watr supplyand poor/fair/sod: condition,due to old ag, cranes, fok-lift fleet municipal swcrge 16%/52%732%;main poor mainteance. need rehabilition, ysm need unoavednetwork rehabilitation;wi 17%117%166%in 194; maintenace.

-MOROCCO Not wck dcteriorated; Geadiy stifi c senmi Bxiitingservices Water supplysytem Conditionsvay; *OforkOCG p vedbrated; t necralysatisfactory, G aintenanckgeneralyf reasonalOlywen- satisfactory;sewerage inadequadioimta maodcin/ oedne k: rabreioanador strizationof m -t, q n a of d sff. dredgng. equiptent betwee thii _;_ccienneemdaFeqd. apexorganizationl, sa the refgonalutilities. NRIERIA Main pavd federal Poor condition;40% of Poor maintenance;lack Sevenrdeterioration of netwods generatingunib non- of spar ; faulty even relativelynew poodhirrood:: aoperationl;poor commoacontrol syem due to inadequat > 23X/15/i984); msintece. equipfen. maintenac Pe

THALAND Main paved twok: Tr dequly Maineac of cxisting Geeally satisfactorily Much of the @%h Soor ma d;lng facilies adequat; s maintained. systm needs 0 21984; systempoor. equipmentobsolete. 4500 km of eidsting haos needw rcsi -sie. .1 *~~~~~~~~~~~~~~~~~~~~~~~~~~~1

11~~fi 1!241~I PI fII *i ii l $qtlj 7 Ii}1| uA'fil 1.[I!§y[ FlilXt

., ,+ ,^ ,..~

[ji '~ U 6 - x uadd Table A.6d Infrastructure - Finanal Perfonrance

COUURY ROA1S RMLWAI'H PrOl8 rOWE I2.LECO%M WAIT SUWLY IWATNI

BANOlADESH Read uwerrevenue exceeds Heavilysubsidized; hig Avenge taiffs cover only Tariffhigh oenpard with to Tanff inres wer t red epee s; volumeof tickede tav. 70% of LRMC. Rul powa otherLDCa at similarincome inflationduring 1980-85; nuemce expendi heavily subsidized.Poor kveb. Rate of return 33%; eating defici chronic. inadequat. Bus paenger billing,heavy gt a . 50% of invesate finaned subidized. from intenul rseare. BRAZIL Road expendiur in recent Heavyannsal loses; Planing, budgetingad Aveage ariff ma 75% Weak;tariff ineas ig Po ct recowry; public ye salmos equalto read psger tffic eurceof costig are weak; tiiffs amt of law il 1989; explcit behid inflatio; eslatig iition costaper hectare user revenues.Now tucks Ie; freight affic relatedto oost. subsdy to rural coom eawting costs, becauseof are much higher ompad cm be charged, adequte to profitble; hugedebts, and and ater ply and higher costaof power, and with prih*t soemes. cmer cos. Publc buapest inadequatetariff. serage cmpanies; inktnal chemia; wy budgey companiesdo net fully fidin negptive; supos high lmhicipa recover cam. ELECTROBAS,the etoal- arrea. Ievl holdingcompany ha been well nmanged. COLOMBIA Road user cbargesexceed Substilal operatingdefwict; F cllly independeat,no Serios f l diffculies: Tariffstoo low, poor No bditin of payg ter nod expendiure, htichare Tariffincras not adeqte budgy ppot from the low toriffs,poorretum, low collectionperfornancc; chargs on publicshemes. belowrqui ; Hevy to crer infbfimncst. gNL Opetig reo declmed in l find ge , hih internalfunding of ivestment Water ches coveredIt" vehicle do aot fullypay. in 1982-84,and is epctd pow luea, heavy arm only 15%. than 30% of O&M in 1967. to declineto 50% by 1969- from publicsetor. Cost reovery in Bank 90. as_yd prct gneal V-d.

HUNGARY Freightrevenue 1ot Overal poit hlthy; rae Satis y; venagetariff FPnncl sitution cover cos; pa r tariff of re declned; but freight highr than LRMC. satisfactry; ate of reun ubsidized; susinwe revene coer cost; 30% in 1990;net inenl spendin dequste - heay p ger rvm cover cash g_nention 94% of vehices pay ess tius their 35% of cosL capitl invesments.

INDIA Read-uer m far in excs Revenues er coses;freight Sectr as a wholecoven Stte elecricy beards sa GenerallysatsfCtasy. Tarffs do not cover O&M of reding; heay vehicle revenes exceedcosts; costs; high pot charges; dole incur leas; situation costs in som.et law pay los. penper revenue am" co subsidieztionof loing Variesfrm ds to tate; colection rate; arrearmfreon only 50% of co. pt. capitl epe finaced l gWW.Certrecovery by debt; avenge tariff 60- about 10% in Punjaband 70% of LWMC;heavy Karata. sbsidy for agriculture. KENYA Read-uer rvaen exceed Chmic loes; declning rel W ening; soe charges Gal cty Tariff in t; heavy read expditue, whichare t riff e iv av tariff les than arte, especiallyfrom govt idquate.; heavy vehicls IRMC; excess subsidyfor agencies. pay s thancost, poor. MOROCCO Read-cur reven exceed Heavybudgeay support. Satlfay; crtoram Irovig; heavya e Satlefactory;no pek off- Tariff sr re adequate; Good os recovery in Bank cots; heavy veicls pay bce sebldisatlon of pest fron govt.dopts. Heavy pek pricing; heavya coecin from public prc*dct. dn os. inuring loses; taff subsidyto law-kine from gv dept. agie poor. srcture i ;equcu NIGERIA Read-uer los a revenuem Unstisftasy; incur hevy F tacodeteo_ie; poor Poorf_iec perfo ne Unsatisfactoy;tao depndnt wmeedco; mauh of to loae. colledien, & aset during 1965467;hevy les, o govt. supr ; peo billing exiudie e con-sc ion, _ntesce, chonoic law tariffs,heay aren. sad collecim. and bls an esntoac opratingpdeficits. THAIIAND Rad-ier chwg cover Operati loseic rasing, 0mity salis-, pr .l pasieic. waound. Cost roecy not denfesd in expndur; B-ki Urban higherlbor and fuei cost; profitableepentiose. Heavydebt, publc pre. Mra pot inusw kudeiqst triff increass. operatingble.

00 Appendix -96 - Page 9 of 9

Table A.7 Telecommunications Sector: Unsatisfied Demand, 1979 and 1988

(Numberin thousands) WaitingList as Percentage MainLines Waidisted of TotalDemand Country 1979 1988 1979 1988 1979 1988 Change 1 Turkey 1,092 4,921 1,535 37 58.4 0.7 -57.7 2 Korea 2,341 10,486 619 51 20.9 0.5 -20.4 3 Malaysia 325 1,248 106 69 24.5 5.3 -19.3 4 Argentina 1,797 2,748 863 789 32.4 22.3 -10.1 5 Thailand 332 1,006 194 377 36.9 27.3 -9.6 6 Kenya 70 157 40 64 36.5 28.9 -7.6 7 Costa Rica 134 257 12 14 8.3 5.1 -3.2 8 Portugal 940 1,849 127 178 11.9 8.8 -3.1 9 Chile 351 626 139 237 28.4 27.4 -0.9 10 Tunisia 100 249 43 104 30.2 29.6 -0.6 11 Uganda 21 28 19 25 47.9 47.6 -0.3 12 China 4,335 8,653 164 484 3.6 5.3 1.7 13 Sudan 44 58 15 22 25.2 27.5 2.3 14 Zambia 30 59 13 29 30. 33.2 2.8 15 Uruguay 212 345 45 93 17.5 21.3 3.8 16 Sri Lanka 54 104 22 51 28.7 33.0 4.3 17 Mexico 2,494 3,821 314 757 11.2 16.5 5.4 18 Senegal 17 29 3 8 15.9 21.3 5.4 19 Poland 1,852 2,953 926 2,189 33.3 42.6 9.2 20 Malawi 11 23 1 6 10.3 19.9 9.6 21 Morocco 154 287 69 205 30.9 41.7 10.7 22 Algeria 260 697 118 524 31.2 42.9 11.7 23 Ghana 36 40 16 30 30.6 42.3 11.7 24 India 1,868 3,488 241 1,125 11.4 24.4 13.0 25 Zimbabwe 93 118 5 32 5.3 21.3 16.0 26 Tanzania 38 66 18 76 32.3 53.5 21.1 27 Indonesia 318 829 43 549 11.9 39.9 28.0 28 Ethiopia 61 113 9 88 13.0 43.8 30.9

Source: InternationalTelecommunications Union Yearbook. Total Demand = Total Lines Connected+ Numberin the Waiting List. - 97 -

References

Alavi, Hamid. 1990. "InternationalCompetitiveness: Determinants and Indicators. " Working Paper 29. World Bank, Industry and Energy Department, Washington, D.C.

Altaf, A. and H. Jamal. 1990. "Household Willingness to Pay for Water in Rural Areas of the Punjab, Pakistan." Draft mimeo. World Bank.

Ambrose, William W., Paul R. Hennemeyer, and Jean-Paul Chapon. 1990. Privatization Telecommunications Systems: Business Opportunities in Developing Countries. International Finance Corporation Discussion Paper 10. Washington, D.C.

Aschauer, David A. 1989. "Public Investment and Productivity Growth in the Group of Seven." Economic Perspectives, 13 No. 5.

Augenblick, Mark, and B. Scott Custer, Jr. 1990. "The Build, Operate, and Transfer ("BOT") Approachto Infrastructure Services Projects in Developing Countries." Working Paper 498. World Bank, Infrastructure and Urban Development Department. Washington, DC.

Bartone, Carl R. 1989a. "EnvironmentalIssues in Urban Waste Management." Draft World Bank Discussion Paper. Infrastructure and Urban Development Department, Washington, D.C.

------1991. Private Sector Participation in Municipal Solid Waste Service: Experience in Latin America. Waste Management and Research (1991) 9, 445-509.

Barwell, L.J., G.K. Hathway, and J.D.G.F Howe. 1982. Guide to Low Cost Vehicles for Rural Communities in Developinz Countries. Vienna: UNIDO.

Baumol, William J., J.C. Panzar, and R.D. Willig. 1982. Contestable Markets and the Theory of Industr Structure. New York: Harcourt.

Bennet, James. 1981. Better Government at Half the Price: Private Production of Public Services. Ottawa, Illinois: Caroline House Publishers.

Brinkerhoff, Derick W. and Richard L. Hopkins. 1989. "Institutional Dimensions of Education Sector Lending: Guidelines for Management and Sustainability Assessment". Paper prepared for the World Bank, Populationand Human ResourcesDivision, Washington, D.C.

Brooke, James. 1991. "Exporters Fight Costs and Rules of Brazil's Ports." The New York Times June 25, 1991.

Butler, Stuart. 1985. Privatizing Federal Spending: A Strategy to Eliminate the Deficit. New York: Universe Books. - 98 -

Carapetis, S., L.Beenhakker, and J.D.F. Howe. 1984. Suply and Ouality of Rural Transport Services in Developing Countries: A Comparative Review. Working Paper 654. World Bank, Washington, D.C.

Cernea, Michael ed. 1991. Putting People First: Sociological Variables in Rural Development. New York: Oxford University Press.

Churchill, Anthony A. 1987. Rural Water Suply and Sanitation: Time for a Change. World Bank Discussion Paper 18. Washington, D.C.

Cook, Cynthia C. 1985. InstitutionalConsiderations in Rural Roads Projects. Working Paper 748. World Bank, Washington, D.C.

Dahlman, Carl J. 1990. "Electronics Development Strategy: The Role of Government.H Working Paper 37. World Bank, Industry and Energy Department, Washington, D.C.

Egan, M. C. and A. Mody. 1990. "Buyer-SellerLinks for Export Development." Working Paper 23. World Bank, Industry and Energy Department, Washington, D.C.

Gray, Cheryl, Lynn S. Khadiagala, and Richard J. Moore. 1990. "Institutional DevelopmentWork in the Bank: Review of 84 Bank Projects. " Working Paper ..... World Bank, Washington, D.C.

Hatry, Harry. 1983. A Review of Private Approaches for Delivery of Public Services. Washington, D.C.: The Urban Institute.

Heggie, Ian G. and M. Quick. 1990. "A Framework for Analyzing Financial Performance of the Transport Sector." Working Paper INU-67. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

------1991. Improving Management and Charting Policies for Roads: an Agenda for Reform. Working Paper INU-92. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

Hickling Corporation. 1990. "Infrastructure in the Developing Countries. A Review of Current Condition and Trends", report prepared for the World Bank, Infrastructure and Urban Development Department. Washington, D.C.

Hirschman, Albert. 1970. Exit. Voice. and Loyalty. Cambridge: Harvard University Press.

Howe, John. 1989. "Social and Economic Impact of Carts and Wheelbarrows in Women. Oxon: I.T. Transport Ltd.

Humplick, Frannie. 1991. "The Reliability of Systems: A Holistic Representation." Draft report prepared for the World Bank, Infrastructure and Urban Development Department. Washington, D.C. - 99 -

International Monetary Fund. Government Finance Statistics Yearbooks. Washington, D.C.: IMP.

International TelecommunicationsUnion Yearbooks.

Israel, Arturo. 1987. Institutional Development: Incentives to Performance. Baltimore: Johns Hopkins University Press.

1990.------. "The Changing Role of the State: Institutional Dimensions." Working Paper 495. World Bank, Policy, Research, and External Affairs, Washington, D.C.

Kaufmann, Daniel. 1991. "The Forgotten Rationale for Policy Reform: The Productivity of Investment Projects." Background paper for World Development Reort 1991. Washington, D.C.: World Bank.

Kirmani, Syed S. 1987. "A Review of Bank Assistanceto the ConstructionIndustry in Developing Countries." World Bank Discussion Paper INU-6, Infrastructure and Urban Development Department, Washington. D.C.

Korten, David C. 1980. "CommunityOrganization and Rural Development: A Learning Process Approach". Public Administration Review September-October.

Kudat, Ayse and Vincy Fon. 1991. "Urban Poor Women and Water Supply in Chittagong: A Case Study." Draft, World Bank.

Lewis, Maureen A. and Ted R. Miller. 1986. "Public-Private Partnership in African Urban Development," prepared for United States Agency for InternationalDevelopment (USAID). Washington, D.C.: USAID.

Lee, Kyu Sik and Alex Anas. 1990. "Impacts of Infrastructure Deficiencies on Nigerian Manufacturing Private Alternatives and Policy Options. A World Bank Research Publication, Washington, DC.

------. 1992. Impacts of infrastructure Deficiencies on Nigerian Manufacturing: Private Alternatives and Policy Options. World Bank Discussion Paper INU-98. Infrastructure and Urban Development Department, Washington, D.C.

Lemer, A. 1971. "Analysis of Systems of Constructed Facilities," PhD Dissertation, Departnent of Civil Engineering, Massachusetts Institute of Technology, Cambridge, Mass.

Martland, C.D. 1974. "Improving Railroad Reliability: A Case Study of the Southern Railway," Studies in Railroad Operations and Economics," Vol 10. Cambridge, Mass: Massachusetts Institute of Technology.

Mason, Melody and Sergio Miquel. 1988. "LAC Road Maintenance Survey." World Bank: LAC. -100-

Moore, Richard. 1987. "Contracting Out: A Study of the Honduran Experience," report prepared for USAID. Washington, D.C.: USAID.

------1989. "PrivatizationExperience in Latin America: A Review and a Case Study." A Paper presented at the Annual Meetings of the American Society for Public Administration. Miami, Florida.

Moore, Edwin A. and George Smith. 1990. "Capital Expenditures for Electric Power in the Developing Countries in the 1990." Working Paper 21. World Bank, Industry and Energy Department, Washington, D.C.

Moyer, Neil and Louis S. Thompson. 1992. "Options for Reshaping the Railway". Draft Working Paper. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

Munasinghe, Mohan. 1979. The Economics of Power System Reliability and Planning. Baltimore: Johns Hopkins University Press.

Nellis, John R. 1989. Contract Plans and Public Enterprise Performance. World Bank Discussion Paper 48. Washington, D.C.

Pagano, Michael A. and Richard J. Moore. 1985. Cities and Fiscal Choices: A New Model of Urban Public Investment. Durham, N.C.: Duke University Press.

Paul, Samuel. 1991 "Accountabilityin Public Services: Exit, Voice and Capture." Working Paper 614. World Bank, Washington, DC.

Paul, Samuel and Arturo Israel. 1991. Nongovernmental Organizations and the World Bank: Cooperation for Development. World Bank Regional and Sectoral Studies, Washington, D.C.

Peters, Hans J. 1990. "India's Growing Conflict Between Trade and Transport." Working Paper INU-69. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

------. 1991. "Trade, Manufacturing and the Service Industries: A Survey of Changing International Practice." Draft World Bank Technical Paper, Infrastructure and Urban Development Departnent, Washington, D.C.

------. 1992. "Trade and Industry Logistics in Developing Countries: A Strategy for Improving Competitivenessin Changing InternationalMarkets." Draft World Bank Policy and Research Series, Infrastructure and Urban DevelopmentDepartment. Washington, D.C.

Ramnassubban,R., Singh, B., J. Briscoe, C. Griffin, and C. Kim. 1992. "Rural Water Supply in Kerala, India: How to Emerge from a Low Level of Equilibrium Trap." Water Resources Research, forthcoming. - 101 -

Rondinelli,Dennis A. 1989. "DecentralizingPublic Services in DevelopingCountries: Issuesand Opportunities."Journal of Social. Political,and EconomicStudies, 14, 1.

Rondinelli, Dennis A. and John R. Nellis. 1986. "AssessingDecentralization Policies in DevelopingCountries: The Case for CautiousOptimism." DevelopmentPolicy Review, 4: 3-23.

Roth, Gabriel. 1987. The Private Provisionof Public Services. New York: Oxford University Press.

Salmen,Lawrence F. 1987. Listento the People: Participant-ObserverEvaluation in Development Eriects. New York: Oxford UniversityPress.

Savas, E. S. 1982. Privatizing the Public Sector: How to Shrink Government. Chatham,New Jersey: ChathamHouse Publishers. Silverman,Jerry. 1990. "Public SectorDecentralization: Economic Policy Reform and Sector InvestmentPrograms." World Bank, Africa TechnicalDepartment, Washington, DC.

Smith,Kirk R. and J. Ramakrishna. 1986. "TraditionalFuels and Health: Social,Economic and TechnicalLinks". ERG MonographNo. 98, InternationalDevelopment Research Center and UnitedNations University, Ottawa (New Delhi, Wiley Eastern, Ltd.).

Stevens, Barbara J. 1984. "ComparingPublic- and Private-SectorProductive Efficiency: An Analysisof Eight Activities." NationalProductivity Review. 3: 395-406,Autumn 1984. Triche, Thelma. 1990. "PrivateParticipation in the Deliveryof Guinea'sWater SupplyServices." Working Paper 477. World Bank, Infrastructureand Urban DevelopmentDepartment, Washington,DC.

The Economist. 1991. "India'sElectronic Industry: SkillsExporter." May 4, p. 70.

United Nations. 1989. Urban TransportDevelopment with Particular Referenceto Developing Countries(New York, UnitedNations Department of InternationalEconomic and SocialAffairs).

------. 1990. Report of the Economic and Social Council on Achievement of the InternationalDrinking Water Supplyand SanitationDecade, 19981-1990. New York: UN.

UnitedNations Development Programme. 1990. Secial Report: The UrbanChallenge. Division of Information. Vol. 2, No. 2 (June).

United NationsEconomic Commission for Europe. 1987. "The Conceptof TransportChains," Report TRANS/GE.36R. - 102 -

Uphoff, Norman. 1985. "People's Participation in Water Management: Gal Oya, Sri Lanka" in Jean Claude Garcia-Zamor, ed., Participation in DevelopmentPlanning. Boulder, Colorado: Westview Press.

Vulysteke, Charles. 1988. "Methods and Implementation, Vol I." Techniques of Privatization of State-Owned Enterprises. Washington, DC: World Bank.

Wellenius, Bjorn. 1990. "Investment in Telecommunications: Lost Opportunities." Mimeo. World Bank.

Whittington, Dale, Donald T. Lauria, and Xinming Mu. 1989. "Paying for Urban Services: A Study of Water Vending and Willingness to Pay for Water in Onitsha, Nigeria." Working Paper INU-40. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

Whittington, Dale, A. Okorafor, A. Okore, and A. McPhail. 1990. "Cost Recovery Strategy for Rural Water Delivery in Nigeria." Working Paper INU-66. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

World Bank. 1988a. Annual Review of Project Performance Results. Operations Evaluation Department. Washington, D.C.

-=----.1988b. Road Deterioration in Developing Countries: Causes and Remedies. A World Bank Policy Study. Washington, D.C.

1988c.-=------. World Development Report 1988. New York: Oxford University Press.

1989a.------. FY89 Sector Review Urban Development Operations: Reaching the Poor through Urban Operations. Report INU-OR-3. World Bank, Infrastructure and Urban Development Department, Washington, D.C.

=---. 1989b. FY90 Annual Sector Review. Energy. Washington, D.C.

------. 1991. World Development Report 1991. New York: Oxford University Press.

------. 1992a. Privatization: The Lessons of Experience. Washington, DC: World Bank.

------. 1992b. World Development Report 1992. New York: Oxford University Press.

------. 1992c. "The Bank's Role in the Electric Power Sector: Policies for Effective Institutional, Regulatory and Financial Reform." Draft World Bank Policy Paper, Industry and Energy Department. Washington, DC.

World Health Organization and UNEP. 1988. Assessment of Urban Air Oualitv, Global Environment Monitoring System, prepared in co-operation with the Monitoring and Assessment Research Center, London. 103 -

Thomson,James T., Ed. Connerley,and James S. Wunsch. 1986. "DecentralizedFinance and Management for Development: A Concept Paper," prepared for US. Agency for InternationalDevelopment. Washington, DC: USAID. Distributors of World Bank Publications

ARGENTINA The Middle East Observer ITALY PORTUGAL Carlos Hirsch, SRL 41, Sherif Street Licosa Comrnissionaria Sansori SPA Livraria Portugal Galeria Guemes Cairo Via Duca Di Calabria, 1/1 Rua Do Carmo 70-74 Florida 165,4th Floor-Ofc.453/465 Casella Postale 552 1200 Lisbon 1333 Buenos Aires FINLAND 50125Firenze Akateeminen Kiriakauppa SAUDI ARABIA, QATAR AUSTRALIA,PAPUA NEW GUINEA, P.O. Box128 JAPAN Jarir BookStore FIJI, SOLOMON ISLANDS, SF-00101Helsinki 10 Easter Book Service P.O. Box 3196 VANUATU, AND WESTERN SAMOA Hongo 3-Chome, Bunkyo-ku113 Riyadh 11471 D.A. Information Services FRANCE Tokyo 648 Whitehorse Road World Bank Publications SINGAPORE, TAIWAN, Mitcham 3132 66, avenue d'lena KENYA MYANMAR,BRUNEI Victoria 75116Paris Africa BookService(EA) Ltd. Cower Asia Pacific Pte Ltd. Quaran House, Mfangano Street Golden Wheel Building AUSTRIA GERMANY P.O. Box45245 41. KaillangPudding, #04-03 Gerold and Co. UNO-Verlag Nairobi Singapore 1334 Graben 31 Poppelsdorfer ADee55 A-1011Wien D-5300Bonn I KOREA, REPUBLICOF SOUTH AFRICA,BOTSWANA Pan Korea Book Corporation For single tiih- BANGLADESH HONG KONG, MACAO P.O. Box101, Kwangwhamun Oxford University Press Micro industries Development Asa 2000 Ltd. Seoul Southern Africa AssisSanceSociety (MIDAS) 46-48Wyndham Street P.O. Box 1141 HouseS, Road 16 Winning Centre Korean Stock BookCentre Cape Town 80O0 Dhanmondi R/Area 2nd Floor P.O. Box34 Dhaka 1209 Central Hong Kong Yeoeido For subscr7tian ordmc: Seoul International Subscription Service Branch : HUNGARY P.O. Box 41095 Pine View, 1StFloor Foundation for Market Econorny MALAYSIA Craighail 100Agrabad Commercial Area Dombovari Ut 17-19 University of Malaya Cooperative Johannesburg 2024 Chittagong 4100 H-1117Budapest Bookhop, Limited P.O. Box1127, jalan Pantai Baru SPAIN BELGIUM INDIA 59700 Kuala Lumpur Mundi-Prensa Libros, SA Jean De Lannoy Allied Publishers PvateLtd. Castello 37 Av. du Roi 202 751Mount Road MEXICO 28001 Madrid 1060 Brussels Madras- 600 002 INFOTEC Apartado Postal 22-60 Libreria IntermacionalAEDCS CANADA Branchoft,.- 14060Tlalpan, MexicoD.F. Consexlde Cent,391 LeDiffuseur 15J.N. HerediaMarg 08009Barcelona 151ABoul. de Mortagne Ballard Estate NETHERLANDS Boucherville,Quebec Bombay - 400 038 De Lindeboom/InOr-Publikaties SRI LANKA AND THE MALDIVES J4B SE6 P.O. Box202 Lake House Bookshop 13/14 Asaf All Road 7480 AE Haaksbergen P.O. Box 244 Renouf Publishing Co. New Delhi -110 002 100,Sir Chittaimpalam A. 1294 Algoma Road NEW ZEALAND Gardiner Mawatha Ottawa, Ontario 17 Chittaranjan Avenue EBSCONZ Ltd. Colombo2 KIB 3W8 Calcutta - 700 072 PrivrateMail Bag 99914 New Market SWEDEN CHXiLE layadeva Hostel Building Auckland For single lili, Invertec ICT SA 5th Main Road, Gandhinagar Fritzes Fackboksforetaget Av. Santa ECara6400 Bangalore -5 009 NIGERIA Regeringsgatan 12. Box 16356 E4jficdoiNTEC, Of. 201 University Press Limitad S-10327 Stockholm Santiago 3-5-1129Kachiguda Three Crowns BuildingJericho Cross Road Private Mail Bag 5095 Forsuicri orrE: CHINA Hyderabad -500 027 Ibadan Wennergren-WilliaznsAB China Financial& Economic P. O. Box 1305 Publishing House Prarthana Flats, 2nd Floor NORWAY S171 25 Solna 8, Da Fo Si Dong Jie Near Thakore Baug,Navrangpura Narvesen Infornation Center Beijing Ahmedabad - 380 009 BookDepartment SWITZERLAND P.O. Box6125 Etterstad For singleetitla: COLOMBIA PatialaHouse N-0602Oslo 6 Librairie Payot Infoenlace Ltda. 16-AAshok Marg Case postale 3212 ApartadoAereo34270 Lucknow-226001 PAKISTAN CH 1002 Lausanne Bogota DE. Mizza BwokAgency Central Bazaar Road 65, Shahrah-e-Quaid-e-Azam For subscotian orien: COTE D'IVO IRE 60 BajajNagar P.O. BoxNo. 729 Librairie Payot Centre d'Edition et de Diffusion Nagpur 440 010 Lahore 54000 Service des Abonnemnents Africaines (CEDA) Case postale 3312 04 B.P.541 INDONESIA PERU CH 1002 Lausanne Abidjan 04 Plateau Pt. Indira Lirnited Editorial Desarrollo SA Jalan Borobudur20 Apartado3824 THAILAND CYPRUS P.O. Box 181 Lima I Central Departmnent Store Center of Applied Research Jakarta 10320 306 Silom Road Cyprus College PHILiPPINES Bangkok 6, DiogenesStreet, Engomi IRAN International BookCanter P.O. Box 2006 Kowkab Publishers Suite 1703,Ctyland 10 TRINIDAD & TOBAGO, ANTIGUA Nicosia P.O. Box 19575-511 Condominium Tower I BARBUDA, BARBADOS, Tehran Ayala Avenue, H.V. dela DOMINICA, GRENADA, CUYANA. DENMARK Costa Extension JAMAICA, MONTSERRAT,ST. SamfundsLitteratur IRELAND Makati, Metro Manila KITTS & NEVIS, ST. LUCIA, Rosenoerns Alef 11 Government S upplies Agency ST. VINCENT & GRENADLYES DK-1970 Frederiksberg C 4-; Harcou r Road POLAND Systematica Studies Unit Dublin 2 International Publishing Service #9 Watts Street DOMINICAN REPUBLIC Ul. Piekna 31/37 Curepe Edit-ra Taller, C por A. ISRAEL 00-677Warzawa Trinidad. West Indies Restauraci6n e Isbel la Cati6ka 309 Yoz-ot Ltera-tre Ltd. Apartado de Correos 2190 Z-1 P.O, Box 56055 For suiscrotion ores: UNITED KINGDOM Santo Domingo Tel Aviv 61560 IPS Journals Microinfo Ltd. Ul. Okrezna 3 P.O. Box 3 EGYPT,ARAB REPUBUC OF 02-916 Warszawa Alton, Hampshire CU34 2PG Al Ahram England Al Gaiaa Street Cairo Recent World Bank Discussion Papers (continued)

No. 140 SocialSecurity and ProspectsforEquity in LatinAmerica. Carmelo Mesa-Lago

No. 141 China's ForeignTrade and ComparativeAdvantage: Prospects, Problems, and PolicyImplications. Alexander J. Yeats

No. 142 RestructuringSocialist Industry: Poland's Experience in 1990.Homi J. Kharas

No. 143 China:Industrial Policiesfor an Economyin Transition.Inderjit Singh

No. 144 ReformingPrices: The Experienceof China, Hungary,and Poland.Anand Rajaram

No. 145 DevelopingMongolia. Shahid Yusuf and Shahidjaved Burki

No. 146 Sino-JapaneseEconomic Relationships: Trade, DirectInvestment, and FutureStrategy. Shuichi Ono

No. 147 The Effectsof EconomicPolicies on AfricanAgriculture: From Past Harm to FutureHope. William K. Jaeger

No. 148 The SectoralFoundations of China'sDevelopment. Shahidjaved Burki and Shahid Yusuf, editors

No. 149 The ConsultingProfession in DevelopingCountries: A StrategyforDevelopment. Syed S. Kirmani and Warren C. Baum

No. 150 SuccessfulRural FinanceInstitutions. Jacob Yaron

No. 151 TransportDevelopment in SouthernChina. Clell G. Harral, editor, and Peter Cook and Edward Holland, principal contributors

No. 152 The UrbanEnvironment and PopulationRelocation. Michael M. Cemea

No. 153 FundingMechanismsfor Higher Education: Financingfor Stability, Efficiency, and Responsiveness.Douglas Albrecht and Adrian Ziderman

No. 154 Eamings, OccupationalChoice, and Mobilityin SegmentedLabor Markets of India.Shahidur R. Khandker

No. 155 ManagingExtemal Debt in DevelopingCountries: Proceedings of aJoint Seminar,Jeddah,May 1990. Thomas M. Klein, editor

No. 156 DevelopingAgricultural Extensionfor WomenFarmers. Katrine A. Saito and Daphne Spurling

No. 157 Awakening the Market: Viet Nam's EconomicTransition. D. M. Leipziger

No. 158 WagePolicy during the Transitionto a Market Economy:Poland 1990-91. Fabrizio Coricelli and Ana Revenga, editors

No. 159 InternationalTrade and the Environment.Patrick Low, editor

No. 160 InternationalMigration and IntemationalTrade. Sharon Stanton Russell and Michael S. Teitelbaum

No. 161 Civil ServiceReform and the WorldBank. Barbara Nunberg andJohn Nellis

No. 162 Rural EnterpriseDevelopment in China, 1986-90. AnthonyJ. Ody

No. 163 The Balancebetween Public and PrivateSector Activities in the Deliveryof LivestockServices. Dina L. Umali, Gershon Feder, and Comelis de Haan

No. 165 FisheriesDevelopment, Fisheries Management, and Extemalities. Richard S. Johnston

No. 166 TheBuliding Blocks of Participation:Testing Bottom-up Planning. Michael M. Cernea

No. 167 Seed System Development:The Appropriate Roles of the Privateand PublicSectors. Stevenjaffee andJitendra Srivastava

No. 168 EnvironmentalManagement and UrbanVulnerability. Alcira Kreimer and Mohan Munasinghe, editors

No. 169 CommonProperty Resources: A MissingDimension of DevelopmentStrategies. N. S. Jodha

No. 170 A ChineseProvince as a ReformExperiment: The Case of Hainan.Paul M. Cadario, Kazuko Ogawa, and Yin-Kann Wen The World Bank Headquarters European Office Tokyo Office H 1818 H Street, N.W. 66, avenue d'l&na Kokusai Building Washington, D.C. 20433, U.SA 75116 Paris, France 1-1 Marunouchi 3-chome Chiyoda-ku, Tokyo 100,Japan Telephone: (202) 477-1234 Telephone: (1) 40.69.30.00 Facsimile: (202) 477-6391 Facsimile:(1) 40.69.30.66 Telephone: (3) 3214-5001 Telex: wULI64145 WOR1DBANK Telex: 640651 Facsimile: (3) 3214-3657 RCA 248423 WORLDBK Telex: 26838 Cable Address: INTBAFRAD WASHINGTONDC

12193 URD 100 0-8213-2193-5 ISSUES INFRASTRUCTUREMG

400000005829 $7 .95

ISBN 0-8213-2193-5