World Development RepOrt 1994 T 1I - IFRASTRUCTURE FOR DEVELOPMENT Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized
268
Embed
World Bank Documentdocuments1.worldbank.org/curated/en/535851468336642118/...Kerr, Kavita Mathur, Dambisa Moyo, and Sarbajit Sinha. The work was carried out under the general direction
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
World DevelopmentRepOrt 1994
T 1I -
IFRASTRUCTURE FOR DEVELOPMENT
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
wb425962
Typewritten Text
13184
World Development Report 1994Infrastructure for Development
Published for the World BankOxford University Press
for Reconstruction and Development / THE WORLD BANK
1818 H Street, N.W., Washington, D.C. 20433, U.S.A.
Published by Oxford University Press, Inc.
200 Madison Avenue, New York, N.Y. 10016
Oxford is a registered trademark of Oxford University Press.
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise, without
the prior permission of Oxford University Press.
Manufactured in the United States of America
First printing June 1994
Photographs: Maurice Asseo, 73; Doug Barnes, 13;
Curt Carnemark, 37, 52, 89, 109.
The boundaries, colors, denominations, and other informationshown on any map in this volume do not imply on the part of the
World Bank Group any judgment on the legal status of any territory
or the endorsement or acceptance of such boundaries.
ISBN 0-19-520991-5 clot hbound
ISBN 0-19-520992-3 paperback
ISSN 0163-5085
S 'I,
Text printed on recycled paper that conforms to
the American National Standard for Permanence of Paper
for Printed Library Material, Z39.48-1984
Foreword
World Development Report 1994, the seventeenth inthis annual series, examines the link between infra-structure and development and explores ways inwhich developing countries can improve both theprovision and the quality of infrastructure services.Like the health and environment topics of the twoprevious reports in this series, infrastructure is anarea in which government policy and finance havean important role to play because of its pervasiveimpact on economic development and human wel-fare.
In recent decades, developing countries havemade substantial investments in infrastructure,achieving dramatic gains for households and pro-ducers by expanding their access to services such assafe water, sanitation, electric power, telecommuni-cations, and transport. Even more infrastructure in-vestment and expansion are needed in order to ex-tend the reach of servicesespecially to peopleliving in rural areas and to the poor.
But as this report shows, the quantity of invest-ment cannot be the exclusive focus of policy. Im-proving the quality of infrastructure service also isvital. Low operating efficiency, inadequate mainte-nance, and lack of attention to the needs of usershave all played a part in reducing the developmentimpact of infrastructure investments in the past.Both quantity and quality improvements are essen-tial to modernize and diversify production, helpcountries compete internationally, and accommo-date rapid urbanization. Future success meansbuilding on lessons learned.
The report identifies the basic cause of poor pastperformance as inadequate institutional incentivesfor improving the provision of infrastructure. To pro-mote more efficient and responsive service delivery,incentives need to be changed through commercialmanagement, competition, and user involvement.
Commercial management including financial au-tonomy, accountability, and well-defined objec-tivesfocuses providers of infrastructure serviceson increasing efficiency and meeting customer de-mand. Competition provides users with choices thatcan better meet their needs and compels providersto become more efficient and accountable. Involve-ment of users and other stakeholders in the design, op-
eration, and maintenance of infrastructure is alsokey to better performance, particularly in areaswhere competition is constrained.
Several trends are helping to improve the perfor-mance of infrastructure. First, innovation in tech-nology and in the regulatory management of mar-kets makes more diversity possible in the supply ofservices. Second, an evaluation of the role of gov-ernment is leading to a shift from direct govern-ment provision of services to increasing private sec-tor provisionand recent experience in manycountries with public-private partnerships is high-lighting new ways to increase efficiency and ex-pand services. Third, increased concern about socialand environmental sustainability has heightenedpublic interest in infrastructure design and perfor-mance.
Differences between and within infrastructuresectors, together with major variations in countryneeds and capacities, mean that the detailed designand implementation of policy reform must be tai-lored to specific cases. But there is no question thatthe overall benefits from improving infrastructureare large. Roughly $200 billion is invested in the sec-tor annually in the developing world, and the sav-ings that would accrue from better provision andperformance would be substantial. More efficient,more accessible, and less costly infrastructure ser-vices are also, of course, essential to more effectivepoverty reduction.
As in the past, World Development Report 1994 in-cludes the World Development Indicators, whichoffer selected social and economic statistics for 132countries. The Report is a study by the Bank's staff,and the judgments made herein do not necessarilyreflect the views of the Board of Directors or of thegovernments they represent.
Lewis T. PrestonPresidentThe World Bank
May 31, 1994
111
This Report has been prepared by a team led by Gregory K. Ingram and comprising John Besant-Jones,Antonio Estache, Christine Kessides, Peter Lanjouw, Ashoka Mody, and Lant Pritchett. Valuable contri-butions and advice were provided by Esra Bennathan, Koji Kashiwaya, Miguel Kiguel, Lyn Squire, andPaulo Vieira Da Cunha. Assisting the team were Ritu Basu, Leslie Citroen, Marianne Fay, ChristineKerr, Kavita Mathur, Dambisa Moyo, and Sarbajit Sinha. The work was carried out under the generaldirection of Michael Bruno.
Many others inside and outside the Bank provided helpful comments and contributions (see theBibliographical note). The International Economics Department contributed to the data appendix andwas responsible for the World Development Indicators. The production staff for the Report includedAnn Beasley, Kathryn Kline Dahi, Stephanie Gerard, Audrey K. Heiligman, Cathe Kocak, Jeffrey N.Lecksell, Nancy Levine, Deirdre T. Murphy, Hugh Nees, Kathy Rosen, Walton Rosenquist, DavidTheis, and Michael Treadway. The support staff was headed by Rhoda Blade-Charest and then RebeccaSugui and included Laitan Alli, Michael Geller, and Paul Holtz. Bruce Ross-Larson provided editorialadvice and assistance. Trinidad S. Angeles served as administrative assistant. Anthony Rowley wasthe principal editor.
Preparation of the Report was greatly aided by background papers and by contributions from par-ticipants in consultation meetings, both of which were supported in part by the Policy and Human Re-sources Development Fund financed by the Japanese government. The names of participants in theconsultation meetings are listed in the Bibliographical note.
iv
Definitions and data notes ix
Overview I
1 Infrastructure: achievements, challenges, and opportunities 13
Infrastructure's impact on development 14
Origins of the public sector role in infrastructure 22
The record of performance 25Diagnosis and directions for change 32
2 Running public entities on commercial principles 37
Lessons of success and failure 38Corporatization 40Focused goals and accountable management 41
Pricing for financial independence 47The need for a political commitment to reform 51
3 Using markets in infrastructure provision 52
Unbundling services for competition 53
The range of market alternatives 55
Paths to market provision 64
Conclusion 71
4 Beyond markets in infrastructure 73
Decentralization and participation: involving users 73
Improving budgetary allocations 79
Subsidies and transfers to the poor 80Addressing externalities 82
Elements of infrastructure planning 84
Conclusion 88
5 Financing needed investments 89
Old ways of financing infrastructureand new 89
The spread of project financing: achievements and lessons 93
Institutions and instruments for resource mobilization 102
Prospects 108
Infrastructure's role and record 2
Diagnosing the causes of poor performance 6
New opportunities and initiatives 7
Options for the future 8
Implementing reform 10
Potential payoffs from reform 11
Contents
V
6 Setting priorities and implementing reform 109
Choosing among alternatives: institutional options and country conditions 109Sectoral agendas for reform 114
Payoffs from reform 121
Bibliographical note 123
Appendix: Infrastructure data 139
World Development Indicators 151
Boxes
1 What is infrastructure? 2
2 Main messages of World Development Report 1994 2
1.1 Returns on infrastructure investment- too good to be true? 151.2 The importance of infrastructure to economic development: an example from China 181.3 Throwing infrastructure overboard 191.4 Infrastructure's direct and indirect effects in rural India 211.5 Households' responses to unreliability of water supply 301.6 Public failures raise private costs 301.7 Infrastructure activities threaten the Black Sea environment 342.1 The right way to run a public utility: a look at Botswana's Water Utility Corporation 382.2 Severance pay eases layoffs in Argentina Railways 402.3 It took ten years to corporatize Indonesia's main ports 412.4 What's special about Korean performance agreements? 432.5 Management contracting in Guinea-Bissau-a success story? 442.6 AGETIPs: involving the private sector in Africa's urban infrastructure 452.7 Designing tariffs to achieve financial autonomy while addressing multiple goals 482.8 Can earmarking improve highways? 503.1 Divided they stand: unbundling railway services 553.2 Regulatory cycles in the United States 573.3 Tailoring concessions to sectors and government objectives 623.4 Success of a lease contract-Guinea's water supply 623.5 Côte d'Ivoire's experience with a concession for water supply 633.6 Telecom privatization: the case of Venezuela 663.7 The evolution of private power in the Philippines 673.8 Regulation and privatization: which comes first? The case of Malaysia 683.9 Development of regulatory capacity in Argentina 693.10 Jamaica's regulatory roller coaster for telecommunications 693.11 Participation as regulation: an initial step in Bangalore 714.1 Mexico's inunicipios help themselves 744.2 Applying innovative approaches to water and sanitation planning 774.3 Power in Purang and roads in Ethiopia 78
4.4 Centralized and decentralized infrastructure planning in Malaysia 794.5 Assessing a project's reach: water in Kathmandu 824.6 Women can benefit from infrastructure, but success lies in the details 854.7 Donor coordination in infrastructure: the African experience 864.8 The World Bank's experience with project evaluation 864.9 Incorporating environmental concerns early in planning: some recent lessons from Sri Lanka 874.10 Population resettlement and project design: Thailand's Pak Mun hydropower project 885.1 Is there a free lunch?-limits to government finance 915.2 Tapping international capital markets 93
vi
5.3 Warning signs from the nineteenth century 94
5.4 A successful first step in Guatemala 975.5 Land grants and eminent domain 98
5.6 Mexico's toll roads: a big push that faltered 995.7 Leveraging through guarantees in Thailand 1005.8 Look before you leap: limiting government exposure to contingent liabilities 1015.9 Successful municipal credit in Colombia 103
6.1 Ingredients of good performance under alternative institutional forms Till
Text figures
1 As a country's income grows, the amount of infrastructure increases 3
2 The composition of infrastructure changes with country income level 43 Infrastructure has expanded tremendously in recent decades 5
4 Urban populations have better access to safe drinking water than rural populations 6
5 Annual gains from eliminating mispricing and inefficiency are large relative to investment 11
1.1 Public infrastructure investment is a large fraction of both total and public investment in developingcountries 14
1.2 Per capita availability of major infrastructure is closely related to income levels 16
1.3 Infrastructure services differ substantially in their economic characteristics across sectors, within sec-tors, and between technologies 25
1.4 The rural-urban gap in access to power and water in developing countries narrowed over the pastdecade 27
1.5 Efficient and effective delivery of infrastructure services does not always accompany increasedavailability 28
1.6 There is very high unmet demand for telephone connection 31
1.7 Walking is a transport mode used frequently by the poor 33
2.1 The adoption of commercial principles in 1984 allowed Togo's water utility to increase coverage andproduction. . . but a performance agreement in 1989 was needed to improve financial outcomes 46
2.2 Costs are seldom fully recovered in infrastructure 473.1 Unbundling activities increases the options for competition and private sector involvement 56
3.2 Leases and concessions in infrastructure sectors are common, even in low-income countries 603.3 Privatizations in telecommunications can lead to large gains 65
4.1 Countries with decentralized road maintenance have better roads 75
4.2 Participation increases water project effectiveness by improving maintenance 76
4.3 In water and sewerage, the better-off often get more subsidies than the poor 804.4 Even in some formerly centrally planned economies, infrastructure subsidies went mainly to the
better-off 81
5.1 Large shares of official development finance for infrastructure go to energy and transport 90
5.2 Official lending for infrastructure has increased, but publicly guaranteed private loans havefallen 90
5.3 Infrastructure is a large share of privatization proceeds; foreign financing of infrastructureprivatization is important in Latin America 105
5.4 Infrastructure equities are contributing to the growth of Argentina's capital market 106
5.5 Infrastructure equities have outperformed other stocks by a huge margin 107
5.6 Options for financing increase with administrative capacity and maturity of domestic capitalmarkets 108
Text tables
1.1 Value added of infrastructure services by country groups 13
1.2 Average economic rates of return on World Bank- supported projects, 1974-92 17
1.3 Expansion of infrastructure coverage in low-, middle-, and high-income economies,recent decades 26
1.4 Percentage of the poorest and richest population quintiles with access to infrastructure, variouscountries 32
2.1 Common management problems in public sector infrastructure entities, 1980-92 39
vii
3.1 Contractual arrangements for private water supply 61
3.2 Value of infrastructure privatizations in developing countries, 1988-92 645.1 Portfolio and foreign direct investment in developing countries, 1990-93 925.2 Infrastructure project financing for projects funded and in the pipeline, October 1993 955.3 Project financing of funded infrastructure projects, by sector, October 1993 956.1 The main institutional options for provision of infrastructure 1106.2 Country infrastructure coverage and performance 1126.3 Feasibility of private sector delivery varies by infrastructure components 1156.4 Options in telecommunications and energy 1166.5 Options in water and waste 1186.6 Options in transport 1196.7 Fiscal burden of underpriced infrastructure 1216.8 Savings from increased efficiency 122
Appendix tablesA.1 Physical measures of infrastructure provision 140A.2 Access to drinking water and sanitation 146A.3 IBRD and IDA commitments 149A.4 Official development finance commitments 149
Definitions and data notes
Selected terms used in this Report
BOT (buildoperatetransfer). A form of concessionusually referring to totally new projects. Typically ina BOT, a private party (or consortium) agrees to fi-nance, construct, operate, and maintain a facility fora specified period and then transfer the facility to agovernment or other public authority Variations in-clude BOOT (buildownoperatetransfer) andBOO (buildown-operate); in the last case, the con-tract accords the right to construct and operate thefacility, but the facility is not transferred back to thepublic sector.
Concession. An arrangement whereby a privateparty leases assets for service provision from a pub-lic authority for an extended period and has respon-sibility for financing specified new fixed invest-ments during the period; these new assets thenrevert to the public sector at expiration of thecontract.
Con testability. The vulnerability of an activity tocompetition from new entrants in a market. The keycriterion for contestability is that costs of entering amarket be recoverable (e.g., through a sale of assets).
Corporatization. The transformation of a state-owned enterprise or agency into a legal entity sub-ject to company law, including formal separation ofownership and management responsibilities, for ex-ample, through a board of directors or other body
Economies of scale. A characteristic of a productiontechnology whereby unit costs decline with increas-ing output over a large range. Economies of scaleare a major source of natural monopoly.
Leasing. An arrangement whereby a private party(lessee) contracts with a public authority for theright to operate a facility (and the right to a flow ofrevenues from providing a specific service) for aspecified period of time. The facility continues to beowned by the public authority. Unlike in a conces-sion, the lessee does not have responsibility for in-
vestments in fixed assets. (A lease may sometimesbe called a "service concession," and a BOT is some-times called a "public works concession.")
Management contract. An arrangement whereby aprivate contractor assumes responsibility for a fullrange of operation and maintenance functions, withauthority to make day-to-day management deci-sions. Compensation may be based partially onservices rendered (as for service contracts) andpartially on performance achieved (as in profitsharing).
Watural monopoly. An economic activity that ismost efficiently carried out by a single producer.
Parastatal (also public or state enterprise). An orga-nization engaged in productive activity that isowned and controlled in majority by the state.
Performance agreement. An agreement negotiatedbetween the government and the public manager ofa public utility or a government department. It usu-ally defines explicit commercial goals (such as de-gree of cost recovery) and may define noncommer-cial goals (such as increases of services to poorneighborhoods). Its main purpose is to increase theaccountability of both the government and the pub-lic managers by sharpening and clarifying the goalsof public entities.
Service contract (or contracting out). An arrange-ment with the private sector to perform particularoperating or maintenance functions for a fixed pe-riod and for specified compensation.
Country groups
For operational and analytical purposes the WorldBank's main criterion for classifying economies isgross national product (GNP) per capita. Everyeconomy is classified as low-income, middle-in-come (subdivided into lower-middle and upper-middle), or high-income. Other analytical groups,based on geographic regions, exports, and levels ofexternal debt, are also used.
ix
Because of changes in GNP per capita, the coun-try composition of each income group may changefrom one edition of World Development Report to thenext. Once the classification is fixed for any edition,all the historical and projected data presented arebased on the same country grouping. The countrygroups used in this edition are defined as follows.
Low-income economies are those with a GNP percapita of $675 or less in 1992.
Middle-income economies are those with a GNPper capita of more than $675 but less than $8,356 in1992. A further division, at GNP per capita of $2,695in 1992, is made between lower-middle-income andupper-middle-income economies.
High-income economies are those with a GNPper capita of $8,356 or more in 1992.
World comprises all economies, includingeconomies with sparse data and those with less thanI million population; these are not shown sepa-rately in the main tables but are presented in TableIa in the technical notes to the World DevelopmentIndicators (WDI).
Low-income and middle-income economies aresometimes referred to as developing economies.The use of the term is convenient; it is not intendedto imply that all economies in the group are experi-encing similar development or that other economieshave reached a preferred or final stage of develop-ment. Classification by income does not necessarilyreflect development status. (In the WDI, high-income economies classified as developing by theUnited Nations or regarded as developing by theirauthorities are identified by the symbol ±). The useof the term "countries" to refer to economies impliesno judgment by the Bank about the legal or otherstatus of a territory.
For some analytical purposes, other overlappingclassifications that are based predominantly on ex-ports or external debt are used, in addition to in-comes or geographic groups. Countries with sparsedata and those with less than 1 million population,although not shown separately, are included ingroup aggregates.
The table "Classification of economies" at theend of the WDI lists countries by the WDI's income,regional, and analytical classifications.
Data notes
Billion is 1,000 million.Trillion is 1,000 billion.Tons are metric tons, equal to 1,000 kilograms,
or 2,204.6 pounds.
x
Dollars are current U.S. dollars unless other-wise specified.
Growth rates are based on constant price dataand, unless otherwise noted, have been computedwith the use of the least-squares method. See thetechnical notes to the WDI for details of thismethod.
The symbol / in dates, as in "1990/91," meansthat the period of time may be less than two yearsbut straddles two calendar years and refers to a cropyear, a survey year, or a fiscal year.
The symbol .. in tables means not available.The symbol - in tables means not applicable.
(In the WDI, a blank is used to mean not applicable.)The number 0 or 0.0 in tables and figures means
zero or a quantity less than half the unit shown andnot known more precisely.
The cutoff date for all data in the World Develop-ment Indicators is April 29, 1994.
Historical data in this Report may differ fromthose in previous editions because of continual up-dating as better data become available, because of achange to a new base year for constant price data, orbecause of changes in country composition in in-come and analytical groups.
Economic and demographic terms are defined inthe technical notes to the WDI.
Acronyms and initials
AGETIPs Agences d'Exécution des Travaux d'In-térêt Public
BOT Build-operate-transferDAC Development Assistance CommitteeGDP Gross domestic productGNP Gross national productIPP Independent power projectNGO Nongovernmental organizationNTC National Telecommunications Commis-
sionOECD Organization for Economic Cooperation
and Development (Australia, Austria,Belgium, Canada, Denmark, Finland,France, Germany, Greece, Iceland, Ire-land, Italy, Japan, Luxembourg, Nether-lands, New Zealand, Norway, Portugal,Spain, Sweden, Switzerland, Turkey,United Kingdom, and United States)
USAID United States Agency for InternationalDevelopment
Overview
Developing countries invest $200 billion a year innew infrastructure 4 percent of their national out-put and a fifth of their total investment. The resulthas been a dramatic increase in infrastructure ser-vicesfor transport, power, water, sanitation, tele-communications, and irrigation. During the past fif-teen years, the share of households with access toclean water has increased by half, and power pro-duction and telephone lines per capita have dou-bled. Such increases do much to raise productivityand improve living standards.
These accomplishments are no reason for com-placency, however. One billion people in the devel-oping world still lack access to clean waterandnearly 2 billion lack adequate sanitation. In ruralareas especially, women and children often spendlong hours fetching water. Already-inadequatetransport networks are deteriorating rapidly inmany countries. Electric power has yet to reach 2billion people, and in many countries unreliablepower constrains output. The demands for telecom-munications to modernize production and enhanceinternational competitiveness far outstrip existingcapacity. On top of all this, population growth andurbanization are increasing the demand for infra-structure.
Coping with infrastructure's future challengesinvolves much more than a simple numbers gameof drawing up inventories of infrastructure stocksand plotting needed investments on the basis ofpast patterns. It involves tackling inefficiency andwasteboth in investment and in delivering ser-vicesand responding more effectively to userdemand. On average, 40 percent of the power-gen-erating capacity in developing countries is unavail-able for production, twice the rate in the best-
performing power sectors in low-, middle-, andhigh-income countries. Half the labor in Africanand Latin American railways is estimated to be re-dundant. And in Africa and elsewhere, costly in-vestments in road construction have been wastedfor lack of maintenance.
This poor performance provides strong reasonsfor doing things differentlyin more effective, lesswasteful ways. In short, the concern needs tobroaden from increasing the quantity of infrastruc-hire stocks to improving the quality of infrastructureservices. Fortunately, the time is ripe for change. Inrecent years, there has been a revolution in thinkingabout who should be responsible for providing in-frastructure stocks and services, and how these ser-vices should be delivered to the user.
Against this background, World Development Re-port 1994 considers new ways of meeting publicneeds for services from infrastructure (as defined inBox 1)ways that are more efficient, more user-responsive, more environment-friendly, and moreresourceful in using both the public and private sec-tors. The report reaches two broad conclusions:
Because past investments in infrastructurehave not had the development impact expected, it isessential to improve the effectiveness of investmentsand the efficiency of service provision.
Innovations in the means of delivering infra-structure servicesalong with new technolo-giespoint to solutions that can improve perfor-mance.
This Report marshals evidence in support ofthese conclusionsidentifying causes of failureand examining alternative approaches. The mainmessages and policy options are summarized inBox 2.
1
Box 1 What is infrastructure?
This Report focuses on economic infrastructure andincludes services from:
Public utilitiespower, telecommunications,piped water supply, sanitation and sewerage, solidwaste collection and disposal, and piped gas.
Public worksroads and major dam andcanal works for irrigation and drainage.
Other transport sectorsurban and interur-ban railways, urban transport, ports and water-ways, and airports.
Infrastructure is an umbrella term for many ac-tivities referred to as "social overhead capital" bysuch development economists as Paul Rosenstein-Rodan, Ragnar Nurkse, and Albert Hirschman.Neither term is precisely defined, but both encom-pass activities that share technical features (such aseconomies of scale) and economic features (such asspillovers from users to nonusers).
Infrastructure's role and record
The adequacy of infrastructure helps determine onecountry's success and another's failurein diversi-fying production, expanding trade, coping withpopulation growth, reducing poverty, or improvingenvironmental conditions. Good infrastructureraises productivity and lowers production costs, butit has to expand fast enough to accommodategrowth. The precise linkages between infrastructureand development are still open to debate. However,infrastructure capacity grows step for step with eco-nomic outputa 1 percent increase in the stock ofinfrastructure is associated with a 1 percent increasein gross domestic product (GDP) across all countries(Figure 1). And as countries develop, infrastructuremust adapt to support changing patterns of de-mand, as the shares of power, roads, and telecom-munications in the total stock of infrastructure in-
Box 2 Main messages of World Development Report 1994
Infrastructure can deliver major benefits in economic growth,poverty alleviation, and environmental sustainabilitybutonly when it provides services that respond to effective demandand does so efficiently. Service is the goal and the measureof development in infrastructure. Major investmentshave been made in infrastructure stocks, but in too manydeveloping countries these assets are not generating thequantity or the quality of services demanded. The costsof this wastein forgone economic growth and lost op-portunities for poverty reduction and environmental im-provementare high and unacceptable.
The causes of past poor performance, and the source of im-proved performance, lie in the incentives facing providers. Toensure efficient, responsive delivery of infrastructureservices, incentives need to be changed through the ap-plication of three instrumentscommercial manage-ment, competition, and stakeholder involvement. Theroles of government and the private sector must betransformed as well. Technological innovation and ex-periments with alternative ways of providing infrastruc-ture indicate the following principles for reform:
Manage infrastructure like a business, not a bureau-cracy. The provision of infrastructure needs to be con-ceived and run as a service industry that responds tocustomer demand. Poor performers typically have a con-fusion of objectives, little financial autonomy or financialdiscipline, and no "bottom line" measured by customersatisfaction. The high willingness to pay for most infra-structure services, even by the poor, provides greater op-portunity for user charges. Private sector involvement inmanagement, financing, or ownership will in most casesbe needed to ensure a commercial orientation in infra-structure.
Introduce competitiondirectly if feasible, indirectly ifnot. Competition gives consumers choices for better
meeting their demands and puts pressure on suppliersto be efficient and accountable to users. Competition canbe introduced directly, by liberalizing entry into activi-ties that have no technological barriers, and indirectly,through competitive bidding for the right to provide ex-clusive service where natural monopoly conditions existand by liberalizing the supply of service substitutes.
Give users and other stakeholders a strong voice andreal responsibility. Where infrastructure activities involveimportant external effects, for good or bad, or wheremarket discipline is insufficient to ensure accountabilityto users and other affected groups, governments need toaddress their concerns through other means. Users andother stakeholders should be represented in the plan-ning and regulation of infrastructure services, and insome cases they should take major initiatives in design,operation, and financing.
Public-private partnerships in financing have promise.Private sector involvement in the financing of new ca-pacity is growing. The lessons of this experience are thatgovernments should start with simpler projects and gainexperience, investors' returns should be linked to projectperformance, and any government guarantees neededshould be carefully scrutinized.
Governments will have a continuing, if changed, role ininfrastructure. In addition to taking steps to improve theperformance of infrastructure provision under their di-rect control, governments are responsible for creatingpolicy and regulatory frameworks that safeguard the in-terests of the poor, improve environmental conditions,and coordinate cross-sectoral interactionswhether ser-vices are produced by public or private providers. Gov-ernments also are responsible for developing legal andregulatory frameworks to support private involvementin the provision of infrastructure services.
2
crease relative to those of such basic services aswater and irrigation (Figure 2).
The kind of infrastructure put in place also deter-mines whether growth does all that it can to reducepoverty. Most of the poor are in rural areas, and thegrowth of farm productivity and nonfarm rural em-ployment is linked closely to infrastructure provi-sion. An important ingredient in China's successwith rural enterprise has been a minimum packageof transport, telecommunications, and power at thevillage level. Rural enterprises in China now em-pioy more than 100 million people (18 percent of the
labor force) and produce more than a third of na-tional output.
Infrastructure services that help the poor alsocontribute to environmental sustainability. Cleanwater and sanitation, nonpolluting sources ofpower, safe disposal of solid waste, and better man-agement of traffic in urban areas provide environ-mental benefits for all income groups. The urbanpoor often benefit most directly from good infra-structure services because the poor are concentratedin settlements subject to unsanitary conditions, haz-ardous emissions, and accident risks. And in many
Figure 1 As a country's income grows, the amount of infrastructure increases.
Infrastructure stocks per capita, 1990 (1985 prices)
9,000
1,200
160
20
Chad
SZambia S S
S SS
Mali S
.
S Rwanda
Paraguay
IBangladesh
S
El SalvadorS
Guatemala
Spain
VenezuelaMexico
SMauritius
NorwayS
'4'IJapan
Australia
150 400 1,000 3,000
GDP per capita, 1990 (PPP dollars)'
Note: Axes are logarithmic; infrastructure includes roads, rail, power, irrigation, and telephones.a. Purchasing power parity (PPP) dollars are valued in Summers and Heston 1985 international prices.Source: Ingram and Fay, background paper; Summers and Heston 1991.
3
Middle East and North Africa Sub-Saharan AfricaLatin America and the Caribbean South Asia
East Asia and Pacific Europe and Central Asia
8,000 22,000
rapidly growing cities, infrastructure expansion islagging behind population growth, causing localenvironments to deteriorate.
In developing countries, governments own, op-erate, and finance nearly all infrastructure, primar-ily because its production characteristics and thepublic interest involved were thought to requiremonopoly and hence governmentprovision.The record of success and failure in infrastructure islargely a story of government's performance.
Infrastructure's past growth has in some respectsbeen spectacular. The percentage of households andbusinesses served has increased dramatically, espe-cially in telephones and power (Figure 3). The percapita provision of infrastructure services has in-creased in all regions; the greatest improvementshave been in East Asia and the smallest in Sub-Saharan Africa, reflecting the strong associationbetween economic growth and infrastructure.
In other important respects, however, the perfor-mance has been disappointing. Infrastructure in-
4
Figure 2 The composition of infrastructurechanges with country income level.
Source: Ingram and Fay, background paper.
vestments have often been misallocated too muchto new investment, not enough to maintenance; toomuch to low-priority projects, not enough to essen-tial services. The delivery of services has been ham-pered by technical inefficiency and outright waste.And too few investment and delivery decisionshave been attentive to meeting the varied demandsof different user groups, or to the consequences forthe environment.
Inadequate maintenance has been an almost uni-versal (and costly) failure of infrastructure pro-viders in developing countries. For example, a well-maintained paved road surface should last for ten tofifteen years before needing resurfacing, but lack ofmaintenance can lead to severe deterioration in halfthat time. The rates of return from World Bankassisted road maintenance projects are nearly twicethose of road construction projects. Timely mainte-nance expenditures of $12 billion would have savedroad reconstruction costs of $45 billion in Africa inthe past decade. On average, inadequate mainte-nance means that power systems in developingcountries have only 60 percent of their generatingcapacity available at a given time, whereas bestpractice would achieve levels over 80 percent. Andit means that water supply systems deliver an aver-age of 70 percent of their output to users, comparedwith best-practice delivery rates of 85 percent. Poormaintenance can also reduce service quality and in-crease the costs for users, some of whom installbackup generators or water storage tanks and pri-vate wells.
Failings in maintenance are often compoundedby ill-advised spending cuts. Curbing capital spend-ing is justified during periods of budgetary auster-ity, but reducing maintenance spending is a falseeconomy. Such cuts have to be compensated forlater by much larger expenditures on rehabilitationor replacement. Because inadequate maintenanceshortens the useful life of infrastructure facilitiesand reduces the capacity available to provide ser-vices, more has to be invested to produce those ser-vices. Donor objectives (such as seeking contractsfor capital-goods supply or consultancy services)may also play a part in the preference for new in-vestment over maintenance. In many low-incomecountries, donor financing underwrites nearly halfof all public investment in infrastructure.
Project investments misa!located by many countrieshave created inappropriate infrastructure or pro-vided services at the wrong standard. Demands ofusers for services of varying quality and affordabil-ity go unmet even when users are willing and ableto pay for them. Low-income communities are not
Low- Middle- High-income income income
1 Sanitation Li Telecom
Water Roads
Railways U PowerLi Irrigation
Figure 3 Infrastructure has expanded tremendously in recent decades.
Power(production per capita)
Telecom(main lines per capita)
Sanitation(share of population with access)
Paved roads(kilometers per capita)
Water(Share of population with access)
Middle-income countries
Average decade rate of growth (percent)0 10 20 30
/
/
/
Liii Low-income countries
Note: Based on telecom, sanitation, and water data for 1975-90, and road and power data for 1960-90.Source: Appendix tables A.1 and A.2.
40 50 60
offered suitable transport and sanitation optionsthat provide services they value and can afford. Pre-mature investments in capacityespecially inwater supply, railways, power, ports, and irriga-tionhave often absorbed resources that could oth-erwise have been devoted to maintenance, modern-ization, or improvements in service quality. Becausemany infrastructure investments are immobile andserve local markets, excess capacity cannot serveother marketsand it remains underused. In somecases, large public projects have been overambi-tious, placing a costly burden on the economy.
Waste and inefficiency claim a large share of re-sources that could be used for delivering infrastruc-ture services. A review of power utilities in fifty-onedeveloping countries showed that technical effi-ciency has actually declined over the past twentyyears. Older power plants consume between 18 and44 percent more fuel per kilowatt-hour than doplants in power systems operating at best-practice
levelsand have transmission and distributionlosses two to four times greater. Port facilities in de-veloping countries, on average, move cargo fromship to shore at only 40 percent the speed of themost efficient ports. Labor misallocations presentanother source of inefficiency. Overstaffing is far toocommon in many activities, especially railways,while others, such as road maintenance, warrantgreater use of labor-based methods.
These failings in investment and operating effi-ciency are not compensated for by success in ad-dressing poverty or environmental concernsforhere, too, the infrastructure record is weak. Badlydesigned and managed infrastructure is a majorsource of environmental degradation in both urbanand rural areas. The poor often consume fewer in-frastructure services and pay higher prices than dothe nonpoor. For example, households obtainingwater from vendors pay much more than thosehouseholds connected to water systems. In most
5
Figure 4 Urban populations have betteraccess to safe drinking water thanrural populations.
Percentage of ruralpopulation served100
80
60
40
20
20 40 60
Percentage of urbanpopulation served
Middle East and North AfricaLatin America and the CaribbeanEast Asia and Pacific
Sub-Saharan AfricaSouth AsiaEurope and Central Asia
Source: Appendix table A2.
countries, rural areas receive fewer infrastructureservices than do urban areas (with the obvious ex-ception of irrigation), even in such essential servicesas drinking water (Figure 4). But countries thathave made concerted efforts to provide infrastruc-ture in rural areasfor example, Indonesia andMalaysiahave succeeded in reducing povertydramatically.
Given this mixed performance, improvements ininvestment and operation are required as a matterof urgency. In addition, the demands on infrastruc-ture are growing. More competitive global trade re-quires more reliable and sophisticated transport,power, and telecommunications. Governments fac-ing increased fiscal stringency can no longer sustainopen-ended financing of infrastructure. And soci-eties today hold infrastructure to higher environ-mental standards, as evidenced by sections ofAgenda 21, the primary policy document agreed toby countries at the 1992 United Nations Conferenceon Environment and Development.6
Diagnosing the causes of poor performance
The problems of insufficient maintenance, misallo-cated investment, unresponsiveness to users, andtechnical inefficiencies present daunting challengesfor future reforms - challenges compounded bynew demands and constrained resources. The solu-tions lie in the successes and failures of policy andin the lessons from recent policy experiments.
There is great variation both within and acrosscountries in the efficiency of providing infrastruc-ture services. Moreover, good performance by acountry in one infrastructure sector is not necessar-ily associated with good performance in other sec-tors. Some developing countriesand not alwaysthe richer ones--perform at high levels. Côted'Ivoire meets the 85 percent best-practice standardin water supply, while in Manila only about 50 per-cent of treated water is delivered to customers. Inrailways, the availability of locomotives is highwhere maintenance is good: at any given time, Indiahas 90 percent of its locomotives available. Avail-ability is low where maintenance is neglected: 50percent in Romania and 35 percent in Colombia,compared with a developing country average ofabout 70 percent. For telephones, call completionrates are 99 percent in the best-performing coun-tries, 70 percent in the average developing country,and far lower in some. These findings indicate thatthe performance of infrastructure derives not fromgeneral conditions of economic growth and devel-opment but from the institutional environment,which often varies across sectors within individualcountries.
Therefore, to understand what accounts for goodperformance - and badrequires understandingthe institutional arrangements for providing infra-structure services and the incentives governing theirdelivery. This Report identifies three reasons forpoor performance.
First, the delivery of infrastructure services usu-ally takes place in a market structure with one dom-inating characteristic: the absence of competition.Most infrastructure services in the developingworld are provided by centrally managed monopo-listic public enterprises or government depart-ments. Almost all irrigation, water supply, sanita-tion, and transport infrastructure is provided in thismanner. Until a few years ago, telephone services inmost countries were the responsibility of a state-owned post, telephone, and telegraph enterprise.The bulk of power has also been provided by a pub-lic monopoly. As a result, the pressure that competi-tion can exert on all parties to perform at maximumefficiency has been lacking.
Second, those charged with responsibility for de-livering infrastructure services are rarely given themanagerial and financial autonomy they need to dotheir work properly. Managers are often expected tomeet objectives at variance with what should betheir primary functionthe efficient delivery ofhigh-quality services. Public entities are required toserve as employer of last resort or to provide pa-tronage. They are compelled to deliver servicesbelow costoften by not being allowed to adjustprices for inflation. The other side of the coin is thatpublic providers are rarely held accountable fortheir actions. Few countries set well-specified per-formance measures for public providers of infra-structure services, and inefficiency is all too oftencompensated for by budgetary transfers rather thanmet with disapproval.
Third, the users of infrastructureboth actualand potentialare not well positioned to maketheir demands felt. When prices reflect costs, thestrength of consumer demand is a clear signal ofwhat should be supplied. Through the price mecha-nism, consumers can influence investment and pro-duction decisions in line with their preferences. Butprices of infrastructure services typically do not re-flect costs, and this valuable source of informationabout consumer needs is lost. For example, powerprices in developing countries have generally fallen,while costs have not. As a result, prices now coveronly half the supply costs, on average. Watercharges and rail passenger fares typically cover onlya third of costs. Excess consumer demand based onbelow-cost prices is not a reliable indicator that ser-vices should be expanded, although often it is takenas such.
Users can express preferences in other ways,such as local participation in planning and imple-menting new infrastructure investments. But theyseldom are asked, and investment decisions are alltoo often based on extrapolations of past consump-tion rather than on true assessments of effective de-mand and affordability.
Individually, each of these three points is impor-tant. Together, they go a long way toward explain-ing the disappointing past performance of muchinfrastructure. Rival suppliers and infrastructureusers might have exerted pressure for better ser-vices, but they were prevented from doing so. Gov-ernmentsby confusing their roles as owner, regu-lator, and operatorhave failed to improve servicedelivery.
New opportunities and initiatives
Creating the institutional and organizational condi-tions that oblige suppliers of infrastructure services
to be more efficient and more responsive to theneeds of users is clearly the challenge. But is it pos-sible? Three converging forces are opening a win-dow of opportunity for fundamental changes in theway business is done. First, important innovationshave occurred in technology and in the regulatorymanagement of markets. Second, a consensus isemerging on a larger role for the private sector in in-frastructure provision, based in part on recent expe-rience with new initiatives. Third, greater concernnow exists for environmental sustainability and forpoverty reduction.
New technology and changes in the regulatorymanagement of markets create new scope for intro-ducing competition into many infrastructure sec-tors. In telecommunications, satellite and micro-wave systems are replacing long distance cablenetworks, and cellular systems are an emergingalternative to local distribution networks. Thesechanges erode the network-based monopoly intelecommunications and make competition possi-ble. In power generation, too, combined-cycle gasturbine generators operate efficiently at lower out-put levels, while other innovations are reducingcosts. New technology makes competition amongsuppliers technically feasible, and changes in regu-lations are making competition a reality by allowingcompetitive entry in activities such as cellularphone service or power generation. Technical andregulatory change in other infrastructure sectorsranging from transportation to water supply anddrainage and irrigation also make them moreopen to new forms of ownership and provision.
Alongside such changes are new perceptions ofthe role of government in infrastructure. An aware-ness is growing in many countries that governmentprovision has been inadequate. Brownouts andblackouts in power systems, intermittent water sup-plies from municipal systems, long waiting periodsfor telephone service connection, and increasingtraffic congestion provoke strong reactions. Reformsin some industrial countries have increased thecompetition in telecommunications, in road freightand airline transport, and in power generationproving that alternative approaches are possible.The poor performance of planned economies hasalso provoked a reassessment of the state's role ineconomic activity.
These developments have led governments tosearch for new ways to act in partnership with theprivate sector in providing infrastructure services.Most dramatic have been the privatizations of suchenterprises as the telephone system in Mexico andthe power system in Chile. Elsewhere, variousforms of partnership between government and the
7
private sector have evolved. Port facilities havebeen leased to private operatorsthe Kelang con-tainer facility in Malaysia being among the first.Concessions have been granted to private firms,particularly in water supply; Côte d'Ivoire is one ofthe earliest examples. Contracting out services, asKenya has done with road maintenance, is wellunder way in many countries. Private financing ofnew investment has grown rapidly through build-operate-transfer (BOT) arrangements under whichprivate firms construct an infrastructure facilityand then operate it under franchise for a period ofyears on behalf of a public sector client. This ap-proach has been used to finance the construction oftoll roads in Mexico and power-generating plantsin China and the Philippines.
An increasing regard for the environmental sus-tainability of development strategies and a deepen-ing concern for poverty reduction after a decade ofstagnation in many regions of the world also giveimpetus to infrastructure reform. Creating pressuresfor change, environmental issues are coming to thefore in transport (traffic congestion and pollution),irrigation (increased waterlogging and salinity ofagricultural land), water supply (depleted re-
sources), sanitation (insufficient treatment), andpower (growing emissions). At the same time, adecade of reduced economic growthespecially inLatin America and Sub-Saharan Africashowsthat poverty reduction is not automatic and thatcare must be taken to ensure that infrastructure bothaccommodates growth and protects the interests ofthe poor.
Options for the future
To reform the provision of infrastructure services,this Report advocates three measures: the wider ap-plication of commercial principles to service pro-viders, the broader use of competition, and the in-creased involvement of users where commercialand competitive behavior is constrained.
Applying commercial principles of operation in-volves giving service providers focused and explicitperformance objectives, well-defined budgets basedon revenues from users, and managerial and finan-cial autonomywhile also holding them account-able for their performance. It implies that govern-ments should refrain from ad hoc interventions inmanagement but should provide explicit transfers,where needed, to meet social objectives such as pub-lic service obligations.
Broadening competition means arranging for sup-pliers to compete for an entire market (e.g., firms
8
bidding for the exclusive right to operate a port forten years), for customers within a market (tele-phone companies competing to serve users), andfor contracts to provide inputs to a service provid-er (firms bidding to provide power to an electricutility).
Involving users more in project design and opera-tion of infrastructure activities where commercialand competitive behavior is constrained providesthe information needed to make suppliers more ac-countable to their customers. Users and other stake-holders can be involved in consultation during pro-ject planning, direct participation in operation ormaintenance, and monitoring. Development pro-grams are more successful when service users or theaffected community has been involved in projectformulation. User participation creates the appro-priate incentives to ensure that maintenance is car-ried out in community-based projects.
These elements apply whether infrastructure ser-vices are provided by the public sector, the privatesector, or a public-private partnership. To this ex-tent, they are indifferent to ownership. However,numerous examples of past failures in public provi-sion, combined with growing evidence of more effi-cient and user-responsive private provision, arguefor a significant increase in private involvement infinancing, operation, andin many casesowner-ship.
All countries will not be able to increase privateinvolvement at the same rate. Much depends on thestrength of the private sector, the administrative ca-pacity of the government to regulate private suppli-ers, the performance of public sector providers, andthe political consensus for private provision. Withthis in mind, the Report sets out a menu of fourmain options for ownership and provision:
Option A. Public ownership and operation byenterprise or department
Option B. Public ownership with operation con-tracted to the private sector
Option C. Private ownership and operation,often with regulation
Option D. Community and user provision.
Far from exhaustive, these four options merely illus-trate possible points in a broader array of alterna-tives.
Option A: Public ownership and public operation.Public provision by a government department, pub-lic enterprise, or parastatal authority is the mostcommon form of infrastructure ownership and op-eration. Successful public entities run on commer-
cial principles and give managers control over oper-ations and freedom from political interference, butthey also hold managers accountable, often throughperformance agreements or management contracts.And they follow sound business practices and aresubject to the same regulatory, labor law, account-ing, and compensation standards and practices asprivate firms. Tariffs are set to cover costs, and anysubsidies to the enterprise are given for specific ser-vices and in fixed amounts. Water authorities inBotswana and Togo and national power companiesin Barbados and Thailand perform well. The high-way authorities in Ghana and Sierra Leone and therestructured road agency in Tanzania are promisingexamples of this approach. But few successful ex-amples of Option A persist because they are vulner-able to changes in governmental support. Manypublic entities perform well for a time and then fallvictim to political interference.
Option B: Public ownership with private operation.This option is typically implemented through leasecontracts for full operation and maintenance of pub-licly owned infrastructure facilities, or through con-cessions, which include responsibility for construc-tion and financing of new capacity. Arrangementsbetween the owner (government) and the operator(firm) are set out in a contract that includes any reg-ulatory provisions. The private operator typicallyassumes all commercial risk of operation and sharesin investment risk under concessions. Leases andconcessions are working well for railways in Ar-gentina; for water supply in Buenos Aires andGuinea; and for port facilities in Colombia, Ghana,and the Philippines. Concessions also include con-tracts to build and operate new facilities under theBOT arrangement and its variants. Proliferating inrecent years, concessions to build and operate facili-ties include toll roads in China, Malaysia, and SouthAfrica; power plants in Colombia, Guatemala, andSri Lanka; water and sanitation facilities in Malaysiaand Mexico; and telephone facilities in Indonesia,Sri Lanka, and Thailand. Each has brought privatefinancing to support new investments.
Option C: Private ownership and private operation.The private ownership and operation of infrastruc-ture facilities is increasingboth through new entryby private firms in infrastructure markets andthrough divestiture of public ownership of entiresystems. Private ownership is straightforward whenservices can be provided competitively, and, inmany infrastructure sectors, it is possible to identifysuch activities and to allow private provision. Forexample, twenty-seven developing countries allowcellular telephone service to be competitively pro-
vided, and many others allow private firms to con-struct electricity-generating plants and sell power tothe national power grid. Where competition amongsuppliers is possible, private ownership and opera-tion require little or no economic regulation beyondthat applied to all private firms. The necessary com-petition can also occur across sectorsbetweenroad and rail, or between electricity and gas. For ex-ample, because it competes with suppliers of otherenergy sources, the private gas company in HongKong has no special economic regulation.
Where systems are being fully or partly priva-tized and there is no cross-sectoral competition, reg-ulation of both private and public providers may berequired to prevent the abuse of monopoly power.Experience with regulation and with systemwideprivatization in developing countries is still verynew. The Chilean form of regulation, which involvesregular, automatic price adjustments and a well-specified arbitration system, appears to be workingwell. And systems that have been privatized havebeen very successful at expanding service. Venezue-la's telephone company expanded its network by 35percent in the first two years after its privatization;Chile's by 25 percent a year, Argentina's by 13 per-cent a year, and Mexico's by 12 percent a year.
Option D: Community and user provision. Commu-nity and user provision is most common for local,small-scale infrastructuresuch as rural feederroads, community water supply and sanitation, dis-tribution canals for irrigation, and maintenance oflocal drainage systemsand it often complementscentral or provincial services. Successful communityprovision requires user involvement in decision-making, especially to set priorities for expendituresand to ensure an equitable and agreed sharing of thebenefits and costs of service provision. Technical as-sistance, training, and compensation of service oper-ators are also very important. When these elementsare present, community self-help programs can suc-ceed over long periods. A community organizationin Ethiopia devoted mainly to maintaining roads(the Gurage Roads Construction Organization) hasworked well since 1962 because it sets its own prior-ities and allocates its own financial and in-kind re-sources.
Financing: essential for all options. Implementingthe foregoing institutional options and mobilizingfunds to expand and improve services require care-fully designed financing strategies. Foreign and do-mestic sources of finance will need to be tapped, butthere are limits to the capacity of any economy toobtain funds from abroad, especially debt finance.Balance of payments constraints, and the limited
9
tradability of infrastructure services, mean that formost countries an ongoing infrastructure programhas to be sustained by a strategy for mobilizing do-mestic funds.
Private financing in one form or another at pres-ent accounts for about 7 percent of total infrastruc-ture financing in developing countries (the sharemay double by the year 2000), while bilateral andmultilateral foreign aid accounts for around another12 percent. Although an increasing share of the do-mestic savings needed to finance infrastructure pro-vision can come from private sources, governmentswill continue to be a major source of funds for infra-structure, as well as a conduit for resources from thedonor community. As transitional measures to pro-vide long-term financing where sufficient privatesupport is not likely to be forthcoming, govern-ments are revitalizing existing lending institutionsfor infrastructure and creating specialized infra-structure funds.
In the future, governments will often need to bepartners with private entrepreneurs. The task forboth the public and private sectors is to find ways toroute private savings directly to those privateriskbearers that are making long-term investmentsin infrastructure projectsprojects that have vary-ing characteristics and for which no single financingvehicle is appropriate. Official sources of finance,such as multilateral lending institutions, can facili-tate the process by supporting the policy and insti-tutional reforms needed to mobilize private financ-ing and use it more efficiently.
Implementing reform
Just as the differences across infrastructure sectorsimply that no single option can be applied to all sec-tors, infrastructure provision must be tailored tocountry needs and circumstances, which varywidely. To see how, consider a middle-income coun-try with a thriving private sector and well-devel-oped institutional capability, and a low-incomecountry with a small private sector and relativelyundeveloped institutional capacity.
Middle-income countries with good capacity. Thefour major options can all work well in these coun-tries. The broad reform instruments for such coun-tries are clear: apply commercial principles, increasecompetition, and involve users. These actions leadto an increase in private involvement and finance,and to a reduction (or decentralization) of activitiesremaining with government. Some countries are fol-lowing this path for a wide range of sectors, andmany more for only a few sectors, especially tele-communications, power, and roads.10
Activities that can be competitively providedshould be separated and opened to private suppli-ers and contractors. Where possible, entire sectorstelecommunications, railways, power generationcan be privatized, but with regulatory oversight.Sectors that are unlikely to be privatized (such asroads) can be operated on commercial principles,using contracting for construction and periodicmaintenance. Leasing or concessions can be used tooperate facilities that may be difficult to privatizefor strategic reasons, such as ports or airports. More-over, technical and managerial capacity at theprovincial and local level is likely to be sufficient torealize the benefits of decentralization. Responsibil-ity for local servicessuch as urban transport,water supply, sanitation, and local roadscan beturned over to local governments.
Low-income countries with modest capacity. In thesecountries, commercial principles of operation canform the basis for reform in several sectors. Com-mercial approaches can be supplemented with re-forms in procurement and contracting practices thatfoster competition and develop the domestic con-struction industry. Many activities (such as roadmaintenance and the collection of solid waste) canbe contracted out to the private sector. Contractingcan have a salutary effect on all infrastructure be-cause, as experience shows, public providers be-come more efficient when they are exposed to com-petition from private contractors.
Concessions or leasing arrangements are provenways for a low-income country to draw on foreignexpertise, as are the various BOT options that can beused to increase the capacity of systems. Conces-sions and leases have been widely used in watersupply, ports, and transport sectors. BOT schemeshave been extensively used in middle-income coun-tries, and their application is now spreading to low-income countries. These arrangements help developlocal expertise and foster the transfer of new tech-nology, but they do not require the establishment ofindependent regulatory bodies because regulatoryprocedures are specified in the underlying contract.
Community approaches, with technical and fi-nancial support, can be efficient and sustainable insupplying services using intermediate technologiesin rural areas and in the low-income settlementsthat often develop outside existing urban serviceareas. Competition is possible in many activities butmay be impeded by unnecessary regulations. Truck-ing and many types of urban passenger transportcan be provided privately, under regulations thatdeal only with safety and service standards.
Some countries may benefit from arrangementsthat increase the effectiveness of aid by coordinating
the efforts of donors to focus on common objectives.For example, the Sub-Saharan Africa Transport Pol-icy Program coordinates donor assistance for roadmaintenance and in several countries has supportedthe establishment of road boards that oversee execu-tion of road maintenance. More generally, externalassistance should aim to build institutional capacityin those countries where it poses a serious constraint.Well-designed programs of training and technicalcooperation, as well as efforts to collect and dissem-inate information on policy options and perfor-mance across countries, can supplement donors'advice and financial assistance in creating an appro-priate enabling environment for successful reformand development of infrastructure.
Potential payoffs from reform
Because of the great variation in performance, thepayoffs from increasing the efficiency of infrastruc-ture provision will differ from country to countryand from sector to sector. But the rewards are poten-tially large across the spectrum, making the commit-ment to reform imperative and worthwhile.
Reform will produce three types of gains: reduc-tion in subsidies, technical gains to suppliers, andgains to users. It is possible to make rough estimatesof the first two types of gains. The first source ofgains is the reduction in the fiscal burden of serviceprovisioncosts not recovered from users. Al-though a conservative estimate can be made foronly three sectors (power, water, and railways), thetotal savings are nearly $123 billion annuallynearly 10 percent of total government revenues indeveloping countries, 60 percent of annual infra-structure investment, and approximately five timesannual development finance for infrastructure (Fig-ure 5). Eliminating underpricing would not producea net resource savings to the economy (as the costswould be covered by users), but the fiscal reliefwould be substantial.
The second source of gains is the annual savingsto service providers from improved technical effi-ciency. The savings possible from raising operatingefficiency from today's levels to best-practice levelsare estimated at around $55 billion a yearpuresavings equivalent to 1 percent of all developingcountries' GDP, a quarter of annual infrastructureinvestment, and twice annual development financefor infrastructure. Looked at another way, if the an-nual technical losses of $55 billion could be re-directed for three yearsat current costs of roughly$150 per person for water systemsthe I billionpeople without safe water could be served.
Figure 5 Annual gains from eliminatingmispricing and inefficiency are largerelative to investment.
Billions of U.S. dollars250
Subsidiesincurred from
mispricing
Water
LI Railways
LI Roads
Power
Costsincurred from
technicalinefficiency a
Annualinfrastructureinvestment
Development finance
LI All other sources
a. Costs for the water sector are due to leakages; forrailwaysfuel inefficiency, overstaffing, and locomotiveunavailability; for roadsadded investment caused bypoor mainlenance; for powertransmission,distribution, and generation losses.Source: Ingram and Fay, background paper; Appendixtable A.4.
The payoffs from better infrastructure services gobeyond reducing technical inefficiency and financiallosses. Improvements in productivity and pricingwould permit more effective delivery of service inresponse to demand. They would also enhance thegrowth and competitiveness of the economy. Andthey would allow vastly greater mobilization of re-sources for needed new investmentsby generat-ing higher revenues and by creating a policy envi-ronment conducive to the inflow of new investmentresources.
This Report's agenda for reforming the incen-tives and institutional frameworks in infrastructureposes major challengesbut promises major bene-
11
200Investment
200
Fiscal150 burden
123
100 Resourceloss
fits. The way ahead is one of continuing innovationand experimentation, and both industrial and de-veloping countries will learn from each other. Insome countries, the challenge is to keep pace withrapid economic growth and urbanization. In others,it is to restore growth in ways that also providegreater opportunities to the poor. Everywhere, the
12
emphasis needs to be on improving environmentalconditions. Increasingly, infrastructure needs tomatch new demands as developing countries be-come more closely integrated into the global econ-omy. Infrastructure is no longer the gray backdropof economic lifeunderground and out of mind. Itis front and center in development.
IInfrastructure: achievements,challenges, and opportunities
Infrastructure servicesincluding power, trans-port, telecommunications, provision of water andsanitation, and safe disposal of wastesare centralto the activities of households and to economic pro-duction. This reality becomes painfully evidentwhen natural disasters or civil disturbances destroyor disable power stations, roads and bridges, tele-phone lines, canals, and water mains. Major in-frastructure failures quickly and radically reducecommunities' quality of life and productivity. Con-versely, improving infrastructure services enhanceswelfare and fosters economic growth.
Providing infrastructure services to meet thedemands of businesses, households, and other usersis one of the major challenges of economic devel-opment. The availability of infrastructure has in-creased significantly in developing countries overthe past several decades. In many cases, however,the full benefits of past investments are not beingrealized, resulting in a serious waste of resourcesand lost economic opportunities. This outcome isfrequently caused by inadequate incentives embod-ied in the institutional arrangements for providinginfrastructure services. While the special technicaland economic characteristics of infrastructure givegovernment an essential role in its provision, domi-nant and pervasive intervention by governmentshas in many cases failed to promote efficient or re-sponsive delivery of services. Recent changes inthinking and technology have revealed increasedscope for commercial principles in infrastructureprovision. These offer new ways to harness marketforces even where typical competition would fail,and they bring the infrastructure user's perspectiveto the forefront.
This Report focuses on economic infrastructure:the long-lived engineered structures, equipment,
and facilities, and the services they provide that areused in economic production and by households.This infrastructure includes public utilities (power,piped gas, telecommunications, water supply, sani-tation and sewerage, solid waste collection and dis-posal), public works (major dam and canal worksfor irrigation, and roads), and other transport sec-tors (railways, urban transport, ports and water-ways, and airports). Social infrastructure, often en-compassing education and health care, representsan equally important although very different set ofissues that are not analyzed in this Report (see WorldDevelopment Report 1993: Investing in Health).
As defined here, infrastructure covers a complexof distinct sectors that, by any measure, represent alarge share of an economy. Taken togethei the ser-vices associated with the use of infrastructure (mea-sured in terms of value added) account for roughly 7to 11 percent of GDP (Table 1.1), with transportbeing the largest sector. Transport alone commonlyabsorbs 5 to 8 percent of total paid employment. Asample of developing countries shows that infra-
Table 1.1 Value added of infrastructure servicesby country group(percentage of GDP)
Transport,
Note: At market prices. At factor cost (for which fewer observations areavailable), the values are slightly higher. Figures in parentheses arenumber of observations. Data are for 1990 or latest available yearSource: World Bank national accounts data.
13
storage, and 5.34 6.78 9.46communications (9) (26) (3)
Gas, electricity,and water
1.29
(22)
2.24
(36)
1.87
(5)
Low-income Middle-income High-incomeSector countries countries countries
Figure 1.1 Public infrastructure investmentis a large fraction of both total and publicinvestment in developing countries.
Percentage of investment allocatedto infrastructure60
50
40
30
20
10
Total investment Public investment
Low-income countries
Middle-income countries
Sample: Twelve low-income and eight middle-incomecountries; unweighted averages, 1980-89.Source: Easterly and Rebelo 1993.
structure typically represents about 20 percent oftotal investment and 40 to 60 percent of public in-vestment (Figure 1.1). In round figures, public infra-structure investment ranges from 2 to 8 percent (andaverages 4 percent) of GDP. Even these shares un-derstate the social and economic importance of in-frastructure, which has strong links to growth,poverty reduction, and environmental sustainability.
Infrastructure's impact on development
Links to economic growth
Infrastructure represents, if not the engine, then the"wheels" of economic activity. Input-output tablesshow that in the economies of Japan and the UnitedStates, for example, telecommunications, electricity,and water are used in the production process of
14
nearly every sector, and transport is an input forevery commodity. Users demand infrastructure ser-vices not only for direct consumption but also forraising their productivity by, for instance, reducingthe time and effort needed to secure safe water, tobring crops to market, or to commute to work.
Much research in recent years has been devotedto estimating the productivity of infrastructure in-vestments (Box 1.1). Many studies attempting tolink aggregate infrastructure spending to growth ofGDP show very high returns in a time-series analy-sis. Some cross-national studies of economic growthand infrastructure notably, one using public in-vestments in transport and communications and an-other using capital stocks in roads, railways, andtelephonesalso show that infrastructure variablesare positively and significantly correlated withgrowth in developing countries. In both types ofstudies, however, whether infrastructure invest-ment causes growth or growth causes infrastructureinvestment is not fully established. Moreover, theremay be other factors driving the growth of bothGDP and infrastructure that are not fully accountedfor. Neither the time-series nor the cross-sectionalstudies satisfactorily explain the mechanismsthrough which infrastructure may affect growth.
Sectoral studies focusing on rural infrastructure'seffect on the local economy in certain developingcountries have revealed more about the nature ofthe apparent benefits. Studying data over time fromeighty-five districts in thirteen Indian states, re-searchers found that lower transport costs increasedfarmers' access to markets and led to considerableagricultural expansion and that modern irrigationmethods brought higher yields. At the same time,because improved communications (through roads)lowered banks' costs of doing business, banks ex-panded lending to farmers, and farmers used thefunds to buy fertilizer, further increasing yields. Ac-cording to a household- and village-level surveyconducted in Bangladesh, villages classified as"most developed" in terms of access to transport in-frastructure were significantly better off than the"less developed" villagesin terms of agriculturalproduction, incomes and labor demand, and health.(It is difficult, however, to verify whether theBangladesh study took into account all possible in-tervening factors, such as unobserved differencesamong the communities in natural endowments.)
What is evident is that a strong association existsbetween the availability of certain infrastructure -telecommunications (in particular), power, pavedroads, and access to safe waterand per capitaGDP (Figure 1.2). An analysis of the value of infra-
Box 1.1 Returns on infrastructure investmenttoo good to be true?
Recent studies in the United States suggest that the im-pact of infrastructure investments on economic growthrepresents startlingly high rates of return (up to 60 per-cent). Too good to be true? Possibly. The results presentedin Box table 1.1 may overestimate the productivity of in-frastructure for two reasons. First, there may be a com-mon factor that causes growth in both output and infra-structure that is not included in the study. Second, it maybe that growth leads to infrastructure investment, andnot that investment produces growth. A number of stud-ies have found that causation runs in both directions. Yetmore sophisticated estimates that address these issues ei-ther have concluded that the positive results were notmuch affected by different econometric methods or havefound no noticeable impact of infrastructure on growth.Neither findingof an extremely high impact or of anegligible impactis entirely credible, and research ef-forts continue in an attempt to refine the methodology.
An alternative approach estimates the impact of in-
Box table 1.1 Results from studies of infrastructure productivity
United StatesUnited States48 states, United States5 metro areas, United StatesRegions, JapanRegions, FranceTaiwan, China
frastructure on production costs. Studies (summarizedin Aschauer 1993) found that infrastructure significantlyreduces production costs in manufacturing in Germany,Japan, Mexico, Sweden, the United Kingdom, and theUnited States. One estimate suggests that three-quartersof U.S. federal investment in highways in the 1950s and1960s can be justified on the basis of reductions in truck-ing costs alone.
While there is still no consensus on the magnitude oron the exact nature of the impact of infrastructure ongrowth, many studies on the topic have concluded thatthe role of infrastructure in growth is substantial, signifi-cant, and frequently greater than that of investment inother forms of capital. Although the indications to dateare suggestive, there is still a need to explain why thefindings vary so much from study to study. Until thisproblem is resolved, results are neither specific nor solidenough to serve as the basis for designing policies for in-frastructure investment.
Aschauer 1989Munnell 1990Holtz-Eakin 1992Duffy-Deno and Eberts 1991Mera 1973Prud'homme 1993Uchimura and Gao 1993
Uchimura and Gao 1993
Bregman and Marom 1993
Shah 1988, 1992
19 Canning and Fay 199395 Canning and Fay 1993- Baffes and Shah 1993
0.16 63 Easterly and Rebelo 1993Percentage changes in output with respect to a 1 percent change in the level of infrastructure.Ratio of discounted value of increase in dependent variable to discounted value of investment in infrastructure.
Nonmilitary public capitalNonmilitary public capitalPublic capitalPublic capitalIndustrial infrastructurePublic capitalTransportation, water, and
communicationTransportation, water, and
communicationTransportation, power, water, and
sanitationPower, communication, and
transportationTransportationTransportationInfrastructure capital stocks
Transportation and communication
structure stocks indicates that their compositionchanges significantly as incomes rise. For low-in-come countries, more basic infrastructure is impor-tant such as water, irrigation, and (to a lesser ex-tent) transport. As economies mature into themiddle-income stage, most of the basic consump-tion demands for water are met, the share of agri-culture in the economy shrinks, and more transportinfrastructure is provided. The share of power andtelecommunications in investment and infrastruc-
ture stocks becomes even greater in high-incomecountries. Data for 1990 indicate that, while total in-frastructure stocks increase by 1 percent with each1 percent increment in per capita GDP, householdaccess to safe water increases by 0.3 percent, pavedroads increase by 0.8 percent, power by 1.5 percent,and telecommunications by 1.7 percent.
These relationships suggest that infrastructurehas a high potential payoff in terms of economicgrowth, yet they do not provide a basis for prescrib-
15
Implied rateSample Elasticity' of returnb Author/year Infrastructure measure
Figure 1.2 Per capita availability of major infrastructure is closely related to income levels.
Sub-Saharan AfricaSouth AsiaEurope and Central Asia
Note: Axes are logarithmic; infrastructure quantities and GDP are for 1990; purchasing power parity (PPP) dollars are valuedin Summers and Heston 1985 international prices.Source: WDI table 32; Summers and Heston 1991.
ing appropriate levels, or sectoral allocations, for in-frastructure investment. Other evidence confirmsthat investment in infrastructure alone does notguarantee growth. Many studies reveal muchsmaller returns for infrastructure than those sug-gested in Box 1 .1closer, in fact, to the return onprivate investments. These disparities may be due to
16
differences in the efficiency of investment acrosscountries and over time. For example, a study of theeconomic returns to individual World Bank projectsshows that, when overall economic policy condi-tions are poor, the returns to infrastructure invest-ment decline. Returns are lower by 50 percent ormore in countries with restrictive trade policies than
Chad S Bangladesh
Burkina Faso
Indonesia
1001 Sudan
10
0.5
in countries where conditions are more favorable.Infrastructure spending cannot, therefore, overcomea weak climate for economic activity. Nearly twenty-five years ago, the Brookings Transport ResearchProject evaluated the impact of transport projects inseveral developing countries and concluded simi-larly that, although the investments generally hadreasonable rates of return, success depended largelyon economic policy.
Another approach to assessing the economic re-turns from infrastructure investment is to examinethe rates of return in a large sample of completedWorld Bank projects. The average economic returnon infrastructure projects, reestimated after loan dis-bursement (completion of project construction), hasbeen 16 percent over the past decadejust abovethe World Bank project average of 15 percent (Table1.2). Returns have been lowest (and declining) for ir-rigation and drainage, airports (for a very smallsample), railways, power, water supply, and sewer-age. Why should this be so, given the expected ben-efits of such investments in developing countries?
Some of the causes relate to implementationproblems (discussed below under "The record ofperformance") and others to project identificationand design. A common pattern discovered in proj-ect completion reviews of water, railway, and powerprojects is the tendency at the time of appraisal tooverestimate the rate of growth in demand for newproduction capacity and, therefore, of revenues. Forthe power projects in the sample, demand was over-estimated by 20 percent on average over a ten-yearoperating period. In water projects, overestimationof rates of new connections and per capita con-sumption also averaged about 20 percent. In thecase of railways, until recent years projects often as-sumed recovery in demand even where railwayswere continually losing traffic to roads offering bet-ter service. In twenty-nine of thirty-one cases,freight traffic failed to reach its projected level, andin one-third, traffic actually declined.
One important explanation for the misjudgmentsduring appraisal is inadequate procedures for as-sessing demand (including the effects of tariff in-creases). Oversizing and inappropriate design of in-vestments then occur, resulting in financial burdenson the project entities concerned. Although Bankprojects may not be entirely representative, they aresubject to more careful evaluation than many infra-structure investments in developing countries andso may have achieved better performance than av-erage public investments in these sectors.
Infrastructure is a necessary, although not suffi-cient, precondition for growthadequate comple-
Table 1.2 Average economic rates of returnon World Banksupported projects, 1974-92
Not available.a. Rates are financial, not economic, rates of return.Source: World Bank data.
ments of other resources must be present as well.The growth impact of infrastructure investmentsalso depends on the timing and location of addi-tions to capacity, and on the existing imbalance be-tween supply and demand. Because much infra-structure consists of networks, relieving bottlenecksat certain points of the system can produce veryhigh returns. Box 1.2 illustrates the repercussions inChina's economy from critical constraints in thetransport of coal needed for power generation.
Adequate quantity and reliability of infrastruc-ture are key factors in the ability of countries tocompete in international trade, even in traditionalcommodities. In part because of infrastructure prob-lems, shipping costs from Africa to Europe are 30percent higher for plywood (and 70 percent higherfor tuna) than those from Asia to Europe. Thesecosts have to be borne by exporters.
The competition for new export markets is espe-cially dependent on high-quality infrastructure.During the past two decades, increased globaliza-tion of world trade has arisen not only from the lib-eralization of trade policies in many countries butalso from major advances in communications, trans-port, and storage technologies. These advances cen-ter on the management of logistics (the combinationof purchasing, production, and marketing func-tions) to achieve cost savings in inventory andworking capital and to respond more rapidly to cus-tomer demand. About two-thirds of production andsales in the OECD countries are processed directly
17
(percent)
Sector 1974-82 1983-92
Irrigation and drainage 17 13
Telecommunications 20 19
Transport 18 21
Airports 17 13
Highways 20 29
Ports 19 20Railways 16 12
Power 12 11
Urban development 23Water and sanitationa 7 9
Water supply a 8 6Seweragea 12 8
Infrastructure projects 18 16
All Bank operations 17 15
Box 1.2 The importance of infrastructure to economic development: an examplefrom China
The fact that infrastructure provides critical support tothe growth of an economy can be clearly seen when bot-tlenecks arise. One of the most striking examples is thatof China's intercity transport system, with its links to thesupply of raw materials, coal, and electricity.
The coverage of China's intercity transport networksis one of the thinnest in the world: the total route lengthper capita or per unit of arable landfor highways orrailwaysis similar to, or lower than, that in Brazil,India, and Russia. This has resulted mainly from chronicunderinvestment in China's transport infrastructure.China's transport investments amounted to only 1.3 per-cent of GNP annually during 1981-90, a period of rapidgrowth in transport demand.
Since the onset of China's open door policy in 1979,economic growth averaging 9 percent a year has resultedin an unprecedented expansion in intercity trafficwithgrowth averaging 8 percent a year for freight and 12 per-cent a year for passengers. This traffic growth has im-posed tremendous strains on the transport infrastruc-
ture, as manifested by the growth of bottlenecks in therailway network, the severe rationing of transport capac-ity on railway lines, and the poor quality of service expe-rienced by shippers and passengers.
Transport shortages have adversely affected the sup-ply of coal in particular. Coal is the source of some 73percent of China's commercial energy and representsabout 43 percent of the total tonnage of freight handledby the railways. The shortage of coal has in turn ad-versely affected supplies of electricity, about 76 percentof which is generated by thermal plants. In 1989, Chinawas experiencing a shortfall in available power of about20 percent of industrial electricity requirements. Centraland local authorities established quotas for allocatingelectricity and rationed new connections, but power cutshave nevertheless been frequent.
A conservative estimate is that the annual economiccosts of not having adequate transport infrastructure inChina during the past several years amount to about 1percent of China's GNP.
to order, and "just-in-time" delivery of products hasbecome the norm in many sectors. Because about 60percent of their exports are directed to OECD mar-kets, developing countries must meet these stan-dards. Virtually all the improved practices designedto reduce logistics costs, including those in trans-port, have been based on information technologiesusing telecommunications infrastructure. Cost re-ductions and the increased speed of freight move-ments over the past few decades have also been in-creasingly based on multimodal transport involvingcontainerization, which requires intensive coordina-tion by shippers across rail, port, air, and roadfreight modes.
For developing countries wishing to compete inglobal markets, or to participate in "global sourc-ing" (the linking of businesses in several countriesproducing different components for a final prod-uct), not just any kind of transport and telecommu-nications infrastructure will do. Manufacturing as-sembly operations in Mexico and horticulturalexports from Kenya are examples of the diversifica-tion of trade permitted by appropriate logisticalsupport and multimodal facilities. During the 1980s,the proportion of garments, shoes, and handicraftexports shipped by air from northern India quintu-pled because land and ocean transport systemswere no longer able to meet demanding delivery re-quirements. Because India's ports have been slow to
18
adapt to containerization and are subject to regula-tory delays, freight transport to the United States isone-third more expensive from Indian ports thanfrom Bangkok or Singapore.
The availability of infrastructure services valuedby users is also critical for the modernization anddiversification of production. The growth of elec-tronic data exchange involving telecommunica-tionsinformatics--is central to efficient opera-tions in manufacturing, services, the financial sector,arid government. Availability of power allows sub-stantial improvements in workers' productivity (forexample, in the transition from foot-powered toelectrically powered sewing), while internationaltelecommunications, facsimile services, and rapidtransport of goods permit the artisan to produce toorder for a computerized global market. A higherquality of water and sanitation is required to shiftfrom production of raw agricultural commodities toprocessed foods. Surveys of prospective foreigninvestors over a wide range of countries show thatthe quality of infrastructure is an important factorin ranking potential sites for location of directinvestment.
The nature of an economy's infrastructure is cen-tral to its ability to respond to changes in demandand prices or to take advantage of other resources.The formerly socialist countries (particularly thosein Central and Eastern Europe and the former So-
viet Union) provide a clear illustration of how thepatterns of supply and demand imposed by centralplanning affect infrastructure development. Thesecountries showed an extremely high transport andenergy intensity (owing to noneconomic decisionson location of production units, underpricing andinefficient use of energy, and an emphasis on heavyindustry and raw materials production). They alsoshowed a greater reliance on rail than on road trans-port than did countries with similar conditions, andon long- over short-haul public transport facilities.With market reforms, the location and compositionof demand will alter, giving a greater role in theseeconomies to light industry, to services such asdomestic distribution, and to the diversification ofexternal trade. Small enterprises and consumerswill become a more important source of demand.These trends require corresponding modificationsin infrastructure, with greater attention to the qual-ity and variety of services.
Public spending on infrastructure constructionand maintenance can be a valuable policy tool toprovide economic stimulus during recessions. Aslong as quality and cost-effectiveness are not com-promised, labor-based approaches to infrastructuredevelopment can also be an important instrumentfor employment-intensive economic growth. In de-ciding on public spending for infrastructure, policy-makers have frequently not looked sufficientlybeyond the near-term impacts, and many govern-ments have been attracted to the political benefits ofthe highly visible structures created. When publicspending on infrastructure is not wisely deployed, itcan crowd out more productive investment in othersectors. At the same time, short-term fiscal con-straints have often led to disproportionate cutbacksin infrastructure, thereby sacrificing an importantimpetus to renewed growth following adjustment(Box 1.3).
Sometimes the least-cost approach to improvingthe supply of infrastructure services would requireinterregional (cross-country) integration of infra-structure networks, for example, power grids. Suchan agreement would call for not only coordinationof investments but, equally important, cooperationto maintain efficient policies governing the trade inservices. Most countries, however, resist dependingon others for a supply of services deemed to be ofstrategic importance; therefore, importing power tomeet the base load demand is less acceptable thanacquiring only peak load from abroad. Internationalagreements have been more common for cross-bor-der transport, which is a particularly importantissue for landlocked countries. Often, the quality of
I
I
Box 1.3 Throwing infrastructureoverboard
When times are hard, capital spending on infra-structure is the first item to go, and operations andmaintenance are often close behind. Despite thelong-term economic costs of slashing infrastructurespending, governments find it less politically costlythan reducing public employment or wages. Stud-ies of fiscal adjustment and expenditure reductionfind that capital expenditures are cut more thancurrent expenditures, with infrastructure capitalspending often taking the biggest reduction. More-over, within current expenditures, nonwage expen-ditures (which include operations and mainte-nance) are cut by more than the wage bill.
The decline in investment, at least in the initialphases, is not altogether undesirable as it often in-duces a rationalization and strengthening of coun-tries' project portfolios. Cutbacks in operations andmaintenance expenditure, however, are worrisome.A World Bank review of countries' adjustment ex-perience found that reductions in nonwage opera-tions and maintenance and a marked deteriorationin infrastructure services were common. For in-stance, in Costa Rica during the 1980s current non-wage expenditures (principally operations andmaintenance) fell from 1.6 percent of GDP to a mere0.3 percent, and the share of the national and can-tonal road network in poor to very poor conditionrose to 70 percent.
transport infrastructure on an international corridoris less of a problem than are institutional con-straints. For example, one-third of the time requiredto ship freight between landlocked Mali and neigh-boring ports in Lomé (Togo) and Abidjan (Côted'Ivoire) is due to delays in customs clearance. Re-moving inefficient regulation of road transport andprivatizing transport operations, and deregulatingpower generation and distribution (as discussed inlater chapters), may facilitate some international ex-change of services in these sectors.
To summarize, infrastructure investment is notsufficient on its own to generate sustained increasesin economic growth. The demand for infrastructureservices is itself sensitive to economic growth,which is notoriously difficult to predict. The eco-nomic impact of infrastructure investment variesnot only by sector but also by its design, location,and timeliness. The effectiveness of infrastructureinvestmentwhether it provides the kind of ser-vices valued by users (responding to "effective de-mand")depends on characteristics such as quality
19
and reliability, as well as on quantity. Matching sup-ply to what is demanded is essential. Finally, the ef-ficiency with which infrastructure services are pro-vided is also a key to realizing potential returns.
Links to poverty
Infrastructure is important for ensuring thatgrowth is consistent with poverty reduction, a topiccovered extensively in World Development Report1990: Poverty. Access to at least minimal infrastruc-ture services is one of the essential criteria for defin-ing welfare. To a great extent, the poor can be iden-tified as those who are unable to consume a basicquantity of clean water and who are subject to un-sanitary surroundings, with extremely limited mo-bility or communications beyond their immediatesettlement. As a result they have more health prob-lems and fewer employment opportunities. Theburgeoning squatter communities surroundingmost cities in developing countries typically lackformal infrastructure facilities, a condition arisingfrom their nonpermanence of tenure. In India theproportion of the urban population living in slumareas grew during 1981-91, while the share of thepopulation living in poverty (estimated using tradi-tional poverty measures based on income and foodconsumption) declined. The lack of access to infra-structure is a real welfare issue.
Different infrastructure sectors have different ef-fects on improving the quality of life and reducingpoverty. Access to clean water and sanitation hasthe most obvious and direct consumption benefitsin reducing mortality and morbidity. It also in-creases the productive capacity of the poor and canaffect men and women differently. For example, thepoorwomen in particularmust commit largeshares of their income or time to obtaining waterand fuelwood, as well as to carrying crops to mar-ket. This time could otherwise be devoted to high-priority domestic duties, such as childcare, or to in-come-earning activities. Such gender-specific effectsneed to be considered in the evaluation of proposedprojects.
Access to transport and irrigation can contributeto higher and more stable incomes, enabling thepoor to manage risks. Both transport and irrigationinfrastructure have been found to expand the op-portunities for nonfarm employment in rural areas,often in indirect ways (Box 1.4). A seeming develop-ment dilemma is that while rural poverty reductionrequires higher incomes, raising farmgate foodprices could make urban poverty worse. By raisingthe productivity of farms and of rural transport,
20
both an increase in the incomes of rural workers anda reduction in food prices for the urban poor can beachieved. The green revolution (with irrigationplaying a central role) demonstrated that the wagesof, and demand for, low-skilled agricultural laborersrise in step with more intensive cultivation and in-creased yields. Over twenty years, one closely ob-served Indian village saw yields increase almostthreefold and agricultural laborers' wages rise from2.25 to 5 kilograms of wheat a day. Improved ruraltransport can also ease the introduction of improvedfarming practices by lowering the costs of moderninputs such as fertilizer. An adequate transport net-work reduces regional variations in food prices andthe risk of famine by facilitating the movement offood from surplus to deficit areas.
The benefits of transport and communications in-clude the access they provide to other goods andservices, especially in cities. Where the poor are con-centrated on the periphery of urban areas, as inmany developing countries, the costs and availabil-ity of public transport become key factors in theirability to obtain employment. Access to secure andreliable public transport has been identified inhousehold surveys in Ecuador as influential in de-termining the ability of low-income girls andwomen to participate in evening training classes.
The construction and maintenance of some infra-structureespecially roads and waterworkscancontribute to poverty reduction by providing directemployment. Civil works programs (as carried outin Botswana, Cape Verde, and India), which ofteninvolve the provision of infrastructure, have alsobeen important in strengthening famine preventionand providing income.
Links to the environment
Infrastructure provision results from the efforts ofindividuals and communities to modify their physi-cal surroundings or habitat in order to improve theircomfort, productivity, and protection from the ele-ments and to conquer distance. Each sectorwater,power, transport, sanitation, irrigationraises is-sues concerning the interaction between man-madestructures (and the activities they generate) and thenatural environment. Environment-friendly infra-structure services are essential for improving livingstandards and offering public health protection.With sufficient care, providing the infrastructurenecessary for growth and poverty reduction can beconsistent with concern for natural resources andthe global environment (the "green" agenda). At thesame time, well-designed and -managed infrastruc-
Box 1.4 Infrastructure's direct and indirect effects in rural India
A study of two villages in rural Karnataka state, south-ern India, offers a glimpse of the full impact that infra-structure can have on rural living standards. The re-searcher, who studied the Wangala and Dalena villagesin the 1950s and 1970s, described how the two villageshad been similarly poor and backward until a large-scaleirrigation project brought Wangala into a canal networkwhile Dalena's high elevation left it unirrigated.
Although canal irrigation directly promoted rapid in-tensification of cultivation in Wangala, institutions andthe villagers' way of life were relatively unaffected oth-erwise. In contrast, Dalena did not benefit directly fromthe canal. Its villagers were compelled to adjust theirway of life significantly in order to capture the indirecteconomic benefits from the irrigation project. The vil-lagers purchased land outside Dalena, sought positionsin the Public Works Department and a nearby sugar mill,and became involved in the transport of irrigated vil-
lages' sugarcane to the mill. Dalena quickly establisheditself as a service center in the region, and its residentsintegrated themselves into a much wider economicsphere than did those in Wangala.
The research emphasized how many Dalena villagerstraveled daily between their homes in the village andtheir places of work in nearby towns. This observation,echoed in other studies, suggests that the developmentprocess need not entail migration from rural areas tourban centers. In the Uttar Pradesh village of Palanpur,per capita living standards rose between 1957 and 1993in the face of population growth, in part because of ex-panding nonfarm employment. Residents of Palanpurcommute daily to the towns of Chandausi and Morad-abad, largely by rail. This type of rural commuting morecommonly occurs along roads by foot, bicycle, motorcy-cle, bus, or car.
ture can promote the environmental sustainability ofhuman settlements (the "brown" agenda). World De-velopment Report 1992 focuses on environmentalissues, including those of infrastructure sectors, indetail.
The relationship between each infrastructure sec-tor and the environment is complex. The most posi-tive impacts of infrastructure on the environmentconcern the removal and disposal of liquid andsolid wastes. But much depends on how disposal fa-cilities are planned and executed. Underinvestmentin municipal sewerage relative to water supply indensely populated cities such as Jakarta has beenfound to lead to harmful contamination of water re-serves, to exacerbate flooding, and to reduce thehealth benefits from water investments. Provision ofsewerage without wastewater treatment can lead tosevere downstream pollution and public healthproblems where receiving waters are used fordrinking-water supply or for recreation, irrigation,and fisheriesas illustrated by the cholera out-breaks in Peru and neighboring countries in recentyears. Poor management of solid waste complicatesurban street drainage and has been linked with theproliferation of disease-bearing mosquitos in stand-ing water. The growing problem of hazardous andtoxic wastes as countries industrialize poses partic-ular concerns about safe disposal. For example, un-controlled dumping has led to soil contamination inthe Upper Silesian industrial region of Poland andto subsequent food crop contamination.
Power plant and vehicle emissions are importantcontributors to air pollution, so their air quality im-pacts deserve careful analysis when facilities are ex-panded. In developing countries, almost one-thirdof commercial energy is devoted to electricity gener-ation, which is the fastest-growing component of theenergy sector. By the year 2000 Asia may well sur-pass all of Europe in sulfur dioxide emissions, andby 2005 it may surpass Europe and the United Statescombined in power plant emissions. Vehicles are asignificant source of airborne toxic pollutants, ac-counting for up to 95 percent of lead contamination.In Central and Eastern Europe, road transport is es-timated to account for 30 to 40 percent of total emit-ted nitrogen oxides and hydrocarbons. AlthoughOECD countries account for three-quarters of theworld stock of motor vehicles, a rapid increase in ve-hicle use is expected in parts of Central and EasternEurope, East Asia, and South America. In large andgrowing developing country cities, such as Bangkokand Jakarta, vehicle congestion already gives rise toconsiderable environmental and economic costs, ForBangkok, it is estimated that if reduced traffic con-gestion permitted a 5 percent increase in peak-hourvehicle speeds, the value of travel time saved wouldamount to more than $400 million a year. A 20 per-cent improvement in air quality in Bangkok, as a re-suit of a reduction in pollutants related to vehicle orpower plant emissions, would produce annualhealth benefits valued at between $100 and $400 percapita for Bangkok's 6 million residents.
21
Expansion of transport infrastructure can reducetotal pollution loads as congestion falls, average ve-hicle speeds rise, and routes are shortened. But roadimprovements can also encourage vehicle use andincrease emissions. Therefore, additions to infra-structure capacity are oniy part of the solution. Im-proved management of traffic and land use and pro-motion of nonmotorized modes, cleaner fuels, andpublic transport are also needed (see Chapter 4). In-tegrated urban planning and transport policy canlead to more efficient use of both land and transportcapacity with favorable environmental results. Inthe city of Curitiba, Brazil, an emphasis on encour-aging enterprises and residential developments tolocate around carefully designed public transportroutes has contributed to low gasoline consump-tion, low transport costs relative to household in-comes, and very low rates of traffic accidents de-spite one of the highest rates of private vehicleownership in the country.
Beyond urban areas, overuse of water for irriga-tion (which accounts for about 90 percent of waterwithdrawals in most low-income countries) dam-ages soils and severely restricts water availabilityfor industry and households, which often have ahigher willingness to pay for the quantities of waterthey use. The inefficient burning of biomass fuel(plant and animal waste) for household energy con-tributes to deforestation and thus to erosion and lossof soil nutrients, as well as to indoor air pollution.Some infrastructure investments, especially roadconstruction, can put unspoiled natural resources atrisk and threaten indigenous communities. Reser-voirs associated with hydroelectric projects, floodcontrol, or irrigation can give rise to environmentalproblems, both upstream (inundation of land) anddownstream (sedimentation).
Origins of the public sector role in infrastructure
Infrastructure's large and varied potential impactson development derive from certain technologicaland economic characteristics that distinguish itfrom most other goods and services. These charac-teristics make infrastructure subject to special pol-icy attention.
Production characteristics
Historically, society's needs for water supply, irriga-tion and flood control, and transport have led to theconstruction of engineered physical worksmanyof them quite large, elaborately designed, and en-during. Today's distinctively modern infrastructure
22
sectors are the result of a technology-driven "infra-structure revolution" that has changed the way inwhich age-old demands for water, lighting, commu-nications, and waste disposal are met.
Not until the invention of cast-iron pipes andsteam-driven pumps did extensive water infrastruc-ture spread, beginning with a piped water networkin London in the 1850s. This lowered costs (espe-cially in urban areas) and dramatically increaseduse. Before the development of gas networks at thestart of the 1800s, infrastructure for lighting wasrare. The invention of alternating-current transmis-sion near the end of the century lowered costs ofelectricity and led to new and expanded uses ofelectric power, especially in urban transport.
The history of other infrastructure sectors is sim-ilar. The public telegraph and telephones replacedhand-carried messages, and piped sewerage re-placed individual disposal of wastes in many com-munities. Irrigation and transport have for centuriesutilized networks of irrigation canals and roads, al-though development of alternative modes of trans-port (including inland canals and railroads) has pro-ceeded since the early 1800s.
The most general economic characteristic ofmodern infrastructure is the supply of servicesthrough a networked delivery system designed toserve a multitude of users, particularly for publicutilities such as piped water, electric power, gas,telecommunications, sewerage, and rail services.The delivery system is in most cases dedicated, thatis, it carries only one good. Investments in the deliv-ery system (such as underground water pipes orelectric wires) are mostly irrecoverable because theycannot be converted to other uses or moved else-whereunlike the investment in a vehicle, for ex-ample. Once paid, these costs are said to be "sunk."Because the delivery system is networked, coordi-nation of service flows (traffic, electricity, communi-cations signals) along the system is critical to its effi-ciency. This interconnectedness also means that thebenefits from investment at one point in the net-work can depend significantly on service flows andcapacities at other points.
The scope for competitive supply of infrastruc-ture varies greatly across sectors, within sectors, andbetween technologies. Where the unit costs of serv-ing an additional user decline over a wide range ofoutput, economies of scale are createdan impor-tant source of "natural monopoly." This is a com-mon term, although one best used cautiously be-cause many infrastructure monopolies are in factunnatural, driven by policy and not technology. Butsectors differ greatly in the range of declining costs.
For example, the optimal dimensions of a high-volt-age transmission grid may well be national, but thevolume-related unit cost savings for water can berealized at the municipal or submunicipal level.Even within sectors, different production stageshave different characteristics. In power, size savingsfor generation are often exhausted at a capacity thatis small relative to the size of a well-developed mar-ket. Activities also differ in the importance of sunkcosts, another potential source of natural monopoly.In railways and ports, for example, sunk costs areless significant for investments in rolling stock orfreight-handling equipment than for the fixed facili-ties. It is easier for firms to enter and exit activitieswith a relative absence of sunk costs and therebychallenge one another's potential market power.Such activities are said to be "contestable." Techno-logical and economic differences in production cre-ate the possibility of "unbundling" the componentsof a sector that involve natural monopoly fromthose that can be provided more competitively.
Many infrastructure services can be produced byvery different technologies. Sanitation based onimproved latrines or septic tanks provides thesame underlying service as does seweragedis-posal of wastes, but without networked invest-ments. Small-scale irrigationparticularly irriga-tion based on wells or boreholesand small-scalerenewable-energy-based power generation (such asmicro-hydro schemes) also need not involve inter-cormections with large networks but can provideservice highly responsive to users. Telephone ser-vices can be provided over wire-based networks orthrough radio-based systems.
Consumption characteristics
As seen earlier, the demand for infrastructure ser-vices derives from the activities of both industriesand individuals. Ensuring a flow of services of atleast minimum quality and quantity is often consid-ered by governments to be of strategic importance,since any interruption or restriction of supplywould be seen as a threat to society. However, be-cause infrastructure investments are often "lumpy"(new capacity must be created in large increments),it is difficult for planners to match the availability ofsupply with demand at all times. Costly episodes ofover- or undercapacity often result.
Beyond consuming an "essential minimum" ofcertain infrastructure services, users have very di-verse demandsalthough the output of large-scale,monopoly providers is often not sufficiently differ-entiated to meet these demands. For example, a
steel mill and a residential community may both de-rive water from the same supplier, but each usergroup values the quality of the water in quite differ-ent ways. Yet, because many infrastructure facilitiesare locationally fixed and their products are non-tradable, users cannot readily obtain substitute ser-vices that better suit their needs. Moreover, it isoften difficult for users to obtain information aboutservice alternatives or characteristics. They cannot,therefore, "shop around" for the best source of sup-ply and are vulnerable to any abuse of monopolypower. With many infrastructure activities, how-ever, supply can be better tailored to differences indemand once suppliers understand themfor ex-ample, transport can be offered at varying serviceand fare levelsand provided that consumers haveadequate information to declare their choices. Ser-vice markets can also be opened to alternative sup-pliers and technologies in order to provide a differ-entiated product (such as cellular and enhancedservices in telecommunications).
Many infrastructure services are almost (al-though not perfectly) private goods. Private goodscan be defined as those that are both "rival" (con-sumption by one user reduces the supply availableto other users) and "excludable" (a user can be pre-vented from consuming them). In contrast, "publicgoods" are neither rival in consumption nor exclud-able. Markets work best in providing pure privategoods or services. Most of the services that the infra-structure sectors produce are excludable in a specificsensetheir use depends on gaining access to a fa-cility or network, for example by connection to thepiped water, gas, or sewer system, and service usemay be metered and charged for. In the case of rail-ways, ports, and airports, access to the entire infra-structure can be restricted. However, once a user isconnected to the network utility or gains access tothe transport facility, the degree of rivalry with otherusers depends on the costs (including congestion)imposed on existing users or on the service supplierwhen an additional service unit is consumed.
It has been common in many countries not tocharge users for the volume of some utility servicesconsumed because the marginal supply cost wasconsidered negligible, congestion was absent, ortechnological constraints (such as the absence ofwater meters) prevented volume pricing. However,recent developments, such as the increased scarcity(and supply cost) of water, growing congestion asnetwork capacity becomes fully utilized, and techni-cal innovations in metering consumption, havemade it possible and desirable to price these ser-vices like other private goods.
23
Roads are not private goods, although for rea-Sons that differ with the type of road. Rural roads (atypical public good) and uncongested interurbanroads are not completely rival because an additionaldriver does not reduce the value of anyone else'suse of the road. Access to some interurban roadscan be prevented by making them toll roads (a clas-sic "club" good, i.e., a good that is excludable butnonrival). By contrast, urban roads are congestedduring peak periods, but until recently it has beendifficult to exclude users from urban roads or tocharge users different amounts during peak and off-peak periods. New electronic techniques of moni-toring road use may eventually make it technicallyfeasible to treat many urban roads almost as privategoods.
Water outside of piped networks is ofteninpractice and in principlea "common property"resource. While water consumption is rival betweenusers, monitoring the use of groundwater from un-derground aquifers or from other natural sources isdifficult and costly, and therefore groundwater useis rarely excludable. By the same token, controllingthe consumption of common property resources isalso difficult. How much the extraction of water(from aquifers or natural flows) affects other poten-tial users depends on location-specific hydrologicalfeatures that are important in water policy.
Although most infrastructure goods are private,they produce spillovers or external effectsmanyof which (as shown earlier) affect the environment.Ignoring the important negative externality of emis-sions from fossil fuel power generation could leadto excess power being produced with the wrongmix of fuels. By contrast, some cities have neglectedto develop a well-designed public transport system,even though such a system can have positive envi-ronmental effects and also promote social equity. Toensure that society obtains positive benefits suchas public health benefits from water and sanita-tionthe private goods must also be delivered ef-fectively.
Thus, although infrastructure services differ fromother goods, they also differ among themselves(Figure 1.3). The characteristics of various infra-structure activities have important implications forhow services should be provided. To the extent thatspecific infrastructure activities entail natural mo-nopoly or depend on a network characterized bynatural monopoly, they will not be provided effi-ciently by an unfettered market. The network com-ponent can, however, be separated (unbundled)from the more competitive activities of the sector,with regulation to ensure fair access to the network.
24
Infrastructure activities that create externalities orproduce essential services to captive users may alsowarrant some regulation, but this can be narrowlyfocused on these market imperfections while per-mitting wide scope for competition in other compo-nents of the sector.
Certain characteristics of infrastructure also cre-ate challenges in financing. Where a minimum levelof consumption of a particular service (such aswater, heating, or power) can be identified as a "life-line" for some users, society may judge that theyshould not be excluded if they cannot afford to pay.Financing strategies also have to be designed to takeaccount of the risk that arises because many infra-structure investments are large and long-lived,while the revenue stream is often slow to develop.Such characteristics can justify some public financ-ing of infrastructure from general revenues, but tosupplementnot entirely substitute forthe rev-enues obtained from users and commercial sourcesof finance.
Public sector dominance in infrastructure
Infrastructure clearly represents a strong public in-terest, and so merits the attention of governments.However, the special characteristics of infrastruc-ture do not explain or justify the fact that govern-ments and public sector agencies have dominatedalmost all aspects of this sector in developing coun-tries in recent decades. Private participation wasimportant in the nineteenth century and the firsthalf of the twentieth century in many countries -and some pockets of private provision still re-mainbut the overwhelming trend until the early1980s was government or parastatal provision,largely through vertically integrated, monolithic en-tities. By then, only a small percentage of the powersector was in private hands. Virtually no privatetelecommunications firms existed, and most earlyprivate railways had disappeared with nationaliza-tion. Although toll roads played a part in the earlyhistory of many countries, they also became rare,and road construction (and especially maintenance)was executed largely by government employees, orforce account. Other serviceswater, sewerage,waste disposal aisci tended to be both owned andoperated by governments at either the national orthe local level.
The dominant public sector role in infrastructurehas arisen for a number of reasons: recognition ofinfrastructure's economic and political importance;a belief that problems with the supply technologyrequired a highly activist response by governments;
Figure 1.3 Infrastructure services differ substantially in their economic characteristicsacross sectors, within sectors, and between technologies.
Excludable Nonexcludable
10
Interurban highways(toll roads)
Club goods
Lower .i4
Urban bus
Rural sanitation(on-site disposal)
Rail, airport, and port services
High-voltage transmission
Rail, port, and airport facilities
Fossil fuel power generation
Piped water supply
Surface water irrigation
Sanitary landfill
Urban sewerage
Commonproperty
GroundwaterUrban roads
Rural roadsStreet sweepingTraffic signaling
Public goods
Note: Excludable means that a user can be prevented from consuming the good or service. Rival means that consumption byone user reduces the supply available to other users.
and a faith that governments could succeed wheremarkets appeared to fail. Many countries made im-pressive strides in infrastructure expansion underthe earlier stages of this public leadership. But morerecent experience has revealed serious and wide-spread misallocation of resources, as well as a fail-ure to respond to demand. Moreover, the blunt in-struments of public ownership, financing, andoperation have not demonstrated any advantage inachieving poverty reduction goals or environmentalsustainability. These deficiencies in performance arenot happenstancethey are embedded in the pre-vailing system of institutional incentives for thesupply of infrastructure.
The record of performance
Achievements
Although the data are spotty, impressive expansionin infrastructure has been achieved in recentdecades, as measured by stocks and production ofservices (Table 1.3). In low-income economies, tele-communications, sanitation, and water supply reg-istered the highest rates of increase in availabilitybetween 1975 and 1990, starting from a very lowbase in each sector. In middle-income economies,growth in this period was concentrated mainly inthe power and telecommunications sectors, where
25
Externalities Higher
Rival
A
Non-rival
Table 1.3 Expansion of infrastructure coverage in low-, middle-, and high-income economies,recent decades
capacity more than doubled between 1975 and 1990.Even in middle-income economies, however, accessto water and sanitation is still lacking for significantshares of the populationfor water, one-quarter ofthe population in this group remains unserved, andfor sanitation, one-third. The most dramatic expan-sions in paved roads occurred during 1960-75 forboth groups, after which growth slowed.
Infrastructure coverage has increased in bothrural and urban areas. Urban populations are signif-icantly better served than rural populations in ac-cess to drinking water, sanitation, and power. Thegaps in coverage for water and power have beennarrowing (Figure 1.4). Rural and urban areas donot have the same effective demand for infrastruc-ture services and thus may require different rates ofinfrastructure coverage to achieve desired develop-ment benefits. There is an economic case for provid-ing relatively more power and telecommunicationsconnections, and more extensive transport net-works, in locations with a higher density of popula-tion and industry.
Urbanization in itself is an important factor stim-ulating demand for infrastructure. When infrastruc-hire capacity in water supply, sanitation, power,telecommunications, roads, and public transport isinadequate in expanding urban areas, serious con-straints on (environmentally sustainable) economicgrowth and on poverty reduction result. In therapidly growing periurban (and, in many cases,unauthorized) settlements that ring many cities,conventional delivery of formal services is oftenprevented by legal, topographical, or economic con-straints. Projected growth in urbanization in comingdecadesespecially in Africa and South and EastAsiawill inevitably increase pressures for greateraccess to infrastructure. However, some rural-to-
26
urban migration may be forestalled through policiesthat provide appropriate infrastructure in ruralareas and that prevent the degradation of natural re-sources (especially soils, forests, and water sup-plies).
An analysis of how countries measure up on in-frastructure coverage compared with other mea-sures of performance is revealing. Although cover-age tends to be correlated with GDP, efficiency andeffectiveness of infrastructure provision are not.Plots of coverage against performance in water,powel telecommunications, roads, and railwaysshow little relationship across a wide sample of low-and middle-income countries (summarized in Fig-ure 1.5). Moreover, there is no close correlation be-tween a country's efficiency of provision in one sec-tor and its performance in another. These findingsindicate that efficiency and effectiveness of infra-structure provision derive not from general condi-tions of economic growth and development butfrom the institutional environment, which oftenvaries across sectors in individual countries. Thissuggests that changes in the institutional environ-ment can lead to improved performance, even whenincomes are low, because in each sector some low-income countries perform well. As a corollary, a re-cent OECD review of infrastructure noted that evenmany high-income countries encounter the perfor-mance issues described below.
Challenges
To determine future demand for infrastructure, it isnecessary to consider the efficiency with which ex-isting capacity is being used and how well the ser-vices generated are responding to users. Althougheach sector has special problems, there are com-
Sanitation (percentage of population with access) 23 42 3.8 44 68 2.7 95+
Paved roads (kilometers per million persons) 308 396 1.6 1,150 1,335 0.9 10,106
Water (percentage of population with access) 40 62 2.7 54 74 2.0 95+
Note: Percentage increases are compound growth rates.Source: Appendix tables A.1 and A.2.
Figure 1.4 The rural-urban gap in access topower and water in developing countriesnarrowed over the past decade.
Percentage of population withaccess to infrastructure90
60
30
mon patterns operational inefficiencies, inade-quate maintenance, excessive dependence on fiscalresources, lack of responsiveness to users' needs,limited benefits to the poor, and insufficient envi-ronmental responsibility.
INEFFICIENCY OF OPERATIONS. The broadest indica-
tor of inefficient performance by an infrastructuresystem is the extent of output lost in delivery. Unac-counted-for water (that portion of supply for whichconsumption is not recorded, largely because oftechnical and managerial failures) is typically two tothree times higher in developing country systemsthan in countries that achieve the industry stan-dards. In 1987 one-quarter of the power utilities indeveloping countries had losses of electricity in thetransmission and distribution network that weretwice those in efficiently operated systems. In some
African countries, spending $1 million to reduceline losses could save $12 million in generating ca-pacity. Irrigation efficiency (the proportion of waterdelivered to the field) in developing country proj-ects is typically 25 to 30 percent, compared with 40to 45 percent under best practice.
Inefficient use of labor is especially common andcostly in infrastructure. At various periods, two-thirds of the labor in railways in Tanzania and Zaire,80 percent of port staff in Argentina (before recentprivatizations), and one-quarter of highway depart-ment staff in Brazil have been estimated to be re-dundant. The combination of overstaffing and un-derpricing of railway services produced a wage billalmost as large as (and sometimes larger than) totalrailway revenues in Argentina (before recent re-forms) and in Colombia, Egypt, Nigeria, Turkey,and Uruguay. Overstaffing is also common inwater, power, and telecommunications. At thesame time, in the production of public works andrural infrastructure, developing countries oftenuse equipment-based methods of construction andmaintenance rather than employment-intensive ap-proaches that can produce high-quality results,while being more consistent with relative capitaland labor costs.
INADEQUATE MAINTENANCE. Closely related to op-erating inefficiencies is lack of maintenance: roadsdeteriorate, irrigation canals leak, water pumpsbreak down, sanitation systems overflow, installedphone lines fail, and power generators are not avail-able when needed. Capacity is then lost, output de-clines, and substantial additional investment isneeded simply to sustain existing levels of service.
In the road sector, inadequate maintenance im-poses large recurrent and capital costs. The engi-neering and physical properties of paved roads aresuch that, as a road begins to deteriorate, lack of reg-ular routine maintenance will hasten deterioration.Neglect of (relatively inexpensive) routine mainte-nance can compound problems so much that the en-tire surface of a road has to be replaced. Examina-tion of completed Bank highway projects showsthat, on average, estimated returns on projects in-volving primarily maintenance are almost twice ashigh as those on projects involving mainly new con-struction. Yet, in Sub-Saharan Africa, almost $13 bil-lion worth of roadsone-third of those built in thepast twenty years have eroded because of lack ofmaintenance. In Latin America, for every dollar notspent on maintenance, $3 to $4 are estimated to berequired for premature reconstruction. Maintenanceexpenditures often are not allocated by economic
27
1980 1990 1980 1990
Urban areas Rural areas
Power
Water
Sanitation
Source: Israel 1992; WHO 1980, 1990.
Figure 1.5 Efficient and effective delivery of infrastructure services does not alwaysaccompany increased availability.
0 50 100 150
Telephone main lines per thousand persons0 20 40 60 80 100
Percentage of households with electricity
28
Faults per 100 main lines Percentage of power delivered0 100- -
Egypt TurkeySaudi Arabia
.Carneroon Zambia.Zlmbabwe.
Trinidad & TobagoMexico
50 Syria . 80Bolivia #t a
S
SS .
S Colombia.,Panama'
o ArgentinaChile 0 .
100Tunisia
0 SJordan Romania 60 Dominican- Republic
150 Chad40
Mauritius
200.
20Zimbabwe
250 0
100
Percentage
Kilometers of paved roadsper million persons
of accounted-for water
Rail passenger-kilometers +ton-kilometers per million US$ of GDP
Cole divoireNicaragua
80 0 TogoIndonesia Gabon Brazil S
Somalia .0 5Jamaica Middle East and North Africa
60 .Ethiopia
.Nepal
SS
Ecuador . Jordan Latin America and the Caribbean
40 Philippines East Asia and PacificSub-Saharan Africa
20 South AsiaEurope and Central Asia
0
Percentage of paved roads Percentage of dieselin good condition locomotive availability100 100Mauritius' S Algeria
priorities. For example, Cameroon, which still has apredominantly rural population, has neglected its30,000-kilometer unpaved road network over thepast ten years in favor of investment in and mainte-nance of 3,700 kilometers of intercity paved roads.The result is that some 80 percent of the unpavednetwork requires either complete reconstruction orheavy reshaping and compaction.
In railways, inadequate maintenance (as well asother operating deficiencies) is evident in the smallshare of locomotives available for service. In 1991only 60 percent of all locomotives were available forservice in Latin America and 70 percent in the Mid-dle East and North Africa region, compared with 90percent in North America. Such deficiencies causesome railways to turn away freight traffic, which inturn compounds the sector's financial difficulties.
In irrigation, too, poor maintenance is costly andresults in distribution channels filling with silt andweeds, canal linings cracking at an increasing rate,and outlets breaking or being bypassed. Drainagealso fails, causing salt buildup in the soil. In Chinaalmost 1 million irrigated hectares have been takenout of production since 1980, and in the former So-viet Union, even with continuing investment in irri-gation, almost 3 million hectares were lost between1971 and 1985one-quarter of the new irrigatedarea. Worldwide, works covering 60 percent of theirrigated area require upgrading to remain in goodworking condition.
In both rural and urban water supply and in thepower sector, inadequate maintenance is a commonproblem. A study of water and sewerage in Bogotafound that the costs of unaccounted-for waterarising in part from poor maintenance of the distrib-ution systemwere 42 percent of the supplier'stotal operating income. Poor maintenance practicesaccount for some of the low availability of power-generating capacity which averages less than 60percent for thermal plants in many developingcountries, compared with more than 80 percent insystems operated at best-practice standards.
Sometimes problems of operation and mainte-nance are rooted in the initial design or constructionof infrastructure. For example, a recent review ofcompleted World Bank irrigation projects foundthat basic design flaws (such as inappropriate trans-fer of desert technologies to tropical monsoon cli-mates) were widespread. Operations and mainte-nance can be made more difficult by inappropriatedesign standards that increase the requirements forskills in short supply or involve heavy dependenceon imported spare parts where foreign exchange isscarce. Poor construction and design of power and
water treatment plants, or inappropriate location,make it difficult to carry out operations and mainte-nance and to meet environmental objectives. Thereare also many examples of investments that wereeconomically nonviable to begin with and thatshould never have been madesuch as over-designed or "gilt-edged" roads and power plants.
Procurement problems are often a factor in weakoperational performance. Systematic delays in pur-chasing by sector entities and inadequate supervi-sion of contracts are estimated to increase costs ofimported materials to some African countries by 20to 30 percent. Contracting and bidding proceduresmay also favor large-scale enterprises, which tendto use more equipment-based methods of construc-tion and maintenance than is appropriate given rel-ative factor costs. The lack of standardization ofequipment, such as water pumps obtained from di-verse foreign donors, creates delays in repair and in-creases the costs of replacement parts. There is needfor donors to standardize their procurement rules toease the administrative burdens on recipient coun-tries. Donor aid that excludes finance for local costscan also bias the choice of technology for publicworks in favor of capital-intensive methods that areunsustainable for the recipient country.
FINANCIAL INEFFICIENCY AND FISCAL DRAIN. Poor
infrastructure policies and inefficient provision ab-sorb scarce fiscal resources and damage macroeco-nomic stability. Because prices are often held wellbelow costs, the subsidies flowing into public infra-structure enterprises and agencies have been enor-mous in many countries. In Bangladesh, India, In-donesia, Pakistan, and the Philippines, irrigationreceipts have been well below the costs of opera-tions and maintenance. During the 1980s power tar-iffs in developing countries were on average aboutone-half the costs of new supply and were muchlower than in OECD countries. (The record on pric-ing is discussed further in Chapter 2.) In recentyears, 60 percent of Ghana Railway revenues con-sisted of government subsidiesa not-uncommonperformance for this sectorand recurrent subsi-dies to railways have amounted to as much as 1 per-cent of GDP in a number of countries. In Zambia thetotal cash shortfall in transport absorbed 12 percentof the government's current revenue in fiscal 1991.Telecommunications tends to be an exception to thegenerally poor cost recovery elsewhere in infra-structure, although its revenues are often siphonedoff by government for other uses, leaving the sectorunderfunded. Inadequate tariffs are often com-pounded by poor financial management. In a sam-
29
pie of Latin American water utilities, collection ofaccounts receivable took almost four months on av-erage, compared with good-practice standards offour to six weeks. In addition to creating an addedburden on taxpayers, poor financial performance bymany infrastructure providers means a loss ofcreditworthiness for the entity concerned. It alsoresults in a low reliance on internal revenues to fi-nance investment and therefore an inability (andlack of incentive) to expand or improve service.
UNRESPONSIVENESS TO USER DEMAND. The result of
inefficiency and poor maintenance is low-quality,unreliable service, which alienates users. Reliabilityis a critical aspect of user satisfaction that is often ig-nored. Even where users have telephones, high callfailure rates (more than 50 percent in many cases)and high fault rates drastically diminish the value ofthe service. Unreliable quantity or quality of waterleads to enormous investments in alternative sourcesthat are especially costly to those who can least af-
Box 1.5 Households' responses to unreliability of water supply
In 1991, micro-level research on household responses todeficient water supply by public utilities was under-taken in Faisalabad (Pakistan), Istanbul (Turkey), andJamshedpur (India). These surveys revealed that nearlyall households in the three cities are dependent on multi-ple sources of water, including house taps, wells, tube-wells, public taps, rivers, and street vendors. Not all al-ternatives are available to all households. Because accessto a source increases with income, poorer householdsbear a disproportionate share of the burden of deficientinfrastructure. The private expenditures incurred forwater supply indicate consumers' willingness to pay forreliable water.
In Istanbul, the poorest households surveyed spend alarger share of their income (about 5 percent) to supple-ment inadequate water supply than do wealthier ones(which spend about 1 percent). These expenditures oninformal sources of water, including self-provision fromwells or storage facilities, are in addition to the usercharges for publicly supplied water, which amount to 1to 2 percent of annual income.
In Jamshedpur, the connection charges for pipedwater vary between $1.66 and $16.66. The residents ofthe periurban areas, served by the local municipal au-thorities, incur capital costs of $50 to $65 in installingtubewells and $150 to $300 in digging wells to avoid de-pendence on the (unreliable) public water supply. De-spite the existence of a piped water system, at least 17percent of the population meets 90 percent of its waterneeds from wells and handpumps. Over and above themonetary costs that consumers bear, households inJamshedpur spend, on average, two hours a day fetch-ing and storing water. The burden of these activities fallsin nearly all cases on women.
The pattern of private augmentation of the publicwater supply at substantial private costs to consumers isobserved also in Faisalabad, Pakistan. Less than 20 per-cent of the households with piped water use this sourceexclusively; 70 percent have motor pumps and 14 per-cent have handpumps.
Box 1.6 Public failures raise private costs
According to a 1988 study of Nigerian manufacturers, 92percent of the 179 firms surveyed owned electricity gen-erators. In the face of chronically unreliable public ser-vices, many had also acquired radio equipment for com-munications, vehicles to transport personnel and freight,and boreholes to assure their own private water supply.For firms with fifty or more employees that could prac-tice economies of scale, these extra costs amounted tosome 10 percent of the total machinery and equipmentbudget. For small firms, the burden could be as high as25 percent. Yet because Nigerian regulations preventfirms from selling their excess power capacity, busi-nesses both large and small were operating private gen-erators and water systems on average at no more than 25percent of capacity.
Of 306 Indonesian manufacturers recently polled, 64percent had generators and 59 percent (compared withNigeria's 44 percent) had boreholes for their own water
supply. Indonesia's largest companies invested as muchas 18 percent of their capital in private infrastructurealmost twice Nigerian manufacturers' level of 10 per-centyet their generators, too, were underused and op-erating at about 50 percent of capacity.
Today in Indonesia, as in Nigeria, firms too small toafford private power or water are at the mercy of unreli-able public utilities and subject to chronic and costly in-terruptions in service. Yet while the largest Indonesianfirms pay $0.07 per kilowatt-hour to produce electricity(not far above international norms), self-provided elec-tricity costs the smallest firms $1.68 per kilowatt-hourtwenty-four times as much.
Thailandwhere public electric utilities are effi-ciently runhas been able to break this pattern. Of the300 manufacturers polled, only 6 percent had privategenerators and 24 percent had private water supplies.
30
ford it (Box 1.5). In Indonesia and Nigeria, privatebusinesses incur heavy costs in order to guaranteepower supply: 92 percent of firms sampled in Nige-ria and 64 percent in Indonesia had installed privategenerating capacity (Box 1.6); in Thailand, only 6percent of companies needed generators. These largedifferences in self-provision reflect the performanceof the formal suppliers. In Nigeria, only 43 percent ofinstalled capacity was in service by 1990 (despitemassive overinvestment in public power-generatingcapacity throughout the 1980s); in Thailand, thepower utility is efficiently run.
In telecommunications, unmet effective demandcan be roughly measured, because in many coun-tries users must apply for connection, which ofteninvolves paying a heavy initial fee. Of ninety-fivedeveloping countries, more than one-third werefound to have a waiting period of six or more yearsfor a connection, compared with less than a monthin most industrial countries (Figure 1.6). Countriesthat can deliver service in less than a year includesome with little current pressure on available capac-ity (such as Bolivia) as well as others in which in-vestment is proceeding rapidly (Malaysia). In addi-tion to the shortage of basic connections, in manycountries providers fail to offer differentiated ser-vices matching types of use. For example, busi-nesses increasingly require telecommunications fa-cilities that accommodate high-speed datatransmission as well as voice signals. A muchhigher priority could be given in many developingcountries to the provision of pay phones to extendbasic access to improved communications to alarger share of the population.
Excess demand for infrastructure, coupled withvery low rates of compensation to infrastructurestaff, breeds corruption in both service and invest-ment decisions. Where connections are rare andservice is poor, employees often demand side pay-ments from users to install or repair connectionsespecially in telecommunications, irrigation, andwater supply.
NEGLECT OF THE POOR. The poor typically usefewer infrastructure services than the nonpoor, butnot only because of low incomesthey also havevery low access. In Peru, for example, only 31 per-cent of the poorest fifth of households are connectedto a public water network and 12 percent to a publicsewer compared with 82 percent of the top fifthfor water and 70 percent for sewerage. The poorgenerally have less access than the rich in urbanareas as well (Table 1.4).
Many countries have introduced subsidiesthrough low tariffs with the aim of improving the
Figure 1.6 There is very high unmetdemand for telephone connection.
Number of countries25
20
15
10
0
Less 1.0- 3.0-than 1 2.9 5.9
Waiting time for telephone main lines, 1992(years)
Sample: 95 developing countries.Source: ITU 1994.
6.0-- 9.0- 12.0 or8.9 11.9 more
poor's access to infrastructure services, but most ofthese subsidies have been captured by middle- andhigh-income households (as documented in Chap-ter 4). In addition, the providers often are not ade-quately compensated for the subsidies, so that over-all expansion of service is constrained, The structureof tariffs can be an additional barrier. In Brazil, localtelephone call rates are low, but connection chargesare high. This prevents lower-income users fromgetting service. Flat-rate electricity charges in ruralIndia have benefited mainly richer households, be-cause the poor lack the income to purchase thepumps and consumer appliances that account formost electricity use.
While failure to reach the poor has often been as-sociated with flawed infrastructure pricing policies,too little emphasis has been placed on providing thepoor with suitable options for the kinds of servicesof most value to them (and for which they are will-ing to pay). For example, municipal sanitation agen-cies often promote technical designs for conven-tional sewerage that are unaffordable and evenenvironmentally unsuitable in some low-incomesettlements. In large cities such as New Delhi, thereliance of the poor on foot travel is a serious con-
31
straint to their mobility (Figure 1.7) A study oftransport options in Latin American cities foundthat in São Paulo, Brazil, personal travel by the poorhad declined more sharply than for any other in-come group over a decade-partly because publictransport services were ill designed for low-incomeusers. The poorest residents on the periphery of Riode Janeiro spent more of their income than the richfor transport, with longer waits, less frequent ser-vice, and more time spent in crowded vehicles.
Appropriate services for the poor are often lack-ing when decisions on investment and service aredriven by assumptions about a "needs gap" ratherthan by an assessment of effective demand. In theMakete District in Tanzania, a survey of householdsundertaken to determine their transport needs inpreparation for a proposed investment project re-vealed that improvement of the road network alonewould benefit only a few residents and that com-plementary measures were needed-includingsupport to transport services (the introduction ofnonmotorized means of transport to replace head-loading), simple improvements to paths and tracks,and rehabilitation of grinding mills. A retrospectiveevaluation carried out after completion of the proj-ect found that these low-cost improvements werehighly successful-and would likely have been leftout of the project if no inquiry into the actual de-mand of the communities had been undertaken.
NEGLECT OF THE ENVIRONMENT. The impact of in-frastructure on the environment has often been verynegative (Box 1.7 recounts one of many examples,and one where regional cooperation is needed to de-velop a solution). The highly visible effects of cer-tain large-scale facilities-such as dams and roadsin sensitive ecological areas or where resettlementoptions are unsatisfactory to populations -have at-
32
Not available.Source: Glewwe 1987a, b; Glewwe and Twum-Baah 1991; World Bank 1993e.
tracted understandable public attention. Yet equallyserious, and more pervasive, is the damage or lossof potential benefits to the environment because offailure to control unnecessary emissions and waste-ful consumption of water. This is due in particularto the underpricing of power, vehicle fuels, andwater for irrigation and municipal uses and to theneglect of maintenance. Inadequate maintenancepractices leading to inefficient thermal power gen-eration account for a large share of energy-relatedpollution. Neglect of sound environmental manage-ment practices in transport-including safe han-dling of hazardous cargos and appropriate disposalof waste from ships, port dredging, and vehiclemaintenance-is a common failing. Unregulated,badly designed, or poorly managed municipalwater and sanitation infrastructure has often beenone of the biggest sources of urban environmentalpollution. The focus of public spending on urbansolid waste management often stops at collection-few developing country cities meet environmentalstandards for sanitary landfills.
Many of the problems in infrastructure perfor-mance are mutually reinforcing, creating seriouseconomic and financial costs that make it more diffi-cult for countries to achieve greater coverage andmore modern services to better meet social and en-vironmental goals. Systemic problems point to sys-temic causes-and solutions.
Diagnosis and directions for change
The conditions for improved performance: causesand cures
Where infrastructure is operated inefficiently anddelivers poor service, the solution cannot be simplyto tell suppliers to do more maintenance and to
Table 1.4 Percentage of the poorest and richest population quintiles with access to infrastructure,various countries
Country/area
Access to public water supply Access to sewers Access to electricityPoorestquintile
Figure 1.7 Walking is a transport mode used frequently by the poor.
80
60
40
20
Delhi, India
Percentage a
100
Squatters Low-income Middle-income
LII Share who walk
Low-income
LI Share who take bus or paratransit
a. Percentage of urban residents in each group using each mode.Source: Survey data from Jaiswal 1992, as cited in Serageldin 1993.
Kingston, Jamaica
Lower-middle-income
Upper-middle-income
consult users. The weaknesses in infrastructure pro-vision are inherent in the incentives built into cur-rent institutional and organizational arrangements,in which outputs and inputs are not closely mea-sured, monitored, or managed, and suppliers donot depend on user satisfaction for reward. Aproper set of incentives would make managers ac-countable to users and to others who own and fi-nance infrastructure facilities. It would also givemanagers autonomy in making decisionsand re-sponsibility for success or failure. This Report's re-view of experience with infrastructure, in both thepublic and the private sectors, suggests that threeelements are essential in creating the right incen-tives for efficient and responsive delivery of ser-vices. These are management based on commercialprinciples, competition, and involvement of usersand other stakeholders.
COMMERCIAL PRINCIPLES. Infrastructure must be
conceived of as a "service industry," providinggoods that meet customers' demands. Such a com-mercial orientation contrasts sharply with the situa-tion in most government departments and state-owned public utilities, which suffer from multipleand conflicting objectives and inadequate account-
ing for costs or financial risk, and which put littleemphasis on revenues collected and the quality ofservice delivered. Managers have little motivationin such circumstances to satisfy customers or toachieve a reasonable return on assets through effi-cient operation and adequate maintenance. Typicalproviders of infrastructure are subject to pervasiveinterference by political authorities, which ad-versely affects operational decisions on investment,pricing, labor, and technological choices. It is com-mon to view certain infrastructure services (such aspower, water, ports, railways, airports, and telecom-munications) as potentially "commercial" becausethese are the services for which it is easiest to re-cover the costs of provision through user charges ortariffs. In fact, almost all infrastructure (even roadsand sanitation) can be operated with a business ori-entation. The basic conditions for this are limitedand well-focused performance objectives, financialand managerial autonomy (with a hard budget con-straint), and clear accountability both to customersand to providers of capital.
COMPETITION. Competition promotes efficiencyand provides users with options that, in turn, makeinfrastructure providers more accountable. Govern-
33
Box 1.7 Infrastructure activities threaten the Black Sea environment
The Black Sea is fed by a basin of more than 2 millionsquare kilometers, covering parts of seventeen countriesin Central and Eastern Europe, the former Soviet Union,and Turkey. It receives the inflows of several majorrivers, including the Danube, Don, Dnieper, andDniester. As an almost fully enclosed water body, theBlack Sea is especially vulnerable to changes in the quan-tity and quality of inflows from these rivers. The Donand Dnieper, in particular, have been highly developedfor irrigation and other purposes through a chain ofreservoirs.
Increasing pollutant loads from these riversespe-cially the nutrients nitrogen and phosphorushave ledto algal blooms and the destruction of important nurseryareas for fish. In addition, damming of the major riversfor navigation, flood control, water supply, and, aboveall, for irrigation, has considerably altered the seasonalflow patterns of these rivers. The damming has also de-creased the total inflow to the Black Sea, resulting in anincrease in salinity in critical coastal and estuarine areas,especially in the Sea of Azov, which creates further prob-lems for fish breeding. The overall result is a 90 percent
decline in the once-productive Black Sea fishery over thelast thirty years.
With assistance from the Global Environment Facil-ity, the six Black Sea countries (Bulgaria, Georgia, Roma-nia, Russia, Turkey, and Ukraine) have begun a regionalprogram to analyze the causes of observed environmen-tal degradation and to propose solutions. Actions in thebasin to regulate fertilizer use and to control pointsources of pollution are expected to result in reductionsof nutrient inflows. Pilot projects are proposed to restorefish production under the new salinity conditions.Under the Bank-supported Environmental ManagementProject for Russia, a study of the Lower Don Basin willinvestigate ways to alter the operating rules for themajor reservoirs to promote greater fish regenerationdownstream.
Given the size of the problem and the importance ofthese reservoirs in the agricultural economies of Ukraineand Russia, it would be unrealistic to expect dramaticchanges. Nevertheless, recognition of the problem andthe development of mechanisms for regional coopera-tion now make progress much more likely.
ments in most countries have not taken advantageof the potential for competition, even in activitieswhere a natural monopoly does not exist, such asroad freight transport or solid waste collection.Today competition can be used directly in more in-frastructure activities because of technologicalchanges. In telecommunications, satellite, mi-
crowave, and cellular radio transmission of tele-phone signals is revolutionizing the industry, mak-ing the economies of scale with cable-basedtransmission less important. In power generation,combined-cycle gas turbines operate efficiently atlower output levels than other generation technolo-gies. While open competition for users in the marketis still not feasible in many infrastructure areas,there are other ways of obtaining the benefits ofcompetition. For activities with high sunk costs,competing for the right to operate a monopoly cancapture many of these benefits. Even where thenumber of operators is necessarily limited, regula-tion can compel them to compete against perfor-mance benchmarks ("yardstick" competition).
INVOLVEMENT OF USERS AND OTHER STAKEHOLDERS.
In many infrastructure activities, market signalscannot be relied on to provide information aboutdemand or to gauge performance. Where users arelocked into a delivery network, they cannot expresstheir preferences or dissatisfaction through choice.
34
In such circumstances, other means of making sup-pliers accountable to users are needed. Throughvarious mechanisms designed to broaden participa-tion in decisionmaking and to provide wide accessto information on infrastructure provision, usersand other key stakeholders can be represented in(and sometimes take responsibility for) the plan-ning, financing, and delivery of services.
Opportunity knocks
Many of the above notions are not new, and somehave been accepted in principle by policymakers ifnot yet put successfully into practice. Three fac-torstechnological change, more pragmatic atti-tudes, and a greater sensitivity to infrastructure'simplications for poverty and environmental sus-tainabilityhave created a new climate for re-form. Innovative techniques for drawing on pri-vate financing for investment create a furtherchallenge to traditional ways of providing infra-structure. Many countries are now taking advan-tage of all these opportunities to test new ideasand approaches, discussed in later chapters of thisReport.
TECHNOLOGY. Technological changes are creatinga variety of new opportunities for changing the wayinfrastructure is provided in almost every sector-
in particular, by making the unbundling of diverseactivities more feasible. Microelectronic monitoringdevices and nondestructive testing techniques canfacilitate the assessment of infrastructure facilities(at reduced cost), often permitting testing by anagent other than the operatorsuch as the owneror regulator. Remotely controlled devices for in-specting pipe networks and the shift from analog todigital telephone switching have greatly simplifiedand reduced maintenance costs. Electronic informa-tion systems, including geographic mapping, im-prove the planning and design of investments andthe coordination of network operations. Technolo-gies that are clearly more efficient, robust, and flexi-ble than earlier methods enable developing coun-tries to "leapfrog" sectoral transitions experiencedearlier by high-income countries. For example,Brazil based its telecommunications expansion inthe 1970s on emerging digital equipment andthereby facilitated the development of information-based industries. Policy-induced inefficienciesslowed the modernization of the sector in the 1980s,however.
NEW PRAGMATISM. A new attitude, stemmingfrom an enhanced understanding of the relativestrengths and weaknesses of governments and mar-kets, is also creating opportunities for reform of in-frastructure provision. In the 1980s, the efforts ofmany countries to reduce the size of their over-extended public sectors led to a better realization ofwhat governments and markets can and cannot do.Worldwide liberalization of markets and experi-ments with different forms of private sector partici-pation in many sectors have provided a new body ofexperience to reinforce this pragmatic attitude. Theo-retical and institutional advances have also revealedwhen regulation is necessary and how to refine itsapplication. All this leads to two main conclusions.First, there are fewer infrastructure activities requir-ing government intervention than once believed.Second, when required, government interventioncan be exerted through less distorting instruments ofpublic policy than those traditionally used.
RENEWED COMMITMENT TO SOCIAL AND ENVIRON-
MENTAL CONCERNS. Political developmentsinclud-ing the trend in many countries toward democra-tization, pluralism, and decentralizationhavefueled a concern with finding more affordable andenvironmentally friendly solutions in infrastructure.This commitment has led to greater appreciation ofthe need to consult local communities, the poor, andgroups affected by environmental factors. At the
same time, increased efforts are being made to de-volve responsibility for infrastructure provision tolocal governments, to increase participation, and tofoster self-help.
Awareness that the poor (and future generations)are constituencies that must be answered to hasstimulated a search for alternative ways of provid-ing services or managing demands so as to broadenaccess while avoiding environmental problems. Rel-atively simple changes in design parameters forsewerage and improved design of latrines havemade sanitation affordable to low-income commu-nities while permitting private initiatives in financ-ing, maintenance, and manufacture of parts. Anincreasing range of technical, economic, and institu-tional alternatives to conventional wastewater treat-ment can reduce the need for costly filtration plants.Countries are adopting alternatives to large-surfaceschemes in irrigation such as drip, bubble, andsprinkler systems and low-level canals with low-liftpumpsthat are highly responsive to farmers'needs for water and are also environmentally sus-tainable. There is renewed interest in nonmotorizedmeans of transport, including bicycles and handcarts, and simple road improvements that enhancemobility in both rural and urban areas. Recognitionof the need to conserve scarce resources has led toefforts to avoid unnecessary infrastructure invest-mentsfor example, by promoting recycling andrecovery of solid waste materials; reducing wasteand effluents at the source; and managing demandfor water, power, and transport (Chapter 4). Indus-trial and developing countries are learning fromeach other in these areas.
The way ahead: a road map of reform
Awareness of past mistakes, together with new op-portunities, demands that a fresh look be taken atthe roles that governments or other public agenciesand the private sector should play in providing amore efficient and more responsive infrastructure.The challenge is to determine those areas in whichcompetitive market conditions can work and thosethat require public action. Within these broad pa-rameters, there is a menu of institutional optionsthat allow governments, public sector agencies, andprivate groups (both for-profit and nonprofit) to as-sume responsibility for different aspects of serviceprovision. The choices among the options will varyamong countries, on the basis of their economic, in-stitutional, and social characteristics. The spectrumof options is broad, but four main approaches can beidentified:
35
Option A: Public ownership and operation,through a public enterprise or government depart-ment
Option B: Public ownership but with privateresponsibility for all operation (and for financialrisk)
Option C: Private ownership and operationOption D: Community and user provision.
The remainder of this Report discusses howmore efficient and responsive provision of infra-structure can be achieved by improving incen-tivesthrough stronger mechanisms of account-ability and autonomy. Chapter 2 discusses ways tocreate accountability in a public agency or gov-ernment department (Option A) by establishingcommercial principles and through organizationalrestructuring (corporatization). It also reviews con-tracting instruments to permit better monitoringand performance of operations, and appropriatemechanisms for achieving financial autonomy.
Commercial principles are often very difficult toinstill permanently in the absence of effective com-petition. Chapter 3 discusses the scope and tech-niques for marshaling market forces to createaccountability through competition andwherecompetition alone is insufficientregulation. Chap-ter 3 also examines experiences with public owner-ship and private operation (Option B), in whichcompetition for the market is used, as well as pri-vate ownership and operation (Option C). Both of
36
these arrangements require appropriate sectoral re-structuring to maximize the opportunity for compe-tition and to ease the regulatory burden.
Chapter 4 examines issues that neither commer-cialization nor competition alone can addressproblems of externalities (particularly environ-mental), distributional equity, and the need forcoordination of investments. It discusses ap-
proaches for assessing and creating accountabilityto social and environmental concerns, through de-centralization of governmental responsibilities, par-ticipation by users and stakeholders (includingthrough "self-help" schemes, Option D), and plan-ning. Chapter 5 reviews how mechanisms of financ-ing infrastructure can create incentives for efficiencyby providing the disciplinary pressure of private fi-nancial markets. Because different aspects of infra-structure provision involve different kinds of risks,the chapter considers how a suitable packaging offinance using alternative sources and instruments(private and public) can lead to better risk manage-mentin addition to mobilizing increased fundsfor infrastructure investment. Chapter 6 returns tothe menu of options and shows how these can beapplied in different infrastructure sectors and coun-tries. The conditions for successful implementationof these options are also outlined. The chapter closeswith a broad assessment of the economic and finan-cial benefits that countries can gain by following thereform agenda presented in this Report.
2Running public entities oncommercial principles
Successful providers of infrastructure services, inthe public or private sector, are generally run onbusiness lines and have three basic characteristics:
They have clear and coherent goals focused ondelivering services.
Their management is autonomous, and bothmanagers and employees are accountable for re-suits.
They enjoy financial independence.The principles underlying these characteristics
come naturally to a private business, but by nomeans always to organizations in the public sector.Governments are forced to balance many differenteconomic, social, and political objectives, and it iscommonplace for these goals to spill over into theactivities of all public sector organizations, includ-ing infrastructure enterprises. Similarly, manage-ment of public sector employees is often hamperedby numerous restrictions on establishing account-ability and rewarding good performance. In addi-tion, the financial status of public agencies and en-terprises often depends on budgetary decisions thatare unrelated to performance and on pricing deci-sions that are driven by politics. These factors oftenwork against rational management.
Many argue that endemic organizational fail-ures and poor performance are compelling argu-ments for abandoning efforts to reform the publicsector and for relying instead on the private sectorto provide infrastructure services. Increased re-liance on the private sector, discussed in Chapter 3,may be right for some countries and sectors. Nev-ertheless, making the public sector more effective isimportant for (at least) four reasons. First, givencurrent government dominance, the public sectorwill continue to have primary responsibility for in-
frastructure services in most countries and mostsectors in the foreseeable future. In the poorestcountries, today's weak private sector capabilitieswill improve only slowly. Second, even with dy-namic private sector involvement, some sectors -such as road networks and major public workswill remain predominantly in the public domain.Third, only an effective public sector will facilitateprivate sector involvementa dispirited and inef-ficient public works department is unlikely to mo-bilize the will or the ability to contract out roadmaintenance. Fourth, many developing countrygovernments will decide (for strategic, regulatory.or political reasons) to retain much of the responsi-bility for building and operating infrastructure inthe public sector, as many high-income countrieshave done.
Improving the effectiveness of public sector in-frastructure providers (Option A in Chapter 1) isthus critical. It can be done by applying three coreinstruments to reinforce commercial operation inthe public sector:
Corporatization, which establishes the quasi-independence of public entities and insulates infra-structure enterprises from noncommercial pressuresand constraints.
Explicit contracts between governments and(public or private) managers or private entities in-volved in infrastructure services, which increaseautonomy and accountability by specifying perfor-mance objectives that embody government-definedgoals.
A pricing strategy designed to ensure cost re-covery, which creates a desirable form of financialindependence for public utilities and even at timesfor public works.
37
Lessons of success and failure
Although the public sector has invested heavily inexpanding infrastructure stocks (Chapter 1), gov-ernments have done less well in managing the flowof infrastructure services. Experience suggests thatthe key elements present in successful providers,and lacking in troubled ones, are those character-ized above as commercial principles.
What success shows
That many public entities have performed poorlydoes not mean the public sector is incapable of get-ting it right. A recent study of the privatization oftwo previously well-run public power firms inChile shows that the improvements from private
management yielded a productivity increase of only2.1 percent in one case and less than 4 percent in theother. Because these firms were already being runon commercial principles, the gains from privatiza-tion were ten to twenty times less than would other-wise have been the case. There are many otherexamples of successful public provision of infra-structure servicesMexico in power, Korea andSingapore in most or all sectors, and Togo in watersupply to name a few. Until recently, Botswana'swater utility was also run on commercial principles,and it has had an enviable performance record(Box 2.1).
What is the secret of such success? A commonfeature is a high degree of autonomy for the entitiesconcerned. Managerial and organizational auton-omy does not mean complete freedom: all public
38
Box 2.1 The right way to run a public utility: a look at Botswana'sWater Utility Corporation
Created in 1970, Botswana's Water Utility Corporation others in order to retain supply. BWUC does not hesitate(BWUC) has two primary responsibilities: to provide to adjust prices as needed in order to manage demand.potable water to the country's principal urban areas and In 1985-86 charges were raised to counter the effects of ato operate a financially self-supporting service. severe drought. This action effectively reduced demand
BWUC is under the administrative jurisdiction of the to the point where everyone could obtain a minimumMinistry of Mineral Resources and Water Affairs. The quantity of water during the drought and avoided theministry's deputy secretary is chairman of the board, need to cut supplies. Accounts receivable are usually lessand until recently he has been successful in keeping po- than 2 percent of all the amounts collectible, attesting tolitical influence out of the conduct of BWUC's opera- the success of strict billing and collection procedures.tions. This effort has been helped to some extent by con- A family of six people consuming about 100 liters atracting out management (until 1990 mostly to day per capita pays about $8.85 a monthapproxi-expatriates, but increasingly to nationals). The only pos- mately 8 percent of its income. Reducing consumption tosible defect in this arrangement is that contracts are for 80 liters lowers the water bill to about 5 percent of in-two years, which focuses problem solving on short-term come. A wealthier family consuming twice that muchsolutions because managers want to be able to show the would pay about $32.25 a month. These charges are higheffect of their decisions while still under contract. in comparison with those levied by similar utilities in
BWUC maintains a twenty-four-hour supply of Africa, but they have the effect of constraining consump-water to all its service areas, with high-quality treatment. tion and ensuring that the utility does not have to rely onBotswana is therefore one of the few countries in Africa subsidies from the government or from other sectors inwith a safe urban water supply. Water losses are accept- the economy.able, at about 15 percent in the distribution system and A noteworthy achievement is the "one-check" sys-10 percent in the raw-water transmission and treatment tem for government users. The Ministry of Financeprocesses. The overall loss of about 25 percent would be meets all monthly charges for the government andconsidered good by utilities in many industrial coun- deducts them from the cash allocations of each ministrytries. These low losses reflect the good quality of or department. This procedure avoids the accumulationBWUC's engineers, who are attracted by competitive of arrears by government users of infrastructure servicessalaries. that is common elsewhere.
BWUC charges commercially oriented tariffs appro- Recently, however, the utility has begun experiencingpriate for the urban conditions in Botswana, and tariffs problems. After more than twenty years of successfulare increased when necessary. Meters are read and con- operation, BWUC is finding it increasingly difficult tosumers are billed monthly, with thirty days to pay. Sup- adjust its rates as required. Lags in tariff adjustmentsply is cut off immediately if payment is not made, and may yield short-term political gains, but they will alsothere are charges for reconnection. Little evidence exists allow water consumption to grow and increase the risksthat consumers who have been cut off are sharing with of water shortages in this severely water-scarce country.
Table 2.1 Common management problems in public sector infrastructure entities, 1980-92(percentage of World Bank loans in which conditions were imposed to address the problem area)
Source: World Bank database (ALCID).
providers are subject to regulatory oversight bytheir parent ministries. Government sets clear poli-cies and goals while leaving detailed planning andimplementation of services to the providers. Thisdelegation of responsibility and conscious absenceof political intervention are one reason why thesepublic agencies have retained high-quality man-agers and why they enjoy stability in mid-manage-ment and professional structures. Successful publicsector organizations also enjoy financial strength.Tariffs cover (at a minimum) the requirements foroperations and maintenance, while effective cost ac-counting controls expenses. This reliance on cost re-covery from users accounts in part for the emphasison good customer relations. Also common (althoughnot universal) among well-run public organizationsis the use of private contractors and private capitalin infrastructure operation and maintenance.
What failure shows
A survey of forty-four countries with WorldBank-financed projects designed to improve infra-structure performance revealed the most commonproblems in six infrastructure sectors (Table 2.1).Unclear goals, lack of managerial autonomy and ac-countability, financial difficulties, and wage andlabor problems are recurrent problems for the pub-lic sector entities involved.
The goals of public sector infrastructure pro-viders are often hazy and inconsistent. More thansimply financial objectives are necessary in settinggoals for infrastructure providers, especially when alarge share of the population is without access to theservice involved. The goals may include quantita-tive targets like user coverage or capacity expan-sion. In the absence of such goals, public providershave often failed to recognize that some consumergroups - such as the poor and rural consumers -are willing to pay for services and thus should betargeted to receive them. Whether in Africa, Latin
America, or South Asia, water and power entitiesreceive mixed signals from governments aboutwhere to expand their networks. The main victimsof inconsistent official priorities have often beenrural areas, where government failure to improvecoverage is pushing users to search for alternativeforms of service provision (Chapter 4).
A lack of autonomy and accountability underliesmany other problems. Financial problems, overem-ployment, and unfocused goals occur because man-agers do not have control over day-to-day opera-tionsor over decisions on prices, wages,employment, and budgets. Managers in such cir-cumstances seldom have much incentive to tryharder. In Ghana, for example, a 1985 reform madethe chief executive of a utility responsible to itsboard of directors, but amendments graduallyshifted accountability back to the relevant ministry,thereby restoring direct political intervention. Theproblem became even worse when performance-based bonuses, introduced to motivate managersand employers, became an integral part of the salarystructure and thus lost their incentive value.
The third problem, financial difficulties, is com-mon in power and water utilities when politicallymotivated tariff adjustments lag behind cost in-creases. These difficulties reflect a lack of manage-ment autonomy and the use of public infrastructureentities to achieve diverse uncompensated goals -such as keeping tariffs low in order to counter infla-tion. In Brazil, between March 1985 and the end of1989, three freezes on public sector prices caused thereal tariff to drop by 59 percent for port services, 32percent for railways, and 26 percent for telecommu-nications. The results were higher public enterpriselosses that defeated the anti-inflation strategy byfueling the overall public sector deficit.
Problems with wages and employment oftenhave their origins in the first three problems. Manyinfrastructure utilities are overstaffed because gov-ernments use them to create public sector jobs and
Box 2.2 Severance pay eases layoffsin Argentina Railways
By the end of the 1980s, Argentina Railways hadabout 95,000 employees and an annual deficitequivalent to 1 percent of GDP. Since the mid-1970s, the wage bill had consistently exceeded rev-enue. Estimates indicated that cutting personnel byhalf would not affect the level of service.
Major reforms have been introduced over thepast few years. Private sector concessions weregranted to run all freight lines and the Buenos Airesregion passenger service. To curb losses and reduceemployment, intercity passenger service was cutback by two-thirds. The World Bank supported theinitial reform efforts by financing (through an ad-justment loan) severance costs of the voluntary re-tirement of 30,000 rail employees. The severancepay was roughly equivalent to two years of salaryfor each retired employee. Auditors certified thatseverance payments were made only to staff whoselabor contracts were terminated and that paymentsconformed to labor laws and were consistent withseverance pay in other sectors. Measures to preventreemployment were also put in place.
External financing of the initial adjustmentadded credibility to the reform process and reducedthe resistance of unions. It also paved the way forsubsequent employment-reduction cycles financedfrom government resources. Eventually, 60,000workers were retired over a two-year period.
pass the additional cost on to taxpayers or con-sumers. This practice often results in the underfund-ing of maintenance. Overstaffing erodes managerialautonomy, diffuses organizational goals, and createsfinancial problems, especially in transport, althoughalso in other sectors. During the 1980s, one of thelargest water systems in East Asia increased itsbilled services by 132 percent, an increase that nor-mally would yield a decline in per-unit personnelcosts. But staff increased by 166 percent over thesame period, thereby negating the benefit of higherrevenues.
Another employment problem is that, althoughpublic entities are often overstaffed, they seldomuse sufficiently labor-based methods, which can beboth cost-effective and result in high-quality infra-structure in roads, water and sanitation, irrigation,and urban infrastructure. In Sub-Saharan Africapublic agencies have often preferred equipment-and capital-intensive road construction for overde-signed roads that usually require capital-intensivemaintenance. Removing such biases often improves
the use of local resources and can be more consistentwith environmental and poverty objectives. InRwanda, for example, switching to labor-based con-struction of secondary and gravel roads increasedemployment by 240 percent (mostly for low-wageunskilled laborers highly represented among thepoor). It also reduced both total costs and importsby about one-third.
Governments can avoid these four commonproblems and increase the chances of success by cre-ating organizations driven by commercial princi-ples. Corporatization insulates organizations frommany government constraints and pressures. But itdoes not mean that infrastructure providers are ableto set their own agenda and goals. Government, asthe owner of public infrastructure enterprises or cor-porations, continues to set their basic goalsthrough explicit contracts if necessaryand to regu-late their behavior so as to ensure an adequatereturn on society's investments. In addition to man-agerial autonomy and well-focused goals, pricesmust be seteither by the provider or through reg-ulationat levels that ensure financial strength andincentives.
Corporatization
The explicit separation of infrastructure service pro-viders from government starts by changing a gov-ernment department into a public enterprise inorder to increase management autonomy. Manycountries have achieved this changeover in water,power, and railways, although it is a more recentphenomenon in port services. Enterprises areobliged to provide services that match demand, butmany do not have the legal corporate independenceneeded to ensure efficient operation.
Corporatization is the next step, giving the enter-prise an independent status and subjecting it to thesame legal requirements as private firms. Corpora-tization means that the entity is subject to standardcommercial and tax law, accounting criteria, compe-tition rules, and labor law and is less susceptible togovernment interference. In practice, this transfor-mation is not always complete because public orga-nizations do not face adequate competition or donot have solely commercial objectives. For example,corporatization implies the transfer of employeesfrom civil service status to contracts governed byordinary labor law. Yet even under corporate struc-tures, public entities are often reluctant to reduceemployment. The experience of developing coun-tries suggests that the enforcement of ordinarylabor law and the work force cuts needed for suc-
Box 2.3 It took ten years to corporatize Indonesia's main ports
In Indonesia, there are three formal stages in the adop-tion of commercial principles. First, the government de-partment is transformed into a government enterprise.Then the enterprise becomes a corporation that still has acombination of commercial and noncommercial goals.Finally, the corporation is turned into a profit-orientedentity whose ownership can be shared with the privatesector. Ports have just reached this third stage.
The reform of Indonesia's port management began in1983. Before that, the management of all 300 ports wascentralized in the Directorate General of Sea Communi-cations, a government department. Most of these portshad obsolete equipment and failed to meet regionalneeds. In mid-1983 the government decided to decen-tralize management for ninety of its ports by creatingfour new public port corporations, headquartered at thefour largest ports.
Two years elapsed before the government addressedthe overregulation that remained a major impediment tothe success of the new corporations. Moreover, man-agers did not yet have a clear understanding of their re-sponsibilities and accountability and lacked the auton-omy to implement reforms they thought were needed.These problems had been addressed by 1988, when aneffective cost control program lowered expenses by 5percent and increased revenue by 20 percent for thelargest port corporation. Between 1987 and 1992, rev-enue grew almost twice as fast as expenses.
Ten years after the reform process started, the portcorporations face the market test. Competition promisesto be tough: a recent survey of foreign investors rankedIndonesia's port infrastructure at about the same level asAustralia's but below others in the region, such as HongKong, Malaysia, and Singapore.
cessful restructuring are more politically accept-ableand hence more sustainablewhen sever-ance pay accompanies dismissals. This has been theexperience with Argentina's railway reform pro-gram (Box 2.2).
The transformation of a government departmentor ministry into a public enterprise is more difficultfor public works than for utilitiesand roads pre-sent a special challenge. However, converting high-way departments to public utility corporations (asin New Zealand) is attracting interest as a way toimprove performance, especially in the area ofmaintenance. Highway expenditures are budgetedaccording to assessments of traffic-related costs, anduser charges are then calculated to reflect the wearand tear caused by different types of vehicles. Thisexperience is very recent, however, and, although ithas inspired similar approaches (in Tanzania, for in-stance), it is too early to assess its sustainability.
Commercial accounting procedures are an imme-diate benefit of corporatization. Explicit cost ac-counting identifies nonremunerative activities andreveals sources of inefficiencies, making costs andbenefits more transparent in public enterprisesand government departments. In Ghana, for exam-ple, an attempt to reform the main utilities beganwith the development of a good set of accounts forcosts. The government's move to suppress transfersto enterprises that could achieve financial autonomycreated a need for the enterprises to use proper cost-accounting techniques. Within two years, real oper-ating costs in the state transport corporation were
down by 67 percent, allowing its revenue to increasefrom 92 to 111 percent of its full operating costs.
Organizational changes are always simpler onpaper than in practice. It takes time and much effortto convert a government department into a publiccorporation. The introduction and full implementa-tion of standard accounting practices alone can takeup to five years, as many Eastern European policy-makers are finding out. Getting everything elseright is equally difficult. Ghana's utilities have beenundergoing transformation for seven years and stillhave a long way to go. And it took ten years to cor-poratize fully Indonesia's major ports (Box 2.3).
Focused goals and accountable management
Corporatization provides an organizational struc-ture, but by itself it merely transforms the problemof official governance into the more tractable, al-though still difficult, task of corporate governance.Organizational changes alone neither provide cleargoals nor create incentives for managers to meetthese goals. Many governments argue that their de-partments and enterprises are already run on com-mercial principles, but this has not helped managersto be more effective. Many managers argue that theautonomy they do get is too limited to be effectiveand that it is too easily revoked. Many workers arguethat they have little incentive to be effective becausegood and poor performers are treated equally. Andmany users would argue that corporatization hasnot given them access to improved or expanded ser-
41
vices. These concerns are particularly prevalent inAfrica and South Asia, where reorganizations ofpublic utilities and government departments havebeen common but where performance has often re-mained disappointing. Latin American countrieshave preferred a more fundamental shift to privateownership (Chapter 3).
The introduction of market principles can helpsolve the problem of corporate governance. For itspart, government must allow adequate competition,level the regulatory playing field, and instruct man-agers to maximize profits or to achieve set rates ofreturn. Although effective in the long run for somesectors and some services, this solution raises atleast two problems. First, and more obvious, pro-viders in many cases are in the public sector pre-cisely because of the limits on profit maximiza-tioneither because the services are public goods(as with roads) or because governments have objec-tives other than profit. Second, because serviceproviders have monopoly powers, prices have tobe regulated outside the supplying entity (seeChapter 3).
When the market solution cannot be used to ad-dress corporate governance problems in the publicsector, three other approaches might be consideredfor structuring the relationship between govern-ments and infrastructure providers.
Performance agreements retain all decisions inthe public sector. They try to increase the account-ability of employees and managers and to improvethe focus of operations by clarifying performanceexpectations and the roles, responsibilities, and re-wards of all those involved.
Management contracts transfer to private pro-viders the responsibility for managing an operationsuch as a port or a power or water utility. They in-crease the autonomy of management and reduce therisks of political interference in the day-to-day oper-ations of the public entity.
Service contracts transfer to private providersthe responsibility for delivering a specific service atlower costs or obtaining specific skills or expertiselacking in the public sector such as design engi-neering. (Turning all operations over to the privatesector under a lease or concession is discussed inChapter 3.)
Properly designed, these contracts can addressorganizational failures. And they can be just as ef-fective in a public works department as in a publicutility. Many governments are attracted to such con-tracts because they do not involve relinquishingpublic ownership.
42
Performance agreements
Performance agreements negotiated between gov-ernment (the enterprise owner) and managers havebeen tried in most infrastructure sectors. This typeof agreement originated in France, where the mainpurpose was to spell out reciprocal commitments ofgovernment and managers. Korea, which wasamong the early Asian users of performance agree-ments, added explicit performance-based incentivesfor both managers and employees. The focus on in-centives is what most recent contracts are trying toduplicate.
REVEALING INFORMATION TO IMPROVE THE FOCUS. In
order to identify the sources of incentive failures,governments must develop information and evalu-ation systems for performance monitoring. The in-formation component focuses on the developmentof standard financial and cost-accounting proce-dures, as well as detailed quantitative and qualita-tive indicators. In roads, for example, these indica-tors include measures of the condition of thenetwork and its use and management, administra-tion and productivity, and finance. The negotiationof a performance agreement covering most of theseindicators has allowed the Highway Department ofthe State of Santa Catarina in southern Brazil tosharpen its objectives. The result is that prioritieshave changed and focus more on maintenance andrehabilitation of roads than they have in the past.Specific targets have been set for all categories of ex-penditure. The share of paved roads in poor condi-tion is expected to decline from 18 percent in 1991 to4 percent by the end of 1994. Staff needs and skillshave been assessed, supporting a reduction inworkers from 3,149 in 1990 to 1,885 in 1993. Already10 percent of all maintenance work is contracted outto the private sectorand the performance agree-ment requires an increase to 25 percent by 1995.Similar reforms are being introduced in the states ofMaranhão, Piaui, and Tocantins.
BUILDING IN INCENTIVES. This component has sev-eral elements. The first is a promise of increasedmanagerial autonomy for the enterprise as well asrewards for workers and managers in exchange forfulfilling agreed performance targets. Some agree-ments in India, Korea, and Mexico include bonusesof up to 35 percent of total wages. The Koreans con-sider nonpecuniary benefits such as award cere-monies or press coverageto be a key factor intheir success with contracts. Firing nonperformingstaff is one of the sanctions available in Korea (Box
2.4). The second incentive element that can be builtinto these agreements relates to the duration of theagreement. Shorter agreements (one year, as inKorea or Mexico) are more effective because theyallow for more frequent assessments, although theyalso involve time-consuming renegotiations.
The third common incentive is the weight at-tached to various performance indicators after care-ful negotiation between the managers involved andthe government. In Mexico the agreement signed in1989 by the Federal Electricity Commission and thegovernment distributed weights according to itspriorities as follows: 44 percent for improvements inproductivity, 23 percent for better operational effi-ciency, 18 percent for reaching administrative and fi-nancial targets, and 15 percent for improvements inservice quality. These weights were only partly suc-cessful in giving managers and employees a bettersense of priorities and an incentive to focus on whatmatters rather than on what might be easier toachieve. By 1991 the ranking of performance frombest to worst was as follows: efficiency, service qual-ity, productivity, and administrative and financialperformances not quite a match with the prioritiesand weights.
WHAT HAVE PERFORMANCE AGREEMENTS ACCOM-
PLISHED? Performance agreements have often beensuccessful in East Asia, thanks to explicit efforts tobuild incentives for managers and workers into thecontracts and to monitor these incentives. Whenperformance agreements were used, the rate of re-turn on the assets of the Korean Electric Corporationtripled over a period of seven years (Box 2.4). Theseagreements are also proving useful in the reform ofhighway departments, as seen from the supplier'sexperience. Performance agreements have notachieved such impressive results in Africa. Al-though they have often improved noncommercialgoals, such as increases in rural coverage, they haveoften failed to achieve financial targets. In Senegalcost recovery efforts improved initially, but withinthree years costs were back to the level they hadbeen before the introduction of performance agree-ments. In this case, the agreements failed to addressthe lack of performance incentives for managers andworkers. The difficulties that many agreements havehad in differentiating the rewards for performancein the civil service explains why most experts holdlittle hope for such agreements in Africa and suggestrelying more on other alternatives discussed below.
Box 2.4 What's special about Korean performance agreements?
The Korean performance agreements are an outcome ofthe 1983 reform of public enterprises. The agreementsare intended to permit comparative evaluation of theshort- and long-term performance of all managers(rather than focusing on the company), to ensure that in-formation is available for the evaluation, that rewards tomanagers and employees are linked to their perfor-mance, and that the evaluation is done by independentauditors. Korea has been more successful with perfor-mance evaluations than most countries. Despite finan-cial difficulties at some enterprises in recent years, theyhave generally reached their noncommercial goals.
What kind of performance indicators are used? Perfor-mance indicators are selected to measure results againstthe trend and according to agreed targets. The bench-marks are generally based on international experienceand are derived in consultation with independent out-siders to minimize potential conflicts of interest. The tar-gets are set and assessed annually to increase account-ability. Quantitative indicators generally account for 70percent of the final score. The key quantitative indicatorsare profitability and productivity. Other quantitativeindicators are sector-specific, representing such charac-teristics as coverage or physical outputs. Qualitative in-dicators focus on corporate strategy, research and devel-
opment, improvement in management information, andinternal control systems. Indicators are combined into asingle public profitability indicator using a weighted av-erage of performance with respect to each indicator.
What is the information base for the assessment? Koreanow benefits from a sound financial and accountingbasis that provides management with a clear statementof objectives for performance. To some extent, thisspread of standard accounting techniques stems fromtheir introduction as one of the performance indicators.
How is performance related to reward? To increase ac-countability to users of infrastructure services, the per-formance-based ranking of public companies is pub-lished in the press. The best managers get not onlyprestige but also monetary compensation. The annualbonus to staff members and the career prospects of theirmanagers are related to the ranking of their company
The outcome? Within three years, the managementperformance of executive directors, directors, and de-partment chiefs improved substantially in at least 60 per-cent of the enterprises. More dramatically, the rate of re-turn on the assets of the public enterprises (in the case ofthe power and telecommunications companies) rosefrom less than 3 percent before 1984 to more than 10 per-cent by the end of the decade.
43
Management contracting
Management contracting gives responsibility for abroad scope of operations and maintenance to theprivate sectorusually for three to five years. Thisapproach can be more effective than relying on aperformance agreement to achieve similar objec-tives. A management contract signed for the powercompany in Guinea-Bissau is demonstrating thatmanagement contracts may work where many per-formance agreements have failed. There, a newmanagement team succeeded in doubling electricitysales in just three years (Box 2.5).
However, when public agencies prevent a pri-vate contractor from controlling key functions af-fecting productivity and service quality such asstaffing, procurement, or publicly provided work-ing capital the contractor cannot be held account-able for overall performance, and generally the con-tract does not succeed. That is why a recentmanagement contract signed for a power plant inthe Philippines failed within nine months. When thenew managers and the government disagreed onstaffing levels and composition, the contract wasbroken despite the rapid improvements observed inmaintenance following the arrival of the new man-agement team.
WHEN IS IT EFFECTIVE? Management contractingworks better when a contractor is granted significantautonomy in decisionmaking and compensation isbased, at least in part, on performance. In France,where management contracts are common in watersupply and sanitation, the incentive for productivityimprovement links the contractor's payment to suchindicators as reduced leakages and increased con-nections. The contract for the Electricity and WaterCompany of Guinea-Bissau specified that 75 percentof the remuneration was guaranteed but that the re-maining 25 percent was based on performance.Management contracts with fees based on perfor-mance tend to be more successful than those withfixed feessuch as traditional management con-sulting assignments. Fixed-fee arrangements differlittle from technical assistance and are seldom suc-cessful. Relating incentives to performance may notwork, however, where a government can interferewith tariffs. In general, such contracts tend to bemore useful as interim arrangements allowing pri-vate firms and public agencies to gain experiencewith partnerships before engaging in more compre-hensive contracts or while the regulatory frameworkis being developed (both discussed in Chapter 3).
A recent innovative application of managementcontracts is the experience with Agences d'Exécu-
Box 2.5 Management contracting in Guinea-Bissaua success story?
Introducing a five-person management team under aforeign management contract improved the performanceof Guinea-Bissau's national electric utility. Previously,service interruptions had been chronic, and most areashad electricity only a few hours a day. Comparative sta-tistics for 1987 and 1990 show the turnaround. But morerecent experience illustrates the difficulties of manage-ment-government relations.
Box table 2.5 Performance of Guinea-Bissau'snational electric utility
The foreign management contract was implementedunder a joint initiative of the French Ministry of Cooper-ation, the United Nations Development Programme, theAfrican Development Bank, and the World Bank. It re-duced wastage of foreign aid. (In the previous ten years,foreign aid for power was more than three times the esti-mated value of the utility at the end of the period.)
At the beginning of 1994, however, serious problemsbecame evident. Despite economic tariffs the utility wasunable to generate revenue to finance expansionoreven, at times, current operationsleading again toshortages and reductions in service quality. This precari-ous financial condition was due to a large rise in receiv-ables stemming from the utility's difficulty in collectingpayments. The government demanded continued ser-vice for "critical" functions even when its unpaid billswere causing financial distress. And in the private sec-tor fraudulent connections were rampant despite theutility's efforts to prevent them.
Average revenue (dollarsper kilowatt-hour) 0.12 0.25 0.22
Box 2.6 AGETIPs: involving the private sector in Africa's urban infrastructure
If governments do poorly in executing infrastructureprojects, why not leave it to the private sector? That isprecisely what is happening in ten West African coun-tries. The Agences d'Exécution des Travaux d'IntérêtPublic (AGETIPs)nonprofit, nongovernmental agen-cies for executing public worksenter into contractualarrangements with governments to carry out infrastruc-ture projects. The AGETIP in Senegal, which has twentyprofessional staff members, has handled 330 projects inseventy-eight municipalities. It hires consultants to pre-pare designs and bidding documents and to superviseworks, issues calls for bids, evaluates bids and signs con-tracts, assesses progress, pays contractors, and repre-sents the owner at the final handover of the works.
AGETIPs use an integrated approach to designworks that promotes competition while facilitating ac-cess for small contractors. Project designs take intoaccount local constraints, labor markets, the limited out-put potential of small contractors, the weak project-identification capability of local governments, the avail-ability of consultant architects and engineers, and theeconomic and social rationale of subprojects under con-sideration. Project eligibility and selection criteria arespelled out, with particular emphasis on labor-intensivemethods. Open competitive bidding weeds out ineffi-cient operators.
Contracting out promotes the development of localcontracting and consulting firms by creating demand fortheir services. The AGETIP in Senegal now has 980 localcontractors and 260 local consultants on its books. It hasreduced barriers to entry and made life easier for new,weaker firms by paying contractors every ten days; pub-lic entities typically take several months.
The autonomy given to AGETIP managers enablesthem to run efficient, impartial, and transparent opera-tions, and the agencies' protected legal status sheltersthem from political pressures. A strong management in-formation system and institutionalized personal ac-countability enable AGETIP managers to account forevery project, supplier, payment voucher, and outstand-ing bill. All consolidated project accounts are indepen-dently audited eveiy six months. There are also bi-monthly management audits and an annual technicalaudit.
An evaluation of AGETIP activities shows that their"corruption-free procedures" have allowed them tocomplete projects largely on schedule with a cost over-run of only 1.2 percent of the portfolio (cost overruns inpublic procurement average 15 percent of original esti-mates). AGETIPs routinely obtain unit prices 5 to 40 per-cent lower than those obtained by the administrationthrough official bidding.
tion des Travaux d'Intérêt Public (AGETIPs) in WestAfrica. Management responsibilities for urban infra-structure projects have been contracted out to non-profit, nongovernmental agencies that in turn con-tract out the public works involved. Increasedmanagement involvement and accountability haveimproved project performance. The management ofcontracts has improved and so has implementationthrough allowing smaller firms, with more labor-intensive techniques, to participate in governmentcontracts. In Senegal the use of AGETIPs has led to10 to 15 percent reductions in unit costs in local in-frastructure projects (Box 2.6).
Contracting out services
Contracting out services is becoming popular withpublic infrastructure providers. It provides a flexi-ble and cost-effective tool for increasing responsive-ness to users and taps expertise too expensive tomaintain permanently on public payrolls. It alsopermits competition among multiple providers,each with short and specific contracts.
Contracting out is most common for mainte-nance services. Major overhauls of power stations,
for example, are routinely contracted out to plantsuppliers or specialists in most developing coun-tries. Service on contract is also a standardarrangement for the design and construction ofmajor capital works because of the obvious bene-fits from specialized engineering knowledge andconstruction skills. The infrastructure supplier setsthe performance criteria for the contracted ser-vices, evaluates bids from competitive tendering,supervises performance, and pays agreed fees forthe services involved. Contracting out is a versatilemeans for carrying out many other tasks, and thebase of developing country experience is growing.Standard professional services - such as auditing,data processing, and recruitmentare also oftencontracted out. Railways in Pakistan have con-tracted out such activities as ticketing, cleaning,and catering. Private contractors in Kenya do lim-ited locomotive repair and maintenance for thestate railroad. Meter reading and fee collections inthe water supply and sewerage sectors have beenhandled through service contracts in Chile sincethe 1970s. Santiago's public water company evenencouraged employees to leave and compete forservice contracts.
45
Figure 2.1 The adoption of commercialprinciples in 1984 allowed Togo'swater utility to increase coverage andproduction...
Performance index(1984 = 100)
190
170
150
130
110
Number ofconnections per
thousandsubscribers
901984 85
Rate ofreturn
Averagetariff/cost
86 87 88 89 90 91
Formal commercialization period (1984-89)
LI Period with performance agreement (1989-91)
Source: World Bank data.
How EFFECTIVE IS CONTRACTING OUT? Contracting
out tends to be more cost-effective than using publicemployees to handle maintenance (known as forceaccount). Brazil's switch to road maintenance bycontract reduced costs by some 25 percent for equiv-alent service quality. In Colombia rural microenter-prises charge about half the rates implicit in force
46
accounts and achieve a better overall quality of ser-vice. An additional gain from these contracts is thatgovernment force account work groups have be-come more efficient when compelled to competewith private contractors. A survey of contractorssuggests, however, that maintenance contracts oflonger duration and wider scope are needed to jus-tify capitalization and acquisition of specializedequipment by the contractor. Chile, which doesnearly 80 percent of its road maintenance throughcontracts, is now moving toward "global" mainte-nance contracts of longer duration.
Selecting the right type of contract
Which of the three types of contractsperformanceagreements, management contracts, and contractingoutis the right one depends on the infrastructureactivity and the specific cause of poor performancein providing the service. Because the performanceof a public entity depends on the actions of govern-ment, managers, and workers, the best contract isthe one that most effectively alters incentives towhichever of these three performs least well.
If the problem is with the government, the per-formance agreement may be the preferred instru-ment because performance agreements are recipro-cal. For example, a 1989 performance agreement byTogo's water utility illustrates how managers canuse such agreements to get the government to en-dorse needed tariff increases. The performanceagreement was a complement to commercializationin 1984. The utility's managers wanted an explicitperformance agreement to commit the governmentto tariff increases. Although commercialization im-proved performance with respect to noncommercialgoalsa 73.5 percent increase in the number of con-nections in just five yearsit did not help financialperformance because the government did not au-thorize needed tariff increases. By 1989 the cost re-covery ratio was 7 percent lower than in 1984 (Fig-ure 2.1). The performance agreement was needed sothat the government and the utility could agree onthe steps to achieve financial autonomy. Within ayear, the cost recovery ratio was 16 percent higherthan its 1984 value. However, if the problem is oneof weak commitment by the government, no reme-dial instrument short of privatization is likely to bevery effective.
If the problem is with management, the choice ofcontract depends on whether abilities or incentivesare in question. Performance agreements with in-cumbent public managers assume that their capabil-ities are adequate. Thus, in the case of an organiza-
901984 85 86 87 88 89 90 91
. . but a performance agreement in 1989 wasneeded to improve financial outcomes.
190
170
150
130
110
tion with weak management skills, managementcontracts based on performance are more effectivein the short run, as in Guinea-Bissau. For the longerrun, training objectives can be incorporated in bothperformance agreements and management con-tracts.
If the problem with management is one of incen-tives, performance agreements need to make a clearlink between performance and pecuniary and non-pecuniary rewards to managers. This approach hasbeen effective in Korea, where the president of apublic corporation that moved from last place(twenty-fourth) in the ranking of public enterpriseperformance to first place in just one year was pro-moted to deputy minister. Management and servicecontracts have the added advantage of signaling tocivil servants and public managers that, if they failto deliver, alternatives are available in the privatesector. The threat to switch to a private provider hasto be credible to be effective. In Botswana, after longuse of expatriate managers, the water utilityswitched to a domestic manager, but the govern-ment has shown itself willing to rehire expatriates ifperformance deteriorates.
If the problem is one of poorly performing civilservants, incentives must go beyond managers.Governments and managers can agree to build intoa performance agreement a clear link from em-ployee performance to salaries and nonpecuniaryrewards. Yet at the same time, if employees are pro-tected by civil service labor practices, neither perfor-mance agreements nor management contracts maysuffice. A more effective method is to rely systemat-ically on service contracts, the way Chile has doneto improve its road maintenance. This approachguarantees that the job gets done and is an alterna-tive to the use of force account.
Pricing for financial independence
The third element in the successful provision of in-frastructure services on a commercial basis is the es-tablishment of reliable revenue sources that giveproviders more financial autonomy Reliance onrevenues directly related to services delivered willincrease the productivity of infrastructure suppliersand also often benefit users. With fewer budgetarytransfers, the government has less occasion to inter-fere, a fact key to managerial autonomy. For publicutilities, smaller subsidies give managers a greaterincentive to focus on cost reductions and to satisfyusers because payments from users have to coverthe cost of the service. In the case of public works, fi-nancing must rely mainly on budgetary transfers. It
is in the interest of both managers and users to en-sure the predictability and stability of these re-sources. More transparency in the process will in-crease the financial autonomy of managers.
Pricing for public utilities
Among public utilities in developing countries,gross revenues typically cover costs only in telecom-munications (Figure 2.2). Even so, local services aretypically underpriced, with the losses made up fromsignificantly above-cost charges for long distanceand international service. This difference betweentariffs and costs is a type of tax on users. In all othersectors the gap between revenues and costs implies agovernment subsidy to users. These subsidies varyfrom 20 percent for gas to 70 percent for water. Thelow ratios of revenues to costs illustrate how little oftheir costs public utilities recover; the financial lossesthus generated are made good by transfers fromgovernment. For public water utilities in LatinAmerica, annual financial losses represent 15 percentor more of the investments needed to supply the en-tire population with adequate services by the turn ofthe century.
Figure 2.2 Costs are seldom fully recoveredin infrastructure.
Ratio of revenues to costs2.0
1.5
1.0
0.5
0.0
Financial autonomy
47
Telecom Gas Power Water
Source: Telecom, ITU 1994; gas, World Bank data; power,Besant-Jones 1993; water, Bhatia and Falkenmark 1993,
The best way of reducing the gap between costsand revenues is to cut costs and achieve productiveefficiencyperhaps the most important lesson ofthe Bank's experience in infrastructure. Costs due topoor debt management are excessive in about one-third of World Banksupported infrastructure proj-ects. Maintenance problems that cause water orpower losses are even more common and costly. InCosta Rica the national water company estimates anannual loss of income from such losses equivalent to24 percent of investment planned for the next fiveyears. In Mexico City at the end of the 1980s, neglectof maintenance and the lags between tariff increasesand cost increases in the water sector required a fed-eral subsidy amounting to about 0.6 percent of GDPa year.
Once costs are controlled, well-established pric-ing principles can help achieve financial autonomyand reduce distortions in the allocation of re-sourcesreflected in the success of countries as dif-ferent as Botswana, Chile, Korea, and Singapore(Box 2.7). The infrastructure pricing strategy inthese countries aims at cost recovery sufficient toguarantee the financial independence of public util-
ities. This pricing strategy focuses on recovering thethree main cost components of most infrastructureutilities: connection, usage, and peak-capacity costs.
The cost of connecting a customer and maintain-ing that connection to distribution or collection net-works is typically levied as a periodic flat fee, oftenlinked to charges based on usage in a two-part tariff.The usage cost is easiest to recover when metering isavailable to measure use and charges are based onactual consumption. Such charges reduce waste andencourage more efficient use. In Bogor, Indonesia,raising tariffs to meet costs reduced water consump-tion by 30 percent in less than a year without anyobvious impact on health or economic production.Where metering has not been introduced, estimatesof usage are the rule. In Colombia and Thailand,fees rise with the diameter of the pipe. In India, thefee increases with the value of the connected prop-erty. These solutions are not perfect and require fre-quent monitoring, but they often are the best optionavailable. The move to metering depends on the pri-ority given to recovering costs. One outcome of theend of subsidies to Ghana's water utility in 1988was an increase in meter coverage from less than 30
Box 2.7 Designing tariffs to achieve financial autonomy while addressing multiple goals
The general principle for pricing public services to re-cover costs without distorting the allocation of resourcesis to st the price equal to all short-run costs incurred inefficiently producing an additional unit of output (for ex-ample, an extra gallon of water or a cubic meter of gas)while keeping productive capacity constantthat is,price equals the short-run marginal cost. However,telecommunications, power, and water systems periodi-cally require large investments. In such cases, averagecosts fall as production is increased, and the efficientprice is below the average cost. Charging that pricewould result in a deficit and hence a loss of financial au-tonomy. But even when there are no such economies ofscale, financial autonomy is at risk when public pro-viders have an obligation to address social concerns(Chapter 4).
Adjustments in the general pricing formula can beused to avoid an operational deficit and minimize thetradeoffs imposed by the need to jointly address equity,efficiency, and financial goals. In general, if financial au-tonomy is a requirement, the public price has to be re-vised to cover the cost of providing the service plus amarkup, often resulting in multipart tariffs and possiblecross-subsidies. Two common options to minimize thedistortions (to efficiency and equity) of achieving finan-cial autonomy are increasing-block tariffs and time-of-use rate structures.
Under an increasing-block tariff, consumption of ser-
vices (usually water or power) is priced at a low initialrate up to a specified volume of use (block) and at ahigher rate per block thereafter. The number of blocksvaries from three to as many as ten. The most effectivestructure is the simplest, in particular when monitoringand administrative capacity are constraining.
Under the time-of-use rate structure, users pay a pre-mium during periods of high demand. This structure en-courages users to shift demand to the off-peak periodand has the added advantage of increasing the overallutilization of capacityand it often increases profits.Time-of-use rates have been applied to railways, urbanbuses, and subways, but they are more common in utili-ties such as power, water, and telecommunications.Time-of-use rates are practical for infrastructure supplynetworks in which the product cannot be stored cheaplyand its use can be partitioned by time slices into multipleproducts. Time-of-use rates often vary by time of day forpower and telecommunications, and by season for nat-ural gas (to reflect seasonal demand for heating) andwater (to reflect seasonal supply, especially in dryseasons).
Tariffs can also be differentiated in other ways. Forinstance, when service costs differ by region, pricesshould reflect these differences. In Nairobi, Kenya, the1975 cost of providing water at higher elevations was 32percent higher than the cost in lower parts of the city.Prices should vary with such differences.
48
percent to 53 percent in 1993, and in revenue collec-tion from less than 50 percent to 91 percent ofbillings.
One aspect of cost recovery that separates goodperformers from poorer ones is that good perform-ers recover the costs of maintaining sufficient capac-ity to meet peak demand by levying a charge basedon potential demand or actual consumption at peak.This method helps avoid power outages and watershortages. In other words, good performers aremuch more careful than others in assessing de-mand. In Colombia, India, and Korea, this capacitycost is charged only to the largest commercial andindustrial users because they tend to be the mainsource of peak demand.
Just as important as the incentive to meet objec-tives negotiated with the government is the finan-cial independence that allows public managers torely on the price system to assess users' willingnessto pay. Reliance on the price mechanism is in the in-terest of users because it directs provision towardpreferences determined by users rather than bu-reaucrats. Users are willing or able to pay moreoften than they are given credit for (Chapter 4).
What keeps so many public utilities from recov-ering costs is political constraints. Low prices arepopular among those who receive a service even ifthey are willing to pay more. In Bangladesh, In-donesia, Pakistan, and the Philippines, receipts fromirrigation user fees are 20 to 90 percent less than thecost of operation and maintenance. This shortfall re-flects the strength of the farmers' lobbies and theirability to get political endorsement for high subsi-dies. Moreover, with subsidies guaranteed, publicmanagers have little incentive to perform well or toimprove their responsiveness to users. Without po-litical support, the needed organizational changessuch as linking managers' rewards to the financialperformance of the department or utilitywill notsuffice.
COST RECOVERY AND THE POOR. Many govern-ments fear that fully recovering costs will hurt thepoor, yet increasing prices to enable cost recovery inthe delivery of services may actually help the poor.They often pay much higher prices per unit for pri-vately provided water and lighting because they arenot connected to public service networks that havelower unit costs, and because they do not benefitfrom subsidies to users of the public systemusu-ally the better-off. Expansion of access benefits thepoor by allowing them to rely on less costly sourcesof water and power. (Cases in which subsidies areneeded are discussed in Chapter 4.)
This effect has been demonstrated most convinc-ingly for water, where the concerns for the poor areproperly strong. In the Brazilian city of Grande Vito-na, Espirito Santo state, the willingness to pay fornew water connections in 1993 was four times thecost of providing the service, while the willingnessto pay for sewage collection and treatment was 2.3times its cost. Without treatment before disposal, thewillingness to pay falls to only 1.4 times the cost be-cause untreated sewage creates health problems andreduces the recreational value (mostly the fishingyield) of the waters into which it is discharged.
The willingness to pay for water is high for goodreason. For the poor, easier access to water can freeup time that can be used to pursue income-earningactivities. In rural Pakistan, women with access toimproved water supply spend nearly 1.5 fewerhours a day fetching water than do women withoutthis access. Such savings are reflected in the valueusers attach to the services. In Haiti a household'swillingness to pay for a new private connection in-creases by as much as 40 percent if the current watersource is at least a kilometer away.
The poor are not simply willing to pay in theory:they are paying in practice. During the mid-1970s tothe early 1980s, people in seventeen cities surveyedwere paying private water vendors an average oftwenty-five times the prices charged by the utility.In Nouakchott, Mauritania, and Port-au-Prince,Haiti, vendors were charging up to a hundred timesthe public utility price. Expanding the public utilitynetwork to give the poor access would mean thatthey would pay less than they are now willing andable to pay private providers.
Public works and financial autonomy
Making public works agencies financially indepen-dent does not mean that the public organizationcollects revenue directly from users to cover its op-erational costs. For public works, it is difficult or im-possible to measureand hence to priceindivid-ual use. Nevertheless, a predictable and transparentflow of revenue is necessary. based on user fees andstandard budgetary allocations from government.To some extent, the goal is one of financial account-ability rather than financial autonomy because themain objective is to achieve predictable and ade-quate financing. The key to the success of Korea'shighway corporation has been making the perfor-mance of the organization more transparent (aprocess described in Box 2.4) and linking budgetarytransfers to performance. But in many developingcountries the budgetary process does not allow for
49
Box 2.8 Can earmarking improve highways?
When budgetary processes work well, they assign fundsto activities with high economic returns or high priority.In such cases, earmarkingthe assignment of revenuefrom a specific fee or tax to a specific activity or expendi-ture, such as road maintenanceshould be avoided be-cause it impedes the ability of the budget process tomove funds from one activity to another. In times ofbudgetary stringency, earmarking shields expendituresin protected sectors and focuses budget reductions onunprotected activities. In countries with narrow taxbases, earmarking can encumber a large share of taxrevenues.
In many countries, however, budgetary processes donot systematically assign funds to activities with high re-turns. In the road sector, high-return maintenance activi-ties are often underfunded because budgetary resourcesare assigned one year at a time. Underfunding happensin spite of the commitment to fund maintenance for mul-tiple consecutive years that is implicit in the assessmentof the investment decision. Rate-of-return calculationsassume a pattern of maintenance that requires minimumfunding year after year. The failure to assign appropriatepriority to road maintenance explains, to a great extent,the deterioration of many national road systems. Ear-marking can ensure that needed road maintenance willbe reliably funded.
For the past few years, road funds have been en-couraged by the Bank in many African countries whereunderfunding and inconsistent flows of funds disruptmaintenance. The case for earmarking there is based on
the high rates of return for maintenance, among thehighest in the public sector. As long as poor budgetarypractices and policies lead to preferences for invest-ment over maintenance, and as long as the rates of re-turn on maintenance remain high, earmarking willavoid the underfunding of maintenance and improvethe allocation of resources in the short run. But this maybe a short-term solution to a long-term problem andneeds to be reviewed periodically.
The establishment of new road funds involves morethan just earmarking revenues to road maintenance. Italso includes reforms to improve the efficiency of roadagencies and the establishment of road boards with tech-nical experts and representatives of the user community,who oversee the allocation of revenues and the setting ofpriorities. Countries in Africa are starting to adopt apromising "commercialization" approach to makingroad fund operations more economically based andmore user-responsive; Tanzania provides a noteworthyexample of best practice. Moreover, the automatic rev-enue flows have been designed to avoid building up afund surplus and hence to discourage wasteful spend-ing. These additional reforms are necessary because ex-perience shows that the mere existence of earmarkedroad funds does not mean that a government is commit-ted to maintenance. Nor does it ensure that maintenancewill be efficient. Colombia had a road fund for more thantwenty years, but abandoned it in 1991 because the re-sources were going to sectors other than roads in manycases.
such a clear link between resources and perfor-mance, and many public works departments havebeen trying to increase their own sources of rev-enue. Doing so is easier for local public works agen-cies than for highway authorities because the bene-ficiaries of local services are more easily identified.
FINANCIAL INDEPENDENCE OF HIGHWAY DEPART-
MENTS. In principle, departments can increase theirshare of own-revenue sources by making beneficia-ries pay, directly or indirectly, for road use. Userspay many road-related fees on vehicle ownership,such as license charges and taxes on vehicle acquisi-tion, registration, and inspection. They also paycharges for use, such as fuel taxes, tolls, or parkingtaxes. Such road-user charges usually fall far shortof costs, however. In Zambia in 1991, road-usercharges (mainly license fees and road tolls) financedonly 10 percent of the total spending on roads, withgeneral budgetary revenue making up the shortfall.
The gap between user payments and expendi-tures arises because road-user charges often do notcover the costs that different types of vehicles im-
50
pose on roads. In Ghana heavy trucks use four tofive times more fuel than cars, but their axle load-ings, often ten times higher than those of cars, causeroad damage many times higher than cars. The wayto handle this difference is through such supple-mentary taxes as annual licensing fees that vary byvehicle weight. In the case of articulated trucks, ap-propriate licensing fees based on weight have beencalculated at $2,550 in Tanzania and $3,000 inTunisia. But road users resist paying such high roadtaxes where roads are in poor condition.
Some countries have taken to financing roadfunds through the allocation of specific user fees(such as tolls or fuel tax revenue) for specific activi-ties such as maintenance. This narrow earmarkingof specific taxes and fees that are closely related touse of facilities helps overcome resistance to taxes.The practice is common in Latin America, theUnited States (for roads), and some Asian countries(special accounts in Japan, Korea, and the Philip-pines). The desirability of such earmarking hingeson practical rather than theoretical issues in mostdeveloping countries. In general, if the budgetary
process works well, earmarking should be avoided(Box 2.8 gives guidelines).
COST RECOVERY FOR LOCAL INFRASTRUCTURE EXPEN-
DITURES. Local governments have been more suc-cessful in recovering costs indirectlyas in Colom-bia, for example, where "valorization" taxes pay forstreet improvements, water supply, and other localpublic services. With valorization, the cost of publicworks is allocated to affected properties in propor-tion to the benefit the work is expected to bring. Im-portant for success are the participation of prospec-tive beneficiaries in planning and managingprojects, care in planning and implementation, aneffective collection system, andin many cases--significant advance financing from general govern-ment revenues so that works may be started ontime. In Korea and North America, local infrastruc-ture development has recently been financed usingexactions, lot levies, development charges, and sim-ilar mechanisms to levy charges on would-be prop-erty developers to cover the added demands theirdevelopment will impose on the urban infrastruc-ture. The success of local taxes in contributing to thefinancing of infrastructure also depends on thequality of a city's institutional infrastructuresuchas its records, valuations, and collections. Each localtax requires technical expertise and political will inits implementation.
The need for a political commitment to reform
This chapter has focused on one essential element inthe effective public provision of infrastructure ser-
vices: the adoption of commercial principles. Abid-ing by these principles will be unsustainable, how-ever, if they do not reflect a political commitment toimprove public sector delivery. Political commit-ment underlies good public sector performance inSingapore and the sustainability of reforms inKorea's public enterprises. It also explains whyBotswana has been willing to search internationally,not just locally, for the best managers of its publicentities.
Explicit or implicit contracts between policymak-ers and managers or operators have been used effec-tively to generate political commitment. The out-standing common element in contracts used by themost successful countries is that they are governedby clear rules. Among contracts that maintain own-ership in the public sector, service contracts seemthe most promising in this respect. Moreover, theytest the capacity of the private sector to contribute tothe provision of infrastructure. Thus, service con-tracts may be the most useful complement to corpo-ratization and may provide a ready means of alter-ing the partnership between the public and theprivate sectors. Performance agreements have beenthe least successful because they often endorse dis-cretionary decisions driven by the many conflictingor evolving government interests.
Simply establishing commercial principles andmaintaining them through political commitment arenot sufficient for the success of commercial enter-prises, however. The missing element for success isthe introduction of competition with appropriateregulation. That is the focus of the next chapter.
-
51
Market forces and competition can improve theproduction and delivery of infrastructure services.That is the consensus emerging from a reevaluationof the sector based on experience, technologicalchange, and new insights into regulatory design.
This new consensus is displacing the long-heldview that infrastructure services are best producedand delivered by monopolies. Because the unit costsof delivering an infrastructure servicea gallon ofwater, a kilowatt-hour of electricity, a local tele-phone calltypically decline as service output in-creases, provision by a single entity seemed to makeeconomic sense. To limit the undesirable exercise ofmarket power, government was expected to be thesole supplier or to closely regulate the privatemonopoly.
Technological change and, even more important,regulatory innovation are making competitionpossible in many forms. The economies from large-scale production and delivery, although still impor-tant in some infrastructure activities, have dimin-ished, especially in telecommunications and powergeneration. Regulatory innovation has made possi-ble the unbundling of activitiesthe separating ofactivities in which economies of scale are not impor-tant from those in which they are. Unbundling pro-motes competition by detaching activities that wereearlier performed in monolithic organizations andopening them to various forms of competitive pro-vision. Even when infrastructure service is providedmost economically by a single suppliermakingcompetition in the market inadvisable or even infea-sible - competition from alternative suppliers forthe right to supply the market can spur efficiency
Market forces do not eliminate the need to regu-late prices and profits to protect consumers. How-
52
Using markets in infrastructureprovision
ever, where extreme underprovision of services iscommon, as in many developing countries, con-cerns about a private monopolist restricting outputto boost prices and profits may have less force thanwhere networks are better developed. Thus, the reg-ulatory apparatus needs to foster efficiency and in-vestment both by eliminating outdated restrictionson the right to provide service and by assuring fairterms of network access to new entrants.
In the move from a government monopoly to amore competitive system, enforceable contracts arerequired to balance the interests of various parties inspecific projects and to provide the stability neededfor long-term investment. Also required are com-prehensive, transparent, and nondiscriminatoryrules of the game. Although these are desirable inthe long run, the evidence shows that the move toprivate supply and competition does not have towait for the rules to be embedded in a fully devel-oped statutory regulatory system.
Regulation itself is imperfect because the"right" regulatory mechanisms are not always evi-dent. It is also imperfect because effective imple-mentation of economic regulation requires an in-formation base and sophistication that are rarely (ifever) attainable. Regulators are therefore vulnera-ble to manipulation. Regulation can also have per-verse, unintended consequences when competitionfrom substitute goods and services is possible. Agreater appreciation of regulatory failure has led toprogress in the design of simple rules to which reg-ulators can precommit and that produce pre-dictable and consistent outcomes. Moreover, in-volvement of other interested parties, especiallyconsumers, can make the regulatory process moreeffective.
Unbundling services for competition
Should one company provide all telephone ser-viceslocal, long distance, cellular, data transmis-sionor should the elements of the telecommu-nications business be unbundled into separateenterprises? Is electric power provided most effi-ciently when generation, transmission, and distribu-tion are coordinated within a single entity or shouldthe stages involved in delivering power be sepa-rated? Should a railway be a monolithic organiza-tion owning all facilities and offering a variety ofpassenger and business services, or should servicesbe operated as separate lines of business, possiblyunder independent ownership?
Central to this discussion is the concept of a nat-ural monopoly, which is said to exist when oneprovider can serve the market at a lower cost thantwo or more providers could. Such is the case whenthe costs of producing and delivering a service de-cline with increasing output (a condition often re-ferred to as economies of scale). In infrastructuresectors, it is also common for providers to supply anumber of services, some of which are natural mo-nopolies and others of which are not. However, anatural monopoly in one service may allow theprovider to gain an advantage in another servicethat can be competitively provided. This occurswhen it is cheaper for a single provider to produceand deliver two or more services jointly than forseparate entities to provide the services individually(and, when that happens, economies of scope aresaid to exist).
By isolating the natural monopoly segments ofan industry, unbundling promotes new entry andcompetition in segments that are potentially com-petitive. Failure to unbundle can constrain an entiresector to monopoly provision even when numerousactivities can be undertaken competitively. In thepast, maintaining sectors in a bundled form hasbeen justified on two counts. First, where economiesof scope are significant, unbundling raises the costsof provision. However, the gains from economies ofscope, where they do exist, need to be weighedagainst the benefits of cost-minimizing behaviorunder competitive pressures. Second, subsidy ofone service by another has been extensively under-taken within enterprises offering multiple servicesand has been the main mechanism for subsidizingservices to poor customers or those in remote areas.Unbundling, however, is desirable because it makescross-subsidies between different lines of businessmore transparent, identifies more precisely the sub-sidies needed to deliver services to the poor, and im-
proves management accountability. The trend is un-mistakable: unbundling of infrastructure services isproceeding at a brisk pace.
Vertical unbundling
The electric power industry illustrates how regula-tory and technological innovation interact. In 1978,the Public Utilities Regulatory Policy Act (PURPA)required electric utilities in the United States topurchase power from independent power produc-ers. This requirement opened up the industry tomore efficient generators, including those that pro-duce power from waste heat in manufacturing op-erations (cogeneration). Combined-cycle gas tur-bines, using clean natural gas and requiring smallinvestments, also became popular, although manyindependent power projects continue to use con-ventional technologies.
Such vertical unbundling separating electricitygeneration from transmission and distributionhassince been effectively adopted in many developingcountries, allowing new, substantial entry in genera-tion. Countries that have operationally independentpower producers include Argentina, Chile, Colom-bia, Guatemala, and the Philippines. Independentpower projects are being constructed or consideredin Côte d'Ivoire, India, the Lao People's DemocraticRepublic, Pakistan, Sri Lanka, and Tanzania. (SeeChapter 5 on the financing of independent powerproducers.) In addition, to facilitate competition inthe distribution of electric power, transmission anddistribution have been separated in several coun-tries. The transmission agency handles the transportfunction, and generators and distributors contractdirectly for power supply. Transmission is likely toremain a natural monopoly. While the physical dis-tribution network will also retain monopoly charac-teristicsit would not be economical to run morethan one distribution line to a home or a businessalternative suppliers can and do compete for theright to supply over the single distribution line.
Similarly, in the natural gas industry, the well-head and the pipeline and local distribution sys-tems can be owned and operated by different enti-ties. In Argentina, Gas del Estado was until recentlyan integrated monopoly in both the transportationand distribution of natural gas, acting as the solegas trader. Today, ten distinct entitiestwo trans-port businesses and eight distribution corpora-tionsprovide these services, as well as gas treat-ment and storage. To demonopolize the natural gasindustry in Hungary, the OKGTa trust that oper-ated the entire oil and gas sectorwas split into six
53
regional gas distributors and an enterprise owningthe refineries, storage facilities, and transportpipelines. The liquid-propane gas operations thathad been part of OKGT's operations were priva-tized separately.
A key part of many rail transport reforms is toseparate track management from railway opera-tions. For example, in 1988 two rail organizationswere created in Sweden: Baverket is in charge oftrack investment and maintenance, while StatensJarnvagar operates the freight concession and pas-senger transport on trunklines. For its track services,Baverket receives a fixed charge per unit of rollingstock plus a variable charge reflecting the socialmarginal costs of operation (including those for pol-lution and accidents). Separation between track andoperation is implicit in many reforms of the rail sec-tor in developing countries, where specific services,such as passenger and freight, are being separated(see the following discussion on horizontal un-bundling). To be successful, such reform requiresthat operators be allowed access on a fair basis totrack outside their jurisdiction.
Horizontal unbundling
The second type of unbundling separates activitiesby marketseither geographically or by servicecategories. In Japan, the national railway was reor-ganized and split into six regional passenger opera-tors and one freight operator that rents track timefrom the regional railways. Gains from restructur-ing have been enormousfreight volumes, whichhad been falling before the restructuring, have risen,while unit costs, which had been rising, have de-clined; the need for government subsidies has con-sequently fallen. Other countries are now emulatingthe Japanese model. Argentina split the monopolyArgentina Railways into five freight concessionsand seven suburban concessions, with the efficiencygains reflected in a substantial reduction of the gov-ernment operating subsidy. The Polish national rail-way is to be divided according to region served andtype of service (Box 3.1).
Telecommunications lends itself to this kind ofunbundling as well. The operation of rapidly grow-ing radio-based cellular services is typically sepa-rated from the provision of traditional telephoneservices. In some cases, horizontal unbundling, ordivestiture, into a number of producers allows di-rect competition; in other cases, as when divestitureleads to regional monopolies, it allows for betterperformance comparisons and therefore more effi-cient regulatory monitoring.
54
But in other segments of telecommunications thedistinction between vertical and horizontal un-bundling is not always sharp. Specialized providerssell information services using communication linksowned by traditional network operators. In suchcases vertical unbundling between the provision ofnetworks and the supply of information services isneeded to allow fair competition between horizon-tally separated service operators.
Practical approaches to unbundling
Constraints on unbundling are both technical andeconomic. Attempting to force activities that areclosely interdependent into distinct boxes can im-pose high transaction costs as the coordination onceachieved smoothly within a single firm becomesmore difficult and less effective when handled be-tween firms. And having separate, vertically linkedmonopolies, each charging a markup over costs,may result in higher charges than with a single, ver-tically integrated firm.
However, that does not mean that the incumbentmonopolists who will always argue that Un-bundling will increase costsshould go unchal-lenged. There may well be options for allowing avertically unintegrated firm (for example, a powergenerator) to compete with a firm whose operationsspan the entire range of activities, although thatwould require a regulatory framework for ensuringinterconnection. As long as competition occurs on afair basis, the market outcome will indicate whetheror not genuinely important economies of scopeexist.
But even where the technology permits un-bundling, the legacy of history and institutionsoften limits the possibilities. In Hungary a telecom-munications law enacted in 1992 separated long dis-tance (including international) services from localtelephone services, which are under the jurisdictionof municipal authorities. Under the law, privateconcessions for local services were to be granted ona competitive basis. But practical problems inter-vened. As in other countries, local calling rates arevery low, attracting few investors to that part of thenetwork. And investors in the long distance servicefaced the prospect of bargaining with group aftergroup of local government officials on terms of in-terconnection to local networks. A compromiseawarded a single franchise for long distance ser-vices and 60 percent of the local network. Competi-tion for the rest of the local network was open tocompanies with demonstrable financial strengthand sound business plans.
Box 3.1 Divided they stand: unbundling railway services
As infrastructure markets, technology, and operatingpractices have evolved, the need for single ownershiphas diminishedeven in such traditionally monolithicoperations as railways.
Argentina. In 1989, following years of bad service,heavy losses, and government subsidies as high as 1 per-cent of GDP (9 percent of the public sector budget), theArgentine railway began to transfer operational respon-sibility for many services to the private sector.
All services were transferred on a concession basis,most loss-making lines and services were dropped, andthe railway's surplus assets were sold. There were fivefreight concessions, seven suburban concessions (includ-ing the Buenos Aires subway), and a remnant of intercitypassenger service that was transferred to provincial gov-ernments. In Buenos Aires the new company establishedto take over Argentina Railways' suburban operationstransferred the relevant lines to the new concessionairesand then regulated and coordinated all transport issues inthe area. A metropolitan authority was also established.
In their first two years of operation, the new railwayscarried about the same traffic as before (a downwardtraffic trend has been reversed), with only 30 percent ofthe labor force. Freight rates are falling, service is im-proving, and the level of annual government subsidieshas fallen from $800 million to $150 million. Some of thefranchises will have to be reconstituted as demand for
services evolves, and not all the commitments made inthe franchise agreements will be honored because somerequirements imposed as part of the franchise award arelikely to be unsustainable.
Poland. Polish Railways (PKP) is restructuring itsmonolithic railroad system along its principal lines ofbusiness: commercial freight (primarily coal), intercitypassenger, international passenger, and local and subur-ban passenger services. Eventually, PKP is expected tohave an infrastructure department servicing institution-ally separated lines of business, with suitable nondis-criminatory compensation for track use paid by each lineof business (in line with European Union directives).Suburban passenger activities will be spun off to localagencies or covered under "contracts" with national orlocal governments to provide unremunerative publicservices in return for adequate compensation. PKP willtransfer its liabilities (mainly surplus labor) and nonrailassets (mainly urban real estate) to a new authority. Itwill also seek to transfer its nonrailway activities to theprivate sector.
This reorganization will separate commercial ser-vices (unregulated and unsupported) and public ser-vices, such as urban and suburban passenger services,rural lines, and certain lines of strategic importance. Thepublic services are to be planned and paid for by publicauthorities at appropriate levels.
The range of market alternatives
Once sectors have been unbundled, competition canbe used to increase efficiency and new investment.In infrastructure services, the choice is not simplybetween unfettered supply in the marketplace andmonopoly government supply. Four intermediatearrangements for market-based provision are possi-ble, and often advisable. Three of them promotecompetition. The fourth, private monopoly, createsthe basis for greater accountability through a harderbudget constraint and more explicit regulation thangovernment monopoly.
Competition from substitutes. The threat of los-ing customers to suppliers of substitute productsprovides motivation and discipline.
Competition in infrastructure markets. Multi-ple providers compete directly with each other,while government regulatory control ensures faircompetition.
Competition for the market. Governments createcompetitive conditions through leases or conces-sions, and firms compete not for individual con-sumers in the market but for the right to supply theentire market.
Privatization of monopolies. Where monopoliespersist, transfer to private ownership generallyyields efficiency gains. Regulatory innovations thatreward performance (such as price caps and otherincentive mechanisms, discussed below) create thebasis for continued productivity growth.
Moving an existing enterprise to more market-based provision can lead to one or more of thesearrangements (Figure 3.1). Competition for themarket is Option B, public ownership and privateoperation (see Chapter 1); the remaining threearrangements are variations on Option C, privateownership and operation.
Competition from substitutes
Competition from substitutes is frequently disre-garded in discussions of natural monopolies in in-frastructure. Failure to take it into account can resultin perverse consequences. Energy and surface trans-port are the two most important areas where com-petition from substitutes brings pressure to bear onthe monopoly supplier.
A natural gas provider may be a monopolist, butnatural gas is only one possible fuel for the genera-
55
56
Figure 3.1 Unbundling activities increases the options for competitionand private sector involvement.
Competition
Unbundling
No unbundling
Competitiveactivities
Monopolyactivities
tion of electricity. Oil and coal can be used as well,and competitive pressures from oil and coal produc-ers can discipline natural gas suppliers. Germanyviews these competitive pressures as strong enoughto justify deregulating the natural gas industry, eventhough some aspects of gas supply have strongeconomies of scale.
Where railways are operated as a monopoly,shippers often have a choice between rail, road, andwater transport. In the United States the 1887 Inter-
Competitionin the marketwith entry of
new firms
Competitionfrom
substitutes
Monopolyfacility
Competition forthe market
via concessionor lease
Object ofregulation
state Commerce Act regulated railroads, but to sus-tain such regulation without undermining the prof-itability of railroads soon required regulation of theotherwise competitive trucking services, limitinggrowth in that industry In the 1970s and 1980s rail-roads and trucking were both deregulated, leadingto rapid growth in productivity.
Hong Kong's experience with urban transportfurther illustrates what can happen when servicesare regulated and substitutes are available. When a
Businesspractices,
environment,safety,
and antitrust
Right of accessto monopolyfacility andaccess price
Initial status Industry structure Options for competition
government-owned subway system began opera-tion, large buses became less profitable, and the rateof return that had formerly been guaranteed to buscompanies by regulation became unviable. Effortsto maintain the rate of return by raising fares onlarge buses caused passengers to abandon bustransport, leading to taxi shortages, overuse of cars,and continuing congestion.
Thus, when substitutes are available, regulationcan have especially perverse effects. To shore up re-turns in the regulated sector, regulators often ex-tend their reach to sectors in which natural monop-oly elements are weak. It is far better in thesecircumstances to allow the competition from sub-stitutes to discipline the conduct of the allegedmonopolist.
Competition in infrastructure markets
Although infrastructure markets with numeroussuppliers are rare, competition among a few rivalproviders can lower costs and prices. The theoryof contestable markets says that even whereeconomies of scale and scope favor a singleprovider, the existence of potential rival suppliers
that can contest the market limits the risks of mo-nopoly abuse. The implication is that, absent com-pelling arguments to the contrary, all new entrantsshould be allowed to provide services, with themarket deciding how many providers can operateprofitably. Potential competition is most effectivewhere new entrants have limited sunk costs of mar-ket entrythat is, when entrants can recover theirinvestments by selling their assets if they decide topull out of the business. Technological change andeasing of regulatory constraints are permittinggreater contestability
Much of the experience with direct competitionin infrastructure is relatively new, but the resultsvalidate the benefits of competition. Systematic evi-dence of efficiency gains from greater competitioncomes mainly from the United States, which, afteryears of regulation, has introduced a number ofmajor deregulatory initiatives over the past twodecades. In virtually all sectors, greater competitionhas led to lower prices or better services for con-sumerswhile efficiency gains and new technolo-gies or business practices have led to sustained prof-itability (Box 3.2).
Box 3.2 Regulatory cycles in the United States
With its long history of private infrastructure provision,the United States exemplifies the changes in regulatorygoals and implementation and the ensuing cycles in reg-ulatory policy. In the late nineteenth century and wellinto the early part of the twentieth century, much compe-tition prevailed, especially in electric power andtelecommunications.
An early instance of economic regulationthe Inter-state Commerce Act of 1887was concerned with mo-nopoly power in railway operations. The bounds ofeconomic regulation were extended gradually, but espe-cially during the 1930s and the Great Depression, to vir-tually all infrastructure sectors and to other areas of pub-lic interest (for example, creating service obligations andinformation disclosure requirements).
Delivery of infrastructure thus came to be based on aparticular social compact. The service provider was typi-cally provided with exclusive rights to specific markets,and, in return, the government took on the public re-sponsibility of ensuring that service obligations were ful-filled at "reasonable and just" prices. Inflationary pres-sures of the early 1970s caused regulators to interveneeven more heavily in the operations of service providers.Health, safety, and environmental regulation also gainedmomentum around this time.
Public dissatisfaction with regulatory outcomes re-sulted in a move to reduce economic regulation in manysectors in the late 1970s and 1980s. According to one es-
timate, 17 percent of the U.S. gross national product(GNP) in 1977 was produced by fully regulated indus-tries; by 1988, this proportion had declined to 6.6 per-cent as large parts of the transportation, communica-tions, energy, and financial sectors were freed ofeconomic regulation. Greater operational freedom andcompetitive threats stimulated service providers toadopt new marketing, technological, and organizationalpractices. The evidence from the United States points tosubstantial economic gains from deregulation, as shownin Box table 3.2.
Box table 3.2 Estimated gains from competitionthrough deregulation of infrastructure sectorsin the United States
Extent ofSector deregulation
Estimated annualgains from deregulation
(billions of 1990 U.S. dollars)
Note: Gains from competition cover net gains to producers (in terms ofprofits), consumers (prices and service quality), and industry employ-ees (wages and employment).Source: Viscusi, Vernon, and Harrington 1992; Winston 1993.
Helped in part by sectoral unbundling, competi-tion in infrastructure sectors has increased in thepast decade. The possibilities and conditions for ef-fective competition are illustrated below for urbantransport, telecommunications, and power.
URBAN BUS TRANSPORT. Competition has stimu-lated both innovation and cost reduction in urbanpublic transport. In Sri Lanka, for example, deregu-lation permitted the profitable operation of smallervehicles by small-scale entrepreneurs, substantiallyimproving service availability. Competitively ten-dered franchises or the granting of overlappingfranchises to competing associations of operators isbeing practiced successfully in several major citiesin Latin America and Africa.
The challenge is to combine competition, for itscost-reducing impulse, with residual controls to en-sure the quality of service and maintain operatingdiscipline. Fragmentation of ownership has in someinstances led to difficulties with route coordinationand, at times, to excessive congestion and unsafepractices. In some countries, at least part of the orga-nizing or regulatory function has been taken over byan operators' association. Experience with such as-sociations shows that, while some aspects of regula-tion can be successfully delegated to the private sec-tor, provisions are needed to ensure that regulatorypowers are not used to prevent new entry. More-over, public scrutiny and regulation on such mattersas passenger safety, service obligations, and pollu-tion are essential in this competitive industry
TELECOMMUNICATIONS. A major competitive ele-ment of special relevance to developing countries isthe advent of radio-based cellular telephone net-works. These networks have relatively low capitalcosts, making their market readily contestable.Radio-based telephones compete with existing localnetworksand in many countries, with one an-other. By 1993 Sri Lanka had licensed four cellularoperators, leading to tariffs that are among the low-est in the world: connection costs of $100 and oper-ating costs of 16 cents a minute. Compare thosecosts with the more typical costs in El Salvador$1,000 and 35 cents a minutewhich has a singleoperator. However, regulation is important to sus-tain competition. For example, in Mexico regulatoryaction was necessary to ensure fair interconnectionby cellular operators into fixed networks.
Long distance services will be the next arena ofcompetition in developing countries. Korea alreadyallows competition in international services. Othercountries are committed to permitting new entry in
58
domestic long distance services (Chile and Mexicoby 1996 and Hong Kong by 1997).
Although transitional issues arise when competi-tion is being introduced, pragmatic solutions can befound. In the past, long distance telephone callswere priced high enough to allow monopoly suppli-ers of telecommunications services to earn reason-able profits while keeping down the price of accessto the network and of local calls. With unbundlingand increased competition, this structure of pricesbecomes unviable, and rate rebalancing is required.But during the transition the incumbent operator issaddled with the old rate structure and serviceobligations. If new entrants are unencumbered bythese obligations, they will flock to sectors with arti-ficially high profitability, a "cream skimming" thatcan be economically inefficient.
Mexico and the Philippines have taken two dif-ferent approaches to resolving such conflicts. InMexico, Teléfonos de Mexico (Telmex) was awardeda six-year monopoly under a concession agreementin 1990. To begin to bring prices in line with costs,rates for local services were raised three or fourtimes over original levels. Telmex was required tofurther rebalance rates during the period of the con-cession; long distance rates have fallen, while ratesfor local services have risen steadily. The Philip-pines chose instead to encourage new entry imme-diately. New operators are prevented from servingonly the lucrative international services market andare required to provide 300 local exchange lines foreach line on their international gateway.
The opposite problem arises when the incumbentoperator acts to limit competition, placing the aspir-ing entrant at a disadvantage. This is especially thecase when the entrant's use of the incumbent'sestablished network is restricted, reducing the en-trant's reach until it has invested in possibly du-plicative network facilities. Such a bottleneck effectin facilities owned by the incumbent is also an issuein other sectors when they are vertically unbun-dledaccess to the railtrack is required by all ser-vice operators, and competitive generators need theright to transmit and distribute electricity over mo-nopoly facilities. Two distinct issues need to be re-solved for efficient interconnection of entrants: thephysical right of access and, at least as important,the price of access. No established norms exist forinterconnection pricing, although a variety of ap-proaches are being tried. Most favorable to the in-cumbent is an arrangement whereby the price of in-terconnection between a point on the network and acustomer is the retail price charged by the incum-bent less direct costs of operating that link. This
maintains the full profits of the incumbent and isalso socially optimal if the network is efficientlypriced and operated. In New Zealand such a rulehas led to new entry, although the rule has beenchallenged by the new entrant as anticompetitive.Other approaches seek to encourage entry by limit-ing interconnection charges to full costs incurred bythe incumbent (excluding profits accruing on thelink). Such charges (e.g., those in Australia) includean element for fixed costs of the network as well ascosts incurred due to universal service obligations.
The interconnection issue is acquiring increasingimportance in developing countries, and especiallyin Eastern Europe where multiple operators havebeen licensed. In Poland, for example, a 1990telecommunications law allowed independent oper-ators the right to develop networks in regions notserved by the government-owned telecommunica-tions provider Telekomunikacja (TP SA). Threelarge independent operators have been licensed toprovide local services, in addition to almost sixtyother small providers. Interconnection between TPSA and the independent operators involves provid-ing access to each company's network and sharingrevenues from this access. To date there is no onestandard interconnection agreement between TP SAand the independents. The telecommunication lawstates that each independent company must negoti-ate its own separate agreement with TP SA. Thislack of standard agreement has prevented the ma-jority of the independents from further pursuing thedevelopment of their local network. Without inter-connection, outside investors are hesitant to commitany resources until a strong and fair contract is es-tablished. Alternative models are being examined toprovide interconnection on fair terms. Developingcountries seeking to expand networks and new ser-vices may wish to consider a pricing system favor-able to entry, effective antimonopoly legislation, andprocedures for implementing both.
POWER. Electricity generation is another area inwhich unbundling can introduce competition.Using similar approaches in electric power genera-tion, Argentina, Chile, Norway, and the UnitedKingdom have created electricity pools that simu-late competitive market conditions. Generators bidfor the right to supply bulk electricity in time slots(as short as half an hour in the United Kingdom) byspecifying a supply schedule of price and quantity.The power pool manager aggregates these offersand arrives at a systemwide price based on esti-mates of demand for the particular slot. All offersbelow this "pool price" are then accepted. Not all
electricity is supplied in this form. Because poolprices tend to be volatile and unpredictable, bothsuppliers and buyers (mainly regional distributors)tend to enter into long-term contracts as well, rely-ing on the spot market for a relatively small share oftransactions. Having a choice of suppliers whencontracts are renegotiated maintains competitivediscipline.
If generating capacity is concentrated in one ortwo firms, they can try to influence the price atwhich electricity is purchased from them. Anti-trust laws can be used to prevent monopolistic orcollusive behavior. Effective competition, how-ever, may require splitting large generators intonew companies.
Competition in electric power is being extendedto retail distribution in the United Kingdom, start-ing with large consumers. Users whose peak de-mand is 100 kilowatts of power or more are not re-stricted to their local distributor, but may contractwith other distributors or directly with generators.About 45,000 businesses are eligible to shop for elec-tricity in this way. All customers will be able to doso by 1998.
In many developing countries, one legacy ofpoor public sector performance is the large under-used generation capacity of many large manufactur-ing firms. The market for electricity can be mademore contestable by allowing large manufacturerswith their own generating capacity to sell electricityto the public grid, creating competitive disciplineand fostering cost reduction. A systematic studyshows that, if firms in Nigeria were allowed to sellpower from their underused generating capacity,the unit costs of electricity produced by these firmswould fall considerably. Informal evidence suggeststhat the same is likely to be the case in many devel-oping countries.
Competition for the market
Where direct competition is not possible, efficiencycan be increased by means of competition managedthrough contractual arrangements, ranging fromsimple contracts for specific services to long-termconcessions that require operation, maintenance,and facility expansion. Although there is only a sin-gle supplier of the service at any point in time, com-petition occurs before the contract is signed and, inprinciple, when the contract (or concession) expiresand is due for renewal. Thus, there is competitionfor the market even though there is no direct compe-tition in the market during the term of the conces-sion. The commitments entered into through thecontract can then, within limits, provide an alterna-
59
Figure 3.2 Leases and concessions ininfrastructure sectors are common, evenin low-income countries.
Number of countries with leasesand concessions40
35
30
25
20
tive to relying on a full-blown independent regula-tory apparatus.
Leases and concessions are increasingly commonin infrastructure. Such arrangements are in full op-eration or under implementation in thirty-sevencountries, including eighteen low-income countries(Figure 3.2). In transport, concessions are primarilyfor large, fixed facilities such as ports and toll roads.Concessions are common in the water sector. Be-cause economies of scale remain important in watersupply, most countries have used mechanisms thatcreate competition for the market (Table 3.1). Evenamong these agreements, there is a wide variety ofarrangements.
The effectiveness of a franchise arrangement de-pends upon a number of factors. The incentives forfranchise holders to operate efficiently depend onthe criteria for awarding the franchise, which inturn vary with sectoral characteristics and govern-ment objectives (Box 3.3). The contractual provisionof services is most likely to succeed when the con-tract increases transparency and accountability byspecifying in detail the terms of operation. How thecontract is awarded is also important to its success,60
as is demonstrated by the successful award of a con-cession in Buenos Aires for water and sewerage, incontrast to a proposed concession in Caracas thatfailed to attract responsive bids. Buenos Aires bene-fited from a number of advantages that Caracas didnot share, including stronger support from govern-ment authorities, better technical and financialpreparation, more attractive initial tariffs, and lowereconomic risks to investors.
In practice, the original franchisee is rarely dis-lodged. In Hong Kong, which uses franchisingmethods extensively for infrastructure provision,only one bus company has lost its franchise in re-cent decades. In France, franchises tend to extendinto perpetuity. The incumbent enjoys significantadvantages in rebidding, which must be factoredinto efforts to make the market contestable.
LEASES. Under a lease, the government suppliesthe major investments for production facilities, anda private contractor then pays for the right to usethe public facilities in providing service. A leasegenerally awards the contractor exclusive rights tothe stream of revenues for a period of six to tenyears. The contractor bears most or all of the com-mercial risks, but not the financial risks associatedwith large investments. Such arrangements are mostpracticable in activities where investments come ininfrequent bursts, so that responsibility for opera-tions can be separated from responsibility for in-vestment. In France leasing has been used fordecades in urban water supply and sewerage, andthe model was recently adopted in Guinea (Box 3.4).
Leases allow a mix of ownership. In "landlordports," the public authority owns the land and in-frastructure facilities, while a private firm owns andoperates the superstructure. In 1986 Malaysia trans-ferred operation of the Port Kelang container termi-nals and berths to two consortia under leases. Theprivate operators, freed of many of the constraintsfacing the public operator, improved productivitysubstantially. Similar successes in Hong Kong,Japan, and Malaysia began a wave of such opera-tions in Asialeasing is now under way in China,the Philippines, and Thailand and is under consid-eration in Korea, Pakistan, and Viet Nam. At times,only parts of the port such as individual berths orcontainer terminalsare leased, leaving arrange-ments for other parts of the port unaffected.
CONCESSIONS. Concessions incorporate all thefeatures of a lease but give the contractor the addedresponsibility of investmentssuch as for specifiedextensions and expansions of capacity or for the re-placement of fixed assets. Concession arrangementsexist for railways, telecommunications, urban trans-
Low- Middle- Totalincome income
Water and sewerage
El Power
N Transport
Source Berg 1993
15
10
5
Table 3.1 Contractual arrangements for private water supply
Contract
Service
Management
Lease
Concession
Source: Triche 1993.
Applications
Meter reading,billing andcollection, andmaintenance ofprivateconnections
All features ofthe leasecontract, plusfinancing ofsome fixedassets
port systems, and water supply and treatment.SODECI, the private water company in Côted'Ivoire, has a well-established and successful con-cession contract (Box 3.5).
Argentina has recently had a flurry of concessionarrangements, some of which were made possibleby prior sectoral unbundling. In addition to the railand water concessions described above, the opera-
tion of the Buenos Aires subway system was offeredunder a concession and awarded on the basis of thelowest subsidy demanded to operate and invest inthe system. Highway maintenance has also beenopened to concessions, and it is funded by revenuesfrom tolls initiated on many highways in 1992.
A possible problem with leasing and concessionarrangements is that they may not provide suffi-
61
Incentives
Permits competitionamong multipleproviders, each withshort and specificcontracts
Contract renewedevery one to threeyears, andremuneration basedon physicalparameters, such asvolume of waterproduced andimprovement incollection rates
Contract bidding,with contractduration of aboutten years; providerassumes operationalrisk
Contract bidding,with contract periodup to thirty years;provider assumesoperational andinvestment risk
Examples
A public water company, EMOS, inSantiago, Chile, encouragedemployees to leave the company in1977 and compete for servicecontracts for tasks previouslyperformed internallyresulting inlarge productivity gains
Electricity and Water Company ofGuinea-Bissau (EAGB); contractawarded to Electricité de France,with about 75 percent of theremuneration guaranteed and apossible additional 25 percentbased on performance
Water supply in Guinea owned bystate enterprise (SONEG) andleased to operating company(SEEG) from 1989 for ten years;achieved large increases in billcollection
Côte dIvoire's urban water supplyconcession went to SODECI, aconsortium of Ivorian and Frenchcomparnes; SODECI receives nooperating subsidies and allinvestments are self-financed
cient incentives to maintain and expand the facili-ties in their charge. A private supplier that does notown the production facilities or is uncertain of con-tract renewal may depreciate assets rapidly for
short-run gain and skimp on routine maintenance.Most of these problems can be avoided. Explicitmaintenance requirements can be written into con-tracts, and compliance can be monitored. Private
Box 3.3 Tailoring concessions to sectors and government objectives
The method of awarding concessions or the right to op-erate is extremely important in determining the incen-tives to private sponsors. When the returns to the spon-sor are unrelated, or only weakly related, to theperformance of the operation, the benefits of privatesponsorship are forgone.
The goal is to ensure an attractive financial returnfor investors while safeguarding public interests. Onekey element of negotiation is the price the investorpays for the right to operate the serviceor the extentof capital or operating subsidy that the governmentmay provide. Other negotiating points are the pricethat will be charged for services, the concession period,and the rights and obligations at the end of the contractperiod.
This is a complex brew, with each element depend-ing on another. There is always a danger that the termsof a concession will allow investors to secure too high arate of return, or will fail to provide sufficient incentivesfor proper maintenance of the assets and provision ofservices.
To simplify matters, certain norms and conventionshave been adopted. The length of concession periods istypically related to the life of the underlying asset. Forexample, thirty-year concessions are common for tollroads, and fifteen years is common for power generationprojects (although for hydroelectric projects, thirty yearsis more likely). Contracts for solid-waste disposal are inthe range of four years, a period in which garbage trucks
depreciate considerably. But because trucks can be soldmore easily than assets underlying roads and powerplants, the contract period may be as short as severalmonths.
An interesting variation is used in telecommunica-tions, although it could be applicable also for indepen-dent power projects. The focus is not on the length of theconcession period, which can be indefinite, but on theperiod of the exclusive concession. In Mexico and Ar-gentina, the newly privatized companies have beengranted exclusive licenses for six to ten years, duringwhich they have certain investment obligations. Afterthe exclusive period, the government is free to allownew entrants.
The method of charging for the right to provide ser-vice can take different forms. In theory, it is most efficientto award a concession to the bidder who offers thelargest lump sum up front. Having paid a large initialfee, the operator will be motivated to operate the facilityin the most efficient manner. For large projects, however,where project costs and revenues are uncertain, revenue-sharing or profit-sharing arrangements can spread therisk (as in the Guangzhou-Shenzen highway in China).Where the government sees itself mainly as a guardianof consumer interest, it may choose to receive no fee butto award the contract on the basis of the lowest pricecharged to the consumer (which can later create prob-lems with quality of service and requires specification ofminimum service standards).
Box 3.4 Success of a lease contractGuinea's water supply
When the Republic of Guinea's water supply sector wasrestructured in 1989, it was one of the least developed inWest Africa. At that time a new autonomous water au-thority, SONEG, took over ownership of the urban watersupply infrastructure and assumed responsibility forsector planning and investment. SEEG, 49 percent gov-ernment-owned and 51 percent owned by a foreign con-sortium, was created to operate and maintain the sys-tem's facilities.
Under the ten-year lease contract signed withSONEG, SEEG operates and maintains the system at itsown commercial risk. Its remuneration is based on usercharges actually collected and fees for new connections.SEEG also benefits from improvements it achieves in thecollection ratio, from reduced operating costs, and fromreductions in unaccounted-for water. Since SONEG hasultimate responsibility for capital financing, it has strongincentives to seek adequate tariffs and to make prudentinvestments based on realistic demand forecasts.
To make sure the necessary tariff increases would beaffordable, the Guinean lease contract included an inno-vative cost-sharing arrangement. Under the agreementnegotiated by the government, the two sector entities,and the external financier (the World Bank), the con-sumer tariff was to be adjusted gradually from the firstto the tenth year of the contract. During this period theWorld Bank agreed to assume a declining share of theforeign exchange expenditures of operation, and the cen-tral government covered a declining share of the debtservice. By the tenth year tariffs were expected to coverthe full cost of water. Tariff increases have to date ex-ceeded the planned schedule, rising from $0.12 per cubicmeter in 1989 to about $0.75 in 1993. Despite higher tar-iffs, the collection ratio for private customers has in-creased dramaticallyfrom less than 20 percent to morethan 75 percent in 1993and technical efficiency andservice coverage have improved.
62
suppliers can be held responsible for documenteddeterioration of the capital stock (although this canbe problematic because some deterioration may bedue to poor construction). Eligibility for renewal canbe made contingent on the observed state of thecapital stock.
Privatization of monopolies
Another way to introduce market principles into in-frastructure is through privatization, which trans-fers assets out of the public sector. Privatizations arespreading rapidly in developing countriesthevalue of transactions reached more than $6 billion inboth 1991 and 1992 (Table 3.2). Privatization hasgone the furthest in telecommunications. Argentina,Chile, Hungary, Jamaica, Malaysia, Mexico, andVenezuela have all undertaken substantial privati-zations of telecommunications services. The powersector, too, has recently seen several large privatiza-tions.
Although privatization of industrial enterpriseshas a relatively long history providing evidence ofits positive effect on performanceprivatizations ininfrastructure are comparatively new. Privatized
public utilities typically undergo major corporate re-structuring, and the immediate gains from privatiza-tion have been impressive. A study of total welfaregains (net monetary gains to producers, consumers,and employees) found that in three cases involvingtelecommunications, the gains (as a proportion ofsales) ranged from 12 percent in the United King-dom to 155 percent in Chile (Figure 3.3). Two yearsafter the privatization in Venezuela, the total net-work had expanded by 50 percent and virtually alltargets for service improvements had been met (Box3.6). Disentangling the effects of privatization and ofincreased competition is not yet possible in manysectors, however, nor have sustained long-termgains in productivity growth yet been demonstrated.
Utility privatizations are often accompanied by arequirement to undertake certain minimum invest-ments. These so-called roll-out obligations are ex-emplified by the service conditions imposed onTelmex, the privatized Mexican telecommunicationsprovider. Network development targets built intothe concession require Telmex to achieve a linegrowth rate of at least 12 percent a yeartwice thegrowth rate achieved during the late 1980s. Tax in-centives reinforce Telmex's contractual investment
Box 3.5 Côte d'Ivoire's experience with a concession for water supply
An excellent example of a private company providingpublic services in West Africa is Côte d'Ivoire's SODECI.SODECI is an Ivorian company whose capital (about $15million) is owned 52 percent by local interests; 46 percentby Saur, the French water distributor; and 2 percent by agovernment investment fund. It started operations withthe Abidjan water supply system thirty years ago andnow manages more than 300 piped water supply sys-tems across the national territory. Until recently, SODECIoperated under concession contract for water productionin Abidjan, the capital city. It was under lease contractfor water production and distribution in all other urbancenters; for water distribution in Abidjan; and for man-agement of the Abidjan sewerage system.
To deal with financial troubles caused by govern-ment policies in the 1980s regarding sectoral investmentand tariffs, the urban water sector was reorganized.SODECI's contract for urban water supply services wastransformed into a concession contract for the entirecountry, with SODECI taking responsibility for both op-erations and investments. Today the company has300,000 individual connections that serve some 70 per-cent of Côte d'Ivoire's 4.5 million urban residents-2million in Abidjan and the rest in settlements of 5,000 to400,000 people. Under a policy to provide low-incomehouseholds with direct access to water, 75 percent ofSODECI's domestic connections have been providedwith no direct connection charge. The number of connec-tions is growing between 5 and 6 percent a year.
Since the early 1970s, full cost recovery has been therule, and revenues from water sales have fully coveredcapital and operation and maintenance costs. During thepast ten years, unaccounted-for water has never ex-ceeded 15 percent, and collection from private con-sumers has never fallen below 98 percent (collectionfrom government agencies is more problematic). More-over, despite the dispersion of operations, there are onlyfour staff per thousand connections, reflecting best-practice standards. The company has also succeeded inreducing expatriate staff while expanding operations.
SODECI retains part of the rates collected to cover itsoperating costs, depreciate its assets, extend and rehabil-itate distribution networks, and pay dividends to share-holders. It also pays the government a rental fee to ser-vice the debt attached to earlier projects financed by thegovernment.
SODECI provides service close to the standards of in-dustrial countries. Yet the cost to consumers is no higherthan in neighboring countries in similar economic condi-tions or in members of the CFA franc zone, where tariffsrarely cover capital and operation and maintenancecosts, and service lags behind. Private Ivorian interestsnow own a majority of SODECI's shares. Its bonds areone of the main items traded on Abidjan's financial mar-ket, and it has distributed dividends to its shareholders.The company has also paid taxes since its inception.
63
Table 3.2 Value of infrastructure privatizations in developing countries, 1988-92
obligations. In addition to line growth require-ments, the concession requires improvements in ser-vice quality Telmex has more than met the targetsand has announced plans to invest $13 billion overfive years to upgrade equipment, add access lines,and improve service.
Underpinning these requirements is the concernthat a monopoly service providersuch asTelmexmay restrict output below socially desir-able levels. While this may be a legitimate concernin the longer term, it sits uneasily with the currentsituation in many, if not most, developing countries.Levels of service provision are now so low that evenan unfettered monopolist would face strong incen-tives to expandand to do so at lower cost than thepublic sector providers of the past. Roll-out require-ments may consequently be unnecessary and, whenused to secure the provision of services on uneco-nomic terms to particular areas or consumers, canperpetuate pricing distortions.
Paths to market provision
The move from government monopoly to competi-tive market provision has taken many routes, but,
64
Note: Countries undertaking infrastructure privatizations:1988: powerMexico; telecomBelize, Chile, Jamaica, Turkey; airlinesArgentina, Mexico.1989: powerKorea; telecomChile, Jamaica; atrlinesChile.1990: powerMalaysia, Turkey; telecomArgentina, Belize, Chile, Jamaica, Malaysia, Mexico, Poland; madsArgentina; airlinesArgentina, Brazil, Mexico, Pakistan.1991: power generationChile, Hungaiy; power distributionPhilippines; railmadsArgentina; telecomArgentina, Barbados, Belize,Hungary, Jamaica, Mexico, Peru, Philippines, Venezuela; airlinesHonduras, Hungary, Panama, Turkey, Venezuela; shippingMalaysia; roadtransportTogo.1992: power generationArgentina, Belize, Malaysia, Poland; power distributionArgentina, Philippines; gas distributionArgentina,Turkey; telecomArgentina, Estonia, Malaysia, Turkey; railroadsArgentina; portsColombia, Pakistan; waterArgentina, Malaysia; air-linesCzechoslovakia, Hungary, Malaysia, Mexico, Panama, Philippines, Thailand; shippingSri Lanka; mad transportChina, Peru.
Source: Sader 1993.
whatever the path, success requires a sustainedcommitment to private entry. The transitionalphase can be effectively managed through enforce-able contracts that create incentives for the entre-preneur to be efficient while also embodying thepublic interest.
A statutory regulatory system that provides forclear and open enforcement of the terms of the con-tracts is also required, although its absence has notheld up private entry The design of such regulationmay well benefit from contractual experience withearly entrants. Effective statutory regulation re-quires predictable and nondiscriminatory rules andthe creation of consumer constituencies.
Transitions in market structures
Should the move to a market-based system occurin a single step, or can it be achieved more gradu-ally? There are no simple answers. What is impor-tant is that the shift to market provision be credi-ble. Without that, private entrepreneurs are notlikely to take on new investments. Commitmentsfrom governments are most credible when all theenabling measures needed for private entry and
(millions of U.S. dollars)
Total, Number ofSubsector 1988 1989 1990 1991 1992 1988-92 countries
Total developing country privatizations 2,587 5,188 8,618 22,049 23,187 61,629 25
market provision are adopted within a short spanof time as part of a consistently designed program.Where institutional legaciesconcerns aboutlabor redundancy, for example prevent immedi-ate privatization, opening the sector up to substan-tial new entry may be a strong sign of governmentcommitment to sector reform.
One recommended sequencing strategy is to startwith the design of statutory regulation that sets therules of the game. This is to be followed by the de-termination of the appropriate industry structure(the degree of unbundling, extent of new entry, andsplit of existing providers to prevent economicdominance) and privatization. Chile comes closestto having implemented this sequence over the pe-riod of a decade, although industry structure hascontinued to evolve after privatization. Other coun-tries have followed pragmatic strategies dictated bytheir circumstances, with impressive results. Threeexamples illustrate transitional options and issues.
ARGENTINA. Argentina has adopted the most far-reaching privatization program, designed to createcompetitive conditions in the economy. All majorinfrastructure providers were privatized between1989 and 1993, and activities were unbundled to fos-ter competition. In the electric power sector, genera-tion, transmission, and distribution were separated;two telecommunications franchises were awardedto serve the north and the south; and railways wereseparated along different lines of business.
Although privatization has occurred rapidly, thecapacity for regulatory oversight has lagged (as ithas in most developing countries other than Chile,where sophisticated regulatory capabilities wereput in place prior to privatization). The absence ofregulatory oversight has not been an impediment sofar; however, where market forces do not provideadequate discipline, efficient functioning will re-quire regulation. Antitrust regulations will needparticular attention in view of the heavy concentra-tion of ownership. The Chilean experience, with oneprivate firm owning 65 percent of generating capac-ity, shows that a dominant provider can influencemarket outcomes. Also in Chile, concerns have beenexpressed that the large installed base of the localtelephone company may prevent fair competitionwhen the company begins to provide long-distanceservices. And everywhere, market provision will re-quire greater information disclosure and publicfeedback.
PHILIPPINES. In the Philippine power sector, pri-vate provision was based entirely on the entry of
Figure 3.3 Privatizations intelecommunications can lead to large gains.
Total welfare gains as a percentageof annual sales180
160
140
120
100
80
60
40
20
0
new generators. Opening generation to new pro-viders required the elimination of the monopoly en-joyed by the National Power Corporation, a govern-ment-owned utility that has not been privatized.
These reforms came in response to an almostcrippling power shortage. The urgency was so greatthat new entry had to be based on contractual agree-ments between the government and private genera-tors, since reform of the Electricity Regulation Boardwould have taken too long. By August 1993, sevennew projects with a combined capacity of 800megawatts had been completed, and five additionalgenerators were placed under private contracts forrehabilitation and operation. Fifteen more projects(2,000 megawatts of capacity) are under negotiation.
65
Telecom Telmex British(Chile) (Mexico) Telecom
Note: Welfare gains are the sum of gains accruing to allpartiesenterprises, workers, and consumers.Source: Galal, Jones, Tandon, and Vogelsarig forthcoming.
Box 3.6 Telecom privatization: the case of Venezuela
When Venezuela privatized its state-owned telephonecompany (CANTV) in December 1991, it had 1.6 millionlines in service (8.2 lines per 100 people as comparedwith 35 lines per 100 people in Korea. An eight-year waitfor a new telephone was common, and completion ratesfor international calls were less than 20 percent.
The government sought to expand and improve basicservice rapidly by turning the company over to a privateoperator with first-class international experience. Al-though it recognized the need to increase local rates sub-stantially, the government was concerned with the po-tential political fallout from "rate shock" Consequently,it decided to phase the rate rebalancing over nine years.During this period the new operator was granted an ex-clusive franchise for local, long distance, and interna-tional service. The profits from international servicewould be used to cross-subsidize local service and fi-nance the desired network expansion. The concessioncontract included annual obligations to expand and im-prove basic service (including the installation of 3.6 mil-lion additional lines over nine years) and a cap on the in-crease in prices for basic telephone services. All otherservices were open to competitionincluding cellularservice, private lines, information services, and equip-ment. This model resembled telecom privatizations inMexico and Argentina, where the privatized operators
were granted a limited monopoly on basic service (sixyears in Mexico; seven years extendable to ten in Ar-gentina).
In Venezuela the process culminated with the suc-cessful public tender sale of a 40 percent share (but withmajority voting control) to an international operatingconsortium for $1.9 billion. Pending passage of a newtelecom law, the government enacted a series of decreesthat established the regulatory agency, CONATEL, anddefined the regulations for the various types of service.Until the new law is passed, rate increases must be rati-fied by the government.
In the two years following privatization, CANTV in-vested more than $1.1 billion and installed 850,000 newand replacement lines, far exceeding its obligationsunder the concession contract. Virtually all service im-provement targets were met.
Several lessons have emerged from the Venezuelanexperience. Even without a fully defined legislativeframework, telecom privatization can provide immedi-ate benefits from increased investment. Although somerebalancing of tariffs has occurred, sustained tariff in-creases will be needed. With rapidly changing technol-ogy, monopoly rights granted to maintain cross-subsi-dies and to promote service expansion will proveincreasingly difficult to define and enforce.
During this process, new laws and administrativeprocedures have also been put in place (Box 3.7).
Although regulation through individual con-tracts has attracted new investment to the powersector, further development will require sectoralrules to ensure fair competition. As in most devel-oping countries, new generating capacity has beendeveloped without well-agreed principles on inter-connection and dispatch among providers. This ab-sence has not been a problem so far, partly becauseprivate supply was filling large demand gaps. Asthe gaps close, however, the various suppliers willcome closer to being competing sources of power,and the regulatory authority will have to defineclear rules for determining whose power is boughtand on what terms.
MALAYSIA. Malaysia's approach puts it some-where between that of Argentina and the Philip-pines. Utilities have been gradually privatized, andnew entry has been allowed in electric power andwater. Statutory regulatory efforts have lagged, anddiscipline on operations is imposed through con-tractual agreements. The government also hasmaintained direct regulatory supervision of large
66
utilities through continued shareholding or through"golden shares" that give the government vetorights, especially on matters relating to the socialobligation of the utilities (Box 3.8).
Dealing with regulatory imperfections
Regulation must negotiate many potential pitfalls:as it controls the exercise of monopoly power, itmust also ensure service quality safety environ-mental protection, service obligations, and therights to network access (Figure 3.1). The weight ofeach of these objectives varies with industry struc-ture, which evolves over time. Flexibility musttherefore be balanced with commitment to fixedrules. Too much flexibility lets well-organized inter-est groups gain control of the regulatory process, totheir own benefit. Too rigid a regulatory structurelimits the ability to correct mistakes and adapt tochange. It can also stifle initiative. Regulation some-times leads to outcomes worse than those that im-perfect markets could achieve.
Experience argues for keeping regulation to aminimum. Three considerations influence the regu-latory task that accompanies the introduction of pri-vate sector involvement:
Providing sufficient resources, autonomy, andcredibility for the regulator
Where price regulation is necessary, choosinginstruments that encourage cost efficiency in theregulated entity
Creating constituencies in the regulatoryprocess.
REGULATORY RESOURCES, AUTONOMY, AND CREDI-
BILITY. Regulation requires detailed knowledge andcontinual monitoring of the activity concerned. Theregulatory menu includes problem identification,fact finding, rulemaking, and enforcement. Regula-tors need to be able to shift course in order to antici-pate or respond to changing conditions in the indus-try. They also need operational autonomy within abroad policy mandate to ensure their effectiveness.Because doing all of this well requires a detailedworking knowledge of the industry, there is a strongcase for regulatory bodies to be specialized and au-tonomous public agencies, rather than general bu-reaucracies. But because sectorally specialized agen-cies are more susceptible to capture by theindustryand so are more likely to perpetuate reg-ulation that favors incumbents the regulatoryagency must be monitored as well.
Much of the experience with statutory regulationderives from North America, where the private (al-
though often monopoly) provision of infrastructureservices has been the norm. The United States, rely-ing on federal and state commissions, has devel-oped a significant capacity for autonomous regula-tion. Although the process is remarkably open, it isalso characterized by adversarial relationships andlitigation. Europe and Japan have had less experi-ence with explicit regulation, since they rely onpublic monopolies, combined with regulatory andoperational responsibilities. Even when regulatoryinstruments such as price controls, technical stan-dards, and entry licensing have been used, theyhave been implemented by related ministries or in-terministerial committees rather than by specificregulatory agencies. The United Kingdom has re-cently moved toward privatization and indepen-dent regulation, and similar reforms are takingplace elsewhere in Europe. Developing countrieshave virtually no experience with regulation of pri-vate providers because their infrastructure enter-prises have, in the main, been publicly owned andoperated. An exception is Hong Kong, which is wellknown for its encouragement of private initiativebut which has a regulatory system that protectsconsumer interests.
A problem for developing countries is assem-bling experienced professionals to staff a regulatoryagency. Regulators have limited resources and are
Box 3.7 The evolution of private power in the Philippines
The Philippines' evolutionary approach to attracting pri-vate entrepreneurs in power generation is instructive. InJuly 1987 private power generation became a deliberateelement of government policy and effectively signaledthe end of the generating monopoly of the state-ownedNational Power Corporation. Although the first project,Hopewell Navotas 1, was successfully negotiated andcommenced operation in 1991, early dealings with otherprivate proposals were generally not fruitful. A proposalfor a 220-megawatt cogeneration plant did not proceedbeyond the negotiation stage in 1989, in part because ofinadequacies and inconsistencies in administering regu-lations. The lessons from the failed effort helped ongoingefforts to improve regulatory and clearance procedures.
After 1989, Philippine agencies associated with pri-vate power began to work in a more coordinated man-ner. There was greater participation from the NationalEconomic Development Authority (which had played akey role in initiating the private power program) andmore ranking of priorities through the Investment Coor-dinating Committee. A major improvement in the frame-work for reviewing and clearing proposals was the 1990build-operate-transfer law and its accompanying imple-
menting rules and regulationsthe law created a clearerlegal basis for allowing entry by private capital, thoughstill requiring transfer of ownership back to the govern-ment at the end of the concession period.
The Philippines is also seeking to streamline the pri-vate power solicitation process. Under present arrange-ments, the effectiveness of project contracts depends onseveral conditions that must be met after the contractsare signed. Delays or failures to meet certain conditionscan jeopardize a project. The National Power Corpora-tion is seeking to establish model contracts, preapprovedby concerned government agencies, to facilitate privateparticipation. This arrangement is expected to enable in-vestors to proceed immediately from signing the con-tract to finalizing the financing plan.
The urgency in creating new capacity in the Philip-pines led to expensive power generation. Early projectsused "peak-load" plants that can be installed rapidly butoperate at very high cost and are designed to serve onlyfor the few hours in the day when demand is very high.Subsequent projects, prepared under less time pressure,have addressed this concern. At the same time, experi-ence has allowed project size to grow.
67
Box 3.8 Regulation and privatization: which comes first? The case of Malaysia
Privatization of the infrastructure sector has progressedrapidly in Malaysia since the mid-1980s. In all cases, thegovernment department or statutory body that was pre-viously supplying the services has assumed the statu-tory role of regulating the privatized supplier. For exam-ple, the Kelang Port Authority is now the regulatoryagency supervising the two private operating companiesat the port, and the Telecommunications Department isthe regulator of the telecommunications sector. (Thechange in the function of the government department orstatutory body has, in each case, necessitated new em-powering legislation.)
Significant government equity ownership in for-
merly privatized enterprises and the mechanism of the"golden share" also play a role akin to regulation. (Thegolden share, which gives the government veto powerson major policy matters of the privatized firm, was firstintroduced in the privatization of Malaysian Airline Sys-tems and the Malaysian International Shipping Corpora-tion.) Equity ownership by the government and thegolden share are intended to ensure that the policies ofthe privatized firm are in conformity with governmentpolicies and national objectives.
Regulation in Malaysia has really meant the supervi-
sion of tariffs and the maintenance of service standards.The scope of regulatory action in Malaysia, however, isfairly rudimentary. For instance, no clear link exists be-tween the functions of the regulatory agencies and thecreation of incentives for the privatized supplier of infra-structure services to achieve efficiency.
Although regulatory agencies exist, the respectiveministers still appear to have considerable influenceover the policies of the privatized suppliers of infra-structure services. Rate revisions, for example, are notcompletely a matter for the regulatory agency to decideand almost always appear to require ministerial sanc-tion. There is also a distinct possibility that industrymight "capture" the regulatory agency in some cases. Atpresent, considerable ambiguity exists about the inde-pendence of the regulatory agencies from ministerial orpolitical interference.
Although still evolving, the regulatory mechanismdoes not appear to have limited new entry and invest-ment. The lesson from the Malaysian experience is thatmoves toward privatization and private sector provisionof infrastructure services need not wait for the formalcreation of a comprehensive regulatory framework.
often unable to attract qualified people. Even in Ar-gentina, which has a pool of well-qualified people,civil service salary restrictions and tight budgetshave led to weak regulatory agencies (Box 3.9).
Allowing a regulatory agency autonomy whilemaintaining its accountability requires a delicatebalance. If regulators are easily replaced, directlyelected at frequent intervals, or easily influenced byspecial interest groups, they may be unwilling toimplement policies that are socially desirable butpolitically inexpedient. Conversely, a regulator withtoo much discretion can, for example, arbitrarily re-strict new investment. Experience in Jamaica reflectssome of these problems (Box 3.10).
A few principles seem to have general accep-tance. It is important for a regulatory agency to re-port directly to the legislature rather than solely to(or through) a minister. Legislative scrutiny of regu-lators is typically more open, although informalpressures can creep in. The head of the regulatoryagency should be appointed for a fixed term, prefer-ably out of cycle with political elections. Scrutinyshould be regular and should systematically assessan agency's performance in achieving its goals andwhether regulation is well focused. Transparency iscritical to regulatory accountability because only if
68
the process and policies are known and publishedcan assessment of regulation be effective.
The Philippines, responding to the generally in-effective regulation of the past, has recently acted tomake the process more autonomous and account-able. A draft bill in the lower house of Congress de-fines the role of the National TelecommunicationsCommission more clearly, increases the number ofcommissioners, assigns a fixed tenure, and increasesthe commission's access to operational funds.
As regulators become stronger, "regulating theregulators" may be desirable, if experience in indus-trial countries is a guide. In the United Kingdom,for example, the National Audit Office audits regu-lators as part of a mandate to determine "value formoney" in public service, and the Monopolies andMergers Commission hears appeals of decisions bysectoral regulators.
INSTRUMENTS OF REGULATION. While regulatorsseek to maintain "reasonable" and "just" prices inorder to protect consumers, profits must be ade-quate and not subject to political risk or uncertainty.The ubiquitous instrument of regulation used tobalance these goals for sectors ranging from urbantransport to electricity systemshas been "cost-
plus," or rate-of-return regulation, which ensuresthat the financial return received by the providercovers all costs (operations and maintenance, depre-ciation, and taxes) and, in addition, guarantees a ne-gotiated return on investment.
In recent years this instrument has come in formuch criticism. Rate-of-return regulation is difficultto implementobtaining accurate information oncosts of production and the allocation of such costs
between alternative services is a formidable task.Determining an appropriate rate of return is also asource of much contention between the regulatorsand the regulated. These problems encourage mis-representation of information and the adoption ofinefficient technologies that inflate the base onwhich rates of return are calculated; they also fosterunproductive lobbying. Most important, because allcosts are covered and a rate of return is guaranteed,
Box 3.9 Development of regulatory capacity in Argentina
Although a well-defined regulatory framework waslegally in place after the privatization of telecommunica-tions, regulatory practice did not conform to the frame-work. Charged with regulatory responsibilities inNovember 1990, the Comisión Nacional de Telecomuni-caciones (CNT) did little until the end of 1991. No clearregulatory processes were developed, and a backlog ofdecisions began to pile up. Experienced staff were lack-ing, as were resources to hire additional staff or even payexisting staff on a regular basis.
The outcome of these regulatory and staffing gapswas that the development of new telecommunicationsservices proceeded slowly. This was due in part to CNT'sfailure to formulate standards and processes for issuinglicenses, making most of these services uneconomic.Meanwhile, a number of radio operators and telephonecooperatives, faced with little or no regulation, startedoperations without licenses. Consumers also suffered
from CNT's inability to effectively address servicecomplaints.
Since mid-1993almost three years after the begin-ning of the reform processCNT has improved its per-formance, in particular with respect to the concerns ofconsumers. A team of outside consultants working withCNT made progress in developing strategies and proce-dures. Moreover, after some early difficulties in the selec-tion process, CNT's top staff (6 directors) are now inplace. The selection was made by an independent pri-vate recruitment company after a rigorous screening of125 professionals, and its five nominees were retained asdirectors, including the president. The last director wasproposed by the provinces.
Progress in Argentina's telecommunications sectorhas been significant, and privatization has been able tomove ahead in spite of the delays in implementing regu-latory changes.
Box 3.10 Jamaica's regulatory roller coaster for telecommunications
Jamaican telecommunications were initially privatelyrun, then nationalized in 1975, and then reprivatized in1987. Investment under private ownership was stronguntil the 1960s and has been strong again since 1987. Butbetween 1962 and 1975 utility-government relationswere turbulent, and investment levels were low.
Repeated shifts of power between two opposing po-litical parties with divergent views have made it difficultto establish a credible regulatory regime that investorscould rely on with confidence beyond another election.Until 1962 the regulatory regimeincluding precise, en-forceable provisions on the rate of return the utilitycould earnwas built into the utility's operating license.Because of Jamaica's strong, independent judiciary, pri-vate participants were willing to invest, confident thatparliament would not unilaterally change the terms of alicense.
The newly independent Jamaican government de-cided in 1962 that a precisely specified operating licenseunacceptably constrained the democratic process. Usingthe United States as a model, the government estab-
lished the Jamaica Public Utility Commission in 1966.Not only was the commission open to representationsfrom all interested parties; the new system did not set afloor on the returns that the utility could earn. In theUnited States constitutional protections plus well-devel-oped rules of administrative process afford private utili-ties substantial protection, even though the private-utility commission system nominally gives regulatorssubstantial discretion. But Jamaica lacked these founda-tions. It also lacked a cadre of well-trained regulatorsand experience in delegating authority to a quasi-inde-pendent commission. Clashes between the utility andthe commission ensued, culminating in the 1975 nation-alization of telecommunications.
After the 1987 privatization, Jamaica returned to itspre-1962 regulatory system. It wrote into the operating li-cense of the newly privatized utility a guarantee of a 17.5to 20 percent annual rate of return on equity, shieldedfrom change except with the consent of the utility, andenforceable by the judiciary. The result was a surge in in-vestment and substantial welfare gains for Jamaica.
69
private management can become complacent aboutmaking the right investments and keeping costsdown.
The response has been to design new "incentive"regulations in which the prices a provider is al-
lowed to charge do not hinge on costs incurred.Thus, if costs increase, profits are lowered; if costsdecline, the provider and investors enjoy greaterprofits. Incentive regulation therefore seeks to moti-vate providers to use their superior knowledge ofoperating conditions to lower costs and introducenew services.
Price caps. An example of incentive regulation isthe increasingly popular price-cap, or "RPIX,"method for determining permitted increases in ser-vice price. RPI is the percentage increase in the retailprice index (other indexes of costs that the providerdoes not control can also be used), and X is the (pre-determined) expected percentage increase in theprovider's productivity. The infrastructure providerhas an incentive to lower costs, since gains in pro-ductivity greater than the expected X percent con-tribute to increased profits. To maintain incentivesfor efficient production, the X-factor should remainunchanged for a period of several years.
Price caps are diffusing widely to differentcountries and, gradually, to sectors other thantelecommunications, where they originated. TheUnited Kingdom has led the way, using price capsin airports, telecommunications, electricity distri-bution, gas, and water supply. Elsewhere, how-ever, their main application has been in telecom-munications, with electricity distribution a distantsecond. In Mexico, for example, the governmentintroduced price-cap regulation for Telmex in Jan-uary 1992 which applies a price cap to the overallweighted average price of Telmex's services, ratherthan a specific price cap for each service. In theUnited States many state regulatory commissionshave shifted from rate-of-return to price-cap regu-lation. Where comparison is possible, as betweendifferent states in the United States, the evidence isthat price caps lead to lower prices than does rate-of-return regulation.
There are also some early indications that the dif-ference between price-cap and rate-of-return regula-tion may not be as great as originally thought. Pricecaps are rarely observed in their pure form. Mostregulators see a continued need to assess the rate ofreturn and so set the caps on estimates of profitabil-ity, once again increasing the information require-ments for effective regulation. An exception ariseswhen profits are under the control of competitiveforces. For example, in the U.S. market for long dis-
70
tance phone services, price caps on the dominantprovider, AT&T, are thought to be the only instru-ment needed because profits are limited by compet-ing suppliers. But where local monopolies exist (asin local telephone services), rate-of-return consider-ations can reassert themselves so that, over time,price-cap regimes may converge toward their rate-of-return predecessors. Nonetheless, price caps dohave the advantage of shifting a greater part of thefinancial risk onto providers of infrastructure ser-vices, who cannot be sure that the regulator willallow them to recoup excess costs. This threat en-courages tighter self-monitoring of performance.
Yardstick competition. When direct competition orcompetition from producers of substitute productswill not work, competitive forces can be replicatedthrough comparisons with performance elsewhere.A utility in one region can be motivated to performbetter by promises of greater rewards if its perfor-mance exceeds that of a similar utility in another re-gion. However, only if the utilities' input prices,market demand, and government regulationsequate can better performance be attributed to theefforts of the utility.
A number of countries use yardstick competi-tion, formally or informally. In France the contractsof the local water company often depend on thequality of services and their production costs rela-tive to those of other French water companies. Thewater sector regulator in the United Kingdom reliesexplicitly on cost comparisons. The Chilean tele-communications industry uses an important variantof yardstick competition. A hypothetical "efficient"firm, rather than other Chilean firms, is used in set-ting the prices that telecommunications supplierscan charge. International cost and price trends areused to estimate the performance an efficient firmshould achieve, and prices are established based onthis estimate. Within this framework, the more effi-cient the Chilean firm, the larger its financial re-wards. In electric power, reasonable distributioncosts are estimated for three "reference systems,"which vary according to such key determinants asdistribution costs, population density, and peak de-mand. Individual electricity distributors are placedin one of these three systems, and delivery pricesare regulated accordingly. A distributor benefits if itdelivers electricity more cheaply than the averageprovider in its reference system. However, manipu-lation of "reference system" costs by the few suppli-ers in the market has driven the government to ex-plore improvements in its use of benchmarks.
Although yardstick competition is limited by theneed for sufficiently refined and comparable infor-
mation, that constraint is being partly relieved bythe increasing possibilities of international compari-son. Specialized industry organizations and interna-tional development banks can serve a useful func-tion by disseminating data on production costs.Periodic audits can also provide information feed-ing into the regulatory process.
New instruments. The limitations of existing regu-latory instruments (such as rate-of-return, price-cap,and yardstick regulation) have spurred the searchfor new instruments. New instruments have beendesigned to minimize the information required bythe regulator and to increase the responsiveness tothe customer, making them, at least in principle, es-pecially suited to the needs of developing coun-triesalthough many of them have not been fullytested in practice. In the United States, an intermedi-ate form of regulation balances the risk of windfallprofits (or losses) from the selection of an inappro-priate X in the price-cap formula. If the rate of returnexceeds a prespecified limit, the firm has to refundthe difference to customers. If returns fall below thelower limit, price increases greater than those im-plied by the cap are permitted in some cases.
Another regulatory approach offers a provider achoice of regulatory options (a stiff price cap but nomonitoring of profits, or larger price increases withcloser monitoring of profits). The expectation is thatthe regulated entity will, through its choice, revealits ability to undertake significant cost (and, hence,price) reductions, as well as its attitude toward risk.
A form of regulation that is even less restrictivebut that can provide meaningful discipline isknown as "potential regulation." Regulators moni-tor the performance of suppliers and stand ready tointervene should problems arise. As long as cus-tomers are reasonably satisfied with the suppliers'performance, the regulator places no formal restric-tions on the suppliers' activities.
CREATING A CONSUMER CONSTIUJENCY. Consum-ers, both individuals and businesses, are not typi-cally involved much in the regulatory process, eventhough their input can be critical to efficient servicewhere the regulator has only limited means of ac-quiring information. Final consumers are often thebest monitors of service quality. Consumer feedbackcan be employed directly to motivate suppliers toprovide high-quality service. For example, returnsfor suppliers can be linked to consumer ratings ofperformance. Initial steps have been taken in Banga-lore, India, toward creating an information base rel-evant for consumer awareness and decisionmaking(Box 3.11).
Some pointers on consumer involvement in reg-ulation are available from industrial countries. TheUnited Kingdom has ten consumer commissions,one for each of the ten water service jurisdictions.Each is headed by a commissioner who reports tothe Office of Water Services on the needs and theconcerns of consumers, including the results of for-mal surveys and public meetings. In France, wherewater services are controlled through local munici-pal councils with consumer representatives, privateproviders consider good consumer relations essen-tial for maintaining their standing with the munici-pal authorities.
Conclusion
The past decade marks a watershed. Boldly innova-tive measures have been taken to pry open mono-lithic infrastructure sectors. Competition andunbundling of diverse activities are spreading. Tech-nological change (as in telecommunications and
Box 3.11 Participation as regulation:an initial step in Bangalore
A serious handicap facing the individual consumerdealing with a public utility is the lack of knowl-edge of the "rules of the game" and the right to ser-vice. Expectations are often low and incentives forcollective action are often limited.
A random sample of 800 households in the in-dustrial city of Bangalore, India, highlighted dissat-isfaction with the quality of service supplied by thetelephone, electricity, and water utilities. Only 9percent of those sampled were satisfied with theirtelephone service. Even fewer people were satisfiedwith electricity and water services. Problems citedincluded supply shortages, excess billing, inabilityto get errors corrected, and a general lack of com-munication with the service agency.
The conclusions of a broader study of quality ofservice were clear: more competition and better in-formation are needed. The two groups of agenciesthat performed relatively well in consumer assess-mentbanks and hospitalsoperate in a rela-tively competitive environment.
Another conclusion was that consumer "voice,"mobilized through groups such as residents' associ-ations, can be an important force in sectoral reformand reorganization. These associations can providecritical monitoring and feedback to minimizeabuses and hold public officials accountable. Well-publicized intercity comparisons of service qualitywould create an information base on which con-sumer associations could act.
71
electric power generation) has much to do withthese innovations. But more fundamental forces areat work, making the new initiatives relevant to sec-tors as diverse as surface transportation, waste treat-ment and management, and drinking water supply.The weight of evidence is that competition in orfor a market for services is generally more effectivein responding to consumer demands than aremechanisms for making public enterprises moreaccountable. We stand on the cusp of change.Familiar practices are disappearing, but in theirplace are unprecedented opportunities for produc-tivity growth and emergence of new products andservices.
The diffusion of novel ideas such as sector un-bundling, competitive entry and incentive regula-tions from industrial to developing countries has oc-curred at a remarkable speed. Some developingcountries have in fact led the move toward moremarket-based provision of infrastructure, as in pri-vatization of utilities. Continuation along this pathwill bring further dividends. In particular, develop-ing countries need to place greater reliance on newentry and on competition to encourage investmentand efficiency and to mobilize the skills necessary toachieve social goals. As the evidence presented in
72
this chapter shows, where regulatory barriers havebeen lowered, even limited new entry or the credi-ble threat of competition has led to lower prices andsubstantial cost reductions.
Tailoring contracts to attract specific investmentshas been the most common means used to balancethe public interest and private initiative. Contractshave been not only a regulatory instrument, but alsoan essential mechanism for risksharing and hencefor financing private projects (Chapter 5). But ex-pecting individual contracts to bear the continuingburden of policy formulation and regulation, al-though attractive in the short term, raises the possi-bility of misuse because consistency and trans-parency in contract terms are not always easy toensure.
In the long run, what is needed is a statutory reg-ulatory system that clearly defines the rules of thegame in each sector and openly enforces them. Al-though the possibility of abuse cannot be elimi-nated, it can be minimized through a system ofchecks and balances that reinforces the incentivesfor all parties to act in a manner consistent with thesocial good. Using consumer feedback in innovativeways in the regulatory process should be an impor-tant priority for regulators.
4Beyond markets in infrastructure
Commercial and competitive provision of infra-structure can effectively deliver the services neededto meet social goals such as economic growth,poverty reduction, and protection of the environ-ment. But a number of problems arise for whichmarkets cannot guarantee solutions. Many infra-structure services, especially those that resemblepublic goods (as described in Chapter 1), will beundersupplied if markets alone are left to deter-mine their provision. Market outcomes may allo-cate fewer infrastructure services to the poor thansociety desires. Environmental consequences of in-frastructure provision are unlikely to be fully antic-ipated and incorporated in market allocations. Co-ordination within and across sectors may notreceive adequate attention. Although these prob-lems have little in common, government action ap-pears to be the obvious solution in each case. Ad-mittedly, governments often have failed todistinguish themselves in providing adequate pub-lic goods, safeguarding the interests of the poor,protecting the environment, and coordinating sec-tors. But such failure has not been universal. Nor isit inevitable.
A variety of responses and policy initiatives canhelp overcome the limitations of both markets andgovernments. This chapter discusses five such ini-tiatives:
Decentralization and local participation to in-crease the benefits derived from local public goods,such as feeder roads, and improve collective activi-ties, such as maintenance.
Sound budgetary allocations to nationwide spend-ing programs to improve the social value of major in-frastructure networks, such as national trunk roadsand large-scale irrigation.
Narrowly focused subsidies to make services af-fordable to the poor.
Changes in pricing, regulations, and project designto address externalities and to reduce the adverse en-vironmental consequences of infrastructure.
Project-planning techniques to take account ofeconomic, environmental, social, and sectoral con-cerns not addressed in individual commercial orlocal decisions.
Decentralization and participation:involving users
In order for public goods, such as local feeder roads,to be provided, three things must happen. First, theamount and type of infrastructure to be suppliedmust be decided. Since the product will be availableto all, individual choices expressed in the marketcannot be relied on for this decisionmaking. Second,investments must be made and the infrastructuremust be provided. Since user charges that fully re-cover costs are not always feasible, private entitiescannot always be relied on to make the investment.Third, infrastructure facilities must be maintained.Because many infrastructure services benefit thepublic at large, individuals in a market setting can-not be expected to perform this task.
Although the market clearly would fail in thesefunctions, centralized public infrastructure bureau-cracies have not proved particularly adept at per-forming them either. Investment decisions often re-sult in too little infrastructure in rural areas. Whenrural infrastructure is provided, priorities are oftenset centrallyresulting in inadequate responsive-ness to local concerns and inappropriate provisionfor local conditions. For example, road design by
73
transport ministries in Africa is often more sensitiveto technical as opposed to service - considera-tions. This leads to excessive rural road width andcost and hence to fewer roads. Moreover, withoutsufficient local commitment to the infrastructurethat is supplied, investments are not maintainedand thus deteriorate rapidly. Soon after Côted'Ivoire spent $115 million constructing 13,000water supply points, a survey found that barely halfof the handpumps involved were functioninganexperience all too common in the rural water sector.
In most situations, infrastructure provides publicgoods of a localized nature. Decentralized responsi-bility, in which government authority is moved tosubnational levels of government, offers an oppor-tunity to improve the provision of such goods. Pro-vision of local, and to some extent even national,public goods can be more effective when participa-tion provides a voice for infrastructure users andstakeholders.
Decentralization
Mexican experience with a municipal fund programreveals the potential for improving service deliveryby decentralizing government authority to indepen-dent subnational governments. Funds are madeavailable to local governments for projects that arechosen, planned, and executed by local communi-ties. Many of the projects involve infrastructure,
such as roads, bridges, and water supply systems. Areview shows that projects are executed at one-halfto two-thirds the cost incurred by centralized agen-cies. Since 1990, the municipal fund program hasspread to all but two Mexican states (Box 4.1). Be-cause local governments are better placed to deter-mine and respond to local preferences, decentraliza-tion can increase user satisfaction, too.
The group of countries undertaking decentraliza-tion reforms is expanding and is not limited to in-dustrial countries or to large developing countries(such as Brazil and India). A study using compara-ble data from twenty industrial and developingcountries found that decentralized expenditures ac-counted for one-half of infrastructure spending inindustrial countries and one-quarter in developingcountries. While local expenditure has always beencommon in some sectors, such as solid waste dis-posal by municipal authorities, the scope for decen-tralized control extends to other sectors, such asroads and water, especially when responsibility forvarious activities can be divided among national, re-gional (provincial), and local authorities.
DECENTRALIZATION IN ROADS. Since roads in a cityor rural region chiefly benefit local residents, whilethe benefits of primary highway networks are morebroadly spread, decentralization of responsibility forlocal roads is quite natural. Decentralization shouldinclude implementation of maintenance and also fi-
Box 4.1 Mexico's municipios help themselves
Until 1990, Mexico's experience with rural infrastructurewas typical of that in many other countries trying to pro-mote rural development. Projects managed by state andfederal agencies were often poorly selected and de-signed and were implemented with inadequate supervi-sion. Furthermore, there was no commitment to ongoingoperations and maintenance by the agencies, local juris-dictions (municipios), or communities. As a result, expec-tations often outstripped performance.
Many of Mexico's priority projects are relativelysmall and located in inaccessible places. Yet the munici-pal fund program, introduced in 1990, demonstratedthat a locally managed grant fund can become a success-ful alternative for managing rural investment in techni-cally simple infrastructures such as small water supplysystems, rural roads and bridges, and school buildings.
The municipal fund program requires communityparticipation in project selection and execution. Everyyear each municipio receives an allocation to finance proj-ects selected with the participation of its communities.
Execution is usually managed by community commit-tees (Comites de Solidaridad), which hire and superviselocal skilled workers and purchase materials. Communi-ties must also contribute a minimum of 20 percent ofcosts (usually in the form of unskilled labor and localmaterials), which helps to ensure that only projects oflocal priority are selected. Studies have found that mu-nicipal fund projects often cost one-half to two-thirds asmuch as similar projects managed by state or federalagencies. In Mexico this success may be explained inpart by the presence of skilled workers in many commu-nities and a tradition of volunteer community labor.
Currently operating in all but two of Mexico's thirty-one states, the program has financed approximately75,000 projects over the past four years at an averagecost of $11,000 each. Mexico's four poorest states havereceived $32.5 million in municipal fundsan averageinvestment of $8 per capita, spread across 653 ruralmunicipios.
74
nancing to ensure that communities are willing topay for the quality of road service providedif allcosts were borne by higher government levels, localresidents would prefer paved roads. A review offorty-two developing countries found that, whereroad maintenance was decentralized, backlogs werelower and the condition of roads was better (al-though the effect of financing decentralization wasnot included) (Figure 4.1). The decentralized casesalso had higher proportions of paved roadway. Butdecentralization was also associated with higher unitcosts of maintenance (partially reflecting the highershare of paved roads) and with wider differences inquality across regions (reflecting interregional differ-ences in institutional or human capacity).
DECENTRALIZATION IN WATER AND SANITATION. An
analysis of World Bankfunded projects demon-strates that a division of responsibilities, providedthat there is suitable coordination, leads to betterperformance and maintenance in the water andsanitation sectors than would be the case in morecentralized frameworks. Data for a group of devel-oping countries reveal that per capita water produc-tion costs are four times higher in centralized thanin fully decentralized systems and are lowest whendecentralization is combined with centralized coor-dination. Most water sector studies recommend athree-tiered organization, with a national agency re-sponsible for finance, long-term planning, standardsetting, and technical assistance. Under the nationalagency, regional utilities function as operators, mon-itoring compliance with national standards and reg-ulations, supervising local systems, and traininglocal managers and technical staff. The third tierconsists of local agencies that manage the local sys-tem, collect fees, monitor use and maintenance, andplan local budgets. An alternative decentralizedarrangement found in France and Germany (andemerging in Brazil and Poland) moves managementof each activity to the lowest appropriate level. Forexample, water resource managementincludingregulation, emission standard setting, and invest-ment decisionsis at the water basin (rather thanthe national) level, while the provision of services isleft to municipalities.
Of course, technical considerations may dictatecollaboration and planning across government lev-els. For example, water and sanitation investmentdecisions made by regional utilities have to be coor-dinated with local land-use planning. And limita-tions are often imposed by local capacity. In Brazil,although municipalities are constitutionally as-signed responsibility for delivery of urban water, re-
Figure 4.1 Countries with decentralizedroad maintenance have better roads.
Source: World Bank data for 42 developing countries.
1
gional public utilities often take over local functionson contract from those municipalities that lack thenecessary scale of operation to be economic.
Decentralization is not inherently good or bad.As with all arrangements, its success depends onthe incentives it creates, the capabilities it can drawon, and the costs it imposes. To improve incentives,public accountability is essential and can be en-hanced by local choice of leaders, local control of fi-nances, and other forms of local responsibility. Elec-tions are one mechanism for involving citizens inchoiceselectoral reform in Colombia and Vene-zuela has produced a resurgence in local leadership.Newly elected mayors have been able to mobilizeprivate sector financing for investment programs. Inorder to be held accountable, local leaders musthave control. This includes control over revenues,which in turn requires adequate local finance laws(covering budgeting, financial reporting, taxation,contracting, and dispute settlement). In many coun-tries, key responsibilities of local governmentsincluding the ability to tax or to charge user feescan be suspended by the central bureaucracy with-out consultation. This lack of autonomy discourageslocal administrators and contributes to a popularimage of local government inefficiency or even cor-ruption. Accounts and audits are important sources
75
Figure 4.2 Participation increases waterproject effectiveness by improvingmaintenance.
Maintenance Projectlevel effectiveness
001o /o
Lowparticipation
Mediumparticipation
Highparticipation
Maintenance
LI Good
Effectiveness
LI High
44%
17%
83%
LI Bad
LI Low
of information necessary to ensure accountability tolocal citizens.
Imbalances between revenue sources and expen-diture assignments threaten to reduce the perfor-mance of subnational governments. A study of tendeveloping countries, using comparable data, foundthat subnational revenue covers only 55 percent ofexpenditures. Greater effectiveness in raising rev-enues locally getting users who benefit most fromthe local public goods to provide the required re-sourcesis the key to equating revenues with ex-penditures. When national governments maketransfers to the subnational level to offset inter-regional inequalities in resource mobilization capac-ity, these transfers should remain transparent.Transfers that are not clearly publicized to localusers can undermine local government accountabil-ity and jeopardize the improvement in incentivessought from decentralization.
76
Unlocking local effort through decentralizationrequires creating new technical and institutional ca-pacity. Many poor communities lack requisite skillsand cannot take up the opportunities offered by de-centralization. This lack of capacity remains an im-portant constraint. Adequate technical support isneeded, including access to engineering, project de-sign, and administrative skills. Organizations suchas AGETIP (Agences d'Exécution des Travaux d'In-térêt Public) in Africa or the Brazilian-based IBAM(Instituto Brasileiro de Assistencia Municipal) helpdevelop local capacity, prepare projects, and moni-tor project execution and operation.
Participation
The importance of participation in effective deliveryof local public goods is well recognized, and it iscentral to community provision of service (OptionD as presented in Chapter 1). A 1985 World Bank re-view of twenty-five projects (mostly in agricultureand rural development) five to ten years after com-pletion found that participation by beneficiaries andgrass-roots institutions was a key factor in thoseprojects' long-term success. Without local participa-tion, projects often either foundered at the imple-mentation stage or were not maintained and failedto produce sustained benefits. This experience hasnot been unique to World Bank projects; it is mir-rored by other development agencies. Statisticalanalysis reinforces the impression from project re-viewsa 1987 analysis of recent World Bank proj-ects and a 1990 analysis of USAID-funded projectsfound strong evidence for the importance of partici-pation.
Participation in project formulation is particu-larly important for the maintenance of facilities. Astudy of 121 completed rural water supply projectsin Africa, Asia, and Latin America, financed by var-ious agencies, showed that projects with high par-ticipation in project selection and design were muchmore likely to have the water supply maintained ingood condition than would be the case with morecentralized decisionmaking (Figure 4.2). A review ofeight rural water projects in Nepal, comparing gov-ernment-designed projects and those designed with(rather than for) the community, found that the lat-ter were smaller, made greater use of community re-sources, and had more sustained outputs.
There are three keys to using participation to im-prove project performance: involve the beneficiariesdirectly; seek their early consensus on the project;and mobilize cash or in-kind contributions fromthem. Consultation with officials or voluntary orga-
nizations is not a substitute for involving the ulti-mate beneficiaries directly, for example, throughtown meetings. For the water supply projects stud-ied, the effect of increased reliance on intermediarynongovernmental organizations or local govern-ment units that did not involve users directly waseither insignificant or negative, while direct relianceon local organizations whose members includedusers had a positive impact on project performance.It is particularly important to ensure that participa-tory processes involve all groups of beneficiaries, in-cluding women (who are often the primary users ofwater and irrigation facilities) and others who maybe disenfranchised, such as the very poor andlandless.
Reaching consensus on user needs often leads toinfrastructure that is lower in cost, less technologi-cally complex, and more labor-intensive. In Korea
52,400 kilometers of village access roads have beenbuilt since 1971 as part of the saemciul undong move-ment of community self-help. These roads have amodest standardthey are gravel-surfaced andonly 2 to 3 meters wide, with standard designs forculverts and bridges. Brazil and Indonesia haveboth found that using participatory approaches toidentify appropriate low-cost technologies requiresflexibility in planning and engineering, and indonor attitudes as well (Box 4.2).
Improved consensus on a project among in-tended users not only increases their satisfactionand willingness to contribute, but also helps mobi-lize their involvement in construction and mainte-nance. In many rural areas, collective contributionsare often in forms other than cash. In the Banglungdistrict in Nepal, for example, local communitiesconstructed sixty-two suspension bridges using a
Box 4.2 Applying innovative approaches to
Two World Bank-funded projects in Brazil and Indone-sia demonstrate that using demand-oriented planning oflow-cost water and sanitation requires considerable ad-justments by the formal institutions of government, theengineering profession, and external donors (such as theWorld Bank). In Brazil the Water and Sanitation Programfor Low-Income Urban Populations (PROSANEAR)project is investing $100 million to provide water andsanitation infrastructure to about 800,000 people in low-income areas in eleven cities in different regions. In In-donesia the Water Supply and Sanitation Services forLow-Income Communities (WSSSLIC) project is invest-ing about $120 million in similar infrastructure covering1,440 low-income villages in six provinces and affectingabout 1.5 million people in all.
Participation must be tailored to the population. ThePROSANEAR projectnow under way for about twoyearshas taken a variety of approaches to involve ben-eficiaries in the design of subprojects. In one approach,leaders of community organizations are consulted onbasic choices, and the details are then worked out withactual beneficiaries. In another approach, agreement isreached between design engineers and beneficiaries di-rectly, in consultation with community leaders and orga-nizations. In both of these models, conflicts of interestbetween the water company and community-based or-ganizations are resolved through negotiation, with theproject design consultant as facilitator. Preliminary dataindicate that these two approaches have dramaticallylowered per capita investment costs and increased thesense of project ownership among communities.
In yet another approach, Indonesia, which alreadyhas a strong tradition of village organizations providingpublic services, encourages village water and sanitation
water and sanitation planning
committees to act as decentralized water utilities. Thevillage committees can choose from alternative levels ofservice and an array of tested technical solutions, de-pending on how much the village is willing to contributeto basic investment funds provided by the WSSSLICproject.
Engineers need to adapt. In PROSANEAR, the partici-patory process directly affected the kind of engineeringadvice used. For example, water companies were re-quired to award project design consultancies to a con-sortium of engineering firms or firms working withnongovernmental organizations that specialize in com-munity participation. The supervision team at the na-tional level encouraged project design consultants andwater company engineers to discuss plans with benefi-ciaries before agreeing on final proposals. In Indonesianongovernmental organizations with experience in therelevant sector are helping the project management teamand engineering staff to be responsive to the demands oflow-income communities.
Donors have to adjust their practices. The Brazilian andIndonesian projects were approved by the World Bankwithout blueprints of targeted service levels or deliverysystems. Instead, their appraisal reports provided broadprinciples for project execution and indicative targets forbenefits and costs, leaving much of the design to be de-veloped during implementation. The external donormust provide intensive supervision to work out detailsof the subprojects as chosen by the communities and tomonitor and evaluate implementation. Experience so farshows that these learning-intensive, participatory proj-ects can reduce capital costs, although they also entail in-creased investment of staff time from the donor.
77
78
combination of local materials, labor inputs, andgovernment funds. Households unable to partici-pate directly in the construction were asked to con-tribute food or money. Costs to the government to-taled only about $50,000, while the amountsmobilized locally were substantially higher. Similarself-help initiatives supply power in rural Purang,Nepal, and roads in Ethiopia (Box 4.3).
Cash or in-kind contributions from beneficiariesalso enhance project effectiveness by increasinglocal commitment. Statistical evidence from therural water supply projects study mentioned aboveindicates that the larger the share of investmentcosts paid by water users, the more effective theoverall project will be. Until 1990, Mexican irriga-tion operations followed a vicious circlea para-statal organization operated and maintained the fa-cilities poorly, so farmers rarely paid the (highlysubsidized) charges, leaving the operator even morecash-strapped. Service then declined even further,and farmers became even more reluctant to pay.Since 1990, responsibility for more than 2 millionhectares has shifted from the government to water-user associations. In order to improve maintenance,these groups voluntarily raised water charges asmuch as threefold. The higher charges have led to fi-
nancial self-sufficiency in most districts and im-proved the efficiency of water use.
Self-help in the construction and maintenance ofinfrastructure is most feasible with relatively small-scale projects undertaken at the initiative of a well-defined group or community for its direct and ex-clusive benefit. With works that benefit a widerpublic, such as feeder roads, self-help is much moredifficult to sustain over the long term, especially ifheavy reliance is placed on unpaid labor. There arerisks of exploitation of the poor and of low laborproductivity under the banner of self-help and vol-untarism. Moreover, some types of infrastructure,such as dams and major canals, power and telecom-munications systems, trunk highways, and waterand sewer mains, are technologically complex net-works for which local participation cannot ensureadequate design and implementation.
Participation is not a panacea even in the sectorswhere it is most relevant, nor is it costless and with-out risk. Participatory processes take time and oftenrequire the skills of professional intermediaries whointeract with formal sector agencies, explain tech-nology options, and help resolve disputes. Partici-pation works best together with, not in place of,good governance. Special interests, local elites, or
Box 4.3 Power in Purang and roads in Ethiopia
There is a pressing need for electricity in the village of Pu- financial resources for improving and maintaining roadsrang in Nepal's Mustang districtand not just because and other infrastructures overlooked by governments. Athe winters are dark and cold and fuelwood is scarce. nongovernmental organization, GRCO was founded inDuring the winter months, when villagers are house- 1962 to improve and maintain roads and bridges in thebound by bad weather, electric lighting permits indoor Sebat Bet Gurage region southwest of Addis Ababa.income-generating activities, such as carpet making. GRCO mobilized funds from local Gurage villages and
Without initial external assistance or even a bank towns and from Gurage migrants living in Addisloan, Purang has established a 12-kilowatt installation Ababa. Since starting operations, it has financed im-that is owned and managed by the community. The provements on more than 350 kilometers of roads andplant runs twenty-four hours a day and supplies about spent about 7.2 million birr ($3.5 million). In addition,100 houses with, on average, 120 watts each. Consumers members have contributed an estimated 8 million birr inare charged to cover operation and maintenance costs. professional services and labor. In total, GRCO con-Given the icy-cold weather conditions, the heated dis- tributed about 70 percent of the cost, with governmentcharge water is an added benefit. contributing 30 percent through budget allocations to
Why is the Purang project successful? Because of the national roads authority, which carried out the roadcommunity participation, the management of the instal- improvements.lation is well integrated into social, political, and eco- Private citizens' participation in road improvementnomic structures, ensuring that all participants have ac- and maintenance works succeeded in GRCO becausecess to the decisionmaking process. The community not local people were provided not only with adequate in-only owns the installation but also feels responsible for formation but also with the opportunity to set their ownit. Operators are chosen from among the villagers and priorities for development and to contribute both finan-trained by a local firm. cially and in kindthus maintaining their commitment
Ethiopia's Gurage Roads Construction Organization and ownership. Government also supported local initia-(GRCO) is a community organization that has mobilized tives with funds and technical assistance.
powerful minorities can capture the process to theexclusion of others. Finally, local communities can-not be expected automatically to take into accountthe environmental costs they impose on others, anymore than a private firm would.
Improving budgetary allocations
Decentralization and participation can be useful in-struments for overcoming market failure, particu-larly when the public goods provided are local.When the public goods are at the national levelsay, a highway networkthe central governmentmaintains direct involvement in allocating re-sources and in the planning and selection of proj-ects. The process and criteria underlying centralgovernments' decisions on budgetary outlays fornational public goods and for transfers to subna-tional governments are described in this section.Strategic and project planning are discussed in alater section.
In many developing countries, the basic processfor allocating and controlling public funds for capi-tal investment and recurrent operations is often dif-ficult to reconcile with professed development ob-jectives. An analysis of budgetary allocations inUganda revealed that the budgetary process there
largely replicates historical allocations and does notallow for increased emphasis on particular activitiesor the phasing out of others. In Cameroon, Nepal,and Zambia, transport sector allocations have em-phasized the construction of new roads over main-tenance or rehabilitation of existing networks, eventhough the latter is a clear priority.
Comprehensive and centralized medium-termplanning with strong backing from political author-ities was attempted in many developing countriesduring the late 1950s and 1960s, without conspicu-ous success. An excess of ambition spawned largepublic projects, many of which remain a costly bur-den for the economies concerned.
In some economies, including many in East Asia,government decisionmaking of a more intermediatenature has been practiced. In Japan, Korea, Malay-sia, Singapore, and Taiwan, China, authorities focuson directing public expenditures and actively coop-erate with a strong private sector. Flexibility andadaptability to changing circumstances are charac-teristic, with formal plans being indicative ratherthan prescriptive. In Malaysia, government deci-sionmaking involves different levels of government,with each level focusing on those issues for which itis best qualified (Box 4.4).
Box 4.4 Centralized and decentralized infrastructure planning in Malaysia
The Malaysian approach to infrastructure planningblends centralized and decentralized forms. First, at thecentral level, national development objectives and tar-gets are formulated by the National Economic Council (aministerial council chaired by the prime minister) andthe National Development Planning Committee (com-posed of top civil servants from federal ministries).Alongside these two groups, the Economic PlanningUnit, located in the prime minister's department, acts asa coordinating and integrating agency rather than an ini-tiator of sectoral plans.
Following deliberation by these two groups, thefederal government's development policies and sec-toral priorities are conveyed to ministries, statutorybodies, and state governments, which are then invitedto submit their programs for the next five-year plan.This second stage constitutes the decentralized ap-proach to planning. Agencies that are located in thestates are required to discuss their development pro-grams with the appropriate State Economic PlanningUnit before submitting them to the relevant federalministry. This ensures that the state governments areaware of the development proposals of the federalagencies operating within their boundaries. The Na-tional Development Planning Committee has ultimate
jurisdiction over the selection of expenditure programsfor the five-year plans.
The institutional framework for infrastructure devel-opment in Malaysia has been effective in ensuring thatpublic provision of infrastructure has reflected bothbroad national priorities and local needs. Infrastructureprovision was sufficient to sustain strong economicgrowth up to the 1980s.
By that time, however, the emergence of strong pri-vate sector capacity convinced the government that itsdirect involvement in some sectors and activities was nolonger necessary. Fresh approaches were also perceivedto be desirable in dealing with growing infrastructurebottlenecks. The government responded flexibly to thesechanging circumstances, placing infrastructure sectors atthe forefront of its privatization program. To date,eighty-five projects have been partly or completely pri-vatized, including the 900-kilometer North-South High-way, the container terminal in Port Kelang, TelekomMalaysia, and the National Electricity Board. The gov-ernment's "Guidelines on Privatization" issued in 1985and "Privatization Master Plan" formulated in 1989clearly confirm its view of infrastructure privatization asyet another means to achieve its underlying develop-ment strategies.
79
Figure 4.3 In water and sewerage, the better-off often get more subsidies than the poor.
Ratio of public subsidies torichest versus poorest quintile3
Pro-poor
0
i\c' c2
N CO'
Source: Petrei 1987.
çe
crS
Decisions on expenditure allocation within infra-structure sectors as well as across sectors should beguided by consideration of the country's underlyingdevelopment goals. Governments must choose be-tween new construction and maintenance, and be-tween rural and urban sectors among regions. Allo-cating expenditures to different activities on thebasis of social rates of return is an importantmethod of establishing priorities. Analysis of suchreturns in most developing countries reveals thecritical importance of maintenance over new con-struction. A study of irrigation expenditures inIndia identified maintenance of irrigation canals asa top priority with returns as high as 40 percent.Other activities that deserved priority over new in-
80
vestment included drainage and completion of un-finished projects. In Indonesia, rates of return on op-erations and maintenance for irrigation and roadshave been found to be as high as 100 percent, indi-cating that maintenance has been neglected.
In many countries, increasing spending on basicrural infrastructure is an economic priority that maycontribute significantly to poverty reduction. Chinahas been successful in integrating agricultural de-velopment with industrial development by buildingup rural industrial infrastructure. Consequently,rural industries have prospered and rural popula-tions have become employed in industry withoutmajor dislocation. Township and rural enterprisesin China now employ more than 100 million peopleand produce more than one-third of gross nationaloutput. In Indonesia and Malaysia since the late1960s, an important priority for the government hasbeen balancing regional development and reducingpoverty. To this end, infrastructure expendituresparticularly in transport and irrigationhave beendirected to rural areas. In Malaysia in 1965, earthand gravel roads represented 18 percent of the totallength of the road network (15,356 kilometers). By1990 such roads constituted 32 percent of the 50,186-kilometer network in the country. During this pe-riod, poverty in Malaysia fell dramatically. Ruralpoverty, which in 1973 affected 55.3 percent of thepopulation, had fallen to 19.3 percent by 1989. AWorld Bank study of poverty in Malaysia identifiedthe government's programs to raise land productiv-ity as a primary factor in this impressive improve-ment, and noted the importance of rural road andirrigation infrastructure.
Subsidies and transfers to the poor
Although the relationship between infrastructureand poverty is pivotal, infrastructure is neverthelessa blunt instrument for intervening directly on behalfof the poor. Adequate budgetary allocations to par-ticular sectors or to poor regions, removal of pricedistortions which support biases against the poor,and the selection of appropriate standards and de-sign are generally the most effective ways to ensurethat infrastructure realizes its potential for fosteringlabor-intensive growth and helping the poor to par-ticipate in the growth process. Subsidized provisionof infrastructure is often proposed as a means of re-distributing resources from higher-income house-holds to the poor. Yet its effectiveness depends onwhether subsidies actually reach the poor, on theadministrative costs associated with such targeting,and on the scope for allocating budgetary resources
to this purpose without sacrificing other sociallybeneficial public expenditures.
Price subsidies to infrastructure almost alwaysbenefit the nonpoor disproportionately. In develop-ing countries, the poor use kerosene or candlesrather than electricity for lighting, they rely on pri-vate vendors or public standpipes rather than in-house connections for water supply, and they are in-frequently served by sewerage systems. In Ecuadorthe electricity subsidy was found to be $36 a year forthe 37 percent of residential consumers with lowestuse but $500 a year for the better-off householdswith highest use. In Bangladesh subsidies on infra-structure services are roughly six times larger forthe nonpoor than for the poor. Although poor peo-ple generally consume more water and sanitationservices than they do power, a study of five LatinAmerican countries found that water and seweragesubsidies are directed more to richer than to poorerhouseholds (Figure 4.3). Even in formerly centrallyplanned Algeria and Hungary, the rich have re-ceived more than the poor in the way of infrastruc-ture service subsidies (Figure 4.4).
There are, however, ways in which infrastructuresubsidies can be structured to improve their effec-tiveness in reaching the poor. For example, forwater, increasing-block tariffs can be used charg-ing a particularly low "lifeline" rate for the first partof consumption (for example, 25 to 50 liters per per-son per day) and higher rates for additional"blocks" of water. This block tariff links price to vol-ume, and it is more efficient at reaching the poorthan a general subsidy because it limits subsidizedconsumption. Increasing-block tariffs also encour-age water conservation and efficient use by increas-ing charges at higher use. These tariffs are most ef-fective when access is universal. When the poor lackaccess, as is frequently the case, they do not receivethe lifeline rate and typically end up paying muchhigher prices for infrastructure services or their sub-stitutes.
Subsidizing access to public infrastructure ser-vices is often more useful for the poor than pricesubsidies. In Colombia in the early 1980s, water util-ities in Bogota and MedellIn used household surveydata to distinguish between rich and poor house-holds and specifically targeted the poor with sub-sidized connection charges and increasing-blocktariffs. This cross-subsidy scheme resulted in thepoorest 20 percent receiving a subsidy equivalent to3.4 percent of their income, financed by the richestquintile, who paid a "tax" equivalent to 0.1 percentof their income. Many low-income households can-not mobilize the funds needed to pay heavy initial
Figure 4.4 Even in some formerly centrallyplanned economies, infrastructure subsidieswent mainly to the better-off.
Algeria (1991)
Ratio of subsidy to better-offto subsidy to poor0 1 2 3 4
Electricity
Household gas
Urban transport
Heating
Water andsewerage
Transport
Hungary (1989)
Ratio of subsidy to better-offto subsidy to poor0 1 2 3 4
Source: World Bank data, and Hungary and World Bank1989.
connection costs to public services, especially whenpayment is required in advance of connection. Insuch circumstances, access to credit may be moreimportant than subsidized prices. Utilities are oftenuseful conduits for extending loans to finance con-nection costs because they can use their regularbilling procedures to secure repayment. In Bangla-desh the Grameen Bank provides credit to about 2million poor and landless personsmost of themwomen. The Bank combines group lending, whichallows the poor to substitute social collateral basedon peer pressure for financial collateral, with financ-ing mechanisms to extend credit for tubewells andsanitary latrines. In 1993 the Grameen Bank lent $18million for this purpose and since 1992 has pro-vided loans for about 70,000 suction tubewells.
In certain circumstances, programs providingemployment to the poor represent a highly effective
81
way of achieving distributional objectives. Suchschemes work because they mobilize large transfersrapidly, and, by offering relatively low wages in re-turn for unskilled manual labor, they transfer in-come only to those without more attractive options.In India's Maharashtra state, the Employment Guar-antee Scheme, initiated in response to the severedrought in 1972-73, provides unskilled rural em-ployment on demand. The scheme has provided al-most 1.7 billion person-days of employment and iscredited with playing a large part in avertingcalamity during numerous droughts. However, lit-tle evidence exists that such schemes produce themost economically useful infrastructure. Coordinat-ing them with overall infrastructure priorities mightstrengthen their economic impact.
Addressing externalities
Infrastructure often has widespread indirect im-pactsfrequently, on the environmentwhich canbe beneficial or harmful. Irrigation infrastructurecan reduce pressure on land resources by permittinggreater intensity of cultivation on existing plots, butit can also promote excessive water usage, resultingin groundwater salinization and land subsidence.Infrastructure can also reduce or increase publicsafety. Road improvements that raise traffic speedmay expose nonmotorized road users to increasedrisk of accidents; traffic signals can improve pedes-trian safety. Because markets often fail to reflectthese externalities, their management usually falls
82
to government. Environmental sustainability in-volves innovation in technology and organization,as well as improved efficiency in the use of infra-structure services through pricing and regulation.Regulatory efforts are also necessary for infrastruc-ture services to be delivered in compliance withpublic safety standards.
Innovation in design for affordability
Worldwide, roughly I billion people lack access toclean water and more than 1.7 billion do not haveadequate sanitation. Diarrheal disease, often causedby contaminated water, represents one-sixth of theworld's burden of disease (World Development Report1993). The most widespread contaminant of water isdisease-bearing human wastes. The environmentalbenefits of water supply depend not only on deliv-ering safe water for drinking but also on providingenough water to permit good human hygiene.Equally important is reducing contact with humanexcreta by providing pit latrines, toilets, and sewers(Box 4.5).
Although even among the poor the willingnessto pay for water is often sufficient to cover costs, thisis not always so in the case of sewerage, both be-cause conventional sewerage is often expensive andbecause certain costs of inadequate sanitation arenot borne within the household. For limited publicfunding to benefit large numbers, adoption of tech-nical and organizational innovations in low-costsanitation is necessary. A study in Kumasi, Ghana,
Box 4.5 Assessing a project's reach: water in Kathmandu
Evaluating infrastructure projects is difficult at best. En-vironmental costs must be identified and valued, theamount that individuals will pay for service determined,and the effect that service will have on other infrastruc-ture sectors assessed. Water supply, sewage treatment,sanitation, solid and hazardous waste handling, and am-bient water quality are all interrelated. A weakness inany one will affect infrastructure requirements else-where.
In the Nepalese capital of Kathmandu, officials as-sessed the effects of improving the water service usingan extension of traditional cost-benefit analysisthe"service-level" approach to valuation. This approach rec-ognizes that environmental services are valued differ-ently by different users and also attempts to assess indi-rect effects of water provision.
Kathmandu has 1.1 million inhabitants. Based on esti-mates using narrowly defined project appraisal tech-
niques, benefits from the city's new $150 million waterdistribution system included a direct financial savings of$500,000 annually from lower maintenance costs, plussubstantial annual benefits (based on willingness-to-payestimates for different users ranging from a low of $10for standpipe users to a high of $250 for business users).Total benefits were estimated to be $19.1 million per year.At a 12 percent discount rate, the project showed a mar-ginally positive net benefit of $5.2 million.
Using the more detailed service-level approach toproject appraisal, however, it was determined that insome cases health benefits from a reduction in coliformcontamination of the water approached $1,000 per unitserviced. An education program that improved wateruse led to further reductions in health and transportcosts. After these indirect benefits were factored in, theproject showed a positive net benefit of about $275million.
found that, although households were unwilling topay for the delivery of conventional sewerage ser-vices, only modest subsidies would be required toachieve relatively high levels of coverage with ven-tilated improved pit (VIP) latrines.
During the 1980s the Orangi Pilot Project inKarachi, Pakistan, mobilized poor people to con-struct, finance, and maintain their own water-bornesewers. This action resulted in the provision of sew-erage to 600,000 people at a cost of less than $50 perhousehold. The low cost was due to innovativetechnical solutions combined with a participatoryapproach in which corruption was reduced andcommunities contributed their own resources. Asimilar story comes from northern Brazil, where theuse of technically innovative condominial sewer-agea collective connection system provided bycommunity-based organizationslowered capitalcosts by up to 40 percent over conventional systems.
Motivation of user efficiency
Efforts to mitigate environmental impacts throughconsumer investments in energy saving are ham-pered by the low consumer prices and subsidies de-scribed in Chapter 2. On average, developing coun-tries use 20 percent more electricity than they wouldif users paid the incremental cost of supply. Onceeconomic pricing is established, governments areable to promote the use of more energy-efficienttechnologies.
Similar price increases are merited in transportbut are more difficult to implement. Cars using citycenters at rush hour impose congestion costs manytimes higher than they do in off-peak periods, andthe environmental costs of vehicle use are greater inurban than in rural areas. Urban car users can bemade aware of such costs through the introductionof parking fees, area licensing, and tolls. Growingenvironmental consciousness and technologicalchange are likely to increase the use of tolls and feesin the near future, which will encourage travelers touse public transit or nonmotorized modes.
Important user efficiency problems in the watersector stem from the underpricing of water. Domes-tic consumption, sanitation, irrigation, hydroelec-tric generation, and transport all create water de-mands and raise problems of overall supply andsectoral allocation. In India in 1985, 94 percent of allwater used went to agriculture. Conflicts betweenindustry and irrigation have emerged in someareas, and in cities such as Bombay, Delhi, andMadras problems of water scarcity have arisen. Inmany countries, raising the price of water to reflect
scarcity levels (particularly in agriculture) and link-ing price to usage are important first steps in deal-ing with water scarcity as well as with problems ofsalinization, increasing fluoride concentrations,and land subsidence. Influencing demand throughpricing allows the user to decide how much waterto use and how to achieve conservation.
Regulation
Regulation is an additional means of reducing ad-verse environmental consequences. It is also impor-tant for securing infrastructure service delivery thatmeets public safety requirements. The two principalregulatory approaches are command-and-controlmeasures and regulation based on economic incen-tives. Command-and-control measuresdirect reg-ulation along with monitoring and enforcement sys-temsare by far the most widely used technique indeveloping countries. An advantage is that theyprovide the regulator with a degree of certaintyabout, for example, how much pollution levels willbe reduced. But they have the disadvantage of pro-viding little incentive for innovation in pollutioncontrol technology once standards are achieved. Inrecent years, many countries have also adopted eco-nomic instruments. Setting prices to reflect full costs(the "polluter pays" principle) is the most powerfuland obvious of such instruments. In some countries,experiments are under way using additional regula-tory instruments, such as pollution charges, mar-ketable permits, subsidies, deposit-and-return sys-tems, and enforcement incentives, to introducemore flexibility, efficiency, and cost-effectivenessinto pollution control measures. Some of these ef-forts appear promising.
Environmental regulation begins by specifyingabatement standards based on the technical optionsavailable. For example, for power generation, tech-nologies are emerging that effectively reduce nox-ious pollutants from coalregulation can thus sub-stantially reduce emissions. But clean technologiesalmost always add to the cost of coal-fired thermalpower (by 10 to 20 percent on capital costs and 5percent on operating costs). Consequently, suchtechnologies are still far from universally used indeveloping countries. Where switching to gas is aneconomically viable alternative, there are many en-vironmental advantages. Poland provides an exam-ple of market-based incentives to reduce noxiousemissions. Its National Environment Fund, set up in1980, levies charges on all polluters and imposesadditional fines on owners of industries that violateregion-specific abatement standards. The proceeds
83
are bundled into low-cost loans to industries to pur-chase pollution-reducing equipment. In 1992 thefund's income was $188.5 million, double theamount in 1991. Although collection rates for pollu-tion charges and fines increased during the 1980sand early 1990s, a recent decline in compliance ratesis raising concern.
Serious problems are posed by vehicle transportin Central and Eastern Europe, despite a per capitavehicle population only one-third to one-half thelevel in Western Europe. The legacy of fuel and ve-hicle underpricing, the high average age of vehicles,obsolete designs, inadequate pollution controls,dirty fuels, and poorly maintained vehiclesallare factors producing environmental degradation.This situation has prompted suggestions that thecountries take direct measures to restrict road trans-port in favor of railways or river transport. A studyof Hungary undertaken for the World Bank sug-gests, however, that alternative approaches can re-duce vehicular emissions. If all new vehicles wereto comply with available best-practice emissionstandards, the traffic growth accompanying eco-nomic growth (as far forward as the year 2020)could be accommodated at absolute emission levelsbelow those presently experienced. However, limit-ing traffic growth may be necessary to control con-gestion.
In the Netherlands a transport sector strategyaimed at minimizing environmental stress andavoiding unnecessary investment mixes regulatoryand market-based measuresfor example, intro-ducing pollution premiums on road users, encour-aging the use of bicycles and public transport, creat-ing vehicle-free precincts for pedestrians, providingincentives for higher vehicle occupancy rates, andinstituting parking controls. In Japan and in severaldeveloping countries, including China, Ghana, andIndonesia, similar schemes to encourage nonmotor-ized traffic and pedestrian facilities are being con-sidered.
Regulation to preserve safety standards in infra-structure service provision and delivery is an im-portant priority. Studies have shown that road acci-dents are the first or second most important cause ofdeath in many developing countries. Addressingroad safety involves not only restricting speed andtraffic flows, but introducing safety considerationsinto the design and collection of information formonitoring and analyzing safety conditions. Facilityconstruction also requires special consideration. Be-cause construction exposes workers to a high risk ofinjury and death, effective safety standards must beapplied to the construction of facilities, not just totheir operation.84
Elements of infrastructure planning
Because most infrastructure uses geographicallydistributed networks, spatial, sectoral, and intersec-toral coordination and planning are necessary forgovernment activities. In addition, project selection,design, and evaluation are important steps in theoverall decisionmaking process. Incorporation, atthe earliest stages, of the social and environmentalimplications of projects is vital.
Sectoral and cross-sectoral strategies
Because infrastructure investments often havebroad impacts on many groups, planning strategiesshould focus on coordinating the decisions of in-vestors, including donors, while also gaining thebroad acceptance of other stakeholders. Particularattention may be required to ensure that the con-cerns of women are not overlooked (Box 4.6). Usergroups and other interested parties need to be con-sulted by the public officials and technical special-ists who usually lead the process, and mechanismsfor conflict resolution are necessary.
In the case of watershed protection in the SäoPaulo region of Brazil, for example, a workinggroup comprising municipalities, water suppliers,and environmental agencies was set up to solvewater quality problems in the Guarapiringa reser-voir so that it could meet rapidly growing demandsfor water. As part of the consultation process, atown forum was held with more than 120 city andstate government officials, members of nongovern-mental organizations and community groups, aca-demics and researchers, leaders of professional or-ganizations, and the press. Local consultantsprepared an environmental profile of the region andinterviewed city, community, and business leaders.The process resulted in a basin development strat-egy and an action plan that combined public andpolitical commitment.
When an infrastructure system is owned by a sin-gle entity, planning is generally internalized by theowner. Once ownership of a system is unbundled(as described in Chapter 3), however, strategic plan-ning becomes decentralized. To maintain the bene-fits of unbundling, the development of the naturalmonopoly segmentstypically the primary (trunk)facilitiesand the setting of technical standardsshould be coordinated at the sectoral level becauseof the market power that comes with the right tocarry out these functions. In an unbundled network,this responsibility could be entrusted to a coordinat-ing entity made up of representatives from govern-ment, suppliers, and users.
Box 4.6 Women can benefit from infrastructure, but success lies in the details
The beneficial impacts of infrastructure on women canbe profound, often extending beyond the commonlycited impacts of water and sanitation infrastructure onhousehold health or women's time allocation. But ensur-ing such outcomes requires foresight and attention to de-tail during project planning.
Women, as principal producers and marketers offood in many African countries, benefit from the im-proved access to markets that rural roads bring. Yet un-less they can afford to transport their produce by truck,goods must be carried to markets by the farmersthemselves. This sharply diminishes benefits from roadinfrastructure. Intermediate (nonmotorized) means oftransport, such as bicycles and carts, can be attractive al-ternatives to head portage but involve high initial invest-ment costs. In Ghana a pilot component administered byNGOs in the Second Transport Rehabilitation Projectchannels part of the wage earnings from labor-intensive road works to finance hire-purchase programsfor intermediate means of transport.
In many countries, destitute women are eager to par-ticipate in road works programs that offer them opportu-nities to earn cash. In one of Bangladesh's main roadmaintenance programs, women comprise the bulk of theworkforce, but in Kenya's Rural Access Roads Program,one of the oldest and most successful of such programsin Africa, less than 20 percent of the workforce arewomen. Similarly low participation rates for womenhave been observed in other African countries. Althoughit is sometimes argued that low participation by Africanwomen is due to their already oppressive burden of do-mestic duties and subsistence agriculture, evidence from
various countries, including Botswana, Kenya, Lesotho,Madagascar, and Tanzania, reveals that many poorwomen welcome such employment opportunities andare able to perform the same tasks as men for similarwages. To expand women's participation in these proj-ects, eligibility conditions must be extended, and job op-portunities must be advertised more widely. In additionthere should be scope for advancement by women to su-pervisory positions. Where maintenance is contractedout, women's groups should be encouraged to bid forcontracts.
Predicting the impact of infrastructure on women canbe difficult and requires a close understanding of the de-tails of their activities, opportunities, and constraints. Incentral Gambia, agriculture traditionally involved bothwomen and men within a system of coexisting commu-nal and individual cultivation. Men were responsible fororganizing the communal subsistence cultivation of up-land cereals, with both men and women contributinglabor, while women alone were responsible for cultivat-ing and marketing rice from individual plots. A rice irri-gation project was introduced, distributing 1,500hectares of irrigated land to farming households. An ex-plicit intention of the project was to improve the eco-nomic status of the female cultivators by raising their in-comes from higher rice yields. However, male farmersbecame interested in rice cultivation for commercial pur-poses and laid claim to the irrigated land for their com-munally farmed plots. While women did benefit fromthe project through the higher incomes accruing at thehousehold level, their position as producers and mar-keters of rice was undermined.
COORDINATION OF PLANNING. Coordination ofplans for competing or complementary sectors isalso important. Where program and project financ-ing involves many donor agencies, coordinationpreserves overall coherence of activities. In Africaefforts to improve donor coordination in transporthave been embodied in recent initiatives (Box 4.7).With transport, intermodal coordination is oftenrequired. The stress on speed and reliability inmodern-day freight transport is making it increas-ingly vital for shippers to be able to offer door-to-door service, commonly involving many modes. Itis necessary to establish a legal framework that al-lows freight forwarders to accept liability for theentire transport chain. In addition, customs proce-dures in many developing country ports must besimplified to avoid delays that can significantlyraise transport costs and undermine the interna-tional competitiveness of local producers.
Although governments are often tempted to in-tervene in price setting across modes or sectors,
prices that reflect costs provide valuable informa-tion for decisionmaking on sectoral allocations.When the local highway agency decided to expandtrucking cargo capacity to the port of Santos inBrazil, shippers pointed out that rail transport wascheaper, and the railway and the railhead river portcapacity were expanded instead. In China, theHenan Power Company, after evaluating the costsof expanding power generation capacity in the Yan-shi Thermal Power Project, changed its initial pro-posal from locating the coal-fired power stationnear load centers and supplying it with coal by railto siting the station near coal mines and transmit-ting electricity to the load centers.
PROJECT APPRAISAL. Techniques for project ap-praisal are well established and documented, but inpractice they are not widely applied. Although for-mal cost-benefit analysis of projects imposes non-negligible analytical and data demands, these tech-niques bring rational, objective, and, to the extent
85
Box 4.7 Donor coordination in infrastructure: the African experience
The World Bank's Africa region is encouraging donorcoordination through two main routes. First, it has usedregional partnerships of donors to develop policyframeworks and build consensus among those in-volved in the different infrastructure subsectors. Theseinitiatives include the Sub-Saharan Africa TransportPolicy Program (SSATP), which was launched as a jointundertaking by the Bank and the UN Economic Com-mission for Africa (UNECA). The SSATP is supportedby a coalition of donors that provide seconded staff andfinancial support; it involves African institutions suchas the Union of African Railways and the MaritimeConference for West and Central Africa. The SSATP hasbeen particularly effective in developing a common ap-proach among donors regarding road sector reform,railway restructuring, road safety, and improvement ofthe performance of urban public transport. The roadcomponents of the programthe Road MaintenanceInitiative (RMI) and the Rural Travel and TransportProgramresulted in the preparation of a Donor Codeof Conduct for this subsector (currently being ratified)in which participating donors agree to consult with
each other before committing to major new invest-ments.
Second, donor coordination in Africa is translatingthis consensus on policy reforms and investment priori-ties into concerted action through large umbrella pro-jects supported by a coalition of donors. The Bank actsas lead donor for these projects; other donors participateas cofinanciers and sometimes collaborate in prepara-tion. The two largest umbrella projects are the Inte-grated Roads Project in Tanzania (with sixteen partici-pating donors in the first phase and twelve expected tosupport the second phase) and the Roads and CoastalShipping Project in Mozambique (with fifteen partici-pating donors). Both projects have focused on sustain-able road financing, the provision of better qualifiedand higher-paid staff, and the contracting out of roadwork. This integrated project design has improved gov-ernments' efficiency in managing external aid by stan-dardizing their reporting, procurement, accounting, andbudgeting systems. Such approaches are being appliedto the road sectors in Burkina Faso, Cameroon, Kenya,Madagascar, Rwanda, Senegal, and Uganda.
possible, quantitative analysis to the decisionmak-ing process. Project appraisal is important, yet theevaluation of completed projects indicates that bothhigh-quality project appraisal and ongoing monitor-
86
Box 4.8 The World Bank's experience with project evaluation
The World Bank's own experience reveals that projectappraisal alone is not sufficient to ensure the success ofprojects.
During the 1970s and early 1980s, integrated ruraldevelopment projects represented a comprehensive ef-fort to raise rural living standards through, among othercomponents, a set of coordinated infrastructure invest-ments in irrigation, roads, and social services. A reviewof the Bank's experience by the Operations EvaluationDepartment (its internal auditing arm) found that resultswere often disappointing. Among the factors contribut-ing to the relatively low success rates, the report citedoveremphasis in appraisals on the details of projects, atendency to select large and complex projects, and overlyoptimistic projections of project outcomes. The reviewemphasized that a country's implementation capacitywas a critical prerequisite for project success.
A recent review of the Bank's overall project portfolio(the Wapenhans report) documented an increasing num-ber of poorly performing infrastructure projects. One ofthe causes of this increase cited by the report was a ten-dency to concentrate in the appraisal process on loan ap-proval, which can lead to an upward bias in estimating
rates of return. In addition, the report showed that, rela-tive to implementation capacities, projects were often toocomplex. Finally, the report argued that greater attentionto uncertainty and risk was warranted in project prepa-ration.
Both reports draw attention to components of theproject planning process that cannot be addressed by re-fining standard appraisal techniques. The objectivity andinternal consistency that such techniques offer must becomplemented by careful judgments about implementa-tion capacity and the rigorous analysis of project risks. Inaddition, as described in the World Bank's official re-sponse to the Wapenhans report, ensuring that affectedparties are committed to projects increases the likelihoodof project success. Seeking participation by beneficiariesin project identification, design, and implementation,while ensuring intragovernmental coordination andagreement, are useful in establishing such commitmentby stakeholders. Preserving some flexibility in projectcontent and design is also desirable; this requires carefulmonitoring during project implementation and learningfrom experience as the project evolves.
ing of implementation are required for project suc-cess (Box 4.8).
Experience with capital-intensive projects, in-cluding many in infrastructure, shows that manage-
Box 4.9 Incorporating environmental concerns early in planning: some recent lessonsfrom Sri Lanka
Over the past decade, developing countries, and theWorld Bank itself, have begun to require comprehensiveenvironmental assessments (EAs) as a routine compo-nent of project development. This requirement hasforced a better integration of environmental concernsinto project design, with appropriate attention to mitiga-tion options. But a project-level EA is best at dealingwith project-level mitigation issues. Without considera-hon of environmental issues at the long-term planningstage, it is doubtful that project-level EAs can steer thedevelopment of a sector along environmentally sustain-able paths. For example, because the environmental im-pacts of hydroplants are quite different from those ofthermal generation, the question of how air pollutionimpacts are traded off with inundation-related impactsfalls well outside the domain of project EAs. Althoughthe incremental effect of a single plant can be rational-ized quite easily, what matters is the overall impact ofthe sequence of plants in a power sector investmentprogram.
A recent World Bank study of the Sri Lankan powersector examined ways of bringing consideration of envi-ronmental issues into the early stages of power sector in-vestment planning and of dealing with the basic issues ofcomparing very different kinds of environmental im-pacts associated with different technologies. Workingwith the Sri Lankan generation utility and a group of
researchers and environmental experts, the study deter-mined long-term development options for the sector,incorporating environmental concerns. Alternativestrategies were compared, taking into account systemcost, biodiversity, health effects, system reliability, andgreenhouse gas emissions. The technique of multi-attribute decision analysis, which permits analysis oftradeoffs between objectives, is particularly useful insuch assessments when economic valuation of environ-mental externalities proves difficult.
From the analysis, the study identified the set of"nondominated" options that was better than the othersin at least one attribute (such as cost, emissions, reliabil-ity) but no worse in the other attributes. This set repre-sents the options that decisionmakers need to considerand included, for example, not only alternative fuel com-binations in power plants but also supply-side efficiencyimprovements in the transmission and distribution sys-tem and demand-side management options, such as theintroduction of compact fluorescent lighting.
Following this study, such new methods of evalua-tion have begun to be institutionalized in the Sri Lankanutility's planning cycle. In 1993, for the first time, thestudy for planning expansion of generating capacity in-cluded a systematic examination of demand-side man-agement and privatization options, as well as an envi-ronmental overview of conventional supply options.
ment of the economic and financial risks is oftencritical. Because of the inherent uncertainty in fore-casts of future conditions, projects should be Se-lected on the basis of careful sensitivity analysis.Planning forecasts in the World Bank's appraisal ofinfrastructure projects have sometimes overesti-mated demand (Chapter 1). High demand forecastslead to larger facilities, resulting in the selection ofmore capital-intensive investment options. Tech-niques that facilitate risk analysis in complex infra-structure projects, such as applying multicriteriamethods or drawing on financial options theory, arecurrently being developed.
Environmental and social concerns
ASSESSING ENVIRONMENTAL IMPACTS. Environmen-
tal regulation and promotion of the efficient use ofinfrastructure help reduce adverse consequencesfrom existing infrastructure, issues that have beenexplored in detail by World Development Report 1992.More options are available with new projects, al-though investment decisions can be consistent withenvironmental objectives only if environmental im-
pacts are identified and assessed. Experience withenvironmental assessments demonstrates that infra-structure projects are least likely to impose stress onthe environment if such assessments occur earlyand influence the design of individual projects notjust the selection of a particular project from a set ofalternatives. In Sri Lanka a recent power planningstudy involved not only selecting from among vari-ous fossil fuel and other generating options, but alsopaying attention to the need for energy conserva-tion (Box 4.9).
As the scale of infrastructure projects grows,environmental consequences become increasinglysignificant. A study of several large World Bankfunded projects in Brazil (representing total ap-proved Bank financing of $1.15 billion) examinedenvironmental consequences and emphasized thatenvironmental assessments should take a broadperspective capable of recognizing regional effectsand induced economic impacts, as well as the po-tential consequences of broad economic conditionsfor the project. Moreover, even though large invest-ment programs may be broken down into subcom-
87
Box 4.10 Population resettlement andproject design: Thailand's Pak Munhydropower project
Because of its disruptive and impoverishing effects,involuntary displacement of people should beavoided or minimizedoften by modifying the de-sign of a project. Consider the Pak Mun hy-dropower project in Thailand, a project to constructa 300-meter-long, 17-meter-high dam, submergingabout 60 square kilometers. The initial plans for thedamaimed at maximizing the project's powerbenefitswould have flooded about 3,300 house-holds. To reduce this impact, the project's designand siting were modified. The maximum retentionlevel of water was reduced from 112 meters to 108meters, and the dam was moved upstream from theKaeng Tana rapids to Ban Hua Heo. These modifi-cations reduced the generation capacity somewhat,but they also reduced the number of households re-quiring resettlement to 241. Even under the newdesign the Pak Mun project remains in the least-cost development plan of Thailand's ElectricityGenerating Authority. Compared with a 10 percentopportunity cost of capital, the discount rate re-quired before the project becomes unattractive is12.5 percent.
Fostering the success of the Pak Mun projectare four institutional features in the Thai policyenvironment. First, the government and the Elec-tricity Generating Authority of Thailand are com-mitted to sound environmental policies and prac-tices. Second, some local institutions are welltrained in dealing with involuntary resettlement.Third, local and international nongovernmentalorganizations are activeand vocalin repre-senting the interests of affected parties. And fourth,project planners are prepared for public scrutinybefore project implementation.
ponents and implemented in sequence, it is usuallynecessary to conduct the environmental assessmenton the basis of the overall program.
RESETTLEMENT. Physical infrastructure typicallyrequires an extended and undivided site, whetherfor road, rail, power, or water line rights of way orfor water reservoirs. The extreme difficulty of pri-vately bargaining, parcel by parcel, in these cases hasled governments to reserve the right to use eminentdomain in order to force the sale of property. Fre-quently these measures result in the displacement ofpeople. Of the 146 World Bank projects involving re-settlement of people between 1986 and 1993, morethan three-quarters were infrastructure projects.
Resettlement is most likely to be successful ifneeds are addressed early and plans are adopted tominimize avoidable displacement. In Thailand'sPak Mun hydropower project, early incorporation88
of design modifications reduced the number ofhouseholds flooded from 3,300 (20,000 persons) toonly 241 (1,500 persons) while maintaining an ac-ceptable project return (Box 4.10). By contrast, manyprojects are delayed or abandoned as a result of in-adequate resettlement planning. Construction ofColombia's Guatape II Hydro project took threeyears longer at twice the planned cost because offailure to address the resettlement issue early on.Successful resettlement requires monitoring duringand after project completion with flexibility for con-tingencies. In Indonesia, the Saguling and Ciratadams in western Java displaced more than 120,000people in the late 1980s, and despite cash compensa-tion many households saw their longer-run incomesdecline. An enterprise based on reservoir fisherieswas launched to provide employment to 7,500 dis-placed persons. The contribution from this employ-ment to household incomes, and then to the widercommunity, has been substantial. A recent study inCirata found that 59 percent of those who weremoved because of the dam now consider them-selves to be better off than before.
Conclusion
Improving infrastructure performance is often diffi-cultpolitically, technically, organizationally, andadministratively. Without the fundamentals of goodgovernanceaccountability, a predictable and sta-ble legal framework, openness, and transparencyeven the best efforts can go astray. The institutionalapproaches discussed above are not universally ap-plicable, but they do address specific concerns forspecific types of infrastructure. For example, envi-ronmental concerns differ greatly across sectors.Water, sanitation, and power differ in their impacts,and even within the power sector, the environmen-tal implications of fossil fuel generation differ fromthose of hydroelectric generation.
Finally, there is a need to achieve a balance be-tween expert and user, between direct and indirectcontrols, and between broad goals and those nar-rowly defined. The provision of infrastructure ofteninvolves complex, highly engineered systems thatrequire technical expertise but that also must be re-sponsive to user needs to be effective. Direct con-trols, such as plant-specific, quantity-based emis-sion standards, often prove cumbersome and costly,while indirect controls, such as price incentives,may not offer sufficient control. Infrastructureshould contribute to broad social goals, yet it maybe effective only when efforts are narrowly focused.The choice of instruments and approaches must re-flect sectoral needs and the capacities of implement-ing agencies.
5Financing needed investments
Innovative and diverse financing techniques arebeing employed to support an accelerating transi-tion from public to private sector risk bearing in in-frastructure provision. Mechanisms for financingspecific stand-alone projects are contributing to thelearning process as governments shift from beinginfrastructure providers to becoming facilitators,and as private entrepreneurs and lenders take amore direct role. But if there is to be sustained pri-vate risk bearing and investment in infrastructure,parallel and far-reaching actions are required to re-form legal and financial institutions and to developcapital markets that efficiently intermediate savingsinto investment.
Governments at present provide or broker thebulk of infrastructure financing: about 90 percent offinancial flows for infrastructure are channeledthrough a government sponsol which bears almostall project risks. Private financing is needed to easethe burden on government finances, but, more im-portant, it will encourage better risk sharing, ac-countability, monitoring, and management in infra-structure provision. In some sectors, such as poweror telecommunications, the scope for private financ-ing is great. In others, such as road networks, and insome low-income countries, the opportunities aremore limited, although even there increasing roomfor financial market discipline exists.
The challenge for the future is to route privatesavings directly to private risk bearers who makelong-term investments in infrastructure projects.Doing so will require institutions and financing in-struments adapted to the varying needs of in-vestors in different types of projects and at differentstages in a project's life. The benefits of thus financ-ing private initiatives in infrastructure go beyond
the projects involved. Because infrastructure invest-ments command such a large part of total financialflows, improving the efficiency of infrastructure fi-nancing will spur the general development of capi-tal markets. And as governments focus more onbeing facilitators rather than financiers, interna-tional development bankslong the partners ofgovernments in supporting traditional financingsystemswill need to experiment with new waysof doing business.
Old ways of financing infrastructureand new
Governments have been bearing more of the burdenof infrastructure expenditure than they can reason-ably be expected to manage. Under today's system,tax revenues and government borrowings are thepredominant source of infrastructure finance. Bor-rowingwhether from official or private sourcesis backed by a government's full faith and credit,and thus by its tax powers. Under this system, gov-ernments bear virtually all risks associated with in-frastructure financing. Private sponsorship and fi-nancing offer the twin benefits of additional fundsand more efficient provisionespecially valuablebecause substantial new investments are needed tomeet pent-up demand.
Today's financing patterns
Developing countries now spend around $200 bil-lion a year on infrastructure investment, some 90percent or more of it derived from government taxrevenues or intermediated by governments. Theburden on public finances is enormous. On average,half of government investment spending is ac-
89
Figure 5.1 Large shares of officialdevelopment finance for infrastructurego to energy and transport.
Official development finance(billions of U.S. dollars)
30
25
20
15
10
1984 1985 1986 1987 1988
fl Other infrastructure
Communications
Water and sanitation
LI TransportEnergy
Source: Appendix table A.4.
counted for by the infrastructure sectors consideredin this Report. Infrastructure's share of total govern-ment investment is rarely less than 30 percent andsometimes as much as 70 percent (Chapter 1). In ad-dition, maintenance and operating expenditurescommand a high share of current expenditures.
Governments have relied to varying degrees onforeign financing for infrastructure. Official devel-opment finance (including concessional and non-concessional funds from both multilateral and bilat-eral sources) has increased over the past decade andcurrently amounts to nearly $24 billion a year, thusproviding, on average, about 12 percent of total re-sources for investment in these sectors. The over-whelming share of these flows has been directed toenergy and transport (Figure 5.1). In contrast to theincrease in official lending for infrastructure, pub-licly guaranteed commercial financing has declined(Figure 5.2).
90
Figure 5.2 Official lending for infrastructurehas increased, but publicly guaranteedprivate loans have fallen.
Stock of disbursed lending(billions of U.S. dollars)250
200
150
1982 83 84 85 86 87 88 89 90 91 92 93
Total
Note: The loans are for electricity, gas, water,telecommunications, and transport.Source: World Bank Debtor Reporting Service.
External finance is used primarily to importneeded equipment (especially in the electric powerand telecommunications sectors) because most in-frastructure services cannot be exported and so donot directly generate the foreign exchange earningsnecessary to repay foreign currency loans. Externalborrowing, however, often reflects macroeconomicconstraints, and is also used to finance local expen-ditures for construction, equipment, and mainte-nance when public sector savings are limited. TheDominican Republic is one of several countries witha very heavy reliance on foreign funding, which fi-nanced 70 to 80 percent of infrastructure invest-ments in 1991. In the late 1980s the country had alarge public sector deficit (due in part to low pricesof infrastructure services), and a freeze was im-posed on the public sector's use of domestic credit,in order to curb inflationary expectations and per-mit an increase in credit to the private sector.
LIMITATIONS OF THE PRESENT SYSTEM. The main ad-
vantage of the present system is that in most coun-tries the government is the most creditworthy entityand is able to borrow at the lowest rates, makingpossible infrastructure projects that might not other-
100Multilateral
50
Publicly guaranteed private0
I I I I I
Box 5.1 Is there a free lunch?limits to government finance
In infrastructure projects, the cheaper credit available togovernments needs to be weighed against possible inef-ficiencies in channeling funds through government. In-efficiencies arise when financial discipline is relaxed as aresult of government sponsorship.
For a power generation plant, with construction costsaccounting for 70 percent of all costs and a 10 percent in-terest rate, construction cost overruns of 20 percent anddelays in construction of two years each lead to a 15 per-cent increase in unit costs of power produced. The trackrecord for publicly sponsored projects shows that suchcost overruns and time delays are common, leading to acumulative cost increase of about 35 percent. Comparethis with an interest rate advantage for government,which can borrow at, for example, 10 percent rather thanthe 13 percent available to private investors. This 3-per-
centage-point advantage translates into a unit cost reduc-tion of 20 percent. In other words, it would take almost a6-percentage-point interest rate advantage to negate theinefficiencies described.
Consumers would undoubtedly benefit if it were pos-sible to combine low interest rates and efficient provision.But the goal of a free lunch may be illusory. Even credit-worthy governments cannot borrow unlimited amountsat low cost. The evidence is that governments' costs ofraising funds rise with the level of borrowing. Also, highlevels of borrowing at a particular time increase debt lev-els and limit the amount that can be borrowed later,thereby reducing government liquidity These are furtherreasons why governments may be well advised to entrustto private sponsorship those infrastructure investmentsthat can be undertaken by private entrepreneurs.
wise be financially viable. Balanced against this ad-vantage has been the difficulty of maintaining ac-countability, leading often to high costs of provisionfor the consumer (Box 5.1). Moreover, being credit-worthy does not imply that governments have un-limited access to resources.
Governments' ability to spend on infrastructurehas been severely constrained, in part because poorperformance and pricing have strained governmentbudgets, as described in Chapter 2. Where budgetshave been tightened for macroeconomic reasons,the large share that infrastructure represents in gov-ernment investment has led to proportionatelysharp reductions in spending in this sector. In thePhilippines, for example, public investment in infra-structure fell from 5 percent of GDP between 1979and 1983 to less than 2 percent during the remain-der of the 1980s. Such sharp declines are appropri-ate where unnecessary or inefficient spending on in-frastructure is the cause of budgetary problems, orwhere macroeconomic adjustment is needed. How-ever, a continued low level of spending on infra-structure is not sustainable in the longer term; re-newal of economic growth requires accompanyinginvestments in infrastructure.
International donor policies and practices havesometimes reinforced distortions in recipient coun-tries. Many donors have focused on financing newphysical construction rather than on maintaining orimproving existing infrastructure. Like ministriesof public works, donor agencies find it easier tomeasure their achievements in new project ap-provals. Moreover, physical works draw on thewell-practiced technical skills of donor agencies. By
contrast, policy or institutional reforms and prac-tices that build long-term sustainability (such asmaintenance and user participation) require greaterdonor commitment to providing steady support,through longer periods of preparation and imple-mentation.
A World Bank review of urban water supply andsanitation projects identified typical problems. Seri-ous cost overruns (the group of projects as a wholecost 33 percent more than the appraisal estimates)and time overruns (46 percent of the projects re-quired two to four extra years to complete) greatlyincreased costs of service provision. Maintenancewas severely neglected because a lack of funds cre-ated shortages of skilled staff and spare parts. Thereview found that borrowers had often failed tocomply with loan covenants, especially those relat-ing to pricing and financial performance.
In the case of bilateral assistance, a further prob-lem that especially afflicts infrastructure arises fromthe full or partial tying of aid the requirement thatfunds be spent on goods or services purchased onlyfrom specified countries. In recent years betweentwo-thirds and three-quarters of official develop-ment assistance to infrastructure has been fully orpartially tied. By contrast, less than 20 percent of of-ficial development assistance going to areas otherthan infrastructure is tied. By definition, tying aidprecludes international competition in procure-ment. The Principles for Effective Aid agreed on in1992 by the Development Assistance Committee(DAC) of the OECD reaffirmed the superiority ofuntied aid and specified that, except for the least de-veloped countries, tied aid should not be extended
91
to projects that would be commercially viable if fi-nanced on market terms.
THE NEED FOR NEW APPROACHES. In the comingdecade, demand for infrastructure investments willsimultaneously increase in two different sets ofcountries: those that have undertaken macroeco-nomic adjustment with consequent low investmentlevels and, at the other extreme, those whose rapidgrowth is now placing a heavy burden on infra-structure. Infrastructure investments in developingcountries represent, on average, 4 percent of GDP,but they often need to be substantially higher.Where telecommunications or power-supply net-works are expanding rapidly, annual investments ineither sector can be as high as 2 percent of GDP. Aspecial factor increasing investment demand inmany countries is the rapid pace of urbanization, re-quiring investments in water supply as well aswaste treatment and disposal.
In Asia, the share of infrastructure investment inGDP is expected to rise from 4 percent today to morethan 7 percent by the turn of the century, with trans-port and energy likely to demand the most re-sources, followed by telecommunications and envi-ronmental infrastructure. Some of the plannedinvestments are without precedent. China, for exam-ple, has set a target of installing at least 5 milliontelephone lines annually up to 1995 and at least 8million lines per year thereafter, to more than tripleits 1992 base of 18 million lines by the year 2000.
Private entrepreneurship: trends and opportunities
Current efforts to secure increased private sponsor-ship and risktaking in infrastructure projects reflectthese various challenges. After decades of severeregulatory restriction, private entrepreneurship ininfrastructure bounced back in two ways during thelate 1980s: through the privatization of state-ownedutilities and through policy reform that made possi-ble the construction of new facilities in competitionwith, or as a complement to, existing enterprises.
The principal new infrastructure entrepreneursare international firms seeking business in develop-ing countries and operating often in associationwith local companies. These firms bring to bear notonly their management expertise and technicalskills, but also their credit standing and ability to fi-nance investments in developing countries. Majorelectric, telecommunications, and water utilities inindustrial countries face slowly growing demandand increased competition (following deregulation)in their home markets. As a result, they are vigor-
92
ously seeking high-yielding investments in devel-oping countries. Construction conglomerates are ac-tive in toll-road construction and in power projects,where they sometimes take an equity interest. Somecompanies or groups of companies also specialize instand-alone infrastructure projects, putting togetherfinancing packages and overseeing project develop-ment and operation.
Most indicators of infrastructure investmentunder private sponsorship reveal rapid growth. Pri-vatized telecommunications and electricity utilitiesin Latin America and Asia are undertaking largeand growing new investments. The number of theseso-called greenfield projects-especially in the roadand electric power sectors-has grown rapidly (asdiscussed below). Infrastructure investments by theInternational Finance Corporation (IFC), a WorldBank affiliate that invests only in private entities,have experienced a surge, from modest amounts inthe late 1980s to $330 million in fiscal 1993. Theamount invested by the IFC was leveraged morethan ten times, so that, in 1993, IFC participated inprivate investments of $3.5 billion.
The most important development during thepast four years has been the explosion in interna-tional flows of long-term private capital to develop-ing countries, especially in the form of foreign directinvestment and portfolio flows. Aggregate flowsstood at more than $80 billion in 1992 and were pro-jected to reach $112 billion in 1993 (Table 5.1). Infra-
Table 5.1 Portfolio and foreign direct investmentin developing countries, 1990-93
Note: This table records all portfolio and direct investment flows. Sep-arate figures for infrastructure are not available.
1993 figures are estimated or projected.A closed-end fund has a predetermined amount of funding and
sometimes a fixed life.ADR = American depositary receipts; GDR = global depositary
receipts. An ADR is an instrument used by an offshore company toraise equity in the United States without formal listing on a U.S. stockexchange. GDRs are similar instruments used in Europe and elsewhere.
Portfolio investment is the sum of equity and debt.Source: World Bank 1993i, pp. 10, 21.
Several channels exist for tapping international capitalmarkets. The larger private utilities in developing coun-tries have direct access to debt and equity markets. InOctober 1993 Telecom Argentina placed much of its $500million, seven-year bond issue with U.S. and Asian in-vestors; Argentina's Telefonika has also used bond mar-ket placements to raise expansion funds.
Foreign direct investment opens another route intointernational equity markets. General Electric Corpo-ration, an international conglomerate, has an active in-terest in developing infrastructure projects in develop-ing countries. Its subsidiary, the General ElectricCapital Corporation (GECC), issues securities on U.S.and European markets and invests the funds in se-lected projects. GECC has participated as an equity in-vestor, for instance, in the Northern Mindanao powerprojecta 108-megawatt diesel-fired power project inthe Philippines. Backed by the group's total opera-tions, the placement of securities issued by GECC iseasier than it would be for developing country powerprojects alone.
An instrument widely used to tap resources in theU.S. capital market is the American depositary receipt(ADR). ADRs are certificates of deposit that enable for-
eign companies to raise equity on U.S. markets withoutthe need for a listing on a U.S. stock exchange and with-out complex settlement and transfer mechanisms. Theyare issued by a U.S. depository bank, and the underlyingshares of the company are held in trust by a custodianbank in the home country. In 1990 Compania de Telé-fonos de Chile (CTC) raised $92 million on the New YorkStock Exchange through an issue of equity in the form ofADRsthe first major equity issue from Latin Americain three decades.
In April 1990 the U.S. Securities and Exchange Com-mission approved rule 144a, facilitating private place-ment of securities, including those placed as ADRs. Be-fore then, privately placed securities held by qualifiedbuyers (institutions that manage assets worth at least$100 million) could be traded only after a two- or three-year holding period. Rule 144a allows trading to occurimmediately, provided that the new buyer is also quali-fied. Moreover, after three years the securities can besold to all buyers. Rule 144a was used in 1992 to enable a$207.5 million international bond issue for the MexicoCityToluca Toll Road. Since then, other Mexican tollroads and the Subic Bay power plant in the Philippineshave raised funds using rule 144a.
structure has been a significant beneficiary of suchflows (Box 5.2).
Aggregate private investment in infrastructure indeveloping countries is currently about $15 billion ayear, or roughly 7 percent of the $200 billion beingspent annually on infrastructure in these countries.Although small, the fraction of private investmentin infrastructure investment is much larger than itwas some years ago, and there is a strong likelihoodthat private investment will continue to grow, possi-bly doubling its share of the total by the year 2000.One indication is the IFC's current infrastructurepipeline, which is almost as large as all the projectsfinanced to date.
The small overall share of private finance in in-frastructure obscures large regional and sectoral dis-parities. Private finance is proportionately greater inLatin America than in other regions, and larger intelecommunications and electric power generationthan in other sectors. The diffusion of current expe-rience across regions and sectors will raise theglobal share of private sponsorship and finance. Forexample, telecommunications privatization and in-dependent power generation are under discussionin all regions, including Sub-Saharan Africa. Andcontinuing technological and financial innovationswill undoubtedly make private financing more at-
tractive. As an example, electronic methods of iden-tifying vehicles and charging tolls could make roadsmore like a public utility service, and boost theshare of private finance in the highway sector.
Even with the rising share of privately financedinfrastructure, governments will continue to be animportant source of financing. Often, they will needto be partners with private entrepreneurs. Public-private partnerships in some ways represent a re-turn to the nineteenth centur) when infrastructureprojects were privately financed in much of theworld while government support acted as a stimu-lant. But the nineteenth century experience also of-fers important warning signs (Box 5.3).
The spread of project financing: achievementsand lessons
Many new infrastructure projects in the private sec-tor are built by "special-purpose corporations"which bring together private sponsors and other eq-uity holders. Despite their lack of credit history, sev-eral such ventures have successfully attracted eq-uity and loan financeand a huge pipeline of suchprojects bears the promise of decisively shifting thechannels and instruments of infrastructure financ-ing in the future.
93
Box 5.3 Warning signs from the nineteenth century
Throughout the nineteenth century, when infrastructurewas largely in private hands, contemporaries com-plained that many worthwhile projects were neglectedfor lack of financing. Some of the complaints reflectedthe difficulties of financing pioneering transportation(especially railway) projects. Other complaints wereself-serving efforts to shift all risk on to governmentbudgets, and in many cases the financial bankruptcy ofenterprises had severe consequences for governmentfinances.
Governments all over the world provided aid to pri-vate infrastructure projects in various forms, includingdirect subsidies. Two instruments in use then and of cur-rent interest as well are financial guarantees and landgrants.
Guarantees. In India, if a railway company did not at-tain a minimum rate of return of, for example, 5 percent,the government made up the difference under the termsof a guarantee backed by its full powers of taxation. Suchguarantees were also critical in the construction of theCanadian railways. But guarantees removed incentives
Project financing, which permits sponsors toraise funds secured by the revenues and assets of aparticular project, is often used in new ventures thathave no track records. This technique requires aclearer delineation of risk than is the case with tradi-tional public projects. Allocating risk among partici-pants has often been a difficult and time-consumingprocess, but new safeguards and conventions areevolving to deal with project risks and complexities.
Providing funds to a project is an important ob-jective in itself, but the financing process also servesanother important end. Monitoring by financialmarkets and institutions complements regulationand competition in service delivery. As such, it pro-vides another mechanism for investors to imposediscipline. Norms for devising incentive and pen-alty mechanisms to ensure performance by private-sector interests are becoming clearer. Privately spon-sored and financed projects measure their successagainst contractually agreed targets for new capac-ity, construction costs, and time overruns andagainst indicators of service quality.
The continuing role of the government lies in in-suring the private investor against policy-inducedrisks. Moreover, certain types of infrastructurerural roads and, to a lesser extent, sewerage andsanitationmay be unable to finance themselvesthrough user charges. Thus the need for govern-
94
for investors to monitor management performance whileopening the way for promoters to negotiate so-called"sweetheart" deals with construction and supply com-panies. Because many infrastructure projects were one ofa kind, the practice could be readily disguised. It nowappears that bond guarantees led to higher constructioncosts.
Land grants. During the nineteenth century, lands ad-joining railways and canals were often ceded to promot-ers, allowing them to profit from the many side busi-nesses that grew up around their investments. Byproviding collateral that could be used to back bondeddebt, land grantslike interest guaranteescorrectedfor capital market imperfections. In Canada during the1850s and 1860s, defaults on guaranteed bonds drainedgovernment revenues. In 1871, therefore, the CanadianHouse of Commons adopted a policy of land grants as away to subsidize railway construction without having toraise the rate of taxation. Land grants proved most effec-tive in such large speculative ventures as the Indian rail-roads and the transcontinental lines in the United States.
ment support does not disappear. One-time grants,of either capital or land, are the preferred mecha-nism for ensuring efficient operation.
Concepts and trends in project financing
Established companies - such as privatized tele-communications and electric power utilities havea credit history, a customer base, and tangible assetsthat can be offered as security to lenders. New com-paniesas in electric power generation, toll roads,or environmental infrastructurehave only theprospect of a future earnings stream to support bor-rowings. For them, a key issue is what recourselenders have if investments fail to produce the ex-pected returns.
The financing of a project is said to be nonrecoursewhen lenders are repaid only from the cash flowgenerated by the project or, in the event of completefailure, from the value of the project's assets.Lenders may also have limited recourse to the assetsof a parent company sponsoring a project. An im-portant policy question is whether government taxrevenues should be used to provide recourse, in theform of guarantees to lenders.
The use of nonrecourse or limited-recourse fi-nancing, also known as project financing, is a transi-tional response to new needs arising from activitiesrecently brought within the orbit of the private sec-
tor. Financing in this form can be complex and time-consuming, as the interests of various parties haveto be secured through contractual agreements. Theequity stake of private sponsors is typically about 30percent of project costs and usually forms the limitof their liability. Private lenders (especially commer-cial banks) influence project success by demandingperformance guarantees from project sponsors.Where performance depends on government policy,such guarantees are sought from governments. Theexpectation is that projects financed on a limited-recourse basis will, over time, develop a track recordthat will provide comfort for future investments.
ADVANCES IN PROJECT FINANCING. A survey pub-lished in October 1993 provided details of nearly150 private infrastructure projects that had beenfunded worldwide since the early 1980s on limited-recourse terms, at a total cost of more than $60 bil-lion (Table 5.2). Both the number of projects and thefunding involved had doubled compared with anearlier sample (in September 1992). This illustratesthe strong momentum in private projects, whichfive years ago were largely curiosities.
About half the projects surveyed (by number andvalue) were in developing countries, with a heavyconcentration in middle-income countries. The onlylow-income country with more than one fundedproject was China (although many more projects arein the pipeline there). Argentina, Malaysia, Mexico,and the Philippines had the most projects. Alongwith China, they represented 80 percent of the proj-ects for which funding had been committed. Mexicostood out, with the largest number of limited-
Table 5.2 Infrastructure project financing for projects funded and in the pipeline, October 1993(billions of dollars)
Country group
WorldHigh incomeMiddle incomeLow income
Number of projects
Funded Pipeline
Source: Public Works Financing, October 1993.
Table 5.3 Project financing of funded infrastructure projects, by sector, October 1993
Source: Public Works Financing, October 1993.
Total value of projects Average value of projects
Funded
recourse projects. Relative to its size, Malaysia, too,has been a significant user of project finance.
Transportation projects, mainly toll roads, domi-nated the numbers and the value of projects in high-income and developing countries (Table 5.3). Themore than two-thirds share of transport projects inmiddle-income countries reflected the extensivetoll-road programs in Argentina, Malaysia, andMexico. The survey estimated that twelve powerprojects had been funded in middle-income coun-tries (a 16 percent share of all projects). This estimateis already outdated, however, with the number forthe Philippines alone now being eight. The sectoralcomposition of the project pipeline is constantlychanging. For middle- and (especially) low-incomecountries, independent power projects are likely tobe an important focus for future project financing.Water and environmental infrastructure is anothergrowth area projects are being undertaken in mid-dle-income countries (especially for wastewatertreatment), and their diffusion to low-income coun-tries is imminent. A public-private partnership hasmade possible the construction of a chemical wastetreatment and disposal facility south of Jakarta inIndonesia.
The pipeline of projects under serious considera-tion is substantial. Public Works Financing estimatesthat 250 projects are being considered in developingcountriesseventy-two of them in low-incomecountries. The countries of East Asia and the PacificRim are expected to be the biggest users of stand-alone, limited-recourse projects in the next decade.This region has 150 projects in the pipeline, with anestimated total cost of $114 billion. China alone is es-
95
Country group Number of projects funded
Percentage distribution of projects
All projects Power TransportWater and environmental
infrastructure Telecommunications Other
World 148 100 13 60 16 2 10
High income 64 100 8 48 25 2 17
Middle income 77 100 16 69 10 3 3
Low income 7 100 29 57 0 0 14
Pipeline Funded Pipeline
235.4 0.44 0.71
112.0 0.54 1.0577.1 0.33 0.4346.3 0.44 0.64
148 358 63.1
64 107 34.377 179 25.7
7 72 3.1
timated to have as many as fifty projects under way.In the next few years India, Indonesia, and Pakistancould each have more than five projects, and severalprojects are being considered in Sub-Saharan Africaas well.
Adapting project finance techniques
Differences in project, country, and sectoral charac-teristics influence the availability of finance, the in-struments of risk allocation, and the degree and na-ture of government involvement. The main sectoraldivide is between toll roads (and urban transit sys-tems) and all other projects. Toll-road financing re-quires greater government involvement than doother infrastructure projects (see also Chapter 2 onthe unique problems of this sector).
PROJECT SIZE. As project size increases, the com-plexity of risk allocation increases rapidly, requiringmany complicated agreements between equityholders, creditors, input suppliers, and buyers ofservice. The dictum "start small," therefore, has itsattractions. However, many contractual agreementsare required irrespective of project size, and the hightransaction costs entailed often mean that investorsare not interested in projects below a certain size.The average size of projects in low-income countrieshas been $440 million, and that of projects in thepipeline has been even higher, at $640 million (seeTable 5.2). In middle-income countries, averageproject sizes are more than 25 percent smaller. Theinference is that transaction costs in middle-incomecountries are lower, making smaller projects morefeasible.
Large projects can create serious problems inlow-income countries. An early and innovative ef-fort using project finance for power generation isthe $1.8 billion Hub River Project in Pakistan, thecountry's first private power project. When com-pleted, the project will be one of the largest privatepower facilities in the world. It has suffered signifi-cant delays, however, because of complex negotia-tions over the division of responsibilities and risksamong the many parties involved. This experienceappears to support the wisdom of learning throughsmaller projects before moving on to larger ones.
A Sri Lankan power project is a good example ofwhat is needed when a country begins to seek pri-vate infrastructure investment. The project is small(44 megawatts), and the foreign and local entrepre-neurs involved are technically and financiallystrong. The government has guaranteed paymentsby the state-owned power purchaser, a dollar-basedtariff for the first ten years, foreign exchange con-
96
vertibility after the first ten years, and certain forcemajeure risks.
ProElectrica, a 100-megawatt, $70 million gas tur-bine plant near Cartagena, Colombia, financed en-tirely by the private sector, provides another goodexample of carefully structured project finance for asmall project. A group of large industrial consumershas contracted to buy electricity for fourteen yearsfrom ProElectrica. Foreign exchange payments havebeen guaranteed through prepayments into an in-ternational escrow account. In addition to short ne-gotiations and early implementation, the benefits ofProElectrica may extend to the regulatory reform ithas triggered. The Colombian government has re-sponded by creating arrangements to ensure thatthe local transmission utility "wheels" the powerfrom the generator to users, a step that creates aprecedent and a model for further new entry by pri-vate generators.
CREDIBLE CONTRACTS. The credibility of the regu-latory regime determines the bounds of available fi-nance (although success in financing a specific proj-ect creates a body of precedents that itself helps toimprove the regulatory regime). Project financing isa key mechanism for initiating a process of changein countries or sectors with limited track records inprivate infrastructure provision.
The Philippines, as noted in Chapter 3, has sig-nificant experience with privately financed powerprojects. The achievements have been considerable,especially in attracting foreign investment, giventhe obstacles the country faced in mobilizing foreigninvestment in the late 1980s. Although the Philip-pines now has an extensive, and sophisticated, legaland administrative environment for independentpower projects, the country's earlier experienceshows that much can be achieved in less sophisti-cated circumstances, provided that the ability towrite credible contracts exists. This lesson is alsodemonstrated by the experience of a power com-pany in Guatemala (Box 5.4).
An important additional element of contractualeffectiveness is the mechanism for resolving dis-putes. International arbitration procedures are com-monfor example, arbitration may be in a neutraljurisdiction using an internationally recognized setof rules, such as those laid down by the Interna-tional Chamber of Commerce. Sponsors and lendersmay also seek to have key elements of the contractdetermined according to the legal framework of amutually acceptable third country. For example,contracts for the Hopewell Shajiao C power stationin China were drawn up using Hong Kong law.
Box 5.4 A successful first step in Guatemala
In January 1992 Empresa Eléctrica de Guatemala S.A.(EEGSA)the major power distributor in Guatemalasigned a fifteen-year power purchase agreement with alocal power-generating company. Almost immediatelythe company sold its interest in the project to EnronPower Development Corporation, a subsidiary of alarge U.S. natural gas company with interests in severalindependent power projects. The project consists oftwenty 5.5-megawatt generators mounted on a barge atPuerto Quetzal, which operate as a base-load plant. Theproject increases Guatemala's generating capacity by 12percent and its effective capacity by about 15 percent.
The prices in the power purchase agreement are de-nominated in U.S. dollars. The agreement requiresEEGSA to provide the project company, Puerto QuetzalPower Corporation (PQP), with weekly fixed capacitypayments, provided that PQP meets minimum availabil-ity standards; weekly energy payments, with a mini-mum guaranteed purchase of 50 percent of output; andadditional collateral and documentary support to secureEEGSA's obligations to PQP. EEGSA has the option topay PQP in U.S. dollars or quetzales at the prevailing
market rate. When power availability falls below 50 per-cent, PQP will pay EEGSA penalties. The agreement re-quires the project to provide power at a competitiveprice. Under current assumptions of capacity utilization,which allow for deterioration of performance over time,EEGSA will pay an average of $0.07 per kilowatt-hourover the life of the projectwhich is about the long-runmarginal cost of bulk power in Guatemala.
PQP has cut some of its risks by entering into con-tracts for turnkey installation, operations and mainte-nance, and fuel supply. The plant started operating inlate February 1993, on schedule and within budget. A re-view of early operations indicates that PQP has achievedhigh levels of available capacity, that revenues and netincome agree with forecasts, and that converting quetza-les into U.S. dollars has not been a problem. After watch-ing EEGSA's experience with PQP, the Instituto Nacionalde Electrificacióna government-owned enterprise re-sponsible for power generation, transmission, and retaildistribution outside Guatemala Cityhas begun negoti-ating other power purchase agreements with indepen-dent producers.
TOLL ROADS. Today's resurgence in toll-road con-struction reflects practical reality: roads are neededfor economic development, but the financial andmanagerial capacity of the public sector is limited.In the past five years, Mexico has added an impres-sive 4,000 kilometers of new toll roads at a cost of$10 billion. Malaysia has the most expensive pub-lic-private project in the developing world, the $2.3billion North-South Toll Motorway. China is plan-ning many ambitious toll roadsthe 123-kilome-ter, $1 billion Guangzhou-Shenzen superhighwaywill cut through the heart of fast-developingGuangdong Province. Many other smaller tollroads, bridges, and tunnels are also being con-structed.
In most cases, tolls charged directly to users donot cover the full cost of roads. Governments grantland rights to encourage development made viableby the road (for example, shopping centers on free-way exits of the Guangzhou-Shenzen highway; seeBox 5.5). Governments also allow private toll-roadoperators to share in the revenues of existing pub-licly owned toll roads (as is the case for Sydney Har-bor Tunnel and the Bangkok Second Stage Express-way). They can provide capital grants to makeprojects financially attractive to private entrepre-neurs and can offer "shadow tolls" to private opera-tors (tolls paid from government revenues on the
basis of traffic flows), as proposed in Australia andthe United Kingdom.
Governments and the private sector have hadlimited experience in dealing with each other asequal partners on complex toll-road projects. Obli-gations have had to be renegotiated midstreamwhen a project's ambitious original goals were notbacked up by adequate preparation. Sometimes,specific road segments were not viewed as depend-ing on the quality of other roads, and competingministries failed to cooperate.
The Mexican toll-road program illustrates thedangers of launching a major initiative with multi-ple objectives and insufficient preparation. The con-tract terms failed to pin the responsibility for con-struction time and costs on the private projectsponsors, an omission of conditions that have sincebecome the norm. Dc facto flexibility in the conces-sion period allowed sponsors to shift cost increasesonto the consumer or the government. Creditors(mainly state-owned banks) failed to perform theirnormal appraisal and monitoring functions. The re-sulting high tolls have held down road use, al-though measures are now being introduced to in-crease usage (Box 5.6).
However careful the preparation, conflicts canarise. In the Second Stage Expressway in Bangkok,the Japan-led private consortium and the Transport
97
Box 5.5 Land grants and eminent domain
Land grants have proved to be a valuable form of collat-eral for innovative projects that might otherwise nothave been financed because lenders had little experiencewith similar projects. But in implementing a policy ofland grants, there is a risk that the grants might bewasted if they are given to projects that would be built inany case.
Overall, however, land grants have greater meritthan interest guarantees because they represent a one-time infusion of resources and do not reduce the incen-tives for efficient operation of the project. They may beespecially suited for more speculative projects.-such ashigh-speed rail in industrial countries or high-risk trans-port investment in developing countries.
Awarding land grants raises complex questionsabout acquisition procedures and compensation of land-holders. Land acquisition can take several years and
delay infrastructure projects substantially some of thedifficulties of the Second Stage Expressway in Bangkokare related to these delays. Not only is an appropriatelaw of eminent domain required to define the termsunder which the government can acquire the land, but,as was demonstrated in the case of Narita Airport out-side Tokyo, lack of sensitivity in implementation canlead to contentious and expensive delays.
In anticipation of land being acquired, landownershave an incentive to overdevelop their property or un-dertake other measures to overvalue their land. A practi-cal solution is to use prevailing market prices and com-munity standards of land development as a norm. Themore difficult issue is one of compensating those whoseproperty values fall as the flow of business activitychanges because of new infrastructure development. Ingeneral, governments have not compensated such losses.
Authority have disagreed, first over the level of tollsthey had agreed on and then (more seriously) overwho has the right to operate the road. Delays arisingfrom these conflicts and from slow land acquisitionhave affected the viability of the Don Muang Toll-way, intended to link the Second Stage Expresswayto the airport.
The lesson for toll roads, as for electric power,may be that contractual uncertainties are best ironedout in smaller or simpler projects. Argentina has de-veloped an extensive system of private concessionsin which tolls are charged to finance maintenance.There was an initial outcry against tolls on existingroads, and charges had to be loweredbut thegreatly improved quality of the roads has made tollsmore acceptable. In the state of Madhya Pradesh inIndia, an 11.5-kilometer toll road linking an indus-trial park to a national highway was built at a cost of$2 million and commissioned in November 1993.The enabling legislation put in place and the finan-cial mechanisms used are being adopted and re-fined elsewhere in the country.
Risks haring: the lessons learned
At the heart of project financing is a contract thatallocates risks associated with a project and definesthe claims on rewards. While often the cause ofdelay and heavy legal costs, efficient risk allocationhas been central to making projects financeableand has been critical to maintaining incentives toperform. Risks are divided not only between pub-lic and private entities but also among various pri-
98
vate parties. Four kinds of risks can be distin-guished currency, commercial, policy-induced,and countryalthough the distinctions amongthem are not always clear-cut.
CURRENCY RISK. Much recent, privately financedinfrastructure has drawn on foreign capital andtherefore faces the risk of local currency devalua-tion. International lenders rarely assume such risk,preferring instead to denominate their repaymentsin foreign currency terms. In the past, public enter-prises or governments have borne the currency risk,but in the growing move to private finance, the riskof currency depreciation falls on the project sponsoand ultimately on the consumers of the service. Inmany recent private projects, service prices havebeen linked to an international currency.
Independent power generation presents a specialcase. Although most power projects do set theircharges in U.S. dollars, these charges are paid by thetransmission utility, and the final consumer is oftencharged local currency prices untied to movementsin exchange rates. Transmission utilities cannot beexpected to continue to bear currency risk in thelong run.
Countries may wish to promote schemes for in-suring against currency movements (forwardcover) so as to allow for short-term risk manage-ment. In Pakistan, for example, the central bank of-fers forward cover at an average premium of 8 per-cent. In time, private financial institutions may offersimilar schemes. However, even with these arrange-ments, the consumer pays at least in part for ex-
change risk through the passing on of forward-cover premiums.
COMMERCIAL RISK. Two types of commercial riskmay be distinguished, those relating to costs of pro-duction and those arising from uncertainties in de-mand for services. Substantial progress has beenmade in shifting cost-related risks onto privatesponsors and other private parties. Typically, con-tracts include bonuses for early commissioning ofthe project and penalties for late completion. In aproject to construct a power plant in India, the pri-vate sponsor will pay a penalty of $30,000 every daybeyond the agreed commissioning date for the firstsix months and a higher penalty thereafter. A fixedpayment for overall capacity also shifts the risks ofcost overruns to the private sponsor. A contract mayalso specify operational obligations, such as mainte-nance or the availability of capacity. In the case ofutilities, a power or water supplier is sometimes pe-nalized for capacity availability below prespecifiedlevels (see Box 5.4 on the Guatemala power plant).Or the contract may require that a plant be availablein effective working order for a specified period oftime.
Project sponsors are able to transfer some ofthese risks to other private parties. It is common, for
example, to transfer construction risk to specializedconstruction companies through turnkey contracts.Also, sponsors may enter into long-term contractswith input suppliers.
Where sector policy concerns are unimportant,investors also accept market risk, but progress inthis regard has been slower. Tariffs in line withcosts, sector unbundling to permit new entry (as de-scribed in Chapter 3), and access to transmissionnetworks are required in order to enable privatesponsors to assume all market risks. In telecommu-nications projects, the market risk is typically borneby the sponsor. In the electric power and water sec-tors, on the other hand, limitations on assumptionof market risk arise because payments to cover costsare not assured. Also, governments need to deci-sively eliminate the prospect that investors will bebailed out if circumstances are unfavorable. Intransportation projects, such as Mexican toll roadsand certain Argentine rail concessions, govern-ments permitted revisions in contract terms whentraffic levels were lower than expected.
Assumption by private parties of even cost-re-lated risks creates incentives for good perfor-mance. Not only do sponsors have equity holdingsin the project, but lenders are also central to themonitoring process. As part of the contract, several
Box 5.6 Mexico's toll roads: a big push that faltered
Infrastructure projects are often associated with largeconstruction outlays that result in limited productiveuse. This can occur as much under private as under pub-lic enterprise if the right incentives are not in place.
In preparation for an ambitious 6,000-kilometer roadprogram, a Mexican government agency did hasty trafficand cost projections and prepared the road designs. Thequality of these estimates and designs fell far short of re-quirements for such an undertaking. At the same time,state-owned banks lending to toll-road projects did notperform the normal project screening and appraisals.
Although the concessions for road construction andoperation were awarded based on several criteria, in-vestors who promised to transfer the roads back to thegovernment in the shortest time were especially favored.Short concessions were partly motivated by a concernthat only short-term financing would be available. Theattempt to achieve success within a new administration'sterm also created a sense of urgency In turn, investorsnegotiated toll rates that would earn a return within theconcession period. Tolls typically were therefore five toten times higher than those in the United States for com-parable distances.
With tolls that high, traffic failed to materializetheold, free roads were preferred even when travel time wastypically twice as long. Moreover, cost overruns aver-aged more than 50 percent of projected costs. (The High-way of the Sun, from Cuernavaca to Acapulco, for exam-ple, cost $2.1 billion, more than twice the originalestimate.)
To remedy the situation, the Mexican governmenthas taken several steps. In many cases, concession peri-ods have been extended from ten or fifteen years tothirty years. Where joint ventures offer greater prospectsof financial viability, stretches of toll road are being com-bined under single management. Heavier vehicles maybe banned from the old road network as weight limitsare imposed and enforced.
There are signs that the most difficult period is past.In the long run, consolidations of toll roads, longer con-cession periods, and more realistic traffic and cost projec-tions, along with economic growth and greater financialresponsibility on the part of the project's private sponsor,should bring significant returns on this infrastructureinvestment.
99
financial covenants are made. In such situations,commercial banks have a much greater incentivefor supervising projects than do lenders backed bysovereign guarantees.
The evidence, although limited, shows that theassumption of cost-related risks by private sponsorsand the monitoring of performance by banks are ef-fective. Evidence, for example, on private construc-tion is very favorable and reflects the tight contrac-tual conditions and severe penalties for cost andtime overruns. A preliminary review of the IFC's in-frastructure projects shows that time overruns inconstruction have been only seven months on aver-age, and cost performance has been about on target.Such performance, however, is possible only whencommercial risks are truly transferred to privatesponsors. The Mexican toll-road example showsthat when risks can, in practice, be transferred backto the government, incentives for performance aregreatly weakened.
Private investors may wish to insure themselvesagainst commercial risks. The provision of such in-surance is best left to the private sector, althoughgovernments have a role in stimulating domesticguaranty facilities, possibly by taking an initialstake in guaranty funds (Box 5.7). The private mar-ket for risk insurance for international transactionsis small. While short-term insurance for trade creditis available, private insurance for infrastructureprojects is uncommon, although the London insur-
ance market is to provide insurance for traffic riskfor a Mexican toll road.
SECTOR POLICYINDUCED RISK. Especially impor-tant issues arise in the power sector because projectsponsors focus on the credibility and solvency oftheir buyer, typically a government utility thattransmits and distributes power. The instrumentthat protects the power supplier is the "take-or-pay"contract, or power purchase agreement. Under sucha contract, the buyer agrees to pay a specifiedamount regardless of whether the service is used.The government thus provides a contract compli-ance guaranteea useful transitional measurewhile the long-term goal of sector reform is beingaddressed (Box 5.8).
Similar concerns arise with water and other envi-ronmental infrastructure projects (such as watersupply, wastewater treatment, and solid waste dis-posal operations that are typically carried out at themunicipal level by a local monopoly). Here govern-ment agencies (or municipal authorities) are not thedirect purchasers of the service. But they can and doinfluence the ability of the service provider to meter,bill, and collect. Where the municipal authoritiescannot deliver, collection guarantees from the cen-tral government are required.
Thus, in such projects, the "market" risk, or therisk arising from fluctuations in demand, is effec-tively transferred to the government through the
Box 5.7 Leveraging through guarantees in Thailand
To encourage private lending, the Thai government isdeveloping the Thai Guaranty Facility to guaranteeloans made by private financial institutions to munici-palities and private operators of urban environmentalinfrastructure. The facility is planned as a public-privatecorporation with private sector management. The targetdate for initial operation of the facility is June 1994.
Because of limited experience in lending to munici-palities, financial institutions consider them risky bor-rowers. Perceiving high project risks, lenders are reluc-tant to make loans for periods of longer than eightyearstoo short to recoup investment from environ-mental infrastructure.
By providing guarantees to private operators andmunicipalities that help them to secure loans from com-mercial lenders, the government's guaranty facility willcreate longer-term financing. With increased lending tolocal government, it will soon be possible to establishcredit ratings for cities and to allow them to issue bonds.Ten provincial cities, the five cities of the Bangkok Met-
ropolitan Region, and Bangkok itself are expected to bethe primary beneficiaries of loan guarantees for invest-ments in wastewater treatment, solid waste collectionand disposal, and potable water supply.
During its first two years of operation, the guarantyfacility is expected to receive $75 million. Lending willbe five to eight times the level of these guaranty funds.Over a five-year period, it is projected that the facilitywill be funded at a level of $150 million and will lever-age up to $1.2 billion in loans for urban environmentalinfrastructure projects. It will obtain resources princi-pally from the Thai government, from money borrowedin part from the USAID Housing Guaranty Program andin part from Thai financial institutions.
A set of policy initiatives is also being established toensure the effectiveness of this facility, including a movetoward the "polluter-pays" principle, changes in admin-istrative procedures, and greater decentralization ofdecisionmaking.
100
Box 5.8 Look before you leap: limiting government exposure to contingent liabilities
When a guarantee is limited to contractual complianceby government agencies, the government has significantcontrol over events. Such guarantees can be madecallable, for instance, when government agencies inhibitthe supply of inputs to a project, fail to honor purchasecommitments, change pricing rules, or disallow remit-tance of foreign exchange to service a project's privateloans. Unlike blanket loan guarantees, this kind of agree-ment does not commit the government to protect lendersand investors against such commercial risks as cost andtime overruns, adverse movements in exchange rates,and inefficient operations. Contractual compliance guar-antees have the added advantage of creating incentivesfor government agencies to stick by their commitmentsand of limiting government liability to times when gov-ernment agencies are out of compliance.
Governments may also issue guarantees to ensure acertain rate of returna type of guarantee that producesthe worst incentivesor to lengthen maturities of loans.In both cases, the government takes on a commercialrisk. A century of experience shows that, although guar-
antees are sometimes desirable, they also create perverseincentives that can lead to project mismanagement.Guarantees make sense when international investors'perceptions of country risk are poorer than economicconditions warrant, so that the guarantees are a strictlytransitional measure for attracting broad, and ultimatelyself-sustaining, investor interest.
Government guarantees are not always necessary, asdemonstrated by the financing of ProElectrica, theColombian power plant. A significant part of the foreigndirect investment and portfolio flows to developingcountries has not been guaranteedthe underlying eco-nomic environment is what drives the flows.
Thus, when offering guarantees to private lenders,governments need to determine whether such guaran-tees are truly required, what form the guaranty shouldtake, and how they should be accounted for in govern-ment accounts. At the same time, governments need toset policies to enable the development of private insur-ance markets.
take-or-pay formula. This becomes necessary be-cause market risk is intermingled with the dangerthat financially troubled power purchasers (trans-mission utilities) or water users may not honor theircommitments. Overall sector reform is required toeliminate policy-induced risks and thus reveal themarket risk.
COUNTRY RISK. Where governments do provideguarantees against sector policy or even commercialrisks, these may not always be acceptable to privateinternational lenders, who may look instead forguarantees from creditor countries or from multilat-eral banks to insure against "country" risks. Therole of the borrower government does not disappearin such situations, since counterguarantees are typi-cally required.
Export credit agencies in OECD countries offerguarantees against risk of nonrepayment to their na-tional exporters or banks that extend credit to over-seas importers of goods and services. Typically,these agencies underwrite sovereign risk by provid-ing insurance on commercial credits and by extend-ing finance directly. During the period 1983-91,export credit agencies did $53.1 billion worth ofbusiness with a maturity of five years or more. Ofthis, 60 percent applied to infrastructure financelinked principally to the import of capital goods. Intheir most limited form, export credit agency guar-antees or insurance may be extended only against
sovereign risk, with exporters or bankers responsi-ble for commercial risks. In most cases, these guar-antees are extended to both types of risk, in part be-cause it is difficult to distinguish sovereign fromcommercial risks. As the primary motives for set-ting up such insurance schemes are supporting ex-port industries (and thus domestic employment),export credit agency premiums have been highlysubsidized, although they have been increased fol-lowing losses incurred in the 1980s.
The Hopewell-Pagbilao independent power pro-ject in the Philippines marked the first time that aloan from an export-import bank was not backed bya government counterguarantee, placing the bankon the same footing as private lenders. Nonguaran-teed lending by export-import banks exposes themto the same risks as other lenders, which gives themreason to improve their project appraisal, assess-ment of borrower creditworthiness, and monitoring.
To attract international private capital to devel-oping countries, several multilateral developmentbanks, including the World Bank and the AsianDevelopment Bank, have developed guaranteeschemes. The World Bank's capital-market guaran-tees are used to facilitate the access of developingcountries to the international capital markets bylengthening the maturity of related borrowing. Theproceeds from such loans can be used for infrastruc-ture investments. The World Bank also issues guar-antees for project financingunder the Extended
101
Cofinancing Facility (ECO)to cover sovereignrisks associated with infrastructure projects. This fa-cility, designed to improve developing country ac-cess to international capital markets, has been usedfor the Hub River Project in Pakistan and a thermalpower project in China. The Multilateral InvestmentGuarantee Agency (MIGA) another World Bankaffiliatehas also provided guarantees for severalinfrastructure projects.
Institutions and instruments for resourcemobilization
If the trend toward private investment in infra-structure is to continue, financial markets will haveto respond by providing the necessary long-termresources. Paralleling the innovations describedabove in the structuring of contractual agree-mentswhich are critical to making a project fi-nanceable lessons have been learned about deliv-ering long-term finance through alternativeinstitutions and instruments.
Both foreign and domestic sources of capital willneed to be tapped. Reliance on foreign savings re-mains a necessity for many countries with inade-quate domestic savings. But there are limits to thecapacity of any economy to access funds fromabroad, particularly for debt finance. External bor-rowing must be serviced largely by domestic rev-enues. Overall balance of payments constraints andthe sheer size of infrastructure investments imply,for most countries, that a sustained infrastructureprogram will have to be accompanied by a strategyfor mobilizing domestic funds. In turn, an increas-ing share of domestic savings will need to comefrom private sources as governments reduce theirinvolvement in infrastructure.
As the dominant owner and supplier of infra-structure, however, governments will continue to bea major user of funds, as well as a conduit for re-sources from multilateral development banks. Mu-nicipal governments (responsible for large andgrowing urban infrastructure) represent a majorsource of demand for financial resources. To meettheir needs, new initiatives are being tried, includ-ing the revitalizing of existing infrastructure lendinginstitutions. Governments are also creating special-ized infrastructure funds (discussed later) as a tran-sitional measure to make long-term financing avail-able where private financing is not likely to besufficient. Specialized infrastructure banks andfunds are imperfect mechanisms that need increas-ingly to acquire marketlike discipline, and theirvalue needs to be assessed periodically.
Synergistic links can develop between private in-frastructure projects and domestic financial inter-102
mediation through capital markets. Infrastructuredevelopers and private (especially contractual)savers share a long-term horizon. Bringing compati-ble savers and investors together is the task of capi-tal markets. At the same time, the financing of infra-structure projects improves appraisal capabilitiesand expands risk-diversification possibilities forlocal commercial banks, equity and bond markets,and institutional investors such as insurance com-panies and pension funds. Exploitation of theselinks can be promoted through prudent regulation,improved disclosure and reporting standards, andthe development of credit-rating capabilities andcredit risk insurance.
Infrastructure development banks
In many countries, specialized development banksare a conduit for funds used in infrastructure proj-ects, especially for municipal infrastructure such aswater, solid waste collection and disposal, and localroads. For municipalities, borrowing from suchinstitutions supplements local taxes and centralgovernment transfers and is intended to cover fluc-tuations in expenditure or to prevent large shifts inrevenue requirements.
In developing countries, such specialized infra-structure development banks have suffered from allthe negative features associated with governmentownership, such as inefficient targeting and subsi-dization of lending, interference in operations, andcorruption. Inadequate diversification of risk hasalso led to periods of heavy demand followed bysubstantial slack. Moreover, the banks' traditionalfunction as conduits of government funds is incon-sistent with the trend toward less reliance on gov-ernment budgets and increased use of private sav-ings to finance infrastructure.
In industrial countries, with stronger traditionsof autonomy and solid appraisal capabilities, infra-structure banks have performed better. In Japan,postal savings have constituted the primary sourceof long-term funds used by such institutions as theJapan Development Bank (JDB) to finance infra-structure. The JDB has been crucial to past infra-structure development, and even today, with themove toward public-private partnerships, it contin-ues to play a major role in financing, often at highlysubsidized rates. In Europe, municipal banksob-taining their resources from contractual savings in-stitutions and other long-term sourceshave gen-erally performed well where local governmentshave had operational independence.
Few municipal banks in developing countries,however, have shown a capacity for sustained in-vestment, largely because of undercapitalization,
poor financial discipline, and substantial arrears.Although such banks have helped add to the stockof urban infrastructure, they have done little to pro-mote the capacity or commitment of municipalitiesto expand and operate it efficiently. Exceptions in-clude a facility in Colombia that rediscounts lendingby commercial banks to municipal infrastructureprojects (Box 5.9).
Certain pragmatic principles emerge from the ex-perience thus far. A specialized institution is justi-fied only if the value of business warrants it and ifthe concomitant technical and managerial capabili-ties are available. A more practical alternative is todevelop and improve existing commercial and de-velopment banking channels. An effort in Argentinato create a new lending institution (COFAPyS) dedi-cated to the water sector failed. Besides defects indesign that led to limited funding capability, thebank was seen to offer no benefits beyond those ofexisting channels for routing official developmentassistance.
The long-term goal for existing infrastructurebanksin keeping with the shift toward greatercommercial orientation and accountabilitymustbe to diversify their portfolios and operate underprivate ownership and control, possibly as whole-sale banks. In the interim, the discipline underwhich they operate can be improved. For example,efforts are under way in Morocco to reform theFond d'Equippement Communal (FEC), an agencyestablished in 1959 to fund municipal investment.The FEC is being transformed into an autonomousagency subject to supervision by the finance min-istry and by the central bank, with a board compris-ing mainly central government officials. The re-forms provide the FEC with a new set of operationalguidelines, eligibility criteria, and financial targets.This is not an ideal solution, but until capital mar-kets are better developed or alternative financingmechanisms are available, revitalizing institutionsby making them more accountable is a pragmaticway to proceed.
Looking ahead, specialized infrastructure inter-mediaries could play a catalytic role in capital-mar-ket development. In India, the new and innovativeInfrastructure Leasing and Financial Services andthe more traditional Housing and Urban Develop-ment Corporation (which is seeking to redefine itsrole) aim to sell their loans to other private financialinstitutions once project credit histories have beenestablished. They also plan to package securitiesfrom different projects and to offer shares in thesepackages to investors. Another specialized in-frastructure bank, BANOBRAS in Mexico, is alsolooking for new responsibilities consistent with
Box 5.9 Successful municipal credit inColombia
Colombia's experience with its municipal credit in-stitution is a success story, with a history goingback more than twenty years. The municipal creditinstitution has evolved through successive incarna-tions into the Financiera de Desarrollo Territorial(FINDETER), an autonomous agency that operatesunder the finance ministry.
FINDETER does not lend directly to municipalgovernments but operates as a discount agency toprivate sector and state-owned commercial banksthat make the loans, appraise the projects, andmonitor performance. The system's success has de-pended on the quality of HNDETER's staff and thatof the intermediaries through which it lends. Underthe control of the finance ministry, it has been rela-tively insulated from political pressures.
Between 1975 and 1990, more than 1,300 proj-ects with a value of more than $1 billion were fi-nanced, assisting 600 municipalities. The system'sfunding does not rely on government budgetaryappropriations but rather on bonds, recycling of itsloans, and foreign credits from bilateral and multi-lateral sources.
greater privatization of municipal infrastructure.BANOBRAS is playing an important role in facilitat-ing private water and sewerage projects by guaran-teeing that municipalities will pay for services pro-vided (or will allow water billing and collection).At the same time, BANOBRAS is working tostrengthen municipal finances by demanding betteroperational and financial performance as a condi-tion for its support.
Domestic construction capability is crucial to thedevelopment and maintenance of basic infrastruc-ture, but construction contractors are difficult to fi-nance because they have uncertain cash flows andlimited bankable collateral. Typically, banks dis-count no more than 60 percent of the value of pay-ment certificates issued by government depart-ments. Frequent delays in payments by publicauthorities compound the inherently difficult finan-cial position of contractors, who are often forced toresort to high-interest informal financing. The fi-nancing requirements of the construction industrycan be partially met through local development fi-nance companies that on-lend funds to contractorsfor highways and similar civil works.
Finance is also provided through specialized in-frastructure banks. BANOBRAS, for example, pro-vides short-term loans for public works against con-tractors' receivables from the government agency
103
sponsoring a project, a practice that is thought tohave contributed much to the development of theconstruction industry in Mexico. As part of its trustactivities, BANOBRAS also operates a special fundthat can provide up to 25 percent of the full cost of aproject to finance the start-up costs of construction.Such finance is no substitute for regular payment bygovernment agencies to contractors, however.
New infrastructure funds
Two types of infrastructure funds have emerged inrecent years. Government-sponsored infrastructuredevelopment funds are designed as transitionalmechanisms to provide long-term finance until cap-ital markets are better developed. Private funds, ofwhich there are a growing number, serve the com-mercially useful function of diversifying investorrisk. As transitional mechanisms, these funds servetwo purposes. They allow the leveraging of govern-ment resources or official development assistanceby attracting cofinancing from private sources. Theycan also create credit histories for borrowers per-ceived as risky. In time, these borrowers can securedirect access to capital markets.
The Private Sector Energy Development Fund inPakistan and the Private Sector Energy Fund in Ja-maica are designed to catalyze private financing forpower projects. In response to perceived countryrisk and a lack of long-term financing compatiblewith the requirements of the power sector, the Ja-maican government makes long-term financingavailable through the Energy Fund (up to a maxi-mum of 70 percent of project costs) as a means of at-tracting private investments. Investors in the fundinclude the World Bank and the Inter-American De-velopment Bank. Another example of fund leverag-ing in a developing country is the proposed ThaiGuaranty Facility for financing environmental infra-structure (see Box 5.7). This facility will not lend di-rectly to infrastructure projects but will guaranteeprivate loans to municipalities and private opera-tors. The Regional Development Account (RDA) inIndonesia is a transitional credit system designed toshift financing of infrastructure projects from gov-ernment grants to debt instruments, thereby creat-ing a credit history for borrowers, principally localauthorities. The RDA lends at near-market rates.The goal is to give local authorities three to fiveyears to establish measures for cost recovery and todemonstrate adequate financial managementthusenabling them to borrow directly from financial in-stitutions and capital markets.
Good design for such domestic funds requiresthat they price their loans on market benchmarks. Itis also important to incorporate incentives for pri-104
vate sponsors to seek commercial financing or tocommit a larger amount of equity funds and to en-sure that the fund manager or the operating inter-mediary has a stake in the success of projects fi-nanced. Although appropriately designed fundscould be useful instruments while capital marketsare still developing, reform of the financial sectorand improved creditworthiness of borrowersshould be the long-term goals.
A number of private funds have recently been es-tablished to channel international capital for devel-oping country infrastructure. They pool risks acrossprojects and hence increase the availability as wellas lower the costs of finance. These funds mobilizeresources through private placements from institu-tional investors, including pension funds. For exam-ple, a pension fund with little interest in investingdirectly in a toll road in Mexico might be interestedin participating in a fund that invests in a portfolioof such toll roads. As has been the practice of gov-ernment-backed funds, private funds have concen-trated heavily so far on power projects. Continuedflow of resources into such funds will depend on in-vestments being made in sound projects with credi-ble sponsors as well as on the pace at which regula-tory restrictions on institutional investors arerelaxed.
Development of domestic capital markets
The long-term goal must be to broaden and deependomestic capital markets so that they can serve asefficient and reliable conduits for infrastructure fi-nance. Getting there will require broad investor par-ticipation, a variety of market-making players (bro-kers, dealers, underwriters), and a wide range offinancial instruments. In addition, markets requireadequate disclosure of information to ensure effi-ciency, and effective laws to safeguard investors.
In most developing economies, the informationaland contractual preconditions are not in place for ef-ficient private and commercial financing of infra-structure projects. Private institutions such as credit-rating agencies and public ones such as regulatoryagencies are needed to ensure an adequate flow ofinformation to investors, to facilitate monitoring,and to discipline management. Financial liberaliza-tion and policies to encourage the growth of the for-mal financial sector will in time help overcome suchshortcomings.
Experience shows, however, that equity listingsand bond issues by infrastructure companies orprojects can spur capital-market development byincreasing the range of investment options. Thediscussion here highlights how infrastructure devel-opment, private provision strategies, and capital-
Figure 5.3 Infrastructure is a large share of privatization proceeds; foreign financing ofinfrastructure privatization is important in Latin America.
Privatization proceeds (billions of U.S. dollars)
50
40
Foreign 56%
Domestic 44%
Domestic 98%
3_Foreign 2%
Noninfrastructure
! Infrastructure
market development are best considered within anintegrated framework.
PRIVATIZATION. The privatization of infrastructureconcerns has given a boost to local stock markets. Ofthe $61.6 billion of revenue obtained by developingcountries from the privatization of public enter-prises between 1988 and 1992, about one-third ($21billion) came from the privatization of infrastructureentities. Aggregate proceeds from infrastructure pri-vatization have been highest in Latin America, withthe most activity being in telecommunications (Fig-ure 5.3). Some Asian countries, such as Malaysia andKorea, have opted for partial privatization. OutsideLatin America and Asia, however, privatization hasso far had a limited impact.
Techniques for financing privatization have im-plications for the broadening of share ownership onstock markets and for the general development ofcapital markets. Three privatizations in telecom-municationsEmpresa Nacional de Telecomunica-
ciones (ENTeI) in Argentina, CompañIa de Telé-fonos de Chile (CTC) in Chile, and Teléfonos deMexico (Telmex) in Mexicoand one in electricpower generation, Chilgener in Chile, illustrate theimplications of privatization for financial markets.
All except Chilgener sought a strategic (orcore) investor in order to introduce management ex-pertise and to create a commitment to furthergrowth.
A significant proportion of shares was sold tothe general public, and in all cases shares were allo-cated to employees.
Substantial proceeds from the initial stock of-ferings and subsequent rises in share prices havegiven these companies a dominant position in theirdomestic capital markets.
The two Argentine telephone companies consti-tute almost 40 percent of the market capitalizationin Buenos Aires, and Telmex dominates in Mexicowith a 20 percent share. These large capitalizationshave attracted financing from pension funds, creat-
105
Latin America List \jjand the Caribbean
Source: Sader 1993.
ing the basis for long-term capital flows into thecapital markets. Substantial returns (especially fromtelecommunications and electric utilities), risingmarket shares for infrastructure companies, andgrowing investor confidence are mutually reinforc-ing (Figures 5.4 and 5.5).
Explicit and implicit commitments to growthhave led to ambitious investment programs, fi-
nanced in part through new equity offerings, furthersustaining the growth of the domestic capital mar-ket. As the aggregate numbers indicate, such priva-tizations have been a source of substantial foreignexchange inflows in Latin America. The Argentinegovernment used a debt-equity swap mechanism inthe privatization of ENTe1, bringing in cash proceedsof around $2.2 billion and reducing its commercialbank debt (at face value) by roughly 14 percent ofthe total debt to commercial banks and 7.7 percent ofthe total external debt involved. These privatizedcompanies have also attracted significant portfolioinvestment, directly in the form of equity held in thecompanies and indirectly through such instrumentsas American depositary receipts (see Box 5.2).
BOND MARKETS. Bonds can attract to infrastruc-ture financing a whole new class of investors, suchas pension funds and insurance companies seeking
long-term, stable returns. Generally, it has been therole of the government to foster the development ofbond markets. Government bond issues establishthe benchmarksin terms of pricing and maturitystructurefor bond markets overall.
In developing countries, the use of bond financ-ing is in its early stages. Revenue bonds (used forgreenfield projects and paid back from the project'srevenues) are new in infrastructure finance in devel-oping countries. They have been used to help fi-nance toll roads in Mexico and the Subic Bay PowerStation in the Philippines. Corporate or municipalbonds, based on the credit of a company or govern-ment authority, have been used by infrastructureentities, but the bonds have often been placed on in-ternational markets because domestic bond marketsare underdeveloped.
The experience of industrial countries offerssome guidance. In industrial countries, bond financ-ing is widely used to raise funds for municipal in-frastructure. It has also stimulated the developmentof the local bond market. Municipal authoritiesissue bonds directly. They sometimes pooi theirneeds with those of other local governments, partic-ularly when their borrowing requirements are smallor their creditworthiness is poor. For the investor,municipal bonds have been a source of high re-
Figure 5.4 Infrastructure equities are contributing to the growth of Argentina's capital market.
Stock market capitalization Institutional Investor country ranking
Billions of U.S. dollars35
30
25
20
15
10
5
1989 1990 1991 1992 1993Q3
30
25
20
15
10
5
0
Index of creditworthiness35
106
Source: IFC 1993. Source: Institutional Investor.
1992 1993Q3
1989 1990 1991
-LI Noninfrastructure
- Infrastructure
Figure 5.5 Infrastructure equities haveoutperformed other stocks by a huge margin.
U.S. dollars (1984 = 100)2,000
1,500
1,000
500
0
1984 85 86 87 88 89 90 91 92 93
- Infrastructure stocks index
- IFC Latin America index
- IFCG composite index
Source: IFC 1993.
turnsin part because they are often tax-exempt.But risks have also been high, and market liquidityhas often been low. Municipal bond financing canalso be a device to escape budgetary discipline andhence carries the risk that municipalities may bor-row excessively and then default, leaving the centralgovernment to pick up the tab. Closer surveillanceand legal restrictions on municipal borrowing aretherefore needed complements to the discipline thatmarkets impose.
CONTRACTUAL SAVINGS. Infrastructure companiesand projects add to the supply of long-term securi-ties on the capital market. But for the market tofunction well, there must be a matching demand forsuch securities. Contractual savings institutions,such as pension funds and life insurance compa-nies, are particularly suited to making long-term in-vestments. These institutions levy fixed premiums,have steady and predictable cash inflows, and incur
long-term liabilities, making them ideal suppliers ofterm finance for infrastructure projects.
Chile has used its pension fund system to pro-mote the privatization of public utilities, includingthe Santiago subway system, Soquimich (a chemicaland mining concern), and CTC. Holdings by pen-sion funds account for 10 to 35 percent of the equitycapital of these companies, although the pensionfunds hold less than 10 percent of their portfolios inthe form of stocks of private companies.
The Philippine social security system recentlycreated a 4-billion-peso loan fund targeted to Philip-pine power projects and administered by localbanks. This fund is able to provide fifteen-yearloans of up to 200 million pesos to a bank, whichthen on-lends to the power project company. The so-cial security system thus assumes only the bankrisk. The banks handle appraisal and monitoring,and they can also leverage the funds by addingother resources. International insurance companiesoperating in the Philippines, where there is a dearthof local long-term investment opportunities, havealso begun to view private infrastructure projects asa viable option for their lending portfolios. Theynow make loans with maturities of up to fifteenyears (with approval required from the PhilippineInsurance Commissioner).
In the past, government-sponsored pensionfunds have often suffered from mismanagementand misuse. For such funds to play a significant rolein domestic capital markets, they need greater au-tonomy and more professional management. Expe-rience from Latin America shows that, even whenthey are technically autonomous, pension fundswithin the public sector often come under pressureto finance government consumption spending andlow-yielding investments.
The restrictions on pension fund investments areunlikely to disappear. To protect individual contrib-utors, governments guarantee the security of pen-sion funds. In turn, they require that the funds beinvested prudently. Chilean regulations stipulatemaximum investment limitsby instrument andby issuealthough with increasing experience,these limits are being relaxed. The Chilean model ofprivately managed but publicly mandated and reg-ulated pension funds is being adopted more widelyin Latin America.
Also important are the risk-taking attitudes andabilities of pension funds in developed countriesthat have as yet made only limited investments indeveloping country infrastructure. Availability of fi-nance would greatly increase if regulators and su-pervisory agencies in industrial countries were torelax the severe restrictions on the share and type of
107
Figure 5.6 Options for financing increase with administrative capacity and maturity ofdomestic capital markets.
Higher
Developmentof domestic
capital markets
Lower
One-off projects using project Wfinancing techniques (e.g., BOTs)
Domestic credit ratingand risk insurance
Extensive use ofcontractual savings
Bond financing
Reform of domestic banks andspecialized infrastructurefinancial institutions
Economic development --*' Higherand administrative capability
assets that pension funds and other institutional in-vestors can hold in the capital markets in develop-ing countries.
Prospects
Moving from today's still heavy dependence onpublic financing to tomorrow's system of more pri-vate sponsorship is likely to be a long and some-times painful process. In important respects, the tra-ditional style of infrastructure financing has beentoo easy. Money has flowed through channelswhere scrutiny has often been limited because pub-lic sponsorship has provided high levels of comfortto lenders. The move to a more open and transpar-ent system implies greater scrutiny and the needfor more resources to coordinate many diverseinterests. In return, it offers the promise of greateraccountability.
From the menu of new approaches, the optionsavailable to a country depend on its administrativecapability and the state of its capital market (Figure5.6). Project finance represents the first rung of theladder and should, in principle, be reachable in allcountries. Where capabilities lack adequate depth,
108
the structuring of stand-alone projects using projectfinance techniques may require considerable effortand technical assistance from international agencies.Where domestic capital markets are not well devel-oped and financial intermediaries are weak, theonly other option may be to strengthen specializedinfrastructure finance institutions. Once financial in-termediaries are well developed, they can take onthe task of catalyzing the development of domesticcapital markets through appraisal and underwritingfunctions. And once credit rating and public regu-lation of financial markets are in place, other op-tions open up, and the use of long-term savingsof contractual institutions and the developmentof a variety of financial instruments should becomepossible.
The good news is that private enterprise has beenmoving into a wide range of countries and projects.Legal and regulatory reform is already under way.Infrastructure providers are being privatized. Flowsof foreign direct investment by new infrastructureentrepreneurs are on the rise, as are internationalflows of portfolio capital. And the growth of domes-tic capital markets is a source of optimism. Financefollows enterprise.
6Setting priorities andimplementing reform
The potential for improving performance in infra-structure provision and investment is substantial, asis the quantity of resources devoted to infrastruc-ture. Thus, both the need and the broad direction forreform are clear. Additional investment will obvi-ously be neededbut more investment will not initself avoid wasteful inefficiencies, improve mainte-nance, or increase user satisfaction. Achieving theseimprovements will require three broad actions: ap-plying commercial principles to infrastructure oper-ations, encouraging competition from appropriatelyregulated private sector providers, and increasingthe involvement of users and other stakeholders inplanning, providing, and monitoring infrastructureservices. These adjustments call not only for policychanges, but also for fundamental institutionalchanges in the way that the "business" of infrastruc-ture is conducted. Four major options (introducedin Chapter 1) are available for effecting suchchanges to improve infrastructure provision andperformance and to expand the capacity to provideinfrastructure services.
Choosing among alternatives: institutionaloptions and country conditions
Institutional options
The four institutional options represent different al-locations of ownership, financing, and operationand maintenance responsibilities, and also of riskbetween government and the private sector (Table6.1). These options are not exhaustive but are repre-sentative points on an underlying continuum of in-stitutional alternatives.
OPTION A: PUBLIC OWNERSHIP AND PUBLIC OPERA-
TION. In nearly all infrastructure sectors, the mostcommon vehicle for ownership and operation is apublic entitya parastatal, public enterprise, pub-lic authority, or government departmentownedand controlled by the central, regional, or local gov-ernment. As described in Chapter 2, the delivery ofservices is better carried out when public organiza-tions are run on commercial lines, freed from gov-ernment budget and civil service constraints andsubject to normal commercial codes and regula-tions. Competition from private firms pressurespublic providers to improve their performance andshould not be prevented by regulatory or other bar-riers. Using private contractors for clearly definedservices, such as maintenance of public utilities,provides experience that can gradually be extendedto full operation by the private sector throughleases or concessions (Option B), as both public andprivate parties gain familiarity and confidence withpartnership.
OPTION B: PUBLIC OWNERSHIP AND PRIVATE OPERA-
TION. Through concessions or leases, the public sec-tor can delegate the operation of infrastructure facil-ities (along with the commercial risk) and theresponsibility for new investment to the private sec-tor. As detailed below, the rights to use publiclyowned assets or to provide exclusive areas of ser-vice vary by infrastructure sector. Port leases allowthe use of public facilities, while municipal solidwaste collection contracts award service rights butusually not exclusive use of publicly owned facili-ties. Leases and concessions permit private sectormanagement and financing without the disman-
109
110
Table 6.1 The main institutional options for provision of infrastructure
Public enterprise Private operator Private l'rivate
Othercharacteristics:
Managerialauthority
Government Public enterprise Privateoperator Private operator Private Private
Bearer ofcommercialrisk
Government Public enterprise Mainlypublic
Private operator Private Private
Basis ofprivate partycompensation
Not applicableFixed feebased onservicesrendered
Based onservices
and results
Based on results, net offee paid by operator for
use of existing assets
Privatelydetermined
Privatelydetermined
Typicalduration
No limitFewerthan
5 yearsAbout
3-5 years5-10years
10-30years
No limit No limit
Box 6.1 Ingredients of good performance under alternative institutional forms
Option A: Public ownership and public operation
Government roles as owner, regulator, and opera-tor clearly separated.
No government interference in detailed manage-ment.
Public enterprises subject to general commerciallaw and to general accounting and auditing standards(operating on "level playing field" with private enter-prises).
Tariffs set to achieve cost recovery as appropriate,and enterprise subject to hard budget.
Public service obligations, if any, targeted andcompensated explicitly by government transfers.
Managers selected by professional qualificationsand compensated appropriately.
Appropriate mechanisms in place to obtain feed-back from users.
Discrete activities and functions that can be un-bundled open to private entry (for example, through ser-vice contracts).
Private management skills obtained as needed (forexample, through management contracts).
Ownership and control shared with the privatesector (for example, as minority shareholder).
Option B: Public ownership and private operation
Basic legal framework of contract law, includingcredible enforcement mechanism, in place.
Contracts clearly specify monitorable performancetargets, responsibilities of owner and operator, processesfor periodic review (especially to account for unforeseen
changes in input costs), mechanisms to resolve disputes,and sanctions for nonperformance.
Contracts awarded by transparent selectionprocess, preferably competitive bidding.
Option C: Private ownership and private operation
Appropriate competitive restructuring of subsec-tor undertaken.
Practical and statutory barriers to private entry re-moved (for example, restrictions on access to credit andforeign exchange).
Regulation in place to protect public interestwhen competitive discipline is insufficient and to en-sure private entrants access to network facilities whenrelevant.
Option D: Community and user provision
Participation of users or community membersfrom earliest stage of program preparation to ensurewillingness to pay and ownership of scheme.
Participation of beneficiaries ensured through ap-propriate organizational means, and with contributionsin kind or in cash.
User group supported by access to training andtechnical assistance from sectoral agency or nongovern-mental organizations.
Appropriate consideration given to technical re-quirements for interconnection with primary or sec-ondary network infrastructure, if relevant.
Service operators appropriately trained, compen-sated, and held accountable.
tling of existing organizations or the immediatecrafting of an entirely new regulatory framework.
OPTION C: PRIVATE OWNERSHIP AND PRIVATE OPERA-
TION. Private (including cooperative) ownershipand operation are most attractive to the private sec-tor when there is high potential for securing rev-enues from user charges and when commercial riskand political risk are low. This option is likely toapply most readily in activities that lend themselvesto competition, such as telecommunications, power,gas, railways, and ports (with appropriate tariffpolicies); less readily to waste disposal; and least ofall to rural roads.
OPTION D: COMMUNITY AND USER PROVISION. For
municipal and local services, user provision or com-munity self-help arrangements that provide smaller-scale infrastructure such as village feeder roads,water supply and sanitation, and canals, or small-
scale power generation off the national grid canprovide effective and affordable service in manyareas, when those who contribute to the costs arethe primary beneficiaries. Community self-helpschemes must be selected, designed, and imple-mented locallynot imposed from outside. Theymay also offer the only feasible approach in infor-mal periurban settlements and rural areas until themore formal supply systems expand their networkssufficiently.
Previous chapters discussed experience witheach of these institutional arrangements and the fac-tors contributing to success or failure. The main con-ditions for good performance in each are summa-rized in Box 6.1.
Country conditions
Ultimately, what is needed and what is possible to-gether determine each country's reform options.
111
The quantity, coverage, and quality of existing in-frastructure facilities and their effectiveness inmeeting present and future user demands definewhat is needed. What is possible is determined by acountry's institutional capacity for commercial andcompetitive services in infrastructureits manage-rial and technical capabilities in the public and theprivate sectors, the ability of government to createan enabling environment for private activity, andthe private sector's interest and response. The im-plications for reform are illustrated here for fourcountry types: low-income countries, countries intransition from central planning, middle-incomecountries undergoing economic reform (manyemerging from periods of low growth), and high-growth countries.
WHAT IS NEEDED? The current supply of servicesand the projected growth, as well as changes, in de-mand determine priorities in infrastructure. Indica-tors of supply (infrastructure coverage and per-foniiance) and demand (economic growth anddemographic shifts) vary considerably across thefour country types (Table 6.2)and so do their in-frastructure needs. Low-income economies tend tohave both low coverage and poor performance;nearly every indicator of performance is on averagethree times worse than in OECD countries. Andwith an urban population growth rate of 6 percent,
Source: Appendix table Al; WDI tables 1,25,31,32.
112
demand for infrastructure will be growing rapidly.The socialist economies in transition, by contrast,enjoy high coverage and relatively good technicalperformance. The main challenge in these countriesis to reorient supply to meet the changing pattern ofdemand brought about by economic restructuring.Middle-income reforming economies have rela-tively high infrastructure coverage but weaker per-formance, especially in maintenance. Improving ef-ficiency is their highest priority, in order to providethe service necessary to restart growth. High-growth economies have comparatively good cover-age and performance. For them the challenge is tomeet rapidly expanding needs for infrastructure ofall types to sustain their 7 percent average annualgrowth in output and to service the needs of urbanpopulations growing by 4 percent a year.
WHAT IS POSSIBLE? The bounds of the possible aredrawn by each country's capacity to implement re-forms, as defined by three characteristics. First is thecountry's managerial and technical capacity. Pro-viding infrastructure is a technically complex activ-ity requiring engineers and other professionals, aswell as managers who understand the need to meetconsumer demands. Second is the enabling en-vironment for beneficial private sector involve-mentboth attracting investment and channelingit productively. The commitment and integrity of
Table 6.2 Country infrastructure coverage and performance
Indicator
Low-income
economiesTransitioneconomies
Middle-incomereformingeconomies
High-growth
economiesOECD
economies
Coverage of infrastructureMain lines per thousand persons 3 95 73 122 475Households with access to
safe water (percent) 47 95 76 86 99Households with electricity (percent) 21 85 62 61 98
Performance of infrastructureDiesel locomotives unavailable (percent) 55 27 36 26 16Unaccounted-for water (percent) 35 28 37 39 13Paved roads not in good condition
Basic indicatorsGNP per capita, 1991 (U.S. dollars) 293 2,042 1,941 3,145 20,535GNP per capita average annual growth
rate, 1980-91 (percent) 0.2 1 0.6 5 2
Population average annual growth rate,1 980-91 (percent)
Urban 6 1 3 4 1
Total 3 0.3 2 2 0.5
government and the strength of the country's un-derlying institutions determine the enabling envi-ronment. Stability of the macroeconomic and sec-toral policy climate is critical for attracting long-terminvestments, as are supporting institutions andstructures such as well-functioning judicial and fi-nancial systems. The capacity of regulation andinstitutions to promote the public interest (as dis-cussed in Chapters 3 and 4) is an integral compo-nent of a climate for private involvement. Third isthe private sector's capacity and will to assemblethe resources needed to supply services in the con-struction, financing, and operation of infrastructure.
In low-income countries, all three dimensions ofcapacitytechnical capability, the enabling envi-ronment, and private sector interestare typicallylow. But large countries in this group, such as India,have a greater depth of technical capacity and pri-vate sector interest because of their size, a fact thatin some ways makes them distinct. In formerly so-cialist economies in transition, technical capability ishigh, but an enabling environment for market activ-ity is just emerging (and private sector capacity isnewly developing). In middle-income reformingcountries, technical and managerial capabilities aregenerally high, and the enabling environment is rea-sonably well established. However, in many suchcountries, low growth restricts private sector capac-ity. High-growth countrieswith generally strongtechnical capacity a favorable business climate, andkeen interest from the private sectorare poised totake advantage of all institutional options.
Differences in country capacity affect the choiceof reform. Institutional arrangements differ in thedemands they make on government administrativeand regulatory resources, as well as in their degreeof dependence on private sector participation. Asdiscussed in Chapter 3, the choice between conces-sions and privatization depends largely on whetherit is more desirable to regulate private sector in-volvement through contractual arrangements orthrough a regulatory agency. Where the economicenvironment is uncertain or evolving (as in low-income or transitional economies), it may be easierto induce private sector entry through contractsboth because agreements can be detailed in advanceand because ownership does not change. Thatarrangement puts less private capital at risk. At-tracting private investment for system expansion isanother matter. As discussed in Chapter 5, privateownership or long-lived concessions are usuallyneeded to induce private investment. The adminis-trative capacity necessary for contracts or for pri-vate sector ownership depends on how much regu-
lation is required. Where free entry can be allowedwithin a competitive environment, private sector in-volvement would only require regulation to ensurefair business practices and to protect health, safety,and the environment, which are common to allsectors.
The choices are not simple. Poor service provi-sion by the public sector often suggests a need formore private involvement. To the extent that poorpublic sector performance occurs in natural monop-oly activities, private involvement may not be desir-able in the absence of adequate regulatory con-trolsbut if the public sector agencies lack thecapacity to administer regulation or, more funda-mentally, if credible governance is lacking, a well-regulated private sector alternative will not be pos-sible. If the choice must be between highlyimperfect options, countries must weigh the alter-native of a minimally regulated private monopolistthat can expand service and achieve reasonable op-erational efficiency against the alternative of a pub-lic monopolist that delivers inadequate service athigh cost to the public treasury. To minimize risks topublic welfare in the case of concessions and priva-tizations, public scrutiny and transparency are im-portant to avoid the granting of "sweetheart" dealsthat can quickly sour the taste for private involve-ment. The introduction of competition is in manycases the most important step in creating conditionsfor greater efficiency by both private and public op-erators, and the performance of the public sector en-terprises that remain can be further improvedthrough such means as contracting out specific ser-vices. In high-growth countries, public agenciesoften perform quite welland, while the capacityfor private involvement in these countries is greater,the urgency for reform may be less.
The choice of institutional option can affect thedevelopment of domestic capacity. Concessions ormanagement contracts can be used to obtain specificexpertise not available domestically. Thailand hasused foreign expertise in developing its irrigation,railway, and airline capabilities; Côte d'Ivoire hasmanaged a transfer of skills from expatriate to localstaff in its water supply concession. Contracts andcontracting out can also contribute to the develop-ment of a healthy domestic construction and con-sulting industry and draw on the talents of formerpublic employees. With appropriate attention tocontract design and supervision, competence andexperience may often be achieved in the domesticprivate sector even in the poorest countries. For ex-ample, road maintenance is now done privately in alarge number of African countries.
113
Sectoral agendas for reform
Although country characteristics are important, sec-toral characteristics cast the deciding vote amonginstitutional options. The "marketability" of infra-structure activities is determined by the followingcharacteristics: production technology that leads tonatural monopoly; the public nature of consump-tion; constraints on cost recovery; distributionalconcerns; and the importance of spillover effects.Table 6.3 illustrates the differences, both within andbetween sectors, in the marketability of infrastruc-ture activities. Each activity is scored from 1 (leastmarketable) to 3 (most marketable) according to thefive characteristics just specified (the darker theshading, the more marketable the activity). For in-stance, large networked facilities, such as transmis-sion grids, primary irrigation channels, andrailbeds, allow very little competition, while activi-ties such as urban waste collection and urban busservice are potentially quite competitive. Some in-frastructure goods, such as phone service or tertiaryirrigation, are entirely private in consumption whileothers, including many roads, are public goods. Thelast column of Table 6.3 gives an index of mar-ketability potential (the simple average of the fivecolumns).
This exercise suggests that the potential for com-mercialization and competition in infrastructure ismore widespread than is commonly supposed.Some activities, such as long distance telecommuni-cations, urban bus services, or solid waste collec-tion, are adaptable to market provision once theyare unbundled from related activities. Other activi-ties, such as urban piped water and power trans-mission, are intrinsically monopolistic but provideprivate goods amenable to commercial provisionand cost recovery. Rural roads are intrinsically pub-lic infrastructure, being both monopolistic and apublic good with low potential for cost recovery.
The following sections relate the main options toinfrastructure sectors within three major groupstelecommunications and energy, water and waste,and transport. The options suggested are indicativeof what is most relevant, not narrowly prescriptiveof the only, or single best, approach in each countrytype.
Telecommunications and energy
This infrastructure group covers utilities that pro-duce services for which user fees are charged, typi-cally based on direct measures of consumption, andthat generally use large-scale networks for distribu-
114
tion. Most elements of these services are highly mar-ketable and can be provided through approaches in-volving competition within a market or competitionfor the right to serve that market (Table 6.4). Themajor exceptions have strong scale economies(power transmission) or require specific sites andhave significant environmental effects (large-scalehydroelectric generation).
TELECOMMUNICATIONS. The marketability of tele-communication services is high, especially for longdistance and value-added services such as datatransmission. Falling transmission and switchingcosts, technological innovations (such as wirelessservices), and changing patterns of demand havestrongly boosted the competitive potential of thetelecommunications industry for most services, in-cluding in many cases local telephony. Private pro-vision is appropriate in countries that have thecapacity to provide the necessary regulatory frame-work. Concessions with regulatory constraints em-bodied in contracts are attractive alternatives incountries where independent regulatory capacity isunlikely to be effective.
Today, the challenge of meeting the large and ris-ing demand can be met by moving toward a sectorstructure that is plural and competitive, with a mixof service providers private and publicusingvarious technologies and offering services tailoredto different user needs. New entry is the single mostpowerful tool for encouraging telecommunicationsdevelopment because monopolies rarely meet alldemands. Licensing multiple providers is the bestway to accelerate the investment needed to create abroadly based national network. Additional serviceproviders also increase user choice, lower costs, andbring capital and management skills into the sector.Market liberalization, like privatization, puts pres-sure on existing service providers to be more effi-cient and more responsive to consumers.
The transition from state-owned monopoly tomultiple operators requires new attention to regula-tion. Preventing the dominant operator from abus-ing its market power (by restricting output and un-derpricing competitive services) requires properaccounting and disclosure requirements, perfor-mance targets, and incentive-based price controls.Experience shows that new service suppliers willnot be able to interconnect with the incumbent oper-ator on reasonable terms without regulatory aid.Service providers, both public and private, shouldoperate at arm's length from the government and besubject to commercial discipline and to oversight byan independent regulator.
Table 6.3 Feasibility of private sector delivery varies by infrastructure components.
a05)
Key to marketability rating:
= 1.0 (least marketable)
LI =2.0
= 3.0 (most marketable)
Local services
Long distance and value-added
Thermal generation
Transmission
Distribution
Gas production, transmission
Railbed and stations
Rail freight and passenger services
Urban bus
Urban rail
Rural roads
Primary and secondary roads
Urban roads
Port and airport facilities
Port and airport services
Urban piped network
Nonpiped systems
Piped sewerage and treatment
Condominial sewerage
On-site disposal
Collection
Sanitary disposal
Primary and secondary networks
Tertiary (on-farm)
Low
Low
Qub
Due to either absence of scale economies or sunk costs, or existence of service substitutes.Marketability index is average of ratings across each row.Including cargo handling, shipping, and airlines.
Medium
Medium
Few
Medium
Medium
Low l'ublic Low Many High
iligh
High
2.6
3.0
2.6
2.4
2.4
3.0
2.0
2.6
2.4
2.4
1.0
2.4
1.8
2.0
2.6
2.0
2.4
1.8
2.0
2.4
2.8
2.0
1.4
2.4
115
Potential for Public servicePotential Characteristics cost recovery obligations Market-
for of good from user (equity Environmental abilitcompetition a or service charges concerns) externalities index
Table 6.4 Options in telecommunications and energy
Options key:s A - Commercialized public authority
B - Concession or leaseC1 - Private sector with interconnection or access regulation onlyC2 - Private sector with price regulation
The policy options in telecommunications aresimilar for all country types (Table 6.4). In low-income countries, extremely limited access totelecommunications calls for a liberal policy on theentry of private suppliers. These providers can re-spond to users who are willing to pay for regularservice and can offer alternative communicationsteclmologies (radio- or satellite-based) to establishbasic service for provincial areas. In most middle-income countries, the regulatory environment al-lows for entry by new providers and for the privati-zation or commercialization (through concessions)of existing services.
POWER. The potential for competition in thepower sector is greatest for thermal generation anddistributionactivities that can be unbundled fromexisting vertically integrated power utilities and op-erated under concession. Alternatively, these activi-ties can be privately provided. A minimum market
116
size may be necessary before unbundling becomesworthwhile, however, and in the very small marketsof many low-income countries, vertical separationof generation from transmission and distributionmay not produce sufficient efficiency gains to offsetthe additional coordination costs involved. In virtu-ally all countries, large-scale hydroelectric genera-tion (because of unique environmental and risk fea-tures) is likely to remain publicly owned but can beoperated on commercial principles for example,under management contracts. Small-scale hydro-electric facilities can be privately owned.
Sector policies that take advantage of opportuni-ties for competition in the generation of power canimprove efficiency and lower costs. Concessions arean established means of increasing sectoral effi-ciency. In order to compete, private power produc-ers must have access to the national grid and becoordinated by a network manager. For the foresee-able future, national power transmission will retain
Large hydroelectric generation 1.4 A,eB A, B *A,B A,S BTransmission 2.4 A, B B,C1 S B, C1Distribution 2.4 B SB,'C2 S C2
GasProduction/distribution 3.0 B, C1 C1 C1
elements of natural monopoly and must be regu-lated when privatized.
Reform of the dominant entities that will remainin many countries - especially in power transmis-sion should focus on creating financial and man-agerial autonomy and on promoting commercialbehavior. Doing so will often require private partici-pation in ownershipthrough joint ventures or di-vestitureand private management or concessioncontracts, although private companies involved inpower transmission are best kept separate from pri-vate companies involved in power generation. Insti-tutional change is needed to provide incentives forsuppliers to seek economic tariffs, which are neces-sary to promote the self-financing of investment,conservation of energy, and more efficient use of ex-isting capacity. Tariffs must also incorporate any en-vironmental charges paid by power companies, inline with the principle that the polluter pays for anyenvironmental costs it imposes on others.
GAS. Natural gas could potentially be competi-tively supplied in many countries. Often, naturalgas production is vertically integrated with petro-leum production that is under public ownership.Unbundling is required to permit competitive pro-duction under concessions, contracts, or privateownership. The main regulatory issue is to ensurecompetitive access of producers to the transmissionpipeline. That assurance can be handled by a regula-tory body or through contract terms in leases orconcessions. Competition from substitute fuels(when realistically priced) can provide sufficientmarket discipline to obviate the need to regulate gasprices. Private (foreign) investment has consider-able potential to meet investment needs for gas pro-duction and distribution, provided that noncom-mercial risks related to the heavy foreign exchangerequirement of projects can be reduced.
Water and waste
Activities involving water and waste all have strongenvironmental links that make them less marketablethan telecommunications or energy, and their localnature makes some activities natural candidates forcommunity provision (Table 6.5). User fees are com-mon in these sectors, although they rarely cover thefull costs of service.
WATER SUPPLY AND SEWERAGE. Urban pipedwater and sewerage at the municipal or metropoli-tan level should be provided by enterprises run oncommercial principles. Professional management
accountable to users and having clear incentives forproviding high-quality, reliable services and effi-cient asset management is also desirable. The re-sponsibilityof government in such situations is, atminimum, to ensure commercial operation, whichcan be achieved through delegation to a privatecompany via a management, lease, or concessioncontract. Public oversight is necessary to ensure ac-cess for low-income users and to protect publichealth and environmental quality. In countries withmodest technical capacity, concessions can success-fully draw on international expertise. Pricing waterto reflect the full financial, environmental, and eco-nomic costs of supply is essential for generatingfunds to expand service and for promoting efficientuse.
SANITATION. Low-income countries should con-sider a two-pronged approach to developing sanita-tion. First, contracting schemes, such as concessions,can apply commercial management to sanitation fa-cilities in urban areas. Second, in poorer urban andrural communities which are unlikely to be con-nected to the formal supply systems in the foresee-able future, intermediate technology can be adaptedto match users' service requirements and their will-ingness to pay. These lower-cost tertiary systems(facilities directly serving end-users, described inChapter 4) can be chosen, financed, and operated bythe community with technical assistance. The trunkinfrastructure to which the tertiary systems connectand the associated treatment facilities remain the di-rect responsibilityin planning, financing, and op-erationof the sector utilities concerned.
IRRTGATION AND DRAINAGE. The policy agenda forirrigation works also varies according to the charac-teristics and scale of the systems involved, but it ismuch the same across country groups. The opera-tion of trunk and feeder facilities can increasingly behandled by financially autonomous entities, whilethe ownership and operation of tertiary systemsmay be best devolved to user associations or coop-eratives. This solution improves both maintenanceand the collection of water chargestwo perennialproblems in many irrigation systems.
User associations for operation and maintenanceof small-scale irrigation schemes and tertiary canalnetworks have proved successful in countries as di-verse as Argentina, Nepal, the Philippines, and SriLanka. Colombia, Indonesia, and Mexico have suc-cessfully transferred responsibility for operationsand maintenance to farmers, even for larger-scalestate-owned schemes. Careful preparation has been
117
Table 6.5 Options in water and waste
Options key:A - Commercialized public authorityB - Concession or leaseC - Private sector without regulation
s C1 - Private sector with access regulation or regulation of exclusive service contractsD - Local community and user self-help (with technical assistance)
needed to ensure that farmers feel a sense of owner-ship and that problems resulting from deferredmaintenance by the public authority have been re-solved. Economic pricing is essential to createproper incentives for farmers to use water effi-ciently, as has been done in Mexico (described inChapter 4).
SOLID WASTE MANAGEMENT. In many developingcountries, municipal sanitation departments engagein all stages of solid waste management. They ex-pend a major share of local budgets for that pur-pose, yet they typically collect only 50 to 70 percentof solid waste and do not achieve environmentallysafe disposal. In all country groups, the collection of
118
urban solid wastes can be carried out more effi-ciently under contract by the private sector. The ac-tivities of informal groups that have traditionallyundertaken recycling or resource recovery in manycountries can be made safer and more efficientforexample, the traditional scavengers (Zabbaleen) inCairo were transformed into a private companycontractually responsible for collection, transport,and recovery of waste.
Ensuring environmentally safe disposal throughsanitary landfills or incineration requires more di-rect involvement by governments in planning andregulation because disposal has large externalitiesand economies of scale that make competitive pro-vision less viable. Municipalities may collaborate in
MarketabilityLow-
income
Middle-income countries
Sector and activity index countries Transition Reforming High-growth
Water supplyUrban piped network 2.0 SB A,SB A,SBRural or nonpiped 2.4 D SD SD
Sanitation and seweragePiped sewerage and treatment 1.8 i A, S B A, SB A, S B SA, SBCondominial 2.0 SD 41D SDOn-site disposal 2.4 C, S D C, D C, D C, * D
IrrigationPrimary and secondary networks 1.4 A,SB PA,SB A,SBTertiary (on-farm) 2.4 D SD SD SD
solid waste disposal through metropolitan or re-gional entities operated under contracts with theprivate sector, as in Caracas, Sio Paulo, and othercities in Latin America. Concessions are a usefulmeans of obtaining technical expertise in waste dis-posal technology.
Transport
The transport sector allows for a rich mix of optionsfor service provision. Roads offer the least scope forprivate sector involvement (because of pricing prob-lems), and sectors such as railways need some regu-lation because of large sunk costs or to ensure net-work access (Table 6.6).
RAILWAYS. The essential element of reform for arailway is to give it autonomy to operate as a com-mercial activitya commitment that can bestrengthened through an infusion of private equitycapital and private management. Railways in devel-oping countries have typically been heavily regu-lated, structured as monolithic organizations, andsaddled with uneconomic lines and overemploy-ment. The resulting fiscal subsidies and unreliableservice make the railways less able to modernizeand to compete with other modes of transport. Thepresence of intermodal competition for freight andpassenger services calls for a reform strategy thatwould largely remove price regulation from rail ser-vices and grant the railways structural flexibility to
Table 6.6 Options in transport
A - Commercialized public authorityB - Concession or leaseC1 - Private sector with access or route regulation only
t C - Private sector with price regulationD - Local community and user self-help (with technical assistance)
119
MarketabilityLow-
income
Middle-income countries
Sector and activity index countries Transition Reforming High-growth
RailwaysRailbeds and stations 2.0 iA A, B B
Rail freight 2.6 A, C1 A, C1 C1 C1Passenger services 2.6 A,.B A,B *B,i C2 B, C2
Urban transport servicesUrban bus 2.4 C1 C1 C1 C1
Urban rail 2.4 SB B
RoadsPrimary and secondary roads 2.4 A, :B-toll A,S B-toll A,SB-tollRural roads 1.0 SDUrban roads 1.8 A
Ports and airportsFacilities 2.0 A, B A,iaB A, B A, B
Services 2.6 B, C1 i-B, C1 B, C1 B, C1
Options key:
permit them to drop uneconomic lines. An exceptionto rail service deregulation should be made for cap-tive shippers without access to alternative modes.
Vertical separation of track and facilities from railserviceswith the latter operated under contract byentities other than the owners of the rail infrastruc-tureis a strategy being considered or adopted insome countries, including Argentina, Chile, Côted'Ivoire, and Mexico. Such separation requires well-defined access rules and agreements for investmentand maintenance and may not be workable in coun-tries with modest institutional capacity.
The countries in transition are already beginningto tailor their stocks of railway assets to serve the fu-ture needs of a restructured (and geographicallyredirected) industrial sector. These countries alsoneed to increase freight tariffs to cover costs and toimprove the energy efficiency of their locomotivefleet.
URBAN TRANSPORT. Urban transport services canbe supplied by private operators or under conces-sions. Provision of bus and taxi transport in urbanareas is an activity in which entry and exit are rela-tively simple, and competition can flourish acrossall country groups. In addition to regulation to ad-dress safety and environmental concerns, some con-trol over route structure and the allocation of busservices to specific routes may be appropriate. Gen-eral restrictions on entry or fares are usually unnec-essary. Urban rail services lend themselves to leases,concessions, and contracts for service provision.
A comprehensive urban transport strategy re-quires that all available modes be examined, includ-ing subways or other rapid transit, private cars, andnonmotorized transport (which may call for pedes-trian sidewalks and bicycle lanes). Strategic choicesabout the relative roles of personal vehicle transportand public transport require a full assessment ofcosts and benefits, including economic, financial,and environmental impacts and effects on land use.Traffic management policies have high priority be-cause better-moving traffic provides major benefitsin terms of efficiency, safety, and the reduction of en-vironmental pollution. These policies require en-forcement capacity in order to be effective. Eco-nomic pricing of fuels and urban land (especiallyparking space) and management of demandthrough the pricing of road access to urban areas arepolicies with increasing relevance to countries withgrowing urban congestion.
ROADS. The key issue for policy concerningroads is to develop institutions that will manage
120
and maintain them adequately and that have suffi-cient funding to do so. This is of particular impor-tance for network components for which it is diffi-cult to charge usersthat is, the vast majority of thenational, rural, and urban networks that do not lendthemselves to provision through toll roads. Reform-ing the management of these roads involves assign-ing clear responsibility for operation and mainte-nance to appropriate authorities. It also meansdesigning a system of economic road-user charges(ideally, including the axle-load-based costs in-flicted by different users), instituting a financingscheme that links users' payments to maintenanceexpenditure (in order to create accountability ofroad agencies to users), and introducing a mecha-nism for users to influence expenditures on roadmaintenance. The recently restructured road author-ities of Sierra Leone and Tanzania provide goodmodels. Both include representatives of users (suchas chambers of commerce, automobile associations,haulers, and other citizens' groups) as well as engi-neers and government officials.
As discussed in Chapter 2, periodic road mainte-nance (for which performance standards can bemore easily defined and monitored than for routinemaintenance) is increasingly executed more effi-ciently by the private sector under contract than bypublic employees. Rural agricultural feeder roadscan be maintained in part by local organizations andcommunities. The most successful experiences com-bine local control of maintenance with some gov-ernment funding or provision of materials. In low-income countries, attention should be directed topromoting cost-effective labor-based approaches forroad maintenance, and to construction. Privately fi-nanced toll roads can be developed for certain roadlinks, particularly major intercity links where trafficflows are high.
PORTS AND AIRPORTS. Ports and air transport raisemany of the same policy issuesand opportuni-tiesas railways. Although competitive provisionof facilities (port infrastructure and airport runwaysand gates) is not economically efficient (because itinvolves large fixed costs that are sunk), equippingand operating such facilities is a contestable activity.Leases and concessions are appropriate options foroperating ports and airports. The competitive provi-sion of berths within ports is also feasible.
When ports and airports are subject to competi-tion from other traffic modes or from neighboringfacilities, prices for port and airport services can bederegulated. To ensure the high-quality, reliable ser-vice required for international trade, it is equally im-
portant that institutional activities such as customsclearance and international communications at theports should facilitate, not obstruct, the movementof goods.
Payoffs from reform
Implementing reform will not be easy. As discussedin Chapter 2, improving productivity will often re-quire firms to shrink workforces. Creating commer-cial enterprises will also mean that prices will risein many sectors, especially in power and water sup-plyincreases often resisted by the powerfulmiddle-class constituencies that benefit most fromsubsidies. But in many countries, dissatisfactionwith existing services is so strong that initially un-popular measures may become palatable if they areaccompanied by effective efforts to improve ser-vices. This provides astute leaders with an opportu-nity to bring about reform. Experience shows thatsuccess requires both a strong commitment fromgovernment and carefully designed implementationstrategies to reduce the costs of reform.
Donors can play a role beyond financing invest-ments in infrastructure by assisting countries tostrengthen their institutional capacity for undertak-ing sustainable reform and sectoral development.Building institutional capacity involves formulatingappropriate policies and putting in place mecha-nisms for their implementation, creating enforce-able legal and regulatory frameworks, and strength-ening human resourcesincluding managementexpertise in the private sector and administrativeskills in the public sector. Enhancing institutionalcapacity in all of these dimensions implies creatinga positive enabling environment for the efficient, re-sponsive delivery of infrastructure services.
External assistance programs can enhance coun-tries' institutional capacity by making relevantknowledge available; by supplementing policy ad-vice with well-designed programs of technical coop-eration and training; and by providing financial as-sistance for investment and reform. On the firstpoint, donors can help to identify the needs and pri-orities for reform through sectoral analysis and re-search and by disseminating knowledge of bestpractice across countries. They can also sponsor sys-tematic data collection and analysis of informationabout sector performance, both to improve policy-making within countries and to promote learningfrom the successes and failures of others. Second,external assistance can provide training and techni-cal cooperation in support of the efforts of develop-ing countries to design and implement reforms and
to manage infrastructure services. Third, externalassistance programs can provide financial resourcesto support countries' sectoral reforms and to pro-mote sustainable developmentfor example, bygiving greater priority to maintenance and rehabili-tation, and by ensuring that the effective demand ofbeneficiaries and concerns of other stakeholders areassessed early in project identification. Such actionsof institution building will take longer to design andimplement, and will be less predictable in outcome,than traditional development assistance, but theyare essential to foster needed improvements in in-frastructure sectors.
Although countries can acquire the necessaryknowledge, skills, and financial resources for reformfrom outside, the commitment to reform must behomegrown. But the payoffs are potentially large,making the commitment well worthwhile.
Reform will lead to gains from three sources: re-duction in subsidies, technical gains to suppliers,and gains to users. Although the gains will obvi-ously differ from country to country, it is possible todevelop rough estimates of the payoffs from re-forms under the first two headings. Table 6.7 givesestimates of the fiscal burden of service provisioncosts not recovered from users. Even though (con-servative) estimates can be made for only three sec-tors (power, water, and railways), the total is $123billion annuallyrepresenting nearly 10 percent oftotal government revenues in developing countries.For some countries, the losses reach remarkable pro-portions. Before reform, the subsidy to the Argen-tine railway alone reached 9 percent of the totalpublic sector budget, or 1 percent of GDP. Althoughelimination of underpricing would not produce adirect resource saving to the economy (as the costswould be covered by users), the fiscal relief wouldbe enormous.
The second source of gains is the annual savingsto service providers from improving technical effi-
Table 6.7 Fiscal burden of underpricedinfrastructure(billions of U.S. dollars)
SectorSavings frombetter pricing
90
135
15
123
Source: Ingram and Fay 1994.
Source
Underpricing
UnderpricingIllegal connections
Underpriced passengerservice
121
Power
Water
Railways
Total
Table 6.8 Savings from increased efficiency(billions of U.S. dollars)
Sector Savings Source of inefficiency
Roads 15 Annual investment requirementscreated by improper maintenance
Power 30 Transmission, distribution, andgeneration losses
ciency. The savings that could be achieved by rais-ing the efficiency of operation from current levels toattainable best-practice levels are estimated atroughly $55 billion (Table 6.8). These represent pureresource savings to the economy. Although the esti-mates cover only certain sectors and only some ofthe technical losses in those sectors, the efficiencycosts are equivalent to 1 percent of developingcountries' GDP and are more than twice the annualdevelopment assistance flows for infrastructure.One-quarter of the $200 billion annual investment ininfrastructure by developing countries could begenerated just from feasible technical savings. Notonly low-income countries stand to benefit. Al-though access to infrastructure increases as incomesrise, infrastructure efficiency is not closely related to
122
income, and therefore virtually all countries havethe potential to make significant gains.
Passing up such gains translates directly intohuman costs because it means limiting progress inreaching the I billion people who still lack safedrinking water and the nearly 2 billion who lack ac-cess to electricity and adequate sanitation facilities.At current costs of roughly $150 per person forwater systems, the redirection over three years ofjust the annual quantifiable technical losses of $55billion would mean that the 1 billion people withoutsafe drinking water could be served.
Thus, although impossible to quantify globally,the most important potential payoffs almost cer-tainly go beyond limiting financial losses and im-proving technical efficiency and would result ingains both in economic progress and for the poor.Better services improve productivity and well-beingthroughout an economy. Increasing the reliability ofpower and telecommunications will save businesseslost output and redundant investments. Better-maintained roads will lower the costs of vehicle op-eration. Improved rural infrastructure can raise theincomes of the rural poor from farm and nonfarmactivities. Better water and sanitation are critical tothe poor, who spend time and money compensatingfor inadequate infrastructure. All of these improve-ments will contribute to raising living standardsby increasing wages in more productive businesses,lowering prices through more efficient transport,and enhancing the quality of life for individuals.
Bibliographical note
This Report has drawn on a wide range of WorldBank sourcesincluding country economic, sector,and project work and research papersand on nu-merous outside sources. The principal sources arenoted below and are also listed by author or organi-zation in two groups: background papers commis-sioned for this Report and a selected bibliography.The background papers are available on requestthrough the World Development Report office. Theviews they express are not necessarily those of theWorld Bank or of this Report.
In addition to the sources listed below, manypeople, both inside and outside the World Bank,helped with the Report. In particular, the core teamwishes to thank Sri-Ram Aiyer, Gary Bond, JohnBriscoe, Robert Burns, Laurence Carter, MichaelCohen, Jean Doyen, Nissim Ezekiel, Ian Heggie, Ar-turo Israel, Emmanuel Jimenez, Shinichiro Kawa-mata, Johannes Linn, Gobind Nankani, Guy Pfeffer-mann, Louis Pouliquen, Andres Rigo, EverettSantos, Zmarak Shalizi, John Shilling, WarrickSmith, Andrew Steer, Richard Stern, Inder Sud,Vinod Thomas, Louis Thompson, Michael Walton,and Hans Wyss.
Others who provided notes or detailed com-ments include Dennis Anderson, Robert Anderson,Hans Apitz, Ephrem Asebe, Mark Baird, Zeljko Bo-getic, Richard Brun, José Carbajo, Krishna Challa,Armeane Choksi, Anthony Churchill, Sergio Con-treras, Dennis de Tray, Shantayanan Devarajan, 1st-van Dobozi, Gunnar Eskeland, Asif Faiz, John Flora,Louise Fox, Hernan Garcia, Amnon Golan, OrvilleGrimes, Luis Guasch, Jeffrey Gutman, KennethGwilliam, Ricardo Halperin, Roger Heath, NormanHicks, Vijay Jagannathan, Frida Johansen, AuKhadr, Homi Kharas, Michael Klein, Pierre Landell-
Mills, Kyu Sik Lee, Andres Liebenthal, Alain Locus-sol, David Lomax, Millard Long, Sergio Margulis,Costas Michalopoulos, Pradeep Mitra, Mohan Mu-nasinghe, Sheoli Pargall, Anthony Pellegrini, SanjayPradhan, D. C. Rao, John Redwood III, Au Sabeti,Mary Shirley, Jerry Silverman, Martin Staab, PedroTaborga, Mateen Thobani, Thomas Walton, PeterWatson, Steven Webb, Jim Wright, and GuillermoYepes.
Many people outside the World Bank contributedadvice, comments, and material. Contributors andconsultation meeting attendees from governmentsand bilateral aid agencies include Mueen Afzal, Pak-istan Ministry of Finance; Joy Barrett, U.S. PeaceCorps; Henk Bosch, Netherlands Directorate Gen-eral for International Cooperation; Emmanuel deCalan, Mme. Chedeville-Murray, M. Gardin, and M.Perelman, Ministry of Foreign Affairs, France; AnneCharreyron-Perchet and Claude Martinand, Min-istry of Public Works, Transport, and Tourism,France; John Crook, New Zealand Telecom House;Zou Deci, Chinese Academy of Urban Planning andDesign; Gabor Demszky, Mayor of Budapest; Mi-chio Fukai and Koichiro Fukui, Japan DevelopmentBank; Yoshitaro Fuwa, Japan Overseas EconomicCooperation Fund; Tøre Gjos, Norwegian Agencyfor Development Cooperation; Eilif Gundersen,Ministry of Foreign Affairs, Norway; Bruno Gurtner,Swiss Coalition of Development Organizations;Cielito Habito, Philippines National Economic andDevelopment Authority; Ameur Horchani, Ministryof Agriculture, Tunisia; C. K. Hyder, MetropolitanChamber of Commerce and Industry, Bangladesh;Yves Jorlin, Caisse Francaise Developpement; PeterJ. Kalas, Swiss Federal Office for Foreign EconomicAffairs; Patrick Lansman and Jean-Michele Sev-
123
erino, Ministry of Cooperation, France; BoguslawLiberadzki, Minister of Transport and MaritimeEconomy, Poland; Aladar Madrarasz, Counselor,Budapest; Pekka Metso, Ministry for Foreign Af-fairs, Finland; Michael Morfit, U.S. Agency for Inter-national Development; Yukio Nishida, Japan Over-seas Coastal Area Development Institute; Paul Peter,Swiss Development Corporation; Anna MariaPinchera, Ministry of Foreign Affairs, Italy; MasihurRahman, Ministry of Communications, Bangladesh;Gedeon Rajaonson, Ministry of Public Works, Mada-gascar; Prathap Ramanujam, Ministry of PolicyPlanning and Implementation, Sri Lanka; lens ErikBendix Rasmussen, Danish International Develop-ment Association (DANIDA); Jacques Rogozinski,National Bank of Public Works and Services, Mex-ico; Joao Salomão, Minister of Construction andWater, Mozambique; Wongcha-um Sansern, Na-tional Economic and Social Development Board,Thailand; Eduard V. Sjerp, Counselor for Trans-portation, Royal Netherlands Embassy; MikaelSöderbäck, Swedish International Development Au-thority (SIDA); Sugijanto Soegijoko, National Devel-opment Planning Agency, Indonesia; Juha Suonen-lahti, Finnish International Development Agency;Jon Wilmshurst, U.K. Overseas Development Ad-ministration (ODA); and Tony Zeitoun, CanadianInternational Development Agency (CIDA).
Contributors and consultation meeting attendeesfrom multilateral agencies include Fabio Ballerin,OECD; Ananda Covindassamy and Clell Harral,European Bank for Reconstruction and Develop-ment (EBRD); Shashi Desai and M. Oketokoun,African Development Bank (AfDB); Jules A. Frip-piat, UNDP; Lucio Gueratto, European Commis-sion; Frederick Jaspersen, Inter-American Develop-ment Bank; Richard Jolly, Santosh K. Mehrotra, andAshok K. Nigam, UNICEF; Jens Lorentzen, UNCentre for Human Settlements (UNCHS); Stephen J.McCarthy, European Investment Bank; Steven K.Miller and Tom Strandberg, International LabourOrganisation (ILO); Eustace Nonis and NigelRayner, Asian Development Bank (ADB); and J.Bruce Thompson, European Commission.
Contributors and consultation meeting attendeesfrom private and public sector enterprises, universitiesand research institutes, and nongovernmental organiza-tions include Yuzo Akatsuka, Saitama University;Kazumi Asako, Yokohama National University;Iwan Jaya Azis, University of Indonesia; MichaelBeesley, London Business School; William Cos-grove, Ecoconsult, Inc.; Dan Craun-Selka, NationalTelephone Cooperative Association, U.S.A.; HenryErgas, Harvard University; Francois Georges, Elec-
124
tricité de France; Bard Jackson, National Rural Elec-tric Cooperative Association, U.S.A.; Tim Kelly,International Telecommunication Union; KiwhanKim, Kim & Chang, Republic of Korea; David Kin-nersley, Water Aid (U.K.); Pierre Laconte, Interna-tional Union of Public Transport; D. Lorrain, CentreNational de Recherches Scientifiques, France; RolfLuders, Universidad Católica de Chile; John R.Meyer, Harvard University; Bridger Mitchell,RAND Corporation; Rakesh Mohan, United Na-tions University, the Netherlands; Nobuichi No-moto, International Engineering Consultants Asso-ciation; Iqbal Noor Ali and Patricia Schied, Ali KhanFoundation, U.S.A.; Remy Prud'homme, Universitéde Paris; Cohn Relf, Intermediate Technology De-velopment Group and International Forum forRural Transport and Development, U.K.; Annick Sa-lomon, National Wildlife Federation; Ammar Siam-walla, Thai Development Research Institute; Byung-Nak Song, Seoul National University; Tatsu Sunami,Electric Power Development Company, Japan;Hideyuki Suzuki, All Japan Prefectural and Munici-pal Workers Union; Hisao Takahashi, Japan AirportTerminal Company; Kunio Takase, InternationalDevelopment Center of Japan; Yasushi Tanahashi,Japan Freight Railways Company; Kimimasa Tarn-mizu, Tokio Marine and Fire Insurance Company;Marie-Aimée Tourres, SOFRERAIL; Alex Wood,World Wildlife Association; Gordon Wu, HopewehlHoldings Ltd.; Shuichiro Yamanouchi, East JapanRailway Company; and Susumu Yoda, Central Re-search Institute of Electric Power Industry, Japan.
Chapter 1
This chapter draws on a wide range of both pub-lished and unpublished sources, including WorldBank project and sector documents, as well as aca-demic literature. The value-added data in Table 1.1were derived from official national accounts asmaintained by the World Bank. The discussion ofthe importance of infrastructure in an economydraws from Bennathan and Johnson 1987, Galenson1989, Japan 1984, and U.S. Department of Com-merce 1984.
The section on estimating the productivity of in-frastructure investments makes reference to cross-national studies, including Canning and Fay 1993and Easterly and Rebelo 1993. Box 1.1 was draftedby Marianne Fay. For Box 1.1, the studies showingthat causation between infrastructure provision andeconomic growth runs in both directions are Duffy-Deno and Eberts 1991 and Holtz-Eakin 1988. Stud-ies that found no noticeable impact of infrastructure
on growth once more sophisticated econometricmethods were used include Holtz-Eakin 1992; thosethat found their positive results not to be very muchaffected are Bregman and Marom 1993, Duffy-Denoand Eberts 1991, Mera 1973, and Uchimura and Gao1993. A review of the literature on infrastructure'simpact on costs of production is in Aschauer 1993.The trucking study is Keeler and Ying 1988. Otheruseful studies on the economic impacts of infra-structure include Argimon and others 1993; Fordand Poret 1991; Hulten and Schwab 1991 and 1993;Munnell 1990; and Uribe 1993.
The discussion of the effects of rural infrastruc-ture draws from Ahmed and Hossain 1990 andBinswanger, Khandker, and Rosenzweig 1989. Thediscussion on the value and composition of infra-structure stocks is based on World Bank data and onSummers and Heston 1991. The section on returnsto World Bank projects draws from Galenson 1993,Galenson and Thompson forthcoming, Cam 1987,Kaufmann 1991, and Sanghvi, Vernstrom, andBesant-Jones 1989. The Brookings Institution's re-search study is Kresge and Roberts 1971. Table 1.2was obtained from the World Bank Operations Eval-uation Department database.
The discussion of infrastructure's various eco-nomic impacts is based on Doyen 1993, IMF 1993b,Kessides 1993a, Mody and Wang 1994, Mody andYilmaz 1994, Peters 1990 and 1992, Rebelo 1992,Wheeler and Mody 1992, and World Bank 1992a.Box 1.2 was drafted by Thawat Watanatada. Thesection on infrastructure in Central and Eastern Eu-ropean countries draws from various World Banksector studies, including Bennathan and Thompson1992 and Blackshaw and Thompson 1993.
Box 1.3 was drafted by Marianne Fay usingHicks 1991, Meyers 1986, IMF 1993b, World Bank1993a, and data on public sector deficits from East-erly, Rodriguez, and Schmidt-Hebbel forthcoming.
The discussion of poverty in India draws uponLanjouw and Stern 1993 and National HousingBank of India 1992. Box 1.4 is based on Epstein 1962and 1973 and Lanjouw and Stern 1993. The discus-sion of infrastructure's effects on the urban poordraws from Kranton 1991. Caroline 0. N. Mosercontributed information on Ecuador from researchwork in progress. The civil works programs men-tioned are discussed in Drèze and Sen 1989. The sec-tion on environmental linkages draws from theRuitenbeek background paper, Rabinovitch andLeitmann 1993, USAID 1991, and World Bank 1992c,as well as World Bank sector work on Thailand.
References to the historical development of infra-structure and the private sector's role draw from the
background papers by Eichengreen; Jacobson andTarr; and Kirwan. The concept of contestability iselaborated in Baumol, Panzar, and Willig 1988, andits relevance to infrastructure is further developedin Baumol and Lee 1991.
The section on the achievements in coverage ofinfrastructure is based on the data presented in theappendix and in the World Bank's economic andsocial database. OECD 1993 provides a review ofinfrastructure performance issues in OECD coun-tries.
The section on operational inefficiency drawsfrom Galenson 1989, Gyamfi, Gutierrez, and Yepes1992, Howe and Dixon 1993, World Bank 1991b and1993h, the World Bank power sector database, andYepes 1990.
Guy Le Moigne provided information on irriga-tion efficiency. John Nebiker provided data for thediscussion of procurement issues, and relevant in-puts were also provided by Jean-Jacques Raoul andFrancesco Sarno.
The section on maintenance draws from Gyamfi,Gutierrez, and Yepes 1992, Heggie forthcoming,Mason and Thriscutt 1991, Postel 1993, World Bank1988, and the Basu background paper. The referenceto Cameroon was provided by John Schwartz. TheWorld Bank railway database and power sectordatabase were also used.
Details of financial inefficiency were obtainedfrom Besant-Jones 1990b, Galenson and Thompsonforthcoming, Heggie and Quick 1990, World Bank1993h, and Gyamfi, Gutierrez, and Yepes 1992.
On the unresponsiveness to user demand,sources include Besant-Jones 1993, Singh and others1993, and World Bank Water Demand ResearchTeam 1993. Box 1.5 is based on Bell and others forth-coming, Humplick, Kudat, and Madanat 1993,Madanat and Humplick 1993, and Sethi forthcom-ing; Kavita Sethi wrote an early draft of the box. Box1.6 derives from Lee and Anas 1992 and from Lee,Anas, and Verma 1993. Data on telephone fault ratesand waiting time for connection are from the Inter-national Telecommunication Union 1994.
The section on service to the poor draws on Bha-tia 1992, Cámara and Banister 1993, and World Banksector work on Brazil. The example of transport de-mand assessment in Tanzania was provided bySteven K. Miller.
The section on environmental impacts is basedon many environmental studies and assessmentsproduced inside and outside the World Bank. Addi-tional material included Bartone and Bernstein 1992and Bartone and others 1994. Box 1.7 was drafted byPeter Whitford.
125
The section on new opportunities draws frommany academic studies and other sources both pub-lished and unpublished. References to digitalizationin Brazil are from Hobday 1990. Albert Wright andJohn Courtney provided information on alternativetechnologies for sanitation. Riverson and Carapetis1991 and Cohn Reif provided examples of improve-ments in nonmotorized transport.
Valuable suggestions on the drafting of this chap-ter were provided by, among others, Jean Baneth,William Easterly, Harvey A. Cam, and Gregory Sta-ple. Inputs to the sections on environmental linkswere provided by Carl Bartone and Josef Leitmann,and additional comments on this subject came fromCarter Brandon, Maureen Cropper, Alfred Duda,and Rogier van den Brink. Others who providedvery helpful comments on earlier drafts includeMarc Juhel, Guy Le Moigne, Hervé Plusquellec, andYan Wang.
Chapter 2
The data on cross-sectoral comparisons in this chap-ter are from the World Bank database on adjustmentlending conditionality and implementation (ALCID)for structural and sectoral adjustment loans. Mostexamples and anecdotes are from appraisal reports,completion reports, and other evaluations of pro-jects managed by the Bank over approximately thelast twenty years. The recent Bank study of opera-tions and maintenance in Latin America by Gyamfi,Gutierrez, and Yepes 1992 provided invaluable in-formation on the quality, quantity, and nature ofgovernment involvement in infrastructure, particu-larly for roads, power, and water.
The quantification of the gains from privatizationin Chile's power sector is discussed in Galal andothers forthcoming. Box 2.1 draws on a 1992 inter-nal evaluation of the Bank's experience over the lasttwenty years in the water and sanitation sector. Box2.2 was drafted by Stefan Alber. Table 2.1 was com-piled from detailed data extracted from ALCID. Theexamples on Brazil, Ghana, and Indonesia, includ-ing Box 2.3, are from internal World Bank docu-ments. Ian Heggie suggested the discussion of NewZealand and the roads corporations. The data on thelarge water utility in East Asia are from the WorldBank's own 1992 assessment of its experience in thesector. The example on the gains from changes inroad construction technology in Rwanda is fromMartens 1990. The overall discussion of the sectionon corporatization and performance agreementsdraws on Cissé forthcoming, Galal and others forth-coming, Nellis 1988, and Shirley and Nellis 1991.
126
Trivedi 1990 provided useful details on the develop-ing country experience with performance agree-ments, and Debande 1993 and the Debande andDrumaux background paper supplied detailed in-sights on the European experience with perfor-mance agreements.
Many of the conceptual insights in the discussionof the roles of incentives in the organization of gov-ernments were inspired by Laffont and Tirole 1993and Milgrom and Roberts 1992. Box 2.4 and the dis-cussion on Korea throughout the chapter draw onmaterial in Cissé forthcoming, Shirley and Nellis1991, and Trivedi 1990.
The information on Brazil's highway depart-ments draws on internal Bank documents and addi-tional data provided by Jacques Celhier. The data onMexico's Federal Electricity Commission were pro-vided by its staff. The discussion of managementcontracts benefited from the ongoing work by MaziMinovi, Hafeez Shaikh, Thelma Triche, and specificsuggestions by John Nellis and Louis Thompson.Electricité de France, Philippe Durand, and WorldBank 1993h are the sources for Box 2.5. Box 2.6 onAGETIPs draws on Péan 1993. The examples onsubcontracting are from Galenson and Thompsonforthcoming, Miguel and Condron 1991, and Yepes1992. The data for Figure 2.1 on Togo are from inter-nal Bank documents.
The survey of cost recovery and pricing issueshas benefited from the discussion in BahI and Linn1992, Julius and Alicbusan 1989, and from ongoingwork by Carlos Veles on Brazil and by Zmarak Shah-izi on the road sector.
Many of the examples on the consequences offailing to minimize costs are from Gyamfi, Gutier-rez, and Yepes 1992; from Bhatia and Falkenmark1993 for the Asian, Haitian, and Mauritanian exam-ples; and from Yepes 1992 for the Latin Americanexamples. The willingness-to-pay study for EspiritoSanto comes from internal World Bank documents.Heggie forthcoming provided data on road-usercharges in Tanzania. Newbery and others 1988 is thesource for the data on Tunisia. Box 2.7 is based onwork by John Besant-Jones. Box 2.8 reflects exten-sive comments from World Bank staff in the Trans-portation, Water, and Urban Development depart-ment and from infrastructure staff in the Africadepartment. Useful background information wasfound in Altaf, Jamal, and Whittington 1992, Hau1990, Johansen 1989a and 1989b, and Whittingtonand others 1990.
Finally, the following contain additional materialcomplementary to the chapter. Bouttes and Haag1992 discuss the economics of networks in infra-
structure and explain the importance of infrastruc-hire in the context of European integration; Lefèvre1989 provides a wider discussion applicable toOECD countries, focusing on transport. Caillaudand Quinet 1991 and 1992 propose a useful method-ology to assess the effectiveness of incentives in thedesign of various types of contracts between theFrench government and bus operators. Mougeotand Naegellen 1992 extend some of this discussionto more general public procurement policies.Pestieau and Tulkens 1992 survey the determinantsof public enterprise performance. Seabright 1993provides important insights on public provision ofinfrastructure services in South Asia. Tirole 1992presents a more general theory of the internal orga-nization of government and provides explanationsfor some of the issues raised in this chapter. Usefulmaterial illustrating the benefits of appropriate tech-nology choices can be found in Edmonds and deVeen 1992, Gaude and Miller 1992, Guichaoua 1987,and von Braun, Teklu, and Webb 1992. Informationon labor redundancy was obtained from Svejnarand Terrell 1991.
Overall the chapter benefited from detailed com-ments, suggestions, and inputs from Yao Badjo,John Blaxall, José Carbajo, Jacques Cellier, NicholaCissé, Pierre Guislain, Timothy Hau, John Nellis,Zmarak Shalizi, Sudhir Shetty. Vinaya Swaroop,Louis Thompson, Kazuko Uchimura, Joris Van DerVen, and Carlos Velez, within the World Bank, andfrom Jacques Crémer (Institut d'Economie Indus-trielle, Toulouse), Mathias Dewatripont and RichardSchlirf (Université Libre de Bruxelles), Paul Sea-bright (Cambridge University), and Barrie Stevens(OECD).
Chapter 3
This chapter draws on academic sources, back-ground papers, journal publications, World Bankand International Finance Corporation documents,personal communications and comments, and ex-pert consultations both within and outside theWorld Bank.
Sectoral unbundling in the electric power sectoris discussed in Bernstein 1988 (Chile), Littlechild1992 (U.K.), and Tenenbaum, Lock, and Barker 1992.For railways see Moyer and Thompson 1992 andNilsson 1993; for telecommunications see Bruce,Harrell, and Kovacs 1993.
The unintended consequences of regulationwhen substitute services are available are describedin Viscusi, Vernon, and Harrington 1992 for theUnited States and in the Kwong background paperfor Hong Kong.
The prescription for allowing all new entry andeasing barriers to exit was stated by Baumol, Pan-zar, and Willig 1988. The example of competition incellular telephone provision is from the Interna-tional Finance Corporation background paper. Bau-mol and Lee 1991 noted the desirability of allowinglarge manufacturers to sell their excess generatingcapacity. Triche, Mejia, and Idelovitch 1993 pro-vided the examples of concessions in Buenos Airesand Caracas.
The case for competition for the market is articu-lated most forcefully by Demsetz 1968. Williamson1976 cautioned that the franchisee (winner of thecompetition) has incentives to neglect maintenanceof assets toward the end of the contract period.Kuhn, Seabright, and Smith 1992 review research oncompetition.
Gains from privatization are documented byGalal and others forthcoming and Vickers and Yar-row 1988. Informative case studies and reviews ofexperience with privatization and competition arein Alexander and Corti 1993, Baumol and Sidak1994, Fukui 1992, Im, Jalali, and Saghir 1993, Rama-murti and Vernon 1991, Roland and Verdier 1993,and Takano 1992. Links between reform, privatiza-tion, and investment are described in Besant-Jones1990a, Churchill 1993, and Helm and Thompson1991.
The discussion of interconnection financingdraws on the background note by Mitchell, on Bau-mol and Sidak 1994, and on personal communica-tion from Henry Ergas and Dan Craun-Selka.
Much literature exists on the different instru-ments of price and profit regulation. Recent sum-maries of the underlying theory and experience canbe found in Braeutigam and Panzar 1993, Liston1993, and the background paper by Sappington.Willig and Baumol 1987 discuss how competitioncan be used as a guide for regulation. The theory ofyardstick competition is discussed by Shleifer 1985,the Chilean power example is from Covarrubiasand Maia 1993, the Chilean telecom example is fromGalal 1994, and the French example is from Lorrain1992. Reviews of experience with regulation andregulatory reform are in Bennathan, Escobar, andPanagakos 1989, Carbajo 1993, Churchill 1992, Cor-dukes 1990, Guasch and Spiller 1993, and Vogel1986.
For methods of involving consumers in regula-tion in industrialized countries see Triche 1993 and,in a developing country context, Paul 1993. On self-regulation by the industry, see Gwilliam 1993 for thecase of urban transport. Regulation of quality is dis-cussed in Rovizzi and Thompson 1992.
127
Box 3.1 is based on Moyer and Thompson 1992and the Stewart-Smith background paper. Box 3.2and Box table 3.2 are based on Viscusi, Vernon, andHarrington 1992 and Winston 1993. Box 3.3 is byAshoka Mody. The source for Box 3.4 is Triche 1990.Box 3.5 is based on personal communication withAlain Locussol. Box 3.6 was drafted by Robert Tay-lor. Material for Box 3.7 was gathered from the In-ternational Finance Corporation background paper.The source for Box 3.8 is the Naidu and Lee back-ground paper. The material for Box 3.9 was gath-ered from Hill and Abdala 1993, and that for Box3.10 is from Levy and Spiller 1993. The source forBox 3.11 is Paul 1993.
In addition, many individuals contributed valu-able comments to this draft, including, among oth-ers, Veronique Bishop, Robert Bruce, Michael Fin-horn, Ray Hartman, David Haug, Hugh Landzke,Subodh Mathur, Barbara Opper, David Sappington,Mark Schankerman, Richard Scurfield, Mark Segal,Claude Sorel, Martin Stewart-Smith, and ThelmaTriche.
Chapter 4
This chapter draws heavily on numerous internalWorld Bank reports. Useful discussions and com-ments were received from many people both withinand outside the World Bank, including Carter Bran-don, Michael Cernea, David Coady, Maureen Crop-per, Lionel Demery, Jean Drèze, Stephen Howes,William Jack, Valerie Kozel, Jean Lanjouw, HervéPlusquellec, David Steers, Lyn Squire, NicholasStern, Elaine Sun, and Vinaya Swaroop.
In addition, the Canadian International Develop-ment Administration, the International Forum forRural Transport and Development, the InternationalLabour Organisation, the Netherlands Ministry ofOverseas Cooperation, UNICEF, and Water Aid(U.K.) provided useful written material and perti-nent advice.
The section on decentralization draws on a data-base compiled by Frannie Humplick and discussedin Humplick 1992. Data on the evolution of decen-tralization are based on IMF statistics, and the dis-cussion draws on background papers by Bird;Crémer, Estache, and Seabright; and Estache andSinha. The section also benefited from recent infor-mation on decentralization from the European Eco-nomic Commission, provided to the team by HorstReichenbach. Other sources for the section areWorld Bank internal documents and Briscoe 1992,Campbell 1991 and 1992, Dillinger 1993, Narayan
128
forthcoming, and Silverman 1992. Comments fromTim Campbell, Rui Coutinho, Bob Ebel, Jim Hicks,Maureen Lewis, Julio Linares, Remy Prud'homme,David Sewell, Anwar Shah, Sudhir Shetty, AndreaSilverman, Jerry Silverman, Kazuko Uchimura, andYoshine Uchimura, of the World Bank, togetherwith the comments of Richard Bird (University ofToronto), Jacques Crémer (University of Toulouse),and George Zodrow (Rice University) on earlierdrafts, significantly improved the text. Useful re-lated work included Afonso 1989, Castells 1988,Derycke and Gilbert 1988, Kirwan 1989, Kitchen1993, Ostrom, Schroeder, and Wynne 1993,Prud'homme 1992, Rondinelli 1991, and Wunsch1990, 1991a, and 1991b.
The section on participation draws heavily onNarayan forthcoming and on World Bank docu-ments, including Bhatnagar and Williams 1992 anda recent survey by Gerson 1993. Analytical workwas based on a database compiled by Deepa Na-rayan, who also provided comments. In addition,the section benefited from written communicationfrom Allain Ballereau. Messrs. Kroh and Pichke ofthe German development agencies Gesellschaft fürTechnische Zusammenarbeit (GTZ) and Kreditan-stalt für Wiederaufbau (KfW), respectively, pro-vided important background material on theiragencies' experience, as did MUller-Glodde 1991.
Much of the material on budget allocations drewfrom World Bank public expenditure reviews of var-ious countries as well as from other internal docu-ments. The background papers by Asako; Naiduand Lee; Reinfeld; Swaroop; and Uzawa provideduseful material, as did Lacey 1989. Qian and Xu1993 provided evidence on township and rural en-terprises in China. Anand 1983 supplied an analysisof poverty in Malaysia during the 1970s.
Aside from internal documents, the section onsubsidies drew on a study of five Latin Americancountries by Petrei 1987 and on material made avail-able to the team by Gaurav Datt, Richard Jolly andcolleagues at UNICEF, and Carlos Veles.
The section on externalities draws on WorldBank internal documents and on Bakalian and Ja-gannathan 1991, Bernstein 1993, Blackshaw 1992,and Whittington and others 1992. Piotr Wilczynski'sinformation on Poland, Vaandrager's backgroundpaper on the Netherlands' transport sector, and theRuitenbeek background paper on the environmentwere also useful.
The final section on planning draws material notonly from World Bank internal documents, but alsofrom Bartone and Rodriguez 1993, Besant-Jones
1993, Drèze and Stern 1987, Goldstein 1993, Jack1993, Little and Mirrlees 1990, the Meier and Mu-nasinghe background paper, Redwood 1993, Rui-tenbeek and Cartier 1993, Squire 1990, and theRuitenbeek background paper.
Box 4.1 is based on material provided by AndreaSilverman. Box 4.2 was provided by Vijay Jagan-nathan and Albert Wright. Box 4.3 was partiallydrafted by John Riverson (on Ethiopia) and drawson material from Aitken, Cromwell, and Wishart1991 (on Nepal). Box 4.4 is based on the Naidu andLee background paper. The Ruitenbeek backgroundpaper is the source for Box 4.5. Box 4.6 draws onBryceson and Howe 1993, Pankaj 1991, and vonBraun 1988. Ian Heggie, John Roome, and JoelMaweni provided material for Box 4.7. Box 4.8draws on internal reports of the World Bank's Oper-ations and Evaluations Department, the OperationsPolicy Department, and a review of the Bank's pro-ject portfolio. Box 4.9 is taken from the Meier andMunasinghe background paper. Finally, Box 4.10 isbased on internal World Bank reports.
Chapter 5
This chapter draws on academic sources; back-ground papers; journal publications; documentsfrom the IFC, the IMF, the OECD, the U.S. govern-ment, and the World Bank; personal communica-tions and comments; and expert consultations bothwithin and outside the World Bank.
The discussion of the theory that governmentsmight be able to raise financing more cheaply thanprivate investorsbut that these gains also need tobe balanced against greater efficiency of provisionunder private ownershipis from Kay 1993. Lane1992 is the source for the fact that governments facea rising cost of finance and also potential liquidityproblems if excessive debt is accumulated. The dis-cussion of tying of aid is based on OECD 1992 andother documents of the Development AssistanceCommittee of the OECD, as well as on commentsfrom Fabio Ballerin.
Projections of infrastructure investments in Asiaare from CS First Boston 1993. The IFC backgroundpaper is the source for estimates of IFC's infrastruc-ture lending. General descriptions of trends in pri-vate international capital flows, and especially theshift in foreign direct investments toward serviceprovision, are described in World Bank 1993i andIMF 1993a.
General principles of project financing may befound in Nevitt 1989. Discussions of case studies of
risk sharing in project finance are in IFC 1993 andPyle 1994. Material on new projects is taken fromvarious issues of the trade journals Public Works Fi-nancing and Latin Finance. Information on privatetransport projects is based on Gómez-Ibáñez andMeyer 1993. Coverage of country risk, and espe-cially the role of export credit agencies in insuringagainst such risks, is described in the Zhu back-ground paper.
Banks for municipal infrastructure in developingcountries are described in Davey 1988 as well as inpersonal communications from Sergio Contrerasand Myrna Alexander. The case study on FEC inMorocco is from Linares 1993. The financing of con-tractors is discussed in Kirmani 1988. Description ofthe new infrastructure funds was provided throughpersonal communications by Per Ljung (Pakistan)and Krishna Challa (Jamaica).
The links between privatization and capital-market development are described in a backgroundnote prepared by Joyita Mukherjee. Municipal bondmarkets are discussed in U.S. Municipal SecuritiesRulemaking Board 1993, Shilling 1992, and U.S. Se-curities and Exchange Commission 1993. Mesa-Lago 1991 and Vittas and Skully 1991 describe theevolution of contractual savings institutions in de-veloping countries.
The source for Box 5.1 and Box 5.8 is AshokaMody. The material for Box 5.2 and for Box 5.4 isfrom the International Finance Corporation back-ground paper. Box 5.3 is from the Eichengreen back-ground paper. The sources for Box 5.5 are Miceli1991 and Williams 1993. Oks 1993 is the resource forBox 5.6. The material for Box 5.7 is from USAID1993. The source for Box 5.9 is Garzon 1992. Figure5.6 was compiled by Ashoka Mody.
Valuable contributions to this chapter came frommany sources, including Myrna Alexander, MarkAugenblick, Anand Chandavarkar, Stijn Claessens,Ash Demirguc-Kunt, David Haug, John Giraudo,George Kappaz, Sunita Kikeri, Timothy Lane, Ken-neth Lay, Julio Linares, Laurie Mahon, SubodhMathur, Barbara Opper, Robert Palacios, ThomasPyle, William Reinhardt, Jean-Francois Rischard,Han Sankaran, Anita Schwarz, Mark Segal, ClaudeSorel, James Stein, Martin Stewart-Smith, JaneWalker, Al Watkins, and Ning Zhu.
Chapter 6
This chapter draws upon the analysis presented inearlier chapters and the bibliographic referencesused therein. Additional references are noted here.
129
Table 6.1 was derived from antecedents provided inCoyaud 1988 and Kessides 1993b. Box 6.1 on theconditions for good performance of each institu-tional option and the related discussion draw onDia 1993, Lorrain 1992, and Martinand 1993, andcomments by Cohn Relf. Helpful comments andsuggestions on this discussion were provided by,among others, Abhay Deshpande and ThelmaTriche.
For the section on sectoral priorities, sources in-cluded Bartone 1991a and 1991b, Bartone and others1994, Cointreau-Levine 1994, World Bank 1992c,World Bank 1993g. and Wellenius and others 1992.People who contributed to the specific sector agen-das include Carl Bartone and Joe Leitmann (solidwaste); John Briscoe (water supply); AnthonyChurchill (power); Eric Daffern (gas); John Flora,Jeffrey Gutman, Kenneth Gwilliam, Ian Heggie,Zmarak Shalizi, Antti Talvitie, and Louis Thompson(transport); Nikola Holcer, Timothy Nulty, PeterSmith, and Gregory Staples (telecommunications);and Guy Le Moigne and David Steeds (irrigation);
The estimates of gains from increasing efficiencyand correcting mispricing are from the Ingram andFay background paper, except for those for thepower sector. Energy inefficiency, transmission, anddistribution losses for the power sector were basedon estimates from World Bank 1993c, as were the es-timated gains from correcting mispricing in the sec-tor. Additional material was provided by DennisAnderson and Edwin Moore.
Background papers
Asako, Kazumi. "Infrastructure Investment in Japan."Basu, Ritu. "Background Note: Rates of Return for Construc-
tion and Maintenance Projects."Basu, Ritu, and Lant Pritchett. "Background Note: Channels
of Effective Participation."Bird, Richard M. "Decentralizing Infrastructure: For Good or
for Ill?"Chandavarkar, Anand. "Infrastructure Finance: Issues, Insti-
tutions and Policies."Crémer, Jacques, Antonio Estache, and Paul Seabright.
"Lessons of the Theory of the Firm for the Decentraliza-tion of Public Services."
Darbéra, Richard. "Bus Public Transport Franchising inFrench Urban Areas: Efficiency Implications."
de Lucia, Russell J. "Background Note: Poverty and Infra-structure Linkages, Issues and Questions."
Debande, Olivier, and Drumaux, Anne. "Infrastructure Regu-lation Policies in Europe."
Eichengreen, Barry "Financing Infrastructure in DevelopingCountries: An Historical Perspective from the 19th Cen-tury"
Mukherjee, Joyita. "Background Note: Privatization and Cap-ital Market Development."
Naidu, G., and Cassey Lee. "Infrastructure in the EconomicDevelopment of Malaysia."
Peskin, Henry M., and Douglas Barnes. "Background Note:What Is the Value of Electricity Access for Poor UrbanConsumers?"
Reinfeld, William. "Infrastructure and Its Relation to Eco-nomic Development: The Cases of Korea and Taiwan,China."
Ruitenbeek, H. Jack. "Infrastructure and the Environment:Lessons and Directions."
Sappington, David E. M. "Principles of Regulatory Policy De-sign."
Schlirf, Richard. "Background Note: Introduction to the Euro-pean Community Financing Policy for Infrastructure."
Stewart-Smith, Martin. "Industry Structure and Regulation."Swaroop, Vinaya. "The Public Finance of Infrastructure: Is-
sues and Options."Uzawa, Hirofumi. "The Environment and Infrastructure."Vaandrager, René. "A Transport Structure Plan."Wade, Robert. "Public Bureaucracy and the Incentive Prob-
lem: Organizational Determinants of a 'High-Quality CivilService,' India and Korea."
Yuan, Lee Tsao. "The Development of Economic Infrastruc-ture: The Singapore Experience."
Zhu, Ning. "Managing Country Risk: The Role of ExportCredit Agencies."
Selected bibliography
Afonso, José Roberto. 1989. Des pesas Federals corn TransferenciasIntergovernamentais: Urna Revisão de Conceitos, Estatisticas eDiagnostico. Rio de Janeiro: Instituto de PlanejamentoEconômico e Social, Instituto de Pesquisas.
Ahmed, Raisuddin, and Mahabub Hossain. 1990. Developmen-tal Irn pact of Rural Infrastructure in Bangladesh. Research Re-port 83. Washington, D.C.: International Food Policy Re-search Institute.
Aitken, J., C. Cromwell, and C. Wishart. 1991. "Mini- andMicro-Hydropower in Nepal." International Centre for In-tegrated Mountain Development, Kathmandu.
Alexander, Myrna, and Carlos Corti. 1993. "Argentina's Priva-tization Program." CFS Discussion Paper 103. World Bank,Cofinancing and Financial Advisory Services Department,Washington, D.C.
Altaf, Mir Anjum, Haroon Jamal, and Dale Whittington. 1992."Willingness to Pay for Water in Rural Punjab, Pakistan."Water and Sanitation Report 4. World Bank, Transport,Water, and Urban Development Department, UNDP-World Bank Water and Sanitation Program, Washington,D.C.
Anand, Sudhir. 1983. Inequality and Poverty in Malaysia: Mea-surement and Decomposition. New York: Oxford UniversityPress.
Argimon, Isabel, José Manuel Gonzhlez-Páramo, Maria JestisMartin, and José M. Roldán. 1993. "El Papel de las In-fraestructuras en la Producción Privada." BoletIn Econó-mico (Banco de Espana) June.
Aschauer, David Alan. 1989. "Is Public Expenditure Produc-tive?" Journal of Monetary Economics 23: 177-200.
1993. "Public Infrastructure Investment: A Bridge toProductivity Growth?" Public Policy Brief 4. Bard College,Jerome Levy Economic Institute, Annandale-on-Hudson,N.Y.
Baffes, John, and Anwar Shah. 1993. "Productivity of PublicSpending, Sectoral Allocation Choices, and EconomicGrowth." Policy Research Working Paper 1178. WorldBank, Policy Research Department, Washington, D.C.
Bahi, Roy W., and Johannes F Linn. 1992. Urban Public Financein Developing Countries. New York: Oxford UniversityPress.
Bakalian, Alex, and N. Vijay Jagannathan. 1991. "InstitutionalAspects of the Condominial Sewer System." InfrastructureNotes SW-6. World Bank, Infrastructure and Urban Devel-opment Department, Washington, D.C.
Bartone, Carl R. 1991a. "Institutional and Management Ap-proaches to Solid Waste Disposal in Large MetropolitanAreas." Waste Management and Research 9: 525-36.
1991b. "Private Sector Participation in MunicipalSolid Waste Service: Experiences in Latin America." WasteManagement and Research 9: 459-509.
Bartone, Carl R., and Janis D. Bernstein. 1992. "ImprovingMunicipal Solid Waste Management in Third World Coun-tries." Resources, Conservation, and Recycling 8: 43-54.
Bartone, Carl R., Janis Bernstein, Josef Leitmann, and JochenEigen. 1994. "Toward Environmental Strategies for Cities:Policy Considerations for Urban Environmental Manage-ment in Developing Countries." Discussion Paper 18.UNDP/UNCHS/World Bank Urban Management Pro-gram, Washington, D.C.
Bartone, Carl R., and Emilio Rodriguez. 1993. "WatershedProtection in the São Paulo Metropolitan Region: A CaseStudy of an Issue-Specific Urban Environmental Manage-ment Strategy." Infrastructure Notes UE-9. World Bank,Transport, Water, and Urban Development Department,Washington, D.C.
Baumol, William J., and Kyu Sik Lee. 1991. "Contestable Mar-kets, Trade, and Development." World Bank Research Ob-server 6 (1): 1-17.
Baumol, William J., John C. Panzar, and Robert D. Willig.1988. Contestable Markets and the Theory of Industry Struc-ture. San Diego: Harcourt Brace Jovanovich.
Baumol, William J., and J. Gregory Sidak. 1994. Toward Compe-tition in Local Telephony. Cambridge, Mass.: MIT Press.
Bell, Michael, John Boland, Frannie Humplick, Ayse Kudat,Samer Madanat, and Natasha Mukherjee. Forthcoming."Reliability of Urban Water Supply in Developing Coun-tries: The Emperor Has No Clothes." World Bank ResearchObserver.
Bennathan, Esra, and Mark Johnson. 1987. "Transport in theInput-Output System." INU Report 2. World Bank, Infra-structure and Urban Development Department, Washing-ton, D.C.
Bennathan, Esra, and Louis S. Thompson. 1992. PrivatizationProblems at Industry Level: Road Haulage in Central Europe.World Bank Discussion Paper 182. Washington, D.C.
Bennathan, Esra, Luis Escobar, and George Panagakos. 1989.Deregulation of Shipping: What Is to Be Learned from Chile.World Bank Discussion Paper 67. Washington, D.C.
Berg, Elliot. 1993. Privatization in Sub-Saharan Africa: Results,Prospects, and New Approaches. Bethesda, Md.: Develop-ment Alternatives.
Bernstein, Janis D. 1993. "Alternative Approaches to PollutionControl and Waste Management." Discussion Paper 3.UNDP/UNCHS/World Bank Urban Management Pro-gram, Washington, D.C.
Bernstein, Sebastian. 1988. "Competition, Marginal Cost Tar-iffs, and Spot Pricing in the Chilean Electric Power Sector."Energy Policy 16 (August): 369-77.
Besant-Jones, John E., ed. 1990a. "Private Sector Participationin Power through BOOT Schemes." Working Paper 33.World Bank, Industry and Energy Department, Washing-ton, D.C.
1990b. "Review of Electricity Tariffs in DevelopingCountries During the 1980s." Energy Series Paper 32.World Bank, Industry and Energy Department, Washing-ton, D.C.
ed. 1993. "Reforming the Policies for Electric Powerin Developing Countries." World Bank, Industry and En-ergy Department, Washington, D.C.
Bhatia, Bela. 1992. Lush Fields and Parched Throats: The PoliticalEconomy of Groundwater in Gujarat. Working Paper 100.Helsinki: United Nations University, World Institute forDevelopment Economic Research.
Bhatia, Ramesh, and Maim Falkenmark. 1993. "Water Re-source Policies and the Urban Poor: Innovative Ap-proaches and Policy Imperatives." World Bank, Transport,Water, and Urban Development Department, UNDP-World Bank Water and Sanitation Program, Washington,D.C.
Bhatnagar, Bhuvan, and Aubrey C. Williams, eds. 1992. Partic-ipatory Development and the World Bank. World Bank Dis-cussion Paper 183. Washington, D.C.
Binswanger, Hans P., Shahidur R. Khandker, and Mark R.Rosenzweig. 1989. "How Infrastructure and Financial In-stitutions Affect Agricultural Output and Investment inIndia." World Bank Working Paper 163. Washington, D.C.
Blackshaw, Philip W. 1992. "Road Transport and the Environ-ment." Paper presented to the 23rd IRU World Congress,Barcelona, Spain, April 29-May 2.
Blackshaw, Philip W., and Louis S. Thompson. 1993. "RailwayReform in the Central and East European (CEE) Econo-mies." Policy Research Working Paper 1137. World Bank,Transport, Water, and Urban Development Department,Washington, D.C.
131
Bouttes, Jean-Paul, and Denis Haag. 1992. "Economie desRéseaux d'Infrastructure." In N. Curien, ed., Economic etManagement des Enterprises de Réseau. Paris: Economica.
Braeutigam, Ronald R., and John C. Panzar. 1993. "Effects ofthe Change from Rate-of-Return to Price-Cap Regulation."American Economic Review 83 (2): 191-98.
Bregman, A., and A. Marom. 1993. "Growth Factors in Israel'sBusiness Sector, 1958-1988."Bank of Israel, Tel Aviv.
Briscoe, John. 1992. "Poverty and Water Supply: How toMove Forward." Finance and Development 29 (4): 16-19.
Bruce, Robert, Michael Harrell, and Zsuzsa Kovacs. 1993."Who Will Win the Battle for Hungary's Telecoms Com-pany?" International Financial Law Review 7 (5): 25-27.
Bryceson, D., and Howe, J. 1993. "Women and Labor-BasedRoad Works in Sub-Saharan Africa". THE Working PaperIP-4. International Institute for Infrastructural, Hydraulicand Environmental Engineering, Amsterdam.
Caillaud, B., and B. Quinet. 1991. "Les Relations Contract-uelles Etat-SNCF: Une Analyse sous l'Angle de la Théoriedes Incitations." Centre d'Etudes Prospectives d'EconomieMathematique Appliquees a la Planification, Paris.
1992. "Analyse du Caractère Incitatif des Contrats deTransport Urbain." Programme de Recherche et deDéveloppement Technologique dans les Transports Ter-restres, Paris.
Cámara, Paulo, and David Banister. 1993. "Spatial Inequalitiesin the Provision of Public Transport in Latin AmericanCities." Transport Reviews 13 (4): 351-73.
Campbell, Tim. 1991. "Decentralization to Local Governmentin LAC: National Strategies and Local Response in Plan-ning, Spending and Management." Report 5. World Bank,Latin America and the Caribbean Technical Department,Washington, D.C.
1992. "Modes of Accountability in Local Governmentsof LAC." World Bank, Latin America and the CaribbeanTechnical Department, Washington, D.C.
Canning, David, and Marianne Fay. 1993. "The Effect ofTransportation Networks on Economic Growth." Colum-bia University Working Paper. New York.
Carbajo, Jose, ed. 1993. Regulatory Reform in Transport: SomeRecent Experiences. Washington, D.C.: World Bank.
Castells, Antonio. 1988. Hacienda Autono'mica: Una Perspectivade Federalismo Fiscal. Barcelona: Arid Economia.
Cernea, Michael M., ed. 1991. Putting People First: SociologicalVariables in Rural Development. 2d ed. New York: OxfordUniversity Press.
Churchill, Anthony. 1972. Road User Charges in Central Amer-ica. World Bank Staff Occasional Paper 15. Baltimore, Md.:Johns Hopkins University Press.
1992. "Private Power: The Regulatory Implications."Paper presented at the ASEAN Energy Conference, Singa-pore, June 4-5, 1992.
1993. "Private Power Generation: Investment andPricing Problems." Paper presented at the Fourth AnnualJakarta International Energy Conference, October 12.
Cissé, Nichola. Forthcoming. "The Impact of PerformanceContracts on Public Enterprise Performance." Paper pre-sented at the World Bank Working Conference on theChanging Role of the State: Strategies for Reforming Pub-lic Enterprises, Washington, D.C.
Cointreau-Levine, Sandra. 1994. "Private Sector Participationin Municipal Solid Waste Services in Developing Coun-tries." Working Paper 13. UNDP/UNCHS/World BankUrban Management Program, Washington, D.C.
132
Commission of the European Communities. 1993. StableMoney-Sound Finances: Community Public Finance in thePerspective of EMU, vol. 53. Brussels: Directorate-Generalfor European Economic and Financial Affairs.
Cordukes, Peter A. 1990. "A Review of Regulation of thePower Sectors in the Developing Countries." World Bank,Industry and Energy Department, Washington, D.C.
Cotton, A., and R. Franceys. 1993. "Infrastructure for theUrban Poor in Developing Countries." Proceedings of theInstitution of Civil Engineers. Municipal Engineer 98 (Sep-tember): 129-38.
Covarrubias, Alvaro, and Suzanne Maia. 1993. "Reforms andPrivate Participation in the Power Sector of Selected In-dustrialized and Latin American and Caribbean Coun-tries." Latin America Technical Paper Series. World Bank,Latin American and the Caribbean Technical Department,Washington, D.C.
Coyaud, Daniel. 1988. "Private and Public Alternatives forProviding Water Supply and Sewerage Services." INU Re-port 31. World Bank, Infrastructure and Urban Develop-ment Department, Washington, D.C.
CS First Boston. 1993. The Asian Miracle Part II: Reversal of For-tune. Hong Kong.
Davey, Kenneth. 1988. "Municipal Development Funds andIntermediaries." PRE Working Paper 32. World Bank,Washington, D.C.
Deaton, Angus, with Duncan Thomas, Janet Neelin, andNikhilesh Bhattacharya. 1987. "The Demand for PersonalTravel in Developing Countries." INU Discussion Paper 1.World Bank, Infrastructure and Urban Development De-partment, Washington, D.C.
Debande, Olivier. 1993. Formalisation des Contrats de Gestion deIa SNCB et de la RVA/S1\JVA sur Base de Ia Théorie des Incita-tions. Brussels: Ecole de Commerce Solvay, UniversitéLibre de Bruxelles.
Derycke, Pierre-Henri, and Guy Gilbert. 1988. EconomicPublique Locale. Paris: Economica.
Dia, Mamadou. 1993. A Governance Approach to Civil ServiceReform in Sub-Saharan Africa. Washington, D.C.: WorldBank.
Dillinger, William. 1993. "Decentralization and Its Implica-tions for Urban Service Delivery." Discussion Paper 16.UNDP/UNCHS/World Bank Urban Management Pro-gram, Washington, D.C.
Doyen, Jean H. 1993. "Implementation of the Objectives of theSecond Transport Decade: The Primacy of Policy Reformand Local Resource Management." Comments from theUnited Nations Economic Commission for Africa, Confer-ence of African Ministers of Transport, Planning, andCommunications, March 10-12.
Drèze, J. P., and A. P. Sen. 1989. Hunger and Public Action. Ox-ford: Clarendon Press.
Drèze, Jean, and Nicholas Stern. 1987. "The Theory of Cost-Benefit Analysis." In A. J. Auerback and M. Feldstein, eds.,Handbook of Public Economics, 2d ed. Amsterdam: ElsevierScience.
Drumaux, Anne. 1993. Rapport de Recherche Intermediaire: Ob-servatoire des Enterprises Publiques. Brussels: Ecole de Com-merce Solvay, Université Libre de Bruxelles.
Duffy-Deno, Kevin T., and Randall W. Eberts. 1991. "PublicInfrastructure and Regional Economic Development: A Si-multaneous Equations Approach." Journal of Urban Eco-nomics 30: 329-43.
Easterly, William, and Sergio Rebelo. 1993. "Fiscal Policy andEconomic Growth: An Empirical Investigation." Journal ofMonetary Economics 32 (2): 417-58.
Easterly, William, Carlos Rodriguez, and Klaus Schmidt-Hebbel, eds. Forthcoming. Public Sector Deficits and Macro-economic Performance. New York: Oxford University Press.
Edmonds, G. A., and J. J. de Veen. 1992. "A Labour-Based Ap-proach to Roads and Rural Transport in Developing Coun-tries." International Labour Review 131 (1): 95-110.
Epstein, T. Scarlett. 1962. Economic Development and SocialChange in South India. Manchester: Manchester UniversityPress.
1973. South India: Yesterday, Today and Tomorrow. Lon-don: Macmillan.
Ford, Robert, and Pierre Poret. 1991. "Infrastructure and Pri-vate-Sector Productivity." OECD Economic Studies 17:
63-89.Fukui, Koichiro. 1992. Japanese National Railways Privatization
Study. World Bank Discussion Paper 172. Washington,D.C.
Galal, Ahmed. 1994. "Regulation and Commitment in the De-velopment of Telecommunications in Chile." Policy Re-search Working Paper 1278. World Bank, Policy ResearchDepartment, Washington, D.C.
Galal, Ahmed, Leroy P. Jones, Pankaj Tandon, and Ingo Vogel-sang. Forthcoming. Welfare Consequences of Selling PublicEnterprises. New York: Oxford University Press.
Galenson, Alice. 1989. "Labor Redundancy in the TransportSector." INU paper 36. World Bank, Transport, Water, andUrban Development Department, Washington, D.C.
1993. "The Evolution of Bank Lending for Infrastruc-ture." World Bank, Transport, Water, and Urban Develop-ment Department, Washington, D.C.
Galenson, Alice, and Louis Thompson. Forthcoming. "TheBank's Evolving Policy toward Railway Lending." WorldBank, Transport, Water, and Urban Development Depart-ment, Washington, D.C.
Garn, Harvey A. 1987. "Patterns in the Data Reported onCompleted Water Supply Projects." World Bank, Trans-port, Water, and Urban Development Department, Wash-ington, D.C.
Garzon, R. Hernando. 1992. "Municipal Credit Institutions:The Case of Colombia." Working Paper 17. World Bank,Transportation, Water, and Urban Development Depart-ment, Washington, D.C.
Gaude, Jacques, and H. Watzlawick. 1992. "Employment Cre-ation and Poverty Alleviation through Labor-IntensivePublic Works in Least Developed Countries." InternationalLabour Review 131 (1): 3-18.
Gerson, Philip R. 1993. "Popular Participation in EconomicTheory and Practice." Human Resources and OperationsPolicy Working Paper 18. World Bank, Washington, D.C.
Gleick, Peter H., ed. 1993. Water in Crisis: A Guide to the World'sFresh Water Resources. Pacific Institute for Studies in Devel-opment, Environment, and Security. New York: OxfordUniversity Press.
Glewwe, Paul. 1987a. "The Distribution of Welfare in the Re-public of Côte d'Ivoire in 1985." LSMS Working Paper 29.World Bank, Washington, D.C.
198Th. "The Distribution of Welfare in Peru, 1985-86."LSMS Working Paper 42. World Bank, Washington, D.C.
Glewwe, Paul, and Kwaku A. Twum-Baah. 1991. "The Distri-bution of Welfare in Ghana, 1987-88." LSMS WorkingPaper 75. World Bank, Washington, D.C.
Goldstein, Ellen. 1993. The Impact of Rural Infrastructure onRural Poverty: Lessons for South Asia. World Bank Discus-sion Paper 131. Washington, D.C.
Gómez-Ibanez, José, and John R. Meyer. 1993. Going Private:The International Experience with Transport Privatization.Washington, D.C.: Brookings Institution.
Grübler, Arnulf. 1990. The Rise and Fall of Infrastructures: Dy-namics of Evolution and Technological Change in Transport.New York: Springer-Verlag.
Guasch, J. Luis, and Pablo Spiller. 1993. "Utility Regulationand Private Sector Development." World Bank, LatinAmerica and the Caribbean Technical Department Advi-sory Group, Washington, D.C.
Guichaoua, A. 1987. Les Paysans et l'Investissement-Travail auBurundi et au Rwanda. Geneva: International Labour Orga-nisation.
Guislain, Pierre. 1992. Divestiture of State Enterprises. AnOverview of the Legal Framework. Technical Paper 186.Washington, D.C.: World Bank.
Gwilliam, K. M. 1993. "Urban Bus Operators' Associations."Infrastructure Notes. World Bank, Transport, Water, andUrban Development Department, Washington, D.C.
Gyamfi, Peter, Luis Gutierrez, and Guillermo Yepes. 1992. "In-frastructure Maintenance in LAC: The Costs of Neglectand Options for Improvement." 3 vols. Regional StudiesProgram Report 17. World Bank, Latin America and theCaribbean Technical Department, Washington, D.C.
Harral, Clell G., ed. 1992. Transport Development in SouthernChina. World Bank Discussion Paper 151. Washington,D.C.
Hau, Timothy D. 1990. "Electronic Road Pricing: Develop-ments in Hong Kong 1983-1989." Journal of Transport Eco-nomics and Policy 24 (2): 203-14.
Hazell, Peter, and Steven Haggblade. 1993. "Farm-NonfarmGrowth Linkages and the Welfare of the Poor." In MichaelLipton and Jacques van der Gaag, eds., Including the Poor.New York: Oxford University Press.
Heggie, Ian. Forthcoming. "Management and Financing ofRoads: An Agenda for Reform." SSATP Working Paper 8.World Bank and United Nations Economic Commissionfor Africa, Sub-Saharan Africa Transport Policy Program.Washington, D.C.
Heggie, Ian, and Michael Quick. 1990. "A Framework for An-alyzing Financial Performance of the Transport Sector."Working Paper 356. World Bank, Infrastructure and UrbanDevelopment Department, Washington, D.C.
Heidarian, Jamshid, and Gary Wu. 1993. "Power Sector: Sta-tistics of Developing Countries (1987-1991)." World Bank,Industry and Energy Department, Washington, D.C.
Helm, Dieter, and Louis Thompson. 1991. "Privatised Trans-port Infrastructure and Incentives to Invest." Journal ofTransport Economics and Policy 25 (3): 247-58.
Hicks, Norman L. 1991. "Expenditure Reductions in Develop-ing Countries Revisited." Journal of International Develop-ment 3 (1): 29-37.
Hieronymi, 0. 1993. "Decision Making for Infrastructure: En-vironmental and Planning Issues." Paper presented to theOECD Forum for the Future conference on "InfrastructurePolicies for the 1990s," Paris, January 18.
Hill, Alice, and Manuel Angel Abdala. 1993. "Regulation, In-stitutions, and Commitment: Privatization and Regulationin the Argentine Telecommunications Sector." Policy Re-search Working Paper 1216. World Bank, Washington,D.C.
133
Hobday, Michael. 1990. Telecommunications in DevelopingCountries: The Challenge from Brazil. London: Routledge.
Holtz-Eakin, Douglas. 1988. "Private Output, GovernmentCapital, and the Infrastructure Crisis." Discussion Paper394. New York: Columbia University.
1992. "Public-Sector Capital and the Productivity Puz-zle." Working Paper 4122. Cambridge, Mass: National Bu-reau of Economic Research.
Howe, Charles W., and John A. Dixon. 1993. "Inefficiencies inWater Project Design and Operation in the Third World:An Economic Perspective." Water Resources Research 29 (7):1889-94.
Hulten, Charles, and Robert M. Schwab. 1991. Is There Too Lit-tle Public Capital? Washington, D.C.: American EnterpriseInstitute.
1993. Optimal Growth with Public Infrastructure Capital:Implications for Empirical Modeling. College Park, Md.: Uni-versity of Maryland.
Humplick, Frannie. 1992. "Private Ownership, Competition,and Decentralization: Impacts on Infrastructure Perfor-mance." World Bank, Latin America and the CaribbeanDepartment, Washington, D.C.
Humplick, F., A. Kudat, and S. Madanat. 1993. "ModelingHousehold Responses to Water Supply: A Service QualityApproach." Working Paper 4. World Bank, Transport,Water, and Urban Development Department, Washington,D.C.
Hungary, Government of, Central Statistics Office and Min-istry of Finance, and World Bank. 1989. "Incidence Analy-sis: The Impact of Consumer and Housing Subsidies onHousehold Income Distribution." Budapest, Hungary.
Im, Soo J., Robert Jalali, and Jamal Saghir. 1993. "Privatizationin the Republics of the Former Soviet Union." World Bank,Legal Department, Private Sector Development and Priva-tization Group, Washington, D.C.
IMF (International Monetary Fund). 1993a. Private Market Fi-nancing for Developing Countries. World Economic and Fi-nancial Surveys. Washington, D.C.
1993b. Direction of Trade Statistics. Washington, D.C.Various years. Government Finance Statistics. Washing-
ton, D.C.International Road Transport Union (IRTU). Various years.
World Transport Data. Geneva.International Roads Federation (IRF). Various years. World
Road Statistics. Washington, D.C.International Telecommunication Union (ITU). 1994. World
Telecommunications Development Report. Geneva.Israel, Arturo. 1992. Issues for Infrastructure Management in the
1990s. World Bank Discussion Paper 171. Washington, D.C.Jack, William. 1993. "Some Guidelines for the Appraisal of
Large Projects." Discussion Paper 126. World Bank, Officeof the Chief Economist, South Asia Region, Washington,D.C.
Jaiswal, Shailendra N. 1992. "The Role of Transport and Com-munication in Resource Conserving Urban Settlements."Punjab, India.
Japan, Government of. 1984. 1980 Input-Output Tables. Tokyo:Administrative Management Agency.
Jimenez, Emmanuel. Forthcoming. "Human and Physical In-frastructure: Public Investment and Pricing Policies in De-veloping Countries." In J. Behrman and T. N. Srinivasan,eds., Handbook of Development Economics, vol. 3. New York:North Holland.
134
Johansen, Frida. 1989a. "Earmarking, Road Funds and TollRoads." INU Report 45. World Bank, Infrastructure andUrban Development Department, Washington, D.C.
1989b. "Toll Road Characteristics and Toll Road Expe-rience in Selected South East Asia Countries." Trans porta-tion Research 23A (6): 463-66.
Julius, DeAnne S., and Adelaida P. Alicbusan. 1989. "PublicSector Pricing Policies: A Review of Bank Policy and Prac-tice." PRE Working Paper 49. World Bank, Washington,D.C.
Kain, John F 1990. A Critical Assessment of Public Transport In-vestment in Latin America. Washington, D.C.: Inter-Ameri-can Development Bank.
Kaufmann, D. 1991. "The Forgotten Rationale for Policy Re-form: The Productivity of Investment Projects." Back-ground paper for World Development Report 1991. WorldBank, Washington, D.C.
Kay, John. 1993. "Efficiency and Private Capital in the Provi-sion of Infrastructure." Paper presented to OECD Forumfor the Future conference on "Infrastructure Policies forthe 1990s," Paris, January 18.
Keeler, Theodore E., and John S. Ying. 1988. "Measuring theBenefits of a Large Public Investment: The Case of the U.S.Federal-Aid Highway System." Journal of Public Economics6: 69-85.
Kessides, Christine. 1993a. The Contributions of Infrastructure toEconomic Development: A Review of Experience and Policy Im-plications. World Bank Discussion Paper 213. Washington,D.C.
1993b. Institutional Options for the Provision of Infra-structure. World Bank Discussion Paper 212. Washington,D.C.
Kikeri, Sunita, John Nellis, and Mary Shirley 1992. Privatiza-tion: The Lessons of Experience. Washington, D.C.: WorldBank.
Kirmani, Syed 5. 1988. "The Construction Industry in Devel-opment: Issues and Options." INU Report 10. World Bank,Infrastructure and Urban Development Department,Washington, D.C.
Kirwan, R. M. 1989. "Finance for Urban Public Infrastruc-ture." Urban Studies 26: 285-300.
Kitchen, H. 1993. "Efficient Delivery of Local GovernmentServices." Queen's University Government and Competi-tiveness School of Policy Studies, Kingston, Ontario.
Kranton, Rachel E. 1991. "Transport and the Mobility Needsof the Urban Poor." INU Report 86. World Bank, Infra-structure and Urban Development Department, Washing-ton, D.C.
Kresge, David T., and Paul 0. Roberts. 1971. "Systems Analy-sis and Simulation Models." In J. R. Meyer, ed., Techniquesof Transport Planning, vol. 2. Washington, D.C.: BrookingsInstitution.
Kuhn, Kai-Uwe, Paul Seabright, and Alasdair Smith. 1992.Competition Policy Research: Where Do We Stand? CEPR Oc-casional Paper 8. London: Centre for Economic Policy Re-search.
Kurian, G. T. 1991. The New Book of World Ranking. New York:Facts on File.
Lacey, Robert. 1989. "The Management of Public Expendi-tures: An Evolving Bank Approach." Background paperfor World Development Report 1988. World Bank, Washing-ton, D.C.
Laffont, Jean-Jacques, and Jean Tirole. 1993. A Theory of Incen-tives in Procurement and Regulation. Cambridge, Mass.: MITPress.
Lane, Timothy D. 1992. "Market Discipline." IMF Staff Papers40 (1): 53-88. Washington, D.C.
Lanjouw, Peter, and N. H. Stern. 1993. "Agricultural Changeand Inequality in Palanpur, 1957-84." In K. A. Hoff and J.Stiglitz, eds., The Economics of Rural Organization. Oxford:Oxford University Press.
Latin Finance. Various years.Lee, Kyu Sik, and Alex Anas. 1992. "Costs of Deficient Infra-
structure: The Case of Nigerian Manufacturing." UrbanStudies 29 (7): 1071-92.
Lee, Kyu Sik, Alex Anas, and Satyendra Verma. 1993. "Infra-structure Bottlenecks, Private Provision, and IndustrialProductivity: A Study of Indonesian and Thai Cities."World Bank, Transport, Water, and Urban DevelopmentDepartment, Washington, D.C.
Lefèvre, Christian. 1989. La Crise des Transports Publics (France,Etats-Unis, Royaume-Uni, Italic, Pays-Bas). No. 4900. Paris:La Documentation Francaise.
Levy, Brian, and Pablo Spiller. 1993. "Utility Regulation-Get-ting the Fit Right." Outreach 14. World Bank, Policy Re-search Department, Washington, D.C.
Linares, Julio. 1993. "Reforming Municipal Finance: Mo-rocco." Infrastructure Notes FN'I-5. World Bank, Transport,Water, and Urban Development Department, Washington,D.C.
Liston, Catherine. 1993. "Price-Cap versus Rate-of-ReturnRegulation." Journal of Regulatory Economics 5 (1): 25-48.
Little, I. M. D., and J. A. Mirrlees. 1990. "Project Appraisal andPlanning Twenty Years On." In Proceedings of the WorldBank Annual Conference on Development Economics 1989.Washington, D.C.: World Bank.
Littlechild, S. C. 1992. "Competition and Regulation in theBritish Electricity Industry" Utilities Policy 2(4): 270-75.
Lorrain, Dominique. 1992. "The French Model of Urban Ser-vices." West European Politics 15 (2): 77-92.
Madanat, Samer, and Frannie Humplick. 1993. "A Model ofHousehold Choice of Water Supply Systems in Develop-ing Countries." Water Resources Research 29 (5): 1353-58.
Martens, Bertin. 1990. "Etude Comparée de l'EfficacitéEconomique des Techniques a Haute Intensité du Main-d'Oeuvre et a Haute Intensité d'Equipement pour la Con-struction de Routes Secondaires au Rwanda." Geneva: In-ternational Labour Organisation.
Martinand, Claude, ed. 1993. "Private Financing of Public In-frastructure: The French Experience." Ministry of PublicWorks, Transportation, and Tourism, Paris.
Mason, Melody, and Sydney Thriscutt. 1991. "Road Deteriora-tion in Sub-Saharan Africa." In World Bank, The RoadMaintenance Initiative: Building Capacity for Policy Reform.Vol. 2, Readings and Case Studies, EDI Seminar Series.Washington, D.C.
Mera, Koichi. 1973. "Regional Production Functions and So-cial Overhead Capital: An Analysis of the Japanese Case."Regional and Urban Economics 3 (May): 157-85.
Mesa-Lago, Carmelo. 1991. Portfolio Performance of Selected So-cial Institutions in Latin America. World Bank DiscussionPaper 139. Washington, D.C.
Meyers, Kenneth. 1986. "A Reappraisal of the Sectoral Inci-dence of Government Expenditure Cutbacks." CPD Dis-cussion Paper 1986-11. International Finance Corporation,Washington, D.C.
Miceli, Thomas 1. 1991. "Compensation for the Taking of Landunder Eminent Domain." Journal of Institutional and Theo-retical Economics 147 (2): 354-63.
Miguel, Sergio, and James Condron. 1991. "Assessment ofRoad Maintenance by Contract." INU Report 91. WorldBank, Infrastructure and Urban Development Depart-ment, Washington, D.C.
Milgrom, Paul, and John Roberts. 1992. Economics, Organiza-tion and Management. Englewood Cliffs, N.J.: Prentice-Hall.
Mody, Ashoka, and Fang Yi Wang. 1994. "Explaining Indus-trial Growth in Coastal China: Economic Reforms. . . andWhat Else?" World Bank, Private Sector Development De-partment, Washington, D.C.
Mody, Ashoka, and Kamil Yilmaz. 1994. "Is There Persistencein the Growth of Manufactured Exports?" Policy ResearchWorking Paper 1276. World Bank, Washington, D.C.
Moser, Carolyn. 1989. "Community Participation in UrbanProjects in the Third World." Progress in Planning 32:73-134.
Mougeot, Michel, and Florence Naegellen. 1992. MécanismesIncitatifs et Formation des Prix. Paris: Economica.
Moyer, Neil E., and Louis S. Thompson. 1992. "Options forReshaping the Railway." Policy Research Working Paper926. World Bank, Infrastructure and Urban DevelopmentDepartment, Washington, D.C.
Müller-Glodde, Ulrike, ed. 1991. "Where There Is No Partici-pation." Eschborn, Germany: Deutsche Gesellschaft fürTechnische Zusammenarbeit (GTZ).
Munnell, Alicia H. 1990. "Why Has Productivity Declined?Productivity and Public Investment." New England Eco-nomic Review January/February: 3-22.
1992. "Infrastructure Investment and EconomicGrowth." Journal of Economic Perspectives 6 (4): 189-98.
Narayan, Deepa. Forthcoming. "Contribution of People's Par-ticipation: Evidence." ESD Occasional Paper. World Bank,Washington, D.C.
National Housing Bank of India. 1992. Report on Trend andProgress of Housing in India. Central Government ReserveBank of India, New Delhi.
Nellis, John. 1988. "Contract Plans and Public Enterprise Per-formance." Working Paper 118. World Bank, Country Eco-nomics Department, Washington, D.C.
Nevitt, Peter K. 1989. Project Financing. London: EuromoneyNewbery, David, Gordon Hughes, William D. Paterson, and
Esra Bennathan. 1988. Road Transport Taxation in DevelopingCountries: The Design of User Charges and Taxes for Tunisia.World Bank Discussion Paper 26. Washington, D.C.
Nilsson, Jan-Eric. 1993. Regulatory Reform in Swedish Railways:Policy Review with Emphasis on Track Allocation Issues. Stock-holm: Stockholm University, Department of Economics.
OECD (Organization for Economic Cooperation and Devel-opment). 1991. Urban Infrastructure: Finance and Manage-ment. Paris.
1992. DAC Principles for Effective Aid: Development As-sistance Manual. Paris.
1993. Infrastructure Policies for the 1990s. Paris.Oks, Daniel. 1993. "Mexico: Private Sector Participation in In-
frastructure Development." Paper presented at Infrastruc-ture Symposium, World Bank, Washington, D.C., Septem-ber.
Ostrom, Elinor, Larry Schroeder, and Susan Wynne. 1993. In-stitutional Incentives and Sustainable Development: Infrastruc-ture Policies in Perspective. San Francisco: Westview Press.
Pankaj, T. 1991. "Designing Low-Cost Rural Transport Com-ponents to Reach the Poor." Infrastructure Notes RD-3.World Bank, Transport, Water, and Urban DevelopmentDepartment, Washington, D.C.
135
Paul, Samuel. 1991a. "Accountability in Public Services: Exit,Voice and Capture." Policy Research Working Paper 614.World Bank, Washington, D.C.
1991b. Strengthening Public Service Accountability.World Bank Discussion Paper 136. Washington, D.C.
1993. "Bangalore's Public Services: A Report Card."Economic and Political Weekly 28 (52): 2901-09.
Péan, Leslie. 1993. "AGETIP: A New Resource to Meet theUrban Challenge." Infrastructure Notes OU-8 (February).World Bank, Washington, D.C.
Pestieau, Pierre, and Henry Tulkens. 1992. Assessing and Ex-plaining the Performance of Public Enterprises: Some RecentEvidence from the Productive Efficiency Viewpoint. Louvain-la-Neuve, Belgium: Center for Operations Research andEconometrics, Université Catholique de Louvain.
Peters, Hans Jtirgen. 1990. "India's Growing Conflict betweenTrade and Transport: Issues and Options." Policy, Plan-ning, and Research Paper 346. World Bank, Washington,D.C.
1992. "Service: The New Focus in International Manu-facturing and Trade." Policy Research Working Paper 950.World Bank, Infrastructure and Urban Development De-partment, Washington, D.C.
Petrei, A. Humberto. 1987. El Gasto Pdblico Social y sus EfectosDistributivos. Rio de Janeiro: Estudios Conjuntos de Inte-gracSo Econômica da America Latina.
Pickrell, Don H. 1989. "Urban Rail Transit Projects: ForecastVersus Actual Ridership and Costs." U.S. Department ofTransportation, Transport Systems Center, Cambridge,Mass.
Platteau, Jean-Phillipe. 1993. "Sub-Saharan Africa as a SpecialCase: The Crucial Role of Structural Constraints." Univer-sity of Namur, Belgium.
Postel, Sandra. 1993. "Water and Agriculture." In Peter H.Gleick, ed., Water in Crisis: A Guide to the World's FreshWater Resources. Pacific Institute for Studies in Develop-ment, Environment, and Security. New York: Oxford Uni-versity Press.
Prud'homme, Rémy. 1992. "On the Dangers of Decentraliza-tion." World Bank, Infrastructure and Urban DevelopmentDepartment, Washington, D.C.
1993. "Assessing the Role of Infrastructure in Franceby Means of Regionally Estimated Production Functions."Observatoire de l'Economie et des Institutions Locales,Paris.
Public Works Financing. 1993. Various issues.Pyle, Thomas. 1994. Private Financing of Infrastructure: Under-
standing the New Hidden Key to Development Success. Prince-ton, N.J.: Princeton Pacific Group.
Qian, Yingyi, and Chenggang Xu. 1993. Why China's EconomicReforms Differ. Development Economics Research Pro-gramme. London: London School of Economics, Suntory-Toyota International Centre for Economics and RelatedDisciplines.
Rabinovitch, Jonas, and Josef Leitmann. 1993. "EnvironmentalInnovation and Management in Curitiba, Brazil." Discus-sion Paper 1. UNDP/UNCHS/ World Bank Urban Man-agement Program, Washington, D.C.
Ramamurti, Ravi, and Raymond Vernon, eds. 1991. Privatiza-tion and Control of State-Owned Enterprises. EDI Develop-ment Studies. Washington, D.C.: World Bank.
Rebelo, Jorge M. 1992. "Landlocked Countries: Evaluating Al-ternative Routes to the Sea." Infrastructure Notes OT-2.World Bank, Infrastructure and Urban Development De-partment, Washington, D.C.
136
Redwood, John, III. 1993. World Bank Approaches to the Environ-ment in Brazil: A Review of Selected Projects. World Bank Op-erations Evaluation Study. Washington, D.C.
Riverson, John, and Steve Carapetis. 1991. Intermediate Meansof Transport in Sub-Saharan Africa: Its Potential for ImprovingRural Travel and Transport. World Bank Technical Paper 161.Washington, D.C.
Riverson, John, Juan Gaviria, and Sydney Thriscutt. 1991.Rural Roads in Sub-Saharan Africa: Lessons from World BankExperience. World Bank Technical Paper 141. Washington,D.C.
Roland, Gerard, and Thierry Verdier. 1993. Privatisation inEastern Europe: Irreversibility and Critical Mass Effects. Brus-sels: ECARE, Université Libre de Bruxelles.
Rondinelli, D. A. 1991. "Decentralizing Water Supply Servicesin Developing Countries: Factors Affecting the Success ofCommunity Management." Public Administration and De-velopment 11(5): 415-30.
Rovizzi, Laura, and David Thompson. 1992. "The Regulationof Product Quality in the Public Utilities and the Citizen'sCharter." Fiscal Studies 13 (3): 74-95.
Ruitenbeek, H. Jack, and Cynthia M. Cartier. 1993. "A CriticalPerspective on the Evaluation of the Narmada Projectsfrom the Discipline of Ecological Economics." Paper pre-sented to the Narmada Forum: Workshop on the NarmadaSagar and Sardar Sarovar, New Delhi, India, December21-23.
Sader, Frank. 1993. "Privatization and Foreign Investment inthe Developing World, 1988-92." Policy Research WorkingPaper 1202. World Bank, International Economics Depart-ment, Washington, D.C.
Sanghvi, Arun, Robert Vernstrom, and John Besant-Jones.1989. "Review and Evaluation of Historic Electricity Fore-casting Experience (1960-1985)." Energy Series Paper 18.World Bank, Industry and Energy Department, Washing-ton, D.C.
Saunders, Robert J., Jeremy Warford, and Björn Wellenius.Forthcoming. Telecommunications and Economic Develop-ment, 2d ed. Baltimore, Md.: Johns Hopkins UniversityPress.
Seabright, Paul. 1993. "Infrastructure and Industrial Policy inSouth Asia: Achieving the Transition to a New RegulatoryEnvironment." South Asia Regional Seminar Series. WorldBank, Washington, D.C.
Serageldin, Ismail. 1993. "Environmentally Sustainable UrbanTransport: Defining a Global Policy." International Unionof Public Transport, Brussels.
Sethi, Kavita. 1992. "Households' Responses to UnreliableWater Supply in Jamshedpur, India." Draft working paper.World Bank, Transport, Water, and Urban DevelopmentDepartment, Water and Sanitation Division, Washington,D.C.
Shah, Anwar. 1988. "Public Infrastructure and Private SectorProfitability and Productivity in Mexico." Policy, Plan-ning, and Research Working Paper 100. World Bank,Country Economics Department, Washington, D.C.
1992. "Dynamics of Public Infrastructure, IndustrialProductivity and Profitability." Review of Economics andStatistics 74 (1): February.
1994. The Reform of Intergovernmental Fiscal Relations inDeveloping and Emerging Market Economies. Policy and Se-ries Research 23. Washington, D.C.: World Bank.
Shilling, John D., ed. 1992. Beyond Syndicated Loans. WorldBank Technical Paper 163. Cofinancing and Financial Ad-visory Services Department, Washington, D.C.
Shirley, Mary, and John Nellis. 1991. Public Enterprise Reform:The Lessons of Experience. EDT Development Study. Wash-ington, D.C.: World Bank.
Shleifer, Andrei. 1985. "A Theory of Yardstick Competition."Rand Journal of Economics 16 (3): 319-27.
Silverman, Jerry M. 1992. Public Sector Decentralization: Eco-nomic Policy and Sector Investment Programs. World Bank,Technical Paper series, Africa Technical Department Paper188. Washington, D.C.
Singh, Branwar, Radhika Ramasubban, Ramesh Bhatia, JohnBriscoe, Charles Griffin, and Chongchun Kim. 1993."Rural Water Supply in Kerala, India: How to Emergefrom a Low-Level Equilibrium Trap." Water Resources Re-search 29 (7): 1931-42.
Squire, Lyn. 1990. "Comment on 'Project Appraisal and Plan-ning Twenty Years On,' by Little and Mirlees." In Proceed-ings of the World Bank Annual Conference on DevelopmentEconomics 1989. Washington, D.C.: World Bank.
Summers, Robert, and Alan Heston. 1991. "The Penn WorldTable (Mark 5): An Expanded Set of International Compar-isons, 1950-1988." Quarterly Journal of Economics 56 (2)(May).
Svejnar, Jan, and Katherine Terrell. 1991. "Reducing Labor Re-dundancy in State-Owned Enterprises." PRE WorkingPaper 792. World Bank, Infrastructure and Urban Devel-opment Department, Washington, D.C.
Takano, Yoshiro. 1992. Nippon Telegraph and Telephone Privatiza-tion Study: Experiences of Japan and Lessons for DevelopingCountries. World Bank Discussion Paper 179. Washington,D.C.
Tenenbaum, Bernard, Reinier Lock, and James V. Barker. 1992."Electricity Privatization: Structural, Competitive, andRegulatory Options." Energy Policy 20: 1134-60.
Tirole, J. 1992. The Internal Organization of Government. Wash-ington, D.C.: Institute for Policy Reform.
Triche, Thelma. 1990. "Private Participation in the Delivery ofGuinea's Water Supply." Policy Research Working Paper477. World Bank, Transport, Water, and Urban Develop-ment Department, Washington, D.C.
1993. "The Institutional and Regulatory Frameworkfor Water Supply and Sewerage: Public and PrivateRoles." Infrastructure Notes WS-9. World Bank, Transport,Water, and Urban Development Department, Washington,D.C.
Triche, Thelma, Abel Mejia, and Emanuel Idelovitch. 1993."Arranging Concessions for Water Supply and SewerageServices: Lessons Learned from Buenos Aires and Cara-cas." Infrastructure Notes WS-10. World Bank, Transport,Water, and Urban Development Department, Washington,D.C.
Trivedi, Prajapati. 1990. Memorandum of Understanding: An Ap-proach to Improving Public Enterprise Performance. NewDelhi: International Management Publishers.
Uchimura, Kazuko, and Hong Gao. 1993. "The Importance ofInfrastructure on Economic Development." World Bank,Latin America and the Caribbean Regional Office, Wash-ington, D.C.
United Nations. 1991. Energy Statistics Yearbook. New York.United Nations Conference on Environment and Develop-
ment. 1992. Agenda 21. New York.Uribe, José Darió. 1993. "Infrastructura FIsica, Clubs de Con-
USAID (U.S. Agency for International Development). 1991.Cholera in Peru: A Rapid Assessment of the Country's Waterand Sanitation Infrastructure and Its Role in the Epidemic.Field Report 331. Water and Sanitation for Health Project,Washington, D.C.
U.S. Central Intelligence Agency. 1991. World Factbook. Wash-ington, D.C.
U.S. Department of Commerce. 1984. The Detailed In put-Out-put Structure of the U.S. Economy, 1977. Washington, D.C.:Bureau of Economic Analysis.
U.S. Municipal Securities Rulemaking Board. 1993. Hearingon Regulation of the Municipal Securities Market. Sub-committee on Telecommunications and Finance. U.S. Con-gress, Washington, D.C., September 9.
U.S. Securities and Exchange Commission. 1993. Staff Reporton the Municipal Securities Market. Washington, D.C.
Vickers, John, and George Yarrow. 1988. Privatization: An Eco-nomic Analysis. Cambridge, Mass.: MIT Press.
Viscusi, W. Kip, John M. Vernon, and Joseph E. Harrington.1992. Economics of Regulation and Antitrust. Lexington,Mass.: D. C. Heath.
Vittas, Dimitri, and Michael Skully. 1991. "Overview of Con-tractual Savings Institutions." World Bank Working Paper605. Washington, D.C.
Vogel, David. 1986. National Styles of Regulation: EnvironmentalPolicy in Great Britain and the United States. Cornell Studiesin Political Economy. Ithaca, N.Y.: Cornell UniversityPress.
von Braun, J. 1988. "Effects of Technological Change in Agri-culture on Food Consumption and Nutrition: Rice in aWest African Setting." World Development 16 (9): 1083-98.
von Braun, Joachim, Tesfaye Teklu, and Patrick Webb. 1992."Labour-Intensive Public Works for Food Security inAfrica: Past Experience and Future Potential." InternationalLabour Review 131 (1): 19-34.
Wade, Robert. 1987. Village Republics: Economic Conditions forCollective Action in South India. New York: Cambridge Uni-versity Press.
1993. The Operations and Maintenance of Infrastructure:Organizational Issues in Canal Irrigation. Sussex, England:Sussex University, Institute of Development Studies.
Wellenius, Björn, and others. 1992. Telecommunications: WorldBank Experience and Strategy. World Bank Discussion Paper192. Washington, D.C.
Wheeler, David, and Ashoka Mody. 1992. "International In-vestment Location Decisions: The Case of U.S. Firms."Journal of International Economics 33 (August): 57-76.
Whittington, Dale, John Briscoe, Xinming Mu, and WilliamBarron. 1990. "Estimating the Willingness to Pay for WaterServices in Developing Countries: A Case Study of the Useof Contingent Valuation Surveys in Southern Haiti." Eco-nomic Development and Cultural Change 38 (2): 293-312.
Whittington, Dale, Donald T. Lauria, Albert M. Wright,Kyeongae Choe, Jeffrey A. Hughes, and VenkateswarluSwarna. 1992. "Household Demand for Improved Sanita-tion Services: A Case Study of Kumasi, Ghana." Water andSanitation Report 3. UNDP-World Bank Water and Sanita-tion Program, Washington, D.C.
WHO (World Health Organization). 1980 and 1990. The Inter-national Drinking Water Supply and Sanitation Decade series.Geneva.
137
Wiesner Duran, Eduardo. 1982. Memoria del Departamento Na-cional de Planeación, 1978-1980. Bogota, Colombia: Bancode la Repdblica.
Williams, A. W. 1993. "Transport, Rights-of-Way and Com-pensation: Injurious Affection from an Economic Perspec-tive and Some Australian Evidence of Freeway Impacts."International Journal of Transport Economics 20 (3): 285-95.
Williamson, Oliver E. 1976. "Franchise Bidding for NaturalMonopolies-in General and with Respect to CATV." BellJournal of Economics 7(1): 73-104.
Willig, Robert D., and William J. Baumol. 1987. "Using Com-petition as a Guide." Regulation (1): 28-35.
Winston, Clifford. 1993. "Economic Deregulation: Days ofReckoning for Microeconomists." Journal of Economic Liter-ature 31: 1263-89.
World Bank. 1988. Road Deterioration in Developing Countries:Causes and Remedies. World Bank Policy Study. Washing-ton, D.C.
1990. World Development Report 1990. New York: Ox-ford University Press.
1991a. The Reform of Public Sector Management: Lessonsfrom Experience. Policy and Research Series 18. WorldBank, Washington, D.C.
1991b. The Road Maintenance Initiative: Building Capac-ity for Policy Reform. Vol. 2, Reading and Case Studies. EDISeminar Series. Washington, D.C.
1992a. Export Processing Zones. Policy and ResearchSeries 20. Washington, D.C.
1992b. Urban Policy and Economic Development: AnAgenda for the 1990s. Washington, D.C.
1992c. World Development Report 1992. New York: Ox-ford University Press.
1993a. Adjustment Lending and Mobilization of Privateand Public Resources for Growth. Policy and Research Series22. Washington, D.C.
1993b. "The Aral Sea Crisis: Proposed Framework ofActivities." Central Asia Region 3, Washington, D.C.
1993c. Energy Efficiency and Conservation in the Develop-ing World: The World Bank's Role. Washington, D.C.
1993d. "Portfolio Management: Next Steps, a Programof Actions." Operations Policy Department, Washington,D.C.
138
1993e. "Poverty and Income Distribution in LatinAmerica: The Story of the 1980s." Latin America and theCaribbean Technical Department, Washington, D.C.
1993f. "Power Supply in Developing Countries: WillReform Work?" Occasional Paper 1. Industry and EnergyDepartment, Washington, D.C.
1993g. Water Resources Management. World Bank Pol-icy Paper. Washington, D.C.
1993h. The World Bank's Role in the Electric Power Sector:Policies for Effective Institutional, Regulatory, and FinancialReform. World Bank Policy Paper. Washington, D.C.
Forthcoming. "An Agenda for Infrastructure Reformand Development: Responding to the Market for Ser-vices." World Bank Discussion Paper. Transport, Water,and Urban Development Department, Washington, D.C.
World Bank Water Demand Research Team. 1993. "The De-mand for Water in Rural Areas: Determinants and PolicyImplications." World Bank Research Observer 8 (1): 47-70.
World Resources Institute. 1992. World Resources 1992-93.New York: Oxford University Press.
Wunsch, James S., ed. 1990. The Failure of the Centralized State:Institutions and Self-Governance in Africa. Boulder, Cob.:Westview Press.
1991a. "Institutional Analysis and Decentralization:Developing an Analytical Framework for Effective ThirdWorld Administrative Reform." Public Administration andDevelopment 11(5): 431-52.
1991b. "Sustaining Third World Infrastructure Invest-ments: Decentralizing and Alternative Strategies." PublicAdministration and Development 11(1): 5-24.
Yepes, Guillermo. 1990. "Management and Operational Prac-tices of Municipal and Regional Water and SewerageCompanies in Latin America and the Caribbean." WorldBank, Infrastructure and Urban Development Depart-ment, Washington, D.C.
Appendix: Infrastructure data
Table A. I presents summary information on infra-structure stocks as well as electricity production andirrigated land. Table A.2 offers data on access todrinking water and sanitation. The two remainingtables provide data on financial commitments andsupport to infrastructure. Readers should refer tothe "Definitions and data notes" for an explanationof the country groups used in these tables. TablesA.1 and A.2 list economies in the same order as inthe World Development Indicators.
Although the data reported here are drawn fromthe most authoritative sources available, compara-bility may be limited by variations in data collec-tion, statistical methods, and definitions.
Table A.1 Physical measures of infrastructureprovision
Data for paved roads are from Canning and Fay 1993for the years prior to 1990; figures for 1990 werecompiled from the U.S. Central Intelligence Agency1991 (primary source), the International Roads Fed-eration (IRF), various years, or the InternationalRoad Transport Union (IRTU), various years. Where1990 data were not available, figures for 1988 or1989 were used. Quinquennial data are availablefrom 1960 to 1990, but the data are available yearlyfrom the IRF and the IRTU.
Both net installed capacity of electricity-generatingplants and electricity production are from Canningand Fay 1993 for the years prior to 1990. Figures for1990 come from United Nations 1991. Quinquennialdata are available from 1960 to 1990; the data areavailable yearly from the UN source.
A telephone main line is a telephone line that con-nects the subscriber's terminal equipment to thepublic switched network and has a dedicated portin the telephone exchange equipment. This term issynonymous with the term main station, which iscommonly used in telecommunication documents.Data for main lines are from the InternationalTelecommunication Union (ITU) 1994. Quinquen-nial data are available, but the data are availableyearly for the years 1975-92 from the ITU's elec-tronic database.
Information on kilometers of railroad tracks isfrom Canning and Fay 1993 for the years prior to1990. Figures for 1990 are from the World Bank;quinquennial data are available from 1960 to 1990.
Figures for irrigated land were obtained from thedata files of the Food and Agriculture Organization(FAO). The data are available from the FAO for 1961onward.
Table A.2 Access to drinking water and sanitation
Access to drinking water means access to safe waterby either standpost or house connections. Safewater is defined here as treated surface waters oruntreated but uncontaminated waters, such as fromprotected springs, boreholes, and sanitary wells. Ac-cess to sanitation includes access by either sewer con-nection or other means such as septic tanks, com-munal toilets, pit privies, pour-flush latrines, etc.Data are primarily from the World Health Organiza-tion 1980 and 1990, complemented by Gleick 1993and World Resources Institute 1992. (Data fromWHO are provided by governments and are notverified independently) Quinquennial data areavailable from 1970 to 1990 for the total and from1980 to 1990 for rural and urban categories.
Table A.3 World Bank and IDA commitments toinfrastructure
The World Bank's central lending database (ALCID)is the source for annual figures for 1950-93. Infra-structure commitments are included for the fol-lowing sectors: irrigation and drainage; power; tele-communications; water and sanitation; and totaltransport. Total transport includes aviation, high-ways, ports and waterways, railways, and urbantransport, as well as commitments to the overalltransportation sector. Sector adjustment loans (SE-CALs) are included. These data do not includeminor infrastructure components of projects inother sectors, for example, rural development or en-vironmental projects.
Table A.4 Official development finance forinfrastructure
The OECD provided data for the years 1984-92. Thefigures given here are based on total official flows asdefined by the Development Assistance Committeeof the OECD. Total infrastructure includes communi-cations, energy, transport, water supply and sanita-tion, as well as river development and other infra-structure not classified in the previous categories.
139
Table A.1 Physical measures of infrastructure provision
140
Paved roads(kilometers)
Electricity-generating capacity(thousands of kilowatts)
Electricity production(millions of kilowatt-hours)
1-Economies classified by the United Nations or otherwise regarded by their authorities as developing.1990 data refer to 1988; World Resources Institute 1992.World Resources Institute 1992.For range estimates, see map on access to safe water in the intmduction to the WDI.
Access to safe drinking water Access to sanitation
Table A.4 Official development finance commitments(millions of current U.S. dollars)
World Development Indicators
Key 154
Introduction 157
Tables1 Basic indicators 162
Production2 Growth of production 1643 Structure of production 1664 Agriculture and food 1685 Commercial energy 1706 Structure of manufacturing 1727 Manufacturing earnings and output 174
Domestic absorption8 Growth of consumption and investment 1769 Structure of demand 178
Fiscal and monetary accounts10 Central government expenditure 18011 Central government current revenue 182
12 Money and interest rates 184
Core international transactions13 Growth of merchandise trade 186
14 Structure of merchandise imports 18815 Structure of merchandise exports 19016 OECD imports of manufactured goods 19217 Balance of payments and reserves 194
External finance18 Official development assistance from OECD and OPEC members 19619 Official development assistance: receipts 19820 Total external debt 20021 Flow of public and private external capital 20222 Aggregate net resource flows and net transfers 20423 Total external debt ratios 20624 Terms of external public borrowing 208
Human resources development25 Population and labor force 21026 Demography and fertility 21227 Health and nutrition 21428 Education 21629 Gender comparisons 21830 Income distribution and PPP estimates of GNP 220
In each table, economies are listed within theirgroups in ascending order of GNP per capita, exceptthat those for which no GNP per capita can be cal-culated are italicized, in alphabetical order, at theend of their group. The ranking below refers to theorder in the tables.
The key shows the years of the most recent cen-sus and of the latest demographic survey or vitalregistration-based estimates. This information isincluded to show the currentness of the sources ofdemographic indicators, which can be a reflectionof the overall quality of a country's indicators. Be-yond these years, demographic estimates may begenerated by projection models, extrapolation rou-tines, or other methods. Other demographic indica-tors, such as life expectancy, birth and death rates,and under-5 mortality rates, are usually derivedfrom the same sources. Explanations of how WorldBank estimates and projections are derived from
154
the sources, as well as more information on thesources, are given in World Population Projections,1994-95 Edition (forthcoming).
Figures in colored bands in the tables are sum-mary measures for groups of economies.
The letter w means weighted average; m, medianvalue; t, total.
All growth rates are in real terms.Data cutoff date is March 24, 1994.The symbol. . means not available.The figures 0 and 0.0 mean zero or less than half
the unit shown.A blank means not applicable.Figures in italics indicate data that are for years
or periods other than those specified.The symbol ± indicates economies classified by
the United Nations or otherwise regarded by theirauthorities as developing.
Note: Economies with sparse data or with populations of more than 30,000 and fewer than I million are included as part of the country groups in themain tables but are shown in greater detail in Table Ia. For data comparability and coverage throughout the tables, see the technical notes.
Supplemented by more recent official demographic estimates.In all tables data include Eritrea, unless otherwise stated.In all tables data refer to the unified Germany, unless otherwise stated.
IntroductionThis seventeenth edition of the World DevelopmentIndicators provides economic, social, and naturalresource indicators for selected periods or years for207 economies and various analytical and geo-graphic groups of economies. Although most of thedata collected by the World Bank are on low- andmiddle-income economies, comparable data forhigh-income economies are readily available andare also included in the tables. Additional informa-tion may be found in the World Bank Atlas, WorldThbles, World Debt Tables, and Social Indicators ofDevelopment. These data are now also available ondiskette through the World Bank's SocioeconomicTime-series Access and Retrieval System*STARS*.
Changes in this edition
Because of space limitations in the main tables, aneconomy must have reasonable coverage of keysocio-economic indicators to be included. Addi-tional basic indicators for economies with sparsedata (Afghanistan, Angola, Bosnia and Herzegov-ma, Cambodia, Croatia, Cuba, Eritrea, Haiti, Iraq,Dem. Rep. of Korea, Kuwait, Lebanon, Liberia,Libya, Macedonia FYR, Viet Nam, Fed. Rep. of Yu-goslavia, and Zaire) are presented, along with coun-tries with less than I million population, in Table Ia.
Other changes have been made to a number oftables. Although these are described more fully inthe technical notes, an outline of the changes may beof interest.
A new table, Table 32, Infrastructure, has been in-cluded to highlight key indicators of the servicelevel and coverage of infrastructure (see the techni-cal notes).
The table on the structure of consumption hasbeen deleted because updates have not been avail-able for most countries since 1985.
In Table 23, Total external debt ratios, net presentvalue of external debt as a percentage of total ex-
ports of goods and services and of GNP has re-placed total external debt as a percentage of exportsof goods and services and of GNP.
Table 25, Population and labor force, includes popu-lations age 15-64 for 1992 and labor force growthrates for 1970-80, 1980-92 and 1990-2000.
Classification of economies
As in the Report itself, the main criterion used toclassify economies and broadly distinguish differentstages of economic development is GNP per capita.This year the per capita income groups are low-income, $675 or less in 1992 (42 economies); middle-income, $676 to $8,355 (67 economies); and high-income, $8,356 or more (23 economies). Economieswith populations of fewer than 1 million and thosewith sparse data are not shown separately in themain tables but they are included in the aggregates.Basic indicators for these economies may be foundin Table Ia.
Further classification of economies is by geo-graphic location. For a list of economies in eachgroup, see the tables on classification of econo-mies at the back of this book. Aggregates for se-verely indebted middle-income economies arealso presented.
Methodology
The World Bank continually reviews methodologyin an effort to improve the international comparabil-ity and analytical significance of the indicators. Dif-ferences between data in this year's and last year'seditions reflect not only updates for the countriesbut also revisions to historical series and changes inmethodology.
All dollar figures are current U.S. dollars unlessotherwise stated. The various methods used for con-verting from national currency figures are describedin the technical notes.
157
Summary measures
The summary measures in the colored bands oneach table are totals (indicated by t), weighted aver-ages (w), or median values (m) calculated for groupsof economies. Countries for which individual esti-mates are not shown, because of size, nonreporting,or insufficient history, have been implicitly includedby assuming they follow the trend of reportingcountries during such periods. This gives a moreconsistent aggregate measure by standardizingcountry coverage for each period shown. Group ag-gregates include countries for which country-specific data do not appear in the tables. Wheremissing information accounts for a third or more ofthe overall estimate, however, the group measure isreported as not available. The weightings used forcomputing the summary measures are stated ineach technical note.
Terminology and data coverage
In these notes the term "country" does not implypolitical independence but may refer to any terri-tory whose authorities present for it separate socialor economic statistics.
The unified Germany does not yet have a fullymerged statistical system. Throughout the tables,data for Germany are footnoted to explain coverage;most economic data refer to the former Federal Re-public, but demographic and social data generallyrefer to the unified Germany. The data for China donot include Taiwan, China, but footnotes to Tables13, 14, 15, and 17 provide estimates of internationaltransactions for Taiwan, China.
Table content
The indicators in Tables I and Ia give a summaryprofile of economies. Data in the other tables fall
Groups of economiesFor this map, economies are classified byincome group, as they are for the tablesthat follow. Low-income economies are thosewith a GNP per capita of $675 or less in 1992;middle-income, $676-8,355; high-income,$8,356 or more.
Low-income economies
Middle-income economies
High-income economies
Data not available
WesternSamoa
AmessaoSamoa (US)
TOnga
InJdbW
FrhPoS,renia (F,)
4' Halo
JanIca
Go ate m a la.El Salvador
, ) Mcaragua
Coota RI
51100DOrIrarIicao RICO (US)RepublIC
U.SOUtrUn (US)
St Kissand Ness
Nether/a rIds400(55 (NoUn)
Aruba J 0 St Lucialath) 7/ St Vincent C Barbados
ad) thn Srnnadrnen.
a°Geeodoa
The£aharnas
Betmuda(/dl(
Ph,,
yoWa
ChIle Argendna Urugoay
FaraD cay
158
Anogua and Barbuda
&,adelOupe (F,)
Domintca
%Ma,tr,lOue (Ar)
into the following broad areas: production, domes-tic absorption, fiscal and monetary accounts, coreinternational transactions, external finance, humanresources development, and environmentally sus-tainable development. The table format of this edi-tion follows that used in previous years. In eachgroup, economies are listed in ascending order ofGNP per capita, except that those for which no suchfigure can be calculated are italicized and listed inalphabetical order at the end of the group deemedappropriate. This order is used in all tables exceptTable 18, which covers only high-income OPEC andOECD countries. The alphabetical list in the keyshows the reference number for each economy;here, too, italics indicate economies with no currentestimates of GNP per capita. Economies in the high-income group marked by the symbol ± are thoseclassified by the United Nations or otherwise re-garded by their authorities as developing.
Technical notes
The technical notes and the footnotes to tablesshould be referred to in any use of the data. Thenotes outline the methods, concepts, definitions,and data sources used in compiling the tables. Abibliography at the end of the notes lists the datasources, which contain some of the comprehensivedefinitions and descriptions of the concepts used.Country notes to the World Tables provide additionalexplanations of sources used, breaks in comparabil-ity, and other exceptions to standard statistical prac-Ices that World Bank staff have identified in na-tional accounts and international transactions.
Comments and questions relating to the WorldDevelopment Indicators should be addressed to:Socio-Economic Data Division, International Eco-nomics Department, The World Bank, 1818 H Street,N.W., Washington, D.C. 20433.
r,00laird(500)
loelood
Farro,Iola,ots
Netherlands
IsA at Muir (55.
Cnaieo,r Islands (ye)
Luxembourg
Lieu ht005teifl
Andorra
Portugal
dib,attar (U
FarmerSgarnmafl
Sahara
Maureanrago Verde
rGrirblauinea-Bissau Guinea
Sierra LOOflO\\N
LiAenla,
! Bslomanero Lanoa
Lithuania
a Berarus
Ukrame
Bane
BaBaar NeOe
TuBey
GolanlaMB Colpiuo Aaab
Tael Lebanon Meg
Mali NigerChad
BalBoaP550
BarrIoNrgefla
PESO
Equatorial Gurnea
Sec TomB nod Priacipe<
Cerroal AfricanRepublic
SudanEruirru a,ui
Ugaoda Keoya
Rwanda LIZaire Burandi
Ethiopia
Somatic
Zambiaaealami
Mozuorbique
lirrbsboo
Aigharistar
NazasBi
Cofliurou SayAliOlIrr
A ad sg 5 0 A 0 iMauritius
AouolOur (Fat
lagridetar
Pakisrao
Iodua
Aaldiueo
Nepal dhutao
SrrLerka
CArs
Lao People'sBarrgladesA ham. Rep
L'°5RoogKurg (UK)LI"ayarmatLqaoao (Port)
CaornBodPtrrlrppiors
7Brungih_
Siiuou to it
[odor ro 10
N Mariana lola
0'
P..'(US) I
Guam (US)
Federated Staten or Mrorooenia
Psgyo SolomonNow Guinea u
- A,lsloO(e
Varuatu h,1
'NeroCalodaoia
(Fr)
NeroZealand
',JarntFa Garde
Nauru ,,jKeibal(
1''
159
8
Population density
Population persquare kilometer
0-19
I20-49
50-199
200 or more
Data not available
Total fertility
Births per woman
Fertility and mortality
Infant mortality
Deaths per 1,000 live births150
For this map, population density is calculated bydividing a country's population by its total surface area(square kilometers of land and inland water area). SeeTable 1 for the population and area of the 132 economiesin the main tables, and Table Ia for an additional 75economies.
Note: For explanation of terms or methods, see the technical notes for Tables 26, 27, and 29.
Share of agriculture in GDP is calculated by taking thevalue added of an economy's agriculture sector anddividing it by gross domestic product. The shares saynothing about absolute values of production. Foreconomies with high levels of subsistence farming, theshare of agriculture in GDP is difficult to measurebecause of difficulties in assigning subsistence farmingits appropriate value. For more details, see the technicalnote for Table 3.
Population with access to safe water, 1990
Safe water is defined as treated surface waters, and untreated butuncontaminated waters such as from protected springs, boreholes,and sanitary wells. For more details, see the technical note forTable 32.
161
Table 1. Basic indicators
Note: For other economies see Table Ia. For data comparability and coverage, see the Key and the technical notes. Figures in italics are for years other than thosespecified.
162
Population Area GNPper cap itaa Avg. annual Life expect. at Adult tll,teracy (%)
t Economies classified by the United Nations or otherwise regarded by their authorities as developing, a. In all tables GDP and GNP data data cover mainland Tanzaniaonly. b. Estimates for economies of the former Soviet Union are subject to more than usual range of uncertainty and should be regarded as very preliminary. c. Datareflect recent revision of 1992 GNP per capita: from $700 to $670 for Côte d'tvoire, from $2,510 to $2,730 for Chile, and from $2,700 to$2,670 for South Africa. d. Inall tables, data for Jordan cover the East Bank only. e. According to UNESCO, illiteracy is less than 5 percent. f. Data refer to GDP g. Data refer to the FederalRepublic of Germany before unification.
163
Population(millions)
Area
(thousands
GNP per capitaa Avg. annualrate ofinflation (%)
Life expect, atbirth (years)
Adult illiteracy (%)
Dollars Avg. ann. growth Female Totalmid-1992 ofsq. km) 1992 (%), 1980-92 1970-80 1980-92 1992 1990 1990
Low- and middle-income 4,610.1 1 101,6691 1,040w 0.9 w 26.2 w 75.7 w 64w 46w 36 wSub-Salsaran Africa 543.0 I 24,274 I 530 w -0.8 w 13.6 w 15.6 w 52w 62w 50 wEast Asia & Pacific 1,688.8 16,368 I 760 w 6.1 w 16.6 w 6.7 w 68w 34w 24 wSouth Asia 1,177.9 I 5,133 1 310 w 3.0 w 9.7 w 8.5 w 60w 69w 55 wEurope and Central Asia 494.5 I 24,370 1 2,080w 18.7 w 47.5 w 70wMiddle East & N. Africa 252.6 I 11,0151 1,950w -2.3 w 17.0 w 10.1 w 64w 57w 45 wLatin America & Caribbean 453.2 I 20,507 1 2,690 w -0.2 w 46.7 w 229.5 w 68w 18w 15wSeverely indebted 504.6 I 22,483 I 2,470 w -1.0 w 42.1 w 208.0 w 67 w 28w 23w
High-income economies 828.1 I 31,7091 22,160 w 2.3 w 9.1 w 4.3w 77w110 Ireland 3.5 70 12,210 3.4 14.2 5.3 75Ill New Zealand 3.4 271 12,300 0.6 12.5 9.4 76 C
a. Because manufacturing is generally the most dynamic part of the industrial sector, its growth rate is shown separately. b. Services, etc. includes unallocateditems. c. GDP and its components are at purchaser values. d. Data refer to the Federal Republic of Germany before unification.
'165
Average annual growth rate (%)
GDP Agriculture Industry Manufacruringa Services, etc. b
Sub-Saharan Africa 3.6w 1.8 w 1.6w 1.7 w 3.6 w 1.2 w 4.3 w 1.4 w 4.9w 2.3wEast Asia & Pacific 7.7w .. 4.4w .. 9.4w 8.9wSouth Asia 3.5 w 5.2 w 1.8w 3.3 w 4.6w 6.4 w 4.6 w 6.5 w 4.7w 6.2wEurope and Central AsiaMiddle East & N. Africa 2.2w .. 4.7w .. 0.9w 4.5 w 1.4wLatin America & Caribbean 5.4w 1.8 w 3.4w 2.0 w 5.7 w 1.3 w 6.2 w 0.8 w 5.7w 2.1wSeverely indebted 5.8w 1.6w 3.9w 1.8 w 6.5 w 1.2 w 6.3 w 1.1 w 6.1 w 2.2 w
a. Because manufacturing is generally the most dynamic part of the industrial sector, its growth rate is shown separately. b. Services, etc. includes unallocateditems. c. GDP and its components are at purchaser values. d. Data refer to the Federal Republic of Germany before unification.
167
GDP (million$)
Distribution ofgross domestic product (%)
Agriculture Industry Manufacturinga Services, etc.b
Low- and middle-income . . 4,695,645 1Sub-Saharan Africa 57,611 I 269,955 27 w 20 w 28 w w 13w 17w 45 w 46wEast Asia & Pacific . . 1,266,819 1 21 w 38 w 42 wSouth Asia 73,642 1 297,360 44w 32 w 21w 26 w 14w 16w 34 w 42 wEurope and Central Asia . . 1,124,423 1Middle East & N. AfricaLatin America & Caribbean
Low- and middle-income 107,193 I 169,727 8,725 I 13,263 1 558 w 855 w 7.2w 6.7wSub-Saharan Africa 15,4161 54,3351 8,647 1 18,512 1 1,602 1 4,2231 124 w 136 w 6.7w 6.1wEast Asia & Pacific 262,572 I 26,8241 33,291 t 1,525 1 581 1 952 w 2,017 w 12.6 w 10.8 wSouth Asia 32,720 1 94,813 1 4,211 I 7,721 1 2,339 1 2,558 1 328 w 750 w 11.5 w 14.4wEurope and Central Asia 17,172 1 45,153 1 2841 1,639 1 1,322 w 730 w 4.0w 4.1wMiddle East & N. Africa 66,356 1 24,557 I 38,008 1 2,255 1 2,484 I 337 w 654 w 1.9w 1.7wLatin America & Caribbean 20,444 I . . 25,782 I 27,044 1 721 1 1,779 1 495 w 485 w 7.5w 6.7wSeverely indebted 22,2941 37,7981 36,073 1 6951 2,460 t 630 w 426 w 4.6 w 4.8 w
High-income economies 87,444 I 79,798 I 75,933 1 1,293 w 1,160 w 8.4 w 8.6 w
World . . . . 186,991 I 245,660 I 8,742 1 13,263 1 791 w 933 w 7.5 w 7.2 w
a. Value added in agriculture data are at purchaser values. b. Includes Luxembourg. c. Data refer to the Federal Republic of Germany before unification.
169
Table 5. Commercial energy
Note: For data comparability and coverage, see the Key and the technical notes. Figures in italics are for years other than those specified.
170
Average annual growth rate (%) Energy use (oil equivalent)Energy imports
as % ofmerchandise exportsEnergy production Energy consumption Per capita (kg) GDP output per kg ($)
Low-income economies 6.7w 4.8w 6.8w 5.4w 171 w 338w . . 1.1 w 7w 9 wExcluding China & India 5.6 w 3.7 w 6.3 w 5.2 w 81 w 151 w 1.9 w 2.5 w 7 w II wI Mozambique 22.9 -24.7 -1.7 -4.6 103 32 . . 2.0 .
a. Figures for the South African Customs union comprising South Africa, Namibia, Lesotho, Botswana, and Swaziland are included in South African data; trade amongthe component territories is excluded.
171
Average annual growth rate (%) Energy use (oil equivalent)Energy imports
% ofmerchandise exportaEnergy production Energy consumption Per capita (kg) GDP output per kg ($)
Low- and middle-income 3.9 w 6.3 w 6.4 w 7.8 w 321 w 790 w 1.0 w 1.3 w 11 w 11 wSub-Saharan Africa 4.5w 3.8 w 4.1 w 2.9 w 225 w 258 w 1.0 w 1.9 w 5 w 7 wEast Asia & Pacific 7.6 w 5.1 w 7.2 w 5.6 w 271 w 593 w . . 1.3 w 9 w 10 wSouth Asia 5.3w 7.1 w 4.9w 6.8w 100w 209w 1.2 w 1.3w 11 w 23 wEurope and Central Asia 3,179w .. 0.7wMiddle East & N. Africa 2.7w 2.1 w 11.5w 5.7w 411 w 1,109w 1.2w 1.6w 3w 5 wLatin America & Caribbean 2.0w 2.9w 5.7w 2.7w 641 w 923 w 1.1 w 3.1 w 14w lOw
Severely indebted 7.1 w 2.2 w 6.7 w 1.6 w 735 w 976 w 1.2 w 2.7 w 13 w 12 w
High-income economies 1.7w 1.8w 2.0 w 1.5 w 4,407 w 5,101 w 0.8 w 4.4w 12 w 10 w
a. Includes unallocated data; see the technical notes. b. Value added in manufacturing data are at purchaser values. c. Data refer to the Federal Republic of Germanybefore unification.
Low- and middle-income 1,090,6641Sub-Saharan Africa 7,288 I 45,273East Asia & PacificSouth Asia 10,362 50,665 1Europe and Central Asia 422,913 1Middle East & N. Africa 48,566Latin America & Caribbean 41,600t 264,349 1Severely indebted 41,629 z 285,146 1
a. General government consumption figures are not available separately; they are included in pnvate consumption, etc. b. Data refer to the Federal Republic ofGermany before unification.
Low- and middle-incomeSub-Saharan Africa w 1.8w 4.1w 1.3w 5.1w -3.0wEast Asia & PacificSouth Asia 7.4w 3.0w 4.9w 4.6w 5111wEurope and Central AsiaMiddle East & N. AfricaLatin America & Caribbean 6_.i 3.5w 6.2w 2.0w 6.8w -0.5wSeverely indebted 7.7 w 3.4 w 6.5 w 2.0 w 7.9 w -1.3 w
High-income economies 2.7 a' 2.3 w 3.5w 3.0w 2.1w 3.5w
a. General government consumption figures are not available separately; they are included in pnvate consumption, etc. b. Data refer to the Federal Republic ofGermany before unification.
Low- and middle-incomeSub-Saharan Africa 12w 17w 71w 69w 20w 16w 18w 15w 21w 26w -2w -1wEast Asia & PacificSouth Asia 9w 12w 76w 69w 16w 22w 14w 19w 5w 12w -2w -3wEurope and Central AsiaMiddle East & N. AfricaLatin America & Caribbean lOw 69w 22w 20w 13w -2wSeverely indebted 10 w 67 w 23 w 21w 12w
Low-income economies 177,233 1 183,6851 3.3 w 6.9 w 6.0 w 2.7 w 106 m 90mExcluding China & India 72,498 t 80,570 t 1.9 w 3.9 w 5.3 w -0.8 w 106 m 88 m1 Mozambique . . . . . .
a. Data are for the South African Customs Union comprising South Africa, Namibia, Lesotho, Botswana, and Swaziland; trade among the component temtories isexcluded. b. Excludes inter-republic trade. c. Includes Luxembourg. d. Data refer to the Federal Republic of Germany before unification.
187
tradeMerchandise trade (million Average annual growth rate (%) Ter,m of(1987=101))Exports
108 Portugal 18,541 30,482 1.2 11.6 1.0 10.4 85 104109 Saudi Arabia 41,833 32,103 5.7 -2.4 35.9 -6.2 176 83
Low- and middle-income 763,299 1 829,967 1 3.9 w 4.4 w 6.1 w 2.3 w 108 m 95 mSub-Saharan Africa 63,233 60,2191 2.8 w 2.4 w 3.0 w -2.7 w 107 m 88 mEast Asia & Pacific 282,425 1 289,984 1 9.5 w 10.5 w 7.8 w 8.8 w 96m 103 mSouth Asia 31,948 1 38,974 3.6 w 6.8 w 2.7 w 2.1 w 97 m 91 mEurope and Central Asia 141,3441 179,275 1 92 m 101 mMiddle East & N. Africa 116,7441 112,1851 3.9 w w 156 w -2.9 w 129 m 93 mLatin America & Caribbean 127,605 149,330 1 -0.1 w 2.9 w 3.6 w 0.6 w 114 m 95 mSeverely indebted 134,8871 143,669 e 9.5 w 2.8 w 5.9 w -0.3 w 118 m 92 m
Food Fuels commodities transport equip. manufactures
a. Data are for the South African Customs Union comprising South Africa, Namibia, Lesotho, Botswana, and Swaziland; trade among the component temtortes isexcluded. b. Excludes the Canal Zone. c. Includes Luxembourg. d. Data refer to the Federal Republic of Germany before unification.
Low- and middle-income 13w 10 w 9w 10 w 9w 7w 33 w 37 w 35 w 36 wSub-Saharan Africa 11w 12w 4w 8w 4w 4w 38 w 39 w 42 w 37 wEast Asia & Pacific 13w 6w 7w lOw 10 w 9w 33 w 39 w 37 w 36 wSouth Asia 25 w 10 w 7w 19w 13w II w 24w 22 w 31 w 39 wEurope and Central AsiaMiddle East & N. Africa 19 w 16w w 5w 7w 5w w 35 w 39 w wLatin America & Caribbean 11w 11w 11w 10w 7w 5w 35 w 40w 36 w 35 w
Severely indebted 14 w 12 w 10 w 10 w 8 w 5 w 34w 39w 34w 34w
High-income economies 16 w 10 w lOw 9w 16w 6w 25w 35w 33w 41w
Low- and middle-income 37 w 29 w 35 w 18 w 8 w 15 w 19 w 37 w 6 w 14 wSub-Saharan Africa 37 w 44 w 46 w 32 w 2 w 3 w 15 w 21 w 1 w 2 wEast Asia & Pacific 22 w 11 w 45 w 15 w 6 w 25 w 27 w 50 w 13 w 20 wSouth Asia 9 w 6 w 44 w 21 w 3 w 5 w 45 w 69 w 28 w 41 wEurope and Central Asia . . . . . . . . . . . . . . .Middle East & N. Africa 74 w 85 w 18 w 5 w 1 w 1 w 7 w 9 w 3 w 4 wLatin America & Caribbean 43 w 32 w 45 w 30 w 2 w 14 w 9 w 24 w 1 w 3 W
Severely indebted 22 w 34 w 47 w 27 w 13 w 14 w 16 w 25 w 3 w 4 w
a. Data are for the South Afncan Customs Union comprising South Africa, Namibia, Lesotho, Botswana, and Swaziland; trade among the component territories isexcluded. b. Excludes the Canal Zone. c. Includes Luxembourg. d. Data refer to the Federal Republic of Germany before unification.
Note: Data cover high-income OECD countries' imports only. For 1970, these are based on SITC, revision I and revision 2 for 1992. a. Data are for the South AfricanCustoms Union comprising South Africa, Narnibia, Lesotho, Botswana, and Swaziland; trade among the component territories is excluded. b. Excludes the CanalZone. c. Includes Luxembourg. d. Data refer to the Federal Republic of Germany before unification.
193
Value of imports ofmanuf Composition of1992 imports of manufactures (%)by origin (million
Low- and middle-income 5,365 295,078 1 29.1 w 5.7 w 15.1 w 5.4 w 44.7 wSub-Saharan Africa 515 1 7,752 1 16.6 w 9.8 w 1.7 w 3.8 w 68.2 wEast Asia & Pacific 1,0861 149,227 I 27.6 w 3.2 w 18.7 w 2.2 w 48.3 wSouth Asia 760 l 17,994 1 63.2 w 3.6 w 0.8 w 1.1 w 31.3 wEurope and Central Asia 1,406 55,946 35.4 w 9.0 w 8.6 w 7.0 w 39.9 wMiddle East & N. Africa 304 1 11,271 1 44.9 w 13.7 w 6.3 w 5.6 w 29.5 wLatin America & Caribbean 1,2941 52,888 1 13.6 w 7.5 w 20.2 w 14.6 w 44.0 w
Severely indebted 1,421 1 57,100 1 12.8 w 7.5 w 19.3 w 14.4 w 46.0 w
High-income economies 120,1901 1,652,662 1 6.0 w 13.0w 11.2w 19.1 w 50.7 w
Total 0.20 0.13 0.11 0.08 0.08 0.08 0.09 0.08 0.09
a. Organization of Economic Cooperation and Development. b. Data mfer to the Federal Republic of Germany before unification. e. See the technicalnotes. d. Organization of Petroleum Exporting Countries. e. Organization of Arab Petroleum Exporting Countries.
Table 19. Official development assistance: receipts
198
Note. For data comparability and coverage, see the Key and the technical notes. Figures in italics are for years other than those specified.
Net disbursement ofODA from all sources
Millions ofdollars Per capita 1$)
1991
As percentage
ofGNP19911985 1986 1987 1988 1989 1990 1991
Low-income economies 17,065 t 19,038 t 20,9881 24,0041 24,530 r 30,441 1 31,711 1 10.2 w 2.7 wExcluding China & India 14,533 t 15,785 t 17,688 1 19,918 1 20,482 1 26,836 1 27,010 t 25.1 w 7.0 w
Low- and middle-income 26,122 1 28,508 1 31,475 I 33,684 I 34,592 45,898 1 47,246 1 11.7 w 1.4 wSub-Saharan Africa 9,521 I 10,587 I 11,9261 13,470 I 13,848 16,539 1 16,158 1 32.9 w 9.3 wEast Asia & Pacific 4,376 1 4,307 1 5,382 1 6,266 I 6,908 1 7,778 1 7,388 1 4.6 w 0.6 wSouth Asia 4,244 I 5,474 I 5,307 6,236 1 6,101 6,030 1 7,488 6.5 w 2.1 wEurope and Central Asia 247 I 403 I 458 1 359 I 207 1 1,3071 1,8961 24.2 w 1.0wMiddle East & N. Africa 4,710 1 4,474 1 4,700 1 3,6701 3,5171 9,7471 9,3001 38.0 w 2.2 wLatin Anserica & Caribbean 3,024 I 3,262 1 3,701 1 3,682 1 4,0101 4,498 1 5,017 1 11.4w 0.4 w
Severely indebted 3,754 I 4,050 1 4,361 1 3,8091 3,8241 6,3941 6,917 1 15.5 w 0.6 w
High-income economies
110 IrelandIll New Zealand112 jIsmel 1,978 1,937 1,251 1,241 1,192 1,372 1,749 352.5113 Spain114 Hong Kong 20 18 19 22 40 38 36 d.ó
108 Portugal 92.6 102.1 43.9 39.0 18.3 18.3 10.5 6.9 4.4 3.6 5.5 10.1109 Saudi Arabia . . .
Low- and middle-income 166.6 w 154.4 w 33.1 w 33.5 w 20.6 w 18.5 w 10.5 w 7.4 w 17.0 w 19.0 w 8.6 w 15.6 wSub-Saharan Africa 277.9 w 282.0 w 82.4 w 88.2 w 11.5 w 20.0 w 6.0 w 8.9 w 26.3 w 35.6 w 13.0 w 23.7 wEast Asia & Pacific 84.5 w 85.8 w 20.4 w 23.6 w 13.4 w 13.0 w 7.6 w 5.0 w 16.4 w 20.8 w 8.7 w 14.3 wSouth Asia 191.2 w 209.2 w 20.6 w 28.2 w 11.9 w 23.0 w 5.1 w 10.6 w 74.4w 52.0w 25.0 w 37.5 wEurope and Central Asia 124.5 w 133.2 w 23.4 w 31.3 w 15.9w 14.0 w 6.4 w 5.7 w 10.1 w 7.1 w 6.2 w 8.0wMiddle East & N. Africa 198.0 w 145.5 w 46.7 w 40.3 w 16.5 w 22.2 w 7.4w 7.5 w 31.8 w 31.6w 8.3 w 11.9 wLatin America & Caribbean 267.1 w 250.4 w 45.5 w 38.1 w 37.1 w 29.5 w 19.6 w 12.3 w 4.4 w 5.8 w 5.8 w 13.3 w
Severely indebted 272.9 w 266.8 w 46.6 w 41.1 w 34.0 w 29.8 w 17.1 w 11.5 w 6.9 w 10.3 w 5.1 w 10.7 w
High-income economies
Table 24. Terms of external public borrowing
208
Note: For data comparability and coverage, see the Key and the technical notes. Figures in italics are for years other than those specified.
Commitments
(million$)Average interest rate
(%)
Average maturity
(years)
Average grace period
(years)
Public loans withvariable tnt. rates as
% ofpublic debt
1980 1992 1980 1992 1980 1992 1980 1992 1980 1992
Low-income economies 30,083 t 41,8041 6.3 w 5.0 w 23 w 21 w 6 w 6 w 16.6 w 20.8 wExcluding China & India 21,408 t 18,721 t 5.8 w 3.9 w 23 w 26 w 6 w 7 w 17.0 w 19.1 w
Middle-incomeeconomies 67,2881 80,2711 10.6w 6.7w 12w 13w 4w Sw 54.8w 51.6wLower-middle-income 33,155 I 48,419 I 9.5 w 6.3 w 14 w 13 w 4 w 5 W 42.6 W 48.1 w
104 Slovenia105 Puerto Rico106 Korea, Rep. 4,928 4,027 11.3 7.1 i i 36.4 40.3107 Greece108 Portugal 2,015 8,257 10.9 6.6 10 Ii 30.6 26.2109 Saudi Arabia
Low- and middle-income 97,3711 122,075t 9.3 w 6.1 w 16w 16w 5w 5w 45.0w 41.1 wSub-Saharan Africa 13,245 t 8,102 1 7.1 w 2.9 w 17 w 30 w 5 w 8 w 26.4 w 16.5 wEast Asia & Pacific 19,445 35,536 t 9.8 w 6.1 w 16 w 16 w 5 w 6 w 40.0 w 39.5 wSouth Asia 7,872 t 11,449 1 4.6 w 4.3 w 33 w 26 w 7 w 9 w 3.1 w 16.7 wEurope and Central Asia 12,4351 27,561 1 10.9 w 7.0 w 12 w 9 w 4 w 3 w 45.8 w 52.2 H'
Middle East & N. Africa 11,5941 17,4401 6.3 w 5.8 w 18 w 11 w 5 w 4 w 18.2 w 27.9 wLatin America & Caribbean 32,780 1 21,987 1 11.6 w 7.2 w 11 w 15 w 4 w 5 w 68.0 w 57.2 w
Severely indebted 36,825 t 29,088 I 10.6 w 6.7 w 12 w 13 w 4 w 4 w 59.3 w 55.5,
High-income economies
Table 25. Population and labor force
210
Note: For data comparability and coverage, see the Key and the technical notes. Figures in italics are for years other than those specified.
Population a Labor force a
Total (millions) Hypottttl Average annual growth (%) Age 15-64(millions)
a. For the assumptions used in the projections, see the technical notes. b. Based on census data or a demographic estimate 5 years or older; tuning is only one element ofdata quality. See the Key for the latest census year.
Low-and middle-income 4,6101 5,2481 7,201 I 10,5761 2.5w 1.9w 1.6w 2,8071 1,9081 2.3w 2.2w 1.9wSub-Saharan Africa 5431 681 1 1,229 I 2,5651 2.8 w 3.0 w 2.8 w 2871 222 I 2.4 w 2.5 w 2.7 wEast Asia & Pacific 1,6891 1,858 1 2,280 1 2,792 i 1.9 w 1.6 w 1.2 w 1,101 1 9281 2.4 w 2.1 w 1.8 wSouth Asia 1,1781 1,3691 1,913 1 2,7781 2.4 w 2.2 w 1.9 w 682 1 4291 1.8 w 2.1 w 1.9 wEurope and Central Asia 495 1 5161 581 1 672 1 4.3 w 1.0 w 0.5 w 3261 941 1.4 w 1.1 w 0.2 wMiddle East & N. Africa 2531 309 1 509 I 8561 2.8 w 3.1 w 2.5 w 1351 691 3.0 w 3.2 w 3.2 wLatin America & Caribbean 4531 515 1 6901 913 1 2.4 w 2.0 w 1.6 w 276 1 1661 3.1 w 2.5 w 2.3 w
Severely indebted 5051 579 1 815 1 1,191 1 2.3 w 2.0 w 1.7 w 3021 187 1 2.7 w 2.3 w 2.2 w
High-income economies 8281 865 1 922 1 903 1 0.8 w 0.7 w 0.5 w 555 1 3801 1.3 w 0.6 w 0.4 w
a. For assumptions used in the projections, see the technicat notes to Table 25. b. NRR is the net reproduction rate, see the technical notes. c. Data include womenwhose husbands practice contraception; see the technical notes. d. Based on a demographic estimate 5 years or older timing is only one element of data quality. See theKey for the latest year.
213
Crude birih Crude deathMarried women of
rate (per 1000 rate (per 1,000 Percentage ofbirths in Projecled childbearing age
population) population) Total fertility rate /992 to women agedyear of
Low- and middle-income 38 w 27 w 13w 9w 5.6 w 3.3 w 3.0 wSub-Saharan Africa 47 w 44w 20 w 15 w 6.5 w 6.1 w 5.6 wEast Asia & Pacific 35 w 21 w 9w 8w 5.7 w 2.3 w 2.2 wSouth Asia 42 w 31 w 18w lOw 6.0 w 4.0 w 3.3 wEurope and Central Asia 22 w 16w lOw lOw 2.5 w 2.2 2.1 wMiddle East & N. Africa 45 w 34 w 16w 8w 6.8 w 4.9 w 4.2 wLatin America & Caribbean 36 w 26 w 10 w 7w 5.2 w 3.0 w 2.5 w
Severely indebted 36 w 27 w II w 8 w 5.2 w 3.3 w 2.9 w
High-income economies 18w 13 w lOw 9w 2.4w 1.7w 1.8w
a. Based on a demographic estimate 5 years or older; timing is only one element of data quality. See the Key for the latest year. b. Data refer to the Federal Republic ofGermany before unification.
Low- and middle-income 7,630 w 4,810 w 4,700 w 65 w 99 w 88 wSub-Saharan Africa 31,720 w 19,690 w 3,160 w 11i 99w 160 w 179 wEast Asia & Pacific 5,090 w . . 2,720 w 84w 39 w 43 w 55 wSouth Asia 6,120 w 2,930 w 10,150 w 138 w 85 w 111 w 122 wEurope and Central Asia . 410 w . 30 w 34 w 41 wMiddle East & N. Africa 6,410 w 2,240 w 1,940 w 670 w 139 w 58 w 72 w 84wLatin America & Caribbean 2,020 w . . 2,640 w 85 w 44w 52 w 61 w
Severely indebted 3,460 w 2,250 w 2,340 w 86w 52w 65w 76w
Low-income economies 74 w 101 w 93 w 21 w 41 w 35 w 3 w 36 w 38 wExcluding China & India 55 w 79 w 44 w 71 w 13 w 28 w 8 w 25 w 3 w 5 w 74 w 39 w 38 w1 Mozambique 47 63 . 53 5 8 5 0 41 69 552 Ethiopia 16 25 10 21 4 12 2 11 0 1 48 303 Tanzania 34 69 27 68 3 5 2 4 0 0 47 47 364 Sierra Leone 34 48 27 39 8 16 5 12 1 1 32 345 Nepal 26 8 10 3 3 7 22 39
41 Yemen,Rep. 22 76 7 37 3 31 0 51 3742 Zambia 90 92 80 13 8 2 2 47Middle-income economies 93 w 104 w 87 w 99 w 32 w 55 w 26 w 56 w 13 w 18 w 90 w 34 w 25 w
Low- and middle-income 79w 102w 63w 94w 24w 45w 17w 39w 6w 8w 92w 35w 35wSub-Saharan Africa 50w 66w 41w 58w 7w 18w 5w 16w 1w 2w .. 42w 41wEast Asia & Pacific 88w 119w .. 115w 24w SOw .. 47w 4w Sw 100w 30w 24wSouth Asia 67w 89w 50w 76w 25w 39w 14w 29w .. .. 42w 57wEurope and Central AsiaMiddle East & N. Africa 68w 98w SOw 89w 24w 56w 15w 51w lOw 15w 89w 35w 27wLatin America & Caribbean 95 w 106 w 94 w 105 w 28 w 47 w 26 w 54 w IS w 18 w 87 w 34 w 26 w
Low-income economies 54 w 63 w 53 w 61 w 78 w 65 W 36 w 35 wExcluding China & India 47 w 57 w 46 w 55 w 65 w 66 w 74 w 69 w 61 w 77 w 44 w 66 w 32 w 31 wI Mozambique 42 45 36 43 70 61 50 472 Ethiopia 44 50 43 47 57 56 56 56 46 64 32 67 40 373 Tanzania 47 52 44 49 342 82 90 88 89 65 98 38 72 51 474 Sierra Leone 36 45 33 41 67 70 40 56 36 325 Nepal 42 53 43 54 833 18 47 16 35 33
Low- and middle-income 56 w 66 w 54 w 62 w 61w 64w .. 69w 81w 59w 74w 35w 35wSub-Saharan Africa 46 w 53 w 43 w 50 w 66w 69w .. 63w 77w 44w 67w 40w 37wEast Asia & Pacific 60 w 69 w 58 w 66 w 88w .. 76w 41w 42wSouth Asia 48w 61 w 50 w 60w 45 w . . 48 w . . 55 w 69 w 38 w 54 w 26 w 22 H'
Europe and Central Asia 69 w 74 w 64w 66wMiddle East & N. Africa 54 w 66 w 52 w 63 w 83w 90w 87w 92w 54w 79w 41w 72w 10w 16wLatin America & Caribbean 63 w 71w 58 w 65 w 66 w . . 60 w . . 96 w 97 w 101 w 114 w 22 w 27 w
a. Obtained from the regression estimates. b. Data refer to expenditure shares by fractiles of persons. c. Data ranked by household expenditures. d. Extrapolatedfrom 1985 ICP estimates. e. Data ranked by per capita expenditure. f. Data refer to income shares by fractiles of persons. g. Data ranked by per capitaincome. h. These values are subject to more than the usual margin of error (see technical notes). i, j. Data are extrapolated, respectively, from 1980 and 1975 ICPestimates and scaled up by the corresponding US deflator. k. Extrapolated from 1990 ICP estimates. I. Data refer to expenditure shares by fractiles of house-holds. m. Data refer to income shares by fractiles of households. n. Data ranked by household income. o. Data refer to the Federal Republic of Germany beforeunification.
221
PPP estimates of GNP
Percentage share of income or consumption per capitaCurrent intl
Lowest Second Third Fourth Highest Highest United States = 100 dollarsYear 20percent quintile quintile quintile 20 percent lOpercent 1987 1992 1992
108 Portugal 36.0 43.8 10120k109 Saudi Arabia 445a 48.3a 111700
Low- and middle-incomeSub-Saharan AfricaEast Asia & PacificSouth AsiaEurope and Central AsiaMiddle East & N. AfricaLatin America & CaribbeanSeverely indebted
Low- and middle-income 25 w 36 w 3.7 w 3.7 w 16 w 6w 41 w 37 w 11w 14 HSub-Saharan Africa 19 w 29 w 5.1 w 5.0 w 33 w 9w 34w 34w 7w 10 nEast Asia & Pacific 19 w 29 w 3.3 w 4.2 w 12w 4w 46w 37 w 9w 11 HSouth Asia 19 w 25 w 4.1 w 3.5 w Sw 2w 35 w 38 w 7w 9isEurope and Central AsiaMiddle East & N. Africa 41 w w 4.ã 'l.a w 14 w
34w 28 w42 w 41 w
15 w18 w
18 ii22 ii
Latin America & Caribbean 57 w 73 w 3.7 w 2.9 w 24 w 16 w 45 w 46w 26 w 34 5
Severely indebted 53 w 68 w 3.7 w 3.0 w 21 w 14 w 42w 43w 23w 29s
a. For range estimates, see map on access to safe water in the introduction. b. 85 percent or mom of roads are in good condition; see the technical notes. c. Data referto the Federal Republic of Germany before unification.
225
Telecommunications Paved roadsPower Water Railways
Households Systemwith losses
electricity (% of total(% of total) output)
1984 1990
Telephonemainlines(ver 1,000persons)
1990
Faults(per 100
mainlinesper year)
1990
Road density Roads in(km per good cond.million (% ofpa ved
persons) roads)1988 1988
Populationwith access to (%
safe water(% of total) provision)
1990
Lossesof total
water
1986
Rail traffic Diesels(km per in usemillion (% of diesel
a. Negative values represent an increase in forest area. b. Water withdrawal data refer to any year from 1970 to 1992. c. Data for Eritrea, not yet disaggregated, areincluded in Ethiopia. d. Total water resources include river flows from other countries in addition to internal renewable resources. e. See the technical notes foralternative estimates. f. Except for water withdrawal estimates, data foriordan cover the East Bank only. g. Includes other wooded land. h. Closed forest only.
227
Natural forest area
Nationally protected areas,
Freshwater resources: annual withdrawal, 1970-92 b
a. See the technical note for Table 1. b. Estimated as low-income ($675 or less). c. Data for Eritrea, not yet disaggregated, are included in Ethiopia. d. Estimatedas lower-middle-income ($676-$2,695). e. Estimated as upper-middle-income ($2,696-$8,355). f. Estimated as high-income ($8,356 or more).
68 Brunei 273 6.00 f 7469 Channellslands 144 0.19 f 7770 Faeroe Islands 48 0.40 f71 French Polynesia 207 4.00 f 6872 Greenland 58 342.00 f . . . .
73 Kuwait 1,410 18.00 f . . 21.9 75 33 2774 SanMarino 23 0.06 f .
75 Virgin Islands (U.S.) 99 0.34 f 6.9 75
Technical notes
The main criterion for country classification is grossnational product (GNP) per capita. With the addi-tion of the recently independent republics of the for-mer Soviet Union, the main tables now includecountry data for 132 economies, listed in ascendingGNP per capita order. A separate table (Table la)shows basic indicators for 75 more economies withsparse data or with populations of fewer than I mil-lion. Other changes are outlined in the Introduction.
Data reliability
Considerable effort has been made to standardizethe data, but full comparability cannot be ensured,and care must be taken in interpreting the indica-tors. Many factors affect availability and reliability;statistical systems in many developing economiesare still weak; statistical methods, coverage, prac-tices, and definitions differ widely among countries;and cross-country and cross-time comparisons in-volve complex technical problems that cannot beunequivocally resolved. For these reasons, althoughthe data are drawn from the sources thought to bemost authoritative, they should be construed onlyas indicating trends and characterizing major differ-ences among economies rather than offering precisequantitative measures of those differences. In partic-ular, data issues have yet to be resolved for the fif-teen economies of the former Soviet Union. Cover-age is sparse, and the data are subject to more thanthe normal range of uncertainty.
Most social and demographic data from nationalsources are drawn from regular administrative files,although some come from special surveys or peri-odic census inquiries. In the case of survey and cen-sus data, figures for intermediate years have to beinterpolated or otherwise estimated from the basereference statistics. Similarly, because not all dataare updated, some figuresespecially those relat-ing to current periodsmay be extrapolated. Sev-eral estimates (for example, life expectancy) are de-rived from models based on assumptions aboutrecent trends and prevailing conditions. Issues re-lated to the reliability of demographic indicators arereviewed in United Nations, World Population Trendsand Policies. Readers are urged to take these limita-tions into account in interpreting the indicators,particularly when making comparisons acrosseconomies.
Base years
To provide long-term trend analysis, facilitate inter-national comparisons, and include the effects ofchanges in intersectoral relative prices, constantprice data for most economies are partially rebasedto three base years and linked together. The year1970 is the base year for data from 1960 to 1975, 1980for 1976 to 1982, and 1987 for 1983 and beyond.These three periods are "chain-linked" to obtain1987 prices throughout all three periods.
Chain-linking is accomplished for each of thethree subperiods by rescaling; this moves the year inwhich current and constant price versions of thesame time series have the same value, without alter-ing the trend of either. Components of gross domes-tic product (GDP) are individually rescaled andsummed to provide GDP and its subaggregates. Inthis process a rescaling deviation may occur be-tween the constant price GDP by industrial originand the constant price GDP by expenditure. Suchrescaling deviations are absorbed under the headingprivate consumption, etc. on the assumption that GDPby industrial origin is a more reliable estimate thanGDP by expenditure.
Because private consumption is calculated as aresidual, the national accounting identities aremaintained. Rebasing does involve incorporating inprivate consumption whatever statistical discrepan-cies arise for expenditure. The value added in theservices sector also includes a statistical discrep-ancy, as reported by the original source.
Summary measures
The summary measures are calculated by simpleaddition when a variable is expressed in reasonablycomparable units of account. Economic indicatorsthat do not seem naturally additive are usually com-bined by a price-weighting scheme. The summarymeasures for social indicators are weighted by pop-ulation.
The World Development Indicators, unlike theWorld Tables, provide data for, usually, two referencepoints rather than annual time series. For summarymeasures that cover many years, the calculation isbased on the same country composition over timeand across topics. The World Development Indica-tors permit group measures to be compiled only if
229
the country data available for a given year accountfor at least two-thirds of the full group, as definedby the 1987 benchmarks. As long as that criterion ismet, noncurrent reporters (and those not providingample history) are, for years with missing data, as-sumed to behave like the sample of the group thatdoes provide estimates. Readers should keep inmind that the purpose is to maintain an appropriaterelationship across topics, despite myriad problemswith country data, and that nothing meaningful canbe deduced about behavior at the country level byworking back from group indicators. In addition,the weighting process may result in discrepanciesbetween summed subgroup figures and overall to-tals. This is explained more fully in the introductionto the World Tables.
Sources and methods
Data on external debt are compiled directly by theWorld Bank on the basis of reports from its develop-ing member countries through the Debtor Report-ing System (DRS). Other data are drawn mainlyfrom the United Nations (U.N.) and its specializedagencies, the International Monetary Fund (IMF),and country reports to the World Bank. Bank staffestimates are also used to improve currentness orconsistency. For most countries, national accountsestimates are obtained from member governmentsthrough World Bank economic missions. In some in-stances these are adjusted by Bank staff to ensureconformity with international definitions and con-cepts, consistency, and currentness.
Growth rates
For ease of reference, only ratios and rates of growthare usually shown; absolute values are generallyavailable from other World Bank publications, no-tably the 1994 edition of the World Tables. Mostgrowth rates are calculated for two periods, 1970-80and 1980-92, and are computed, unless otherwisenoted, by using the least-squares regression method.Because this method takes into account all observa-tions in a period, the resulting growth rates reflectgeneral trends that are not unduly influenced by ex-ceptional values, particularly at the end points. Toexclude the effects of inflation, constant price eco-nomic indicators are used in calculating growthrates. Details of this methodology are given at thebeginning of the technical notes. Data in italics arefor years or periods other than those specifiedupto two years earlier for economic indicators and upto three years on either side for social indicators,
230
since the latter tend to be collected less regularly andchange less dramatically over short periods of time.
All growth rates shown are calculated from con-stant price series and, unless otherwise noted, havebeen computed using the least-squares method. Theleast-squares growth rate, r, is estimated by fitting aleast-squares linear regression trend line to the loga-rithmic annual values of the variable in the relevantperiod. More specifically, the regression equationtakes the form log X = a + bt + e, where this isequivalent to the logarithmic transformation of thecompound growth rate equation, X = X0 (1 + r)t . Inthese equations, X is the variable, t is time, and a =log X0 and b = log (1 + r) are the parameters to be es-timated; e is the error term. If b* is the least-squaresestimate of b, then the average annual percentagegrowth rate, r, is obtained as [antilog (b*)] _1 and ismultiplied by 100 to express it as a percentage.
Table 1. Basic indicators
For basic indicators for economies with sparse dataor with populations of fewer than 1 million, seeTable Ia.
Population numbers for mid-1992 are World Bankestimates. These are usually projections from themost recent population censuses or surveys; mostare from 1980-92 and, for a few countries, from the1960s or 1970s. Note that refugees not permanentlysettled in the country of asylum are generally con-sidered to be part of the population of their countryof origin.
The data on urea are from the Food and Agricul-hire Organization (FAO). Area is the total surfacearea, measured in square kilometers, comprisingland area and inland waters.
GNP per capita figures in U.S. dollars are calcu-lated according to the World Bank Atlas method,which is described below.
GNP per capita does not, by itself, constitute ormeasure welfare or success in development. It doesnot distinguish between the aims and ultimate usesof a given product, nor does it say whether it merelyoffsets some natural or other obstacle, or harms orcontributes to welfare. For example, GNP is higherin colder countries, where people spend money onheating and warm clothing, than in balmy climates,where people are comfortable wearing light cloth-ing in the open air.
More generally, GNP does not deal adequatelywith environmental issues, particularly natural re-source use. The World Bank has joined with othersto see how national accounts might provide insightsinto these issues. "Satellite" accounts that delve into
practical and conceptual difficulties (such as assign-ing a meaningful economic value to resources thatmarkets do not yet perceive as "scarce" and allocat-ing costs that are essentially global within a frame-work that is inherently national) have been includedin the 1993 revision of the U.N.'s System of NationalAccounts (SNA). This will provide a framework fornational accountants to consider environmental fac-tors in estimating alternative measures of income.
GNP measures the total domestic and foreignvalue added claimed by residents. It comprises GDP(defined in the note for Table 2) plus net factor in-come from abroad, which is the income residents re-ceive from abroad for factor services (labor and cap-ital) less similar payments made to nonresidentswho contributed to the domestic economy.
In estimating GNP per capita, the Bank recog-nizes that perfect cross-country comparability ofGNP per capita estimates cannot be achieved. Be-yond the classic, strictly intractable index numberproblem, two obstacles stand in the way of adequatecomparability. One concerns the GNP and popula-tion estimates themselves. There are differences innational accounting and demographic reporting sys-tems and in the coverage and reliability of underly-ing statistical information among various countries.The other obstacle relates to the use of official ex-change rates for converting GNP data, expressed indifferent national currencies, to a common denomi-nationconventionally the U.S. dollarto com-pare them across countries.
Recognizing that these shortcomings affect thecomparability of the GNP per capita estimates, theWorld Bank has introduced several improvementsin the estimation procedures. Through its regular re-view of member countries' national accounts, theBank systematically evaluates the GNP estimates,focusing on the coverage and concepts employed,and, where appropriate, making adjustments to im-prove comparability. As part of the review, Bankstaff estimates of GNP (and sometimes of popula-tion) may be developed for the most recent period.
The World Bank also systematically assesses theappropriateness of official exchange rates as conver-sion factors. An alternative conversion factor is used(and reported in the World Tables) when the officialexchange rate is judged to diverge by an exception-ally large margin from the rate effectively applied toforeign transactions. This applies to only a smallnumber of countries. For all other countries theBank calculates GNP per capita using the WorldBank Atlas method.
The Atlas conversion factor for any year is the av-erage of a country's exchange rate for that year and
its exchange rates for the two preceding years, afteradjusting them for differences in relative inflationbetween the country and the United States. Thisthree-year average smooths fluctuations in pricesand exchange rates for each country. The resultingGNP in U.S. dollars is divided by the midyear pop-ulation for the latest of the three years to deriveGNP per capita.
Some fifty low- and middle-income economiessuffered declining real GNP per capita in constantprices during the late 1980s and early 1990s. In addi-tion, significant currency and terms of trade fluctua-tions have affected relative income levels. For thisreason the levels and ranking of GNP per capita es-timates, calculated by the Atlas method, have some-times changed in ways not necessarily related to therelative domestic growth performance of theeconomies.
The following formulas describe the proceduresfor computing the conversion factor for year t:
1 P (P1 P(e21) = [e + e,1 -/ ) + el
f-:1 [-I
and for calculating per capita GNP in U.S. dollarsfor year t:
(Y)=(Y1/N1)+ e721
where= current GNP (local currency) for year= GNP deflator for year
et = average annual exchange rate (local currencyto the U.S. dollar) for year
= midyear population for yearP = U.S. GNP deflator for year
Because of problems associated with the avail-ability of comparable data and the determination ofconversion factors, information on GNP per capitais not shown for some economies.
The use of official exchange rates to convert na-tional currency figures to U.S. dollars does not re-flect the relative domestic purchasing powers ofcurrencies. The U. N. International Comparison Pro-gramme (ICP) has developed measures of real GDPon an internationally comparable scale, using pur-chasing power of currencies (PPPs) instead of ex-change rates as conversion factors. Table 30 showsthe most recent PPP-based GNP per capita esti-mates. Information on the ICP has been publishedin four studies and in a number of other reports. Themost recent study is Phase VI, for 1990, a part ofwhich has already been published by the Organiza-
231
tion for Economic Cooperation and Development(OECD).
The ICP figures reported in Table 30 are prelimi-nary and may be revised. The United Nations andits regional economic commissions, as well as otherinternational agencies, such as the European Com-mission (EC), the OECD, and the World Bank, areworking to improve the methodology and to extendannual purchasing power comparisons to all coun-tries. However, exchange rates remain the only gen-erally available means of converting GNP from na-tional currencies to U.S. dollars.
Average annual rate of inflation is measured by thegrowth rate of the GDP implicit deflator for each ofthe periods shown. The GDP deflator is first calcu-lated by dividing, for each year of the period, thevalue of GDP at current values by the value of GDPat constant values, both in national currency. Theleast-squares method is then used to calculate thegrowth rate of the GDP deflator for the period. Thismeasure of inflation, like any other, has limitations.For some purposes, however, it is used as an indica-tor of inflation because it is the most broadly basedmeasure, showing annual price movements for allgoods and services produced in an economy.
Life expectancy at birth indicates the number ofyears a newborn infant would live if prevailing pat-terns of mortality at the time of its birth were to staythe same throughout its life. Figures are World Bankestimates based on data from the U.N. PopulationDivision, the U.N. Statistical Office, and nationalstatistical offices.
Adult illiteracy is defined here as the proportionof the population over the age of fifteen who cannot,with understanding, read and write a short, simplestatement on their everyday life. This is only one ofthree widely accepted definitions, and its applica-tion is subject to qualifiers in a number of countries.The data are from the illiteracy estimates and projec-tions prepared in 1989 by the U.N. Educational, Sci-entific, and Cultural Organization (UNESCO).
The summary measures for GNP per capita, lifeexpectancy, and adult illiteracy in this table areweighted by population. Those for average annualrates of inflation are weighted by the 1987 share ofcountry GDP valued in current U.S. dollars.
Tables 2 and 3. Growth and structure ofproduction
Most of the definitions used are those of the U.N.'sA System of National Accounts (SNA), Series F, No. 2,Revision 3. Revision 4 of the SNA was completedonly in 1993, and it is likely that many countries will
232
still be using the recommendations of Revision 3 forthe next few years. Estimates are obtained from na-tional sources, sometimes reaching the World Bankthrough other international agencies but more oftencollected during World Bank staff missions.
World Bank staff review the quality of nationalaccounts data and in some instances, through mis-sion work or technical assistance, help adjust na-tional series. Because of the sometimes limitedcapabilities of statistical offices and basic data prob-lems, strict international comparability cannot beachieved, especially in economic activities that aredifficult to measure, such as parallel market transac-tions, the informal sector, or subsistence agriculture.
GDP measures the total output of goods and ser-vices for final use produced by residents and non-residents, regardless of the allocation to domesticand foreign claims. It is calculated without makingdeductions for depreciation of "manmade" assets ordepletion and degradation of natural resources. Al-though the SNA envisages estimates of GDP by in-dustrial origin to be at producer prices, many coun-tries still report such details at factor cost.International comparability of the estimates is af-fected by differing country practices in valuationsystems for reporting value added by productionsectors. As a partial solution, GDP estimates areshown at purchaser values if the components are onthis basis, and such instances are footnoted. How-ever, for a few countries in Tables 2 and 3, GDP atpurchaser values has been replaced by GDP at fac-tor cost.
The figures for GDP are U.S. dollar values con-verted from domestic currencies using single-yearofficial exchange rates. For a few countries wherethe official exchange rate does not reflect the rate ef-fectively applied to actual foreign exchange transac-tions, an alternative conversion factor is used (andreported in the World Tables). Note that this tabledoes not use the three-year averaging technique ap-plied to GNP per capita in Table 1.
Agriculture covers forestry, hunting, and fishingas well as agriculture. In developing countries withhigh levels of subsistence farming, much agricul-tural production is either not exchanged or not ex-changed for money. This increases the difficulty ofmeasuring the contribution of agriculture to GDPand reduces the reliability and comparability ofsuch numbers.
Industry comprises value added in mining; manu-facturing (also reported as a separate subgroup);construction; and electricity, water, and gas. Valueadded in all other branches of economic activity, in-cluding imputed bank service charges, import du-
ties, and any statistical discrepancies noted by na-tional compilers, are categorized as services, etc.
Partially rebased, chain-linked 1987 series in do-mestic currencies, as explained at the beginning ofthe technical notes, are used to compute the growthrates in Table 2. The sectoral shares of GDP in Table3 are based on current price series.
In calculating the summary measures for each in-dicator in Table 2, partially rebased constant 1987U.S. dollar values for each economy are calculatedfor each year of the periods covered; the values areaggregated across countries for each year; and theleast-squares procedure is used to compute thegrowth rates. The average sectoral percentageshares in Table 3 are computed from group aggre-gates of sectoral GDP in current U.S. dollars.
Table 4. Agriculture and food
The basic data for value added in agriculture are fromthe World Bank's national accounts series at currentprices in national currencies. Value added in currentprices in national currencies is converted to U.S.dollars by applying the single-year conversion pro-cedure, as described in the technical note for Tables2 and 3.
The figures for the remainder of this table arefrom the FAQ. Cereal imports are measured in grainequivalents and defined as comprising all cereals inthe Standard International Trade Classification (SITC),Revision 2, Groups 041-046. Food aid in cereals coverswheat and flour, bulgur, rice, coarse grains, and thecereal component of blended foods. The figures arenot directly comparable because of reporting andtiming differences. Cereal imports are based on cal-endar-year data reported by recipient countries, andfood aid in cereals is based on data for crop years re-ported by donors and international organizations,including the International Wheat Council and theWorld Food Programme. Furthermore, food aid in-formation from donors may not correspond to ac-tual receipts by beneficiaries during a given periodbecause of delays in transportation and recording orbecause aid is sometimes not reported to the FAQ orother relevant international organizations. Food aidimports may also not show up in customs records.The time reference for food aid is the crop year, Julyto June.
Fertilizer consumption measures the plant nutri-ents used in relation to arable land. Fertilizer prod-ucts cover nitrogenous, potash, and phosphate fer-tilizers (which include ground rock phosphate).Arable land is defined as land under permanentcrops and under temporary crops (double-cropped
areas are counted once), temporary meadows formowing or for pasture, land under market orkitchen gardens, and land temporarily fallow orlying idle. The time reference for fertilizer consump-tion is the crop year, July to June.
Average growth rate of food production per capitahas been computed from the index of food produc-tion per capita. The index relates to the average an-nual growth rate of food produced per capita in1979-92 in relation to the average produced annu-ally in 1979-81 (1979-81 = 100). The estimates arederived by dividing the quantity cf food productionby the total population. For the index, food is de-fined as comprising nuts, pulses, fruits, cereals, veg-etables, sugar cane, sugar beet, starchy roots, edibleoils, livestock, and livestock products. Quantities offood production are measured net of annual feed,seeds for use in agriculture, and food lost in pro-cessing and distribution.
Fish products are measured by the level of dailyprotein supply derived from the consumption offish in relation to total daily protein supply from allfood. This estimate indirectly highlights the relativeimportance or weight of fish in total agriculture, es-pecially since fish is not included in the index offood production.
The summary measures for fertilizer consump-tion are weighted by total arable land area; the sum-mary measures for food production are weighted bypopulation.
Table 5. Commercial energy
The data on energy production and consumptionare primarily from International Energy Agency(lEA) and U.N. sources. They refer to commercialforms of primary energypetroleum (crude oil,natural gas liquids, and oil from nonconventionalsources), natural gas, solid fuels (coal, lignite, andother derived fuels), and primary electricity (nu-clear, hydroelectric, geothermal, and other)allconverted into oil equivalents. For converting nu-clear electricity into oil equivalents, a notional ther-mal efficiency of 33 percent is assumed; hydroelec-tric power is represented at 100 percent efficiency.
Energy consumption refers to domestic primaryenergy supply before transformation to other end-use fuels (such as electricity and refined petroleumproduct) and is calculated as indigenous productionplus imports and stock changes, rriinus exports andinternational marine bunkers. Energy consumptionalso includes products for nonenergy uses, mainlyderived from petroleum. The use of firewood, driedanimal excrement, and other traditional fuels, al-
233
though substantial in some developing countries, isnot taken into account because reliable and comprehensive data are not available.
Energy use is expressed as kilogram oil equiva-lent per capita. The output indicator is the U.S. dol-lar estimate of GDP produced per kilogram of oilequivalent.
Energy imports refer to the dollar value of energyimportsSection 3 in the SITC, Revision 1andare expressed as a percentage of earnings from mer-chandise exports. Because data on energy importsdo not permit a distinction between petroleum im-ports for fuel and those for use in the petrochemi-cals industry, these percentages may overestimatedependence on imported energy.
The summary measures of energy productionand consumption are computed by aggregating therespective volumes for each of the years covered bythe periods and applying the least-squares growthrate procedure. For energy consumption per capita,population weights are used to compute summarymeasures for the specified years.
The summary measures of energy imports as apercentage of merchandise exports are computedfrom group aggregates for energy imports and mer-chandise exports in current dollars.
Table 6. Structure of manufacturing
The basic data for value added in manufacturing arefrom the World Bank's national accounts series atcurrent prices in national currencies. Value added incurrent prices in national currencies is converted toU.S. dollars by applying the single-year conversionprocedure, as described in the technical note for Ta-bles 2 and 3.
The data for distribution of manufacturing valueadded among industries are provided by the UnitedNations Industrial Development Organization(UNIDO), and distribution calculations are from na-tional currencies in current prices.
The classification of manufacturing industries isin accordance with the U.N. International StandardIndustrial Classification of All Economic Activities(ISIC), Revision 2. Food, beverages, and tobacco com-prise ISIC Division 31; textiles and clothing, Division32; machinery and transport equipment, Major Groups382-84; and chemicals, Major Groups 351 and 352.Other comprises wood and related products (Divi-sion 33), paper and related products (Division 34),petroleum and related products (Major Groups353-56), basic metals and mineral products (Divi-sions 36 and 37), fabricated metal products and pro-fessional goods (Major Groups 381 and 385), and
234
other industries (Major Group 390). When data fortextiles, machinery, or chemicals are shown as notavailable, they are also included in other.
Summary measures given for value added inmanufacturing are totals calculated by the aggrega-tion method noted at the beginning of the technicalnotes.
Table 7. Manufacturing earnings and output
Four indicators are shown: two relate to real earn-ings per employee, one to labor's share in totalvalue added generated, and one to labor productiv-ity in the manufacturing sector. The indicators arebased on data from UNIDO; the deflators are fromother sources, as explained below.
Earnings per employee are in constant prices andare derived by deflating nominal earnings per em-ployee by the country's consumer price index (CPI).The CPI is from the IMF's International FinancialStatistics.
Total earnings as a percentage of value added are de-rived by dividing total earnings of employees byvalue added in current prices to show labor's sharein income generated in the manufacturing sector.Gross output per employee is in constant prices and ispresented as an index of overall labor productivityin manufacturing, with 1980 as the base year. To de-rive this indicator, UNIDO data on gross output peremployee in current prices are adjusted using im-plicit deflators for value added in manufacturing orin industry, taken from the World Bank's nationalaccounts data files.
To improve cross-country comparability, UNIDOhas, where possible, standardized the coverageof establishments to those with five or moreemployees.
The concepts and definitions are in accordancewith the International Recommendations for IndustrialStatistics, published by the United Nations. Earn-ings (wages and salaries) cover all remuneration toemployees paid by the employer during the year.The payments include (a) all regular and overtimecash payments and bonuses and cost of living al-lowances; (b) wages and salaries paid during vaca-tion and sick leave; (c) taxes and social insurancecontributions and the like, payable by the employ-ees and deducted by the employer; and (d) pay-ments in kind.
The term "employees" in this table combines twocategories defined by the United Nations: regularemployees and persons engaged. Together thesegroups comprise regular employees, working pro-prietors, active business partners, and unpaid fam-
ily workers; they exclude homeworkers. The datarefer to the average number of employees workingduring the year.
"Value added" is defined as the current value ofgross output less the current cost of (a) materials,fuels, and other supplies consumed; (b) contract andcommission work done by others; (c) repair andmaintenance work done by others; and (d) goodsshipped in the same condition as received.
The value of gross output is estimated on thebasis of either production or shipments. On the pro-duction basis it consists of (a) the value of all prod-ucts of the establishment; (b) the value of industrialservices rendered to others; (c) the value of goodsshipped in the same condition as received; (d) thevalue of electricity sold; and (e) the net change inthe value of work-in-progress between the begin-ning and the end of the reference period. In the caseof estimates compiled on a shipment basis, the netchange between the beginning and the end of thereference period in the value of stocks of finishedgoods is also included.
Tables 8 and 9. Growth of consumption andinvestment; structure of demand
GDP is defined in the note for Tables 2 and 3, buthere it is in purchaser values.
General government consumption includes all cur-rent expenditure for purchases of goods and ser-vices by all levels of government. Capital expendi-ture on national defense and security is regarded asconsumption expenditure.
Private consumption, etc. is the market value of allgoods and services, including durable products(such as cars, washing machines, and home com-puters) purchased or received as income in kind byhouseholds and nonprofit institutions. It excludespurchases of dwellings but includes imputed rentfor owner-occupied dwellings. In practice, it in-
cludes any statistical discrepancy in the use of re-sources. At constant prices, it also includes therescaling deviation from partial rebasing, which isexplained at the beginning of the technical notes.
Gross domestic investment consists of outlays onadditions to the fixed assets of the economy plus netchanges in the level of inventories.
Gross domestic savings are calculated by deductingtotal consumption from GDP.
Exports of goods and nonfactor services represent thevalue of all goods and nonfactor services providedto the rest of the world; they include merchandise,freight, insurance, travel, and other nonfactor ser-vices. The value of factor services, such as invest-
ment income, interest, and labor income, is ex-
cluded. Current transfers are also excluded.The resource balance is the difference between ex-
ports of goods and nonfactor services and importsof goods and nonfactor services.
Partially rebased 1987 series in constant domesticcurrency units are used to compute the indicators inTable 8. Distribution of GDP in Table 9 is calculatedfrom national accounts series in current domesticcurrency units.
The summary measures are calculated by themethod explained in the note for Tables 2 and 3.
Table 10. Central government expenditure
The data on central government finance in Tables 10and 11 are from the IMF, Government Finance Statis-tics Yearbook (1993), and IMF data files. The accountsof each country are reported using the system ofcommon definitions and classifications found inIMF, A Manual on Government Finance Statistics(1986).
For complete and authoritative explanations ofconcepts, definitions, and data sources, see theseIMF sources. The commentary that follows is in-tended mainly to place these data in the context ofthe broad range of indicators reported in this edition.
The shares of total expenditure and current rev-enue by category are calculated from series in na-tional currencies. Because of differences in coverageof available data, the individual components of cen-tral government expenditure and current revenueshown in these tables may not be strictly compara-ble across all economies.
Moreover, inadequate statistical coverage ofstate, provincial, and local governments dictates theuse of central government data; this may seriouslyunderstate or distort the statistical portrayal of theallocation of resources for various purposes, espe-cially in countries where lower levels of govern-ment have considerable autonomy and are respon-sible for many economic and social services. Inaddition, "central government" can mean either oftwo accounting concepts: consolidated or bud-getary. For most countries, central government fi-nance data have been consolidated into one overallaccount, but for others only the budgetary centralgovernment accounts are available. Since budgetaryaccounts do not always include all central govern-ment units, the overall picture of central govern-ment activities is usually incomplete. Countries re-porting budgetary data are footnoted.
Consequently, the data presented, especiallythose for education and health, are not comparable
235
across countries. In many economies, private healthand education services are substantial; in others,public services represent the major component oftotal expenditure but may be financed by lower lev-els of government. Caution should therefore be ex-ercised in using the data for cross-country compar-isons. Central government expenditure comprisesthe expenditure by all government offices, depart-ments, establishments, and other bodies that areagencies or instruments of the central authority of acountry. It includes both current and capital (devel-opment) expenditure.
Defense comprises all expenditure, whether bydefense or other departments, on the maintenanceof military forces, including the purchase of militarysupplies and equipment, construction, recruiting,and training. Also in this category are closely re-lated items such as military aid programs. Defensedoes not include expenditure on public order andsafety, which are classified separately.
Education comprises expenditure on the provi-sion, management, inspection, and support ofpreprimary, primary, and secondary schools; of uni-versities and colleges; and of vocational, technical,and other training institutions. Also included is ex-penditure on the general administration and regula-tion of the education system; on research into its ob-jectives, organization, administration, and methods;and on such subsidiary services as transport, schoolmeals, and school medical and dental services.
Health covers public expenditure on hospitals,maternity and dental centers, and clinics with amajor medical component; on national health andmedical insurance schemes; and on family planningand preventive care.
Housing, amenities, social security, and welfare coverexpenditure on housing (excluding interest subsi-dies, which are usually classified with other) such asincome-related schemes; on provision and supportof housing and slum-clearance activities; on com-munity development; and on sanitation services.These categories also cover compensation for loss ofincome to the sick and temporarily disabled; pay-ments to the elderly, the permanently disabled, andthe unemployed; family, maternity, and child al-lowances; and the cost of welfare services, such ascare of the aged, the disabled, and children. Manyexpenditures relevant to environmental defense,such as pollution abatement, water supply, sanitaryaffairs, and refuse collection, are included indistin-guishably in this category.
Economic services comprise expenditure associ-ated with the regulation, support, and more efficientoperation of business; economic development; re-
236
dress of regional imbalances; and creation of em-ployment opportunities. Research, trade promotion,geological surveys, and inspection and regulation ofparticular industry groups are among the activitiesincluded.
Other covers general public services, interest pay-ments, and items not included elsewhere; for a feweconomies it also includes amounts that could notbe allocated to other components (or adjustmentsfrom accrual to cash accounts).
Total expenditure is more narrowly defined thanthe measure of general government consumptiongiven in Tables 8 and 9 because it excludes con-sumption expenditure by state and local govern-ments. At the same time, central government expen-diture is more broadly defined because it includesgovernment's gross domestic investment and trans-fer payments.
Overall surplus/deficit is defined as current andcapital revenue and official grants received, lesstotal expenditure and lending minus repayments.
Table 11. Central government current revenue
Information on data sources and comparability andthe definition of central government is given in thefirst four paragraphs of the note for Table 10. Cur-rent revenue by source is expressed as a percentageof total current revenue, which is the sum of tax rev-enue and nontax revenue and is calculated from na-tional currencies.
Tax revenue comprises compulsory, unrequited,nonrepayable receipts for public purposes. It in-cludes interest collected on tax arrears and penaltiescollected on nonpayment or late payment of taxesand is shown net of refunds and other correctivetransactions. Taxes on income, profit, and capital gainsare taxes levied on the actual or presumptive net in-come of individuals, on the profits of enterprises,and on capital gains, whether realized on land sales,securities, or other assets. Intragovernmental pay-ments are eliminated in consolidation. Social securitycontributions include employers' and employees'social security contributions as well as those of self-employed and unemployed persons. Taxes on goodsand services cover all domestic taxes including gen-eral sales and turnover or value added taxes, selec-tive excises on goods, selective taxes on services,taxes on the use of goods or property. and profits offiscal monopolies. Taxes on international trade andtransactions include import duties, export duties,profits of export or import monopolies, exchangeprofits, and exchange taxes. Other taxes include em-ployers' payroll or labor taxes, taxes on property
and taxes not allocable to other categories. Theymay include negative values that are adjustments,for instance, for taxes collected on behalf of stateand local governments and not allocable to individ-ual tax categories.
Non tax revenue comprises receipts that are not acompulsory nonrepayable payment for public pur-poses, such as fines, administrative fees, or entre-preneurial income from government ownership ofproperty. Proceeds of grants and borrowing, fundsarising from the repayment of previous lending bygovernments, incurrence of liabilities, and proceedsfrom the sale of capital assets are not included.
Table 12. Money and interest rates
The data on broadly defined money are based on theIMF's International Financial Statistics (IFS). Broadlydefined money comprises most liabilities of a coun-try's monetary institutions to residents other thanthe central government. For most countries, broadlydefined money is the sum of money (IFS line 34)and quasi money (IFS line 35). Money comprises theeconomy's means of payment: currency outsidebanks and demand deposits. Quasi money com-prises time and savings deposits and similar bankaccounts that the issuer can exchange for moneywith little if any delay or penalty. Where nonmone-tary financial institutions are important issuers ofquasi-monetary liabilities, these are often includedin the measure of broadly defined money.
The growth rates for broadly defined money arecalculated from year-end figures, while the averageof the year-end figures for the specified year and theprevious year is used for the ratio of broadly de-fined money to GDP.
The nominal interest rates of banks, also from IFS,represent the rates paid by commercial or similarbanks to holders of their quasi-monetary liabilities(deposit rate) and charged by the banks on loans toprime customers (lending rate). The data are, how-ever, of limited international comparability partlybecause coverage and definitions vary.
Since interest rates (and growth rates for broadlydefined money) are expressed in nominal terms,much of the variation among countries stems fromdifferences in inflation. For easy reference, the Table1 indicator of recent inflation is repeated in this table.
Table 13. Growth of merchandise trade
The main data source for current trade values is theU.N. Commodity Trade (COMTRADE) data file,supplemented by World Bank estimates. The statis-
tics on merchandise trade are based on countries'customs returns.
Merchandise exports and imports, with some ex-ceptions, cover international movements of goodsacross customs borders; trade in services is not in-cluded. Exports are valued f.o.b. (free on board) andimports c.i.f. (cost, insurance, and freight) unlessotherwise specified in the foregoing sources. Thesevalues are in current U. S. dollars.
The growth rates of merchandise exports and im-ports are based on constant price data, which areobtained from export or import value data as de-flated by the corresponding price index. The WorldBank uses its own price indexes, which are based oninternational prices for primary commodities andunit value indexes for manufactures. These price in-dexes are country-specific and disaggregated bybroad commodity groups. This ensures consistencybetween data for a group of countries and those forindividual countries. Such consistency will increaseas the World Bank continues to improve its tradeprice indexes for an increasing number of countries.These growth rates can differ from those derivedfrom national practices because national price in-dexes may use different base years and weightingprocedures from those used by the World Bank.
The terms of trade, or the net barter terms of trade,measure the relative movement of export pricesagainst that of import prices. Calculated as the ratioof a country's index of average export prices to itsaverage import price index, this indicator showschanges over a base year in the level of export pricesas a percentage of import prices. The terms of tradeindex numbers are shown for 1985 and 1992, where1987 = 100. The price indexes are from the sourcecited above for the growth rates of exports andimports.
The summary measures for the growth rates arecalculated by aggregating the 1987 constant U.S.dollar price series for each year and then applyingthe least-squares growth rate procedure for the peri-ods shown.
Tables 14 and 15. Structure of merchandiseimports and exports
The shares in these tables are derived from tradevalues in current dollars reported in the U.N. tradedata system, supplemented by World Bank esti-mates.
Merchandise exports and imports are defined inthe technical note for Table 13.
The categorization of exports and imports fol-lows the Standard International Trade Classification
237
(SITC), Series M, No. 34, Revision 1. For some coun-tries, data for certain commodity categories are un-available and the full breakdown cannot be shown.
In Table 14, food commodities are those in SITCSections 0, 1, and 4 and Division 22 (food and liveanimals, beverages and tobacco, animal and veg-etable oils and fats, oilseeds, oil nuts and oil ker-nels). Fuels are the commodities in SITC Section 3(mineral fuels, and lubricants and related materi-als). Other primary commodities comprise SITC Sec-tion 2 (inedible crude materials, except fuels), lessDivision 22 (oilseeds, oilnuts, and oil kernels) andDivision 68 (nonferrous metals). Machinery andtransport equipment are the commodities in SITC Sec-tion 7. Other manufactures, calculated residuallyfrom the total value of manufactured imports, rep-resent SITC Sections 5 through 9, less Section 7 andDivision 68.
In Table 15, fuels, minerals, and metals are the com-modities in SITC Section 3 (mineral fuels, and lubri-cants and related materials), Divisions 27 and 28(crude fertilizers and crude minerals, excludingcoal, petroleum and precious stones, and metallifer-ous ores and metal scrap), and Division 68 (nonfer-rous metals). Other primary commodities compriseSITC Sections 0, 1, 2, and 4 (food and live animals,beverages and tobacco, inedible crude materials, ex-cept fuels, and animal and vegetable oils and fats),less Divisions 27 and 28. Machinery and transportequipment are the commodities in SITC Section 7.Other manufactures represent SITC Sections 5through 9, less Section 7 and Division 68. Textilesand clothing, representing SITC Divisions 65 and 84(textiles, yarns, fabrics, made-up articles, and re-lated products and clothing), are a subgroup ofother manufactures.
The summary measures in Table 14 are weightedby total merchandise imports of individual coun-tries in current U.S. dollars and those in Table 15 bytotal merchandise exports of individual countries incurrent U.S. dollars. (See the technical note for Table13.)
Table 16. OECD imports of manufactured goods
The data are from the United Nations and were re-ported by high-income OECD economies theOECD members excluding Greece, Portugal, andTurkey.
The table reports the value of imports of manufac-tures of high-income OECD countries by the econ-omy of origin and the composition of such importsby major manufactured product groups. These dataare based on the U.N. COMTRADE database-
238
Revision I SITC for 1970 and Revision 2 SITC for1992.
Manufactured imports of the predominant mar-kets from individual economies are the best avail-able proxy of the magnitude and composition of themanufactured exports of developing economies toall destinations taken together.
Manufactured goods are the commodities in theSITC, Revision 1, Sections 5 through 9 (chemical andrelated products, basic manufactures, manufacturedarticles, machinery and transport equipment, andother manufactured articles and goods not else-where classified), excluding Division 68 (nonferrousmetals). This definition is somewhat broader thanthe one used to define exporters of manufactures.
The major manufactured product groups re-ported are defined as follows: textiles and clothing(SITC Sections 65 and 84), chemicals (SITC Section 5),electrical machinery and electronics (SITC Section 72),transport equipment (SITC Section 73), and other, de-fined as the residual. SITC Revision 1 data are usedfor the year 1970, whereas the equivalent data in Re-vision 2 are used for the year 1992.
Table 17. Balance of payments and reserves
The statistics for this table are mostly as reported bythe IMF but do include recent estimates by WorldBank staff and, in rare instances, the Bank's owncoverage or classification adjustments to enhanceinternational comparability. Values in this table arein U.S. dollars converted at current exchange rates.
The current account balance after official transfers isthe difference between (a) exports of goods and ser-vices (factor and nonfactor), as well as inflows ofunrequited transfers (private and official) and (b)imports of goods and services, as well as all unre-quited transfers to the rest of the world.
The current account balance before official transfers isthe current account balance that treats net officialunrequited transfers as akin to official capital move-ments. The difference between the two balance ofpayments measures is essentially foreign aid in theform of grants, technical assistance, and food aid,which, for most developing countries, tends tomake current account deficits smaller than the fi-nancing requirement.
Net workers' remittances cover payments and re-ceipts of income by migrants who are employed orexpect to be employed for more than a year in theirnew economy, where they are considered residents.These remittances are classified as private unre-quited transfers and are included in the balance ofpayments current account balance, whereas those
derived from shorter-term stays are included in ser-vices as labor income. The distinction accords withinternationally agreed guidelines, but many devel-oping countries classify workers' remittances as afactor income receipt (hence, a component of GNP).The World Bank adheres to international guidelinesin defining GNP and therefore may differ from na-tional practices.
Gross international reserves comprise holdings ofmonetary gold, special drawing rights (SDRs), thereserve position of members in the IMF, and hold-ings of foreign exchange under the control of mone-tary authorities. The data on holdings of interna-tional reserves are from IMF data files. The goldcomponent of these reserves is valued throughout atyear-end (December 31) London prices: that is,$37.37 an ounce in 1970 and $333.25 an ounce in1992. Because of differences in the definition of in-ternational reserves, in the valuation of gold, and inreserve management practices, the levels of reserveholdings published in national sources do not havestrictly comparable significance. The reserve levelsfor 1970 and 1992 refer to the end of the year indi-cated and are in current U.S. dollars at prevailingexchange rates. Reserve holdings at the end of 1992,months of import coverage, are also expressed in termsof the number of months of imports of goods andservices they could pay for.
The summary measures are computed fromgroup aggregates for gross international reservesand total imports of goods and services in currentdollars.
Table 18. Official development assistance fromOECD and OPEC members
Official development assistance (ODA) consists of netdisbursements of loans and grants made on conces-sional financial terms by official agencies of themembers of the Development Assistance Commit-tee (DAC) of the Organization for Economic Coop-eration and Development (OECD) and members ofthe Organization of Petroleum Exporting Countries(OPEC) to promote economic development andwelfare. Although this definition is meant to ex-clude purely military assistance, the borderline issometimes blurred; the definition used by the coun-try of origin usually prevails. ODA also includes thevalue of technical cooperation and assistance. Alldata shown are supplied by the OECD, and all U.S.dollar values are converted at official exchangerates.
Total net flows are net disbursements to develop-ing countries and multilateral institutions. The dis-
bursements to multilateral institutions are now re-ported for all DAC members on the basis of the dateof issue of notes; some DAC members previouslyreported on the basis of the date of encashment.
The nominal values shown in the summary forODA from high-income OECD countries were con-verted at 1987 prices using the dollar GDP deflator.This deflator is based on price increases in OECDcountries (excluding Greece, Portugal, and Turkey)measured in dollars. It takes into account the paritychanges between the dollar and national currencies.For example, when the dollar depreciates, pricechanges measured in national currencies have to beadjusted upward by the amount of the depreciationto obtain price changes in dollars.
The table, in addition to showing totals forOPEC, shows totals for the Organization of ArabPetroleum Exporting Countries (OAPEC). Thedonor members of OAPEC are Algeria, Iraq,Kuwait, Libya, Qatar, Saudi Arabia, and UnitedArab Emirates. ODA data for OPEC and OAPECare also obtained from the OECD.
Table 19. Official development assistance: receipts
Net disbursements of ODA from all sources consist ofloans and grants made on concessional financialterms by all bilateral official agencies and multilat-eral sources to promote economic development andwelfare. They include the value of technical cooper-ation and assistance. The disbursements shown inthis table are not strictly comparable with thoseshown in Table 18 since the receipts are from allsources; disbursements in Table 18 refer only tothose made by high-income members of the OECDand members of OPEC. Net disbursements equalgross disbursements less payments to the origina-tors of aid for amortization of past aid receipts. Netdisbursements of ODA are shown per capita and asa percentage of GNP.
The summary measures of per capita ODA arecomputed from group aggregates for populationand for ODA. Summary measures for ODA as a per-centage of GNP are computed from group totals forODA and for GNP in current U.S. dollars.
Table 20. Total external debt
The data on debt in this and successive tables arefrom the World Bank Debtor Reporting System,supplemented by World Bank estimates. The systemis concerned solely with developing economies anddoes not collect data on external debt for othergroups of borrowers or from economies that are not
239
members of the World Bank. The dollar figures ondebt shown in Tables 20 through 24 are in U.S. dol-lars converted at official exchange rates.
The data on debt include private nonguaranteeddebt reported by thirty developing countries andcomplete or partial estimates for an additionaltwenty that do not report but for which this type ofdebt is known to be significant.
Long-term debt has three components: public,publicly guaranteed and private nonguaranteedloans. Public loans are external obligations of publicdebtors, including the national government, itsagencies, and autonomous public bodies. Publiclyguaranteed loans are external obligations of privatedebtors that are guaranteed for repayment by apublic entity. These two categories are aggregated inthe tables. Private nonguaranteed loans are externalobligations of private debtors that are not guaran-teed for repayment by a public entity.
Use of IMF credit denotes repurchase obligationsto the IMF for all uses of IMF resources, excludingthose resulting from drawings in the reservetranche. It is shown for the end of the year specified.It comprises purchases outstanding under the credittranches, including enlarged access resources, andall special facilities (the buffer stock, compensatoryfinancing, extended fund, and oil facilities), trustfund loans, and operations under the enhancedstructural adjustment facilities. Use of IMF creditoutstanding at year-end (a stock) is converted toU.S. dollars at the dollar-SDR exchange rate in effectat year-end.
Short-term debt is debt with an original maturityof one year or less. Available data permit no distinc-tions between public and private nonguaranteedshort-term debt.
Total external debt is defined here as the sum ofpublic, publicly guaranteed, and private nonguar-anteed long-term debt, use of IMF credit, and short-term debt.
Total arrears on long-term debt outstanding and dis-bursed (LDOD) denotes principal and interest duebut not paid.
Ratio of present value to nominal value is the dis-counted value of the future debt service paymentsdivided by the face value of debt.
Table 21. Flow of public and private externalcapital
Data on disbursements, repayment of principal(amortization), and payment of interest are for pub-lic, publicly guaranteed, and private nonguaranteedlong-term loans.
240
Disbursements are drawings on long-term loancommitments during the year specified.
Repayments of principal are actual amount of prin-cipal (amortization) paid in foreign currency, goods,or services in the year specified.
Interest payments are actual amounts of interestpaid in foreign currency, goods, or services in theyear specified.
Table 22. Aggregate net resource flows and nettransfers
Total net flows on long-term debt are disbursementsless the repayment of principal on public, publiclyguaranteed, and private nonguaranteed long-termdebt. Official grants are transfers made by an officialagency in cash or in kind in respect of which nolegal debt is incurred by the recipient. Data on offi-cial grants exclude grants for technical assistance.
Net foreign direct investment (FDI) in the reportingeconomy is defined as investment that is made to ac-quire a lasting interest (usually 10 percent of the vot-ing stock) in an enterprise operating in a countryother than that of the investor (defined according toresidency), the investor's purpose being an effectivevoice in the management of the enterprise.
Portfolio equity flows is the sum of the countryfunds (note that the sum of regional or income-group flows does not add up to the total due to theglobal funds), depository receipts (American orglobal), and direct purchases of shares by foreign in-vestors
Aggregate net resource flows are the sum of netflows on long-term debt (excluding use of IMFcredit), plus official grants (excluding technical as-sistance) and net foreign direct investment. Aggre-gate net transfers are equal to aggregate net resourceflows minus interest payments on long-term loansand remittance of all profits.
Table 23. Total external debt ratios
Net present value of total external debt as a percentage ofexports of goods and services is the discounted value offuture debt service to exports of goods and services.
The present value can be higher or lower than thenominal value of debt. The determining factor forthe present value being above or below par are theinterest rates of loans and the discount rate used inthe present value calculation. A loan with an inter-est rate higher than the discount rate yields a pre-sent value that is larger than the nominal value ofdebt: the opposite holds for loans with an interestrate lower than the discount rate. Throughout this
table, goods and services include workers' remit-tances. For estimating net present value of total exter-nal debt as a percentage of GNP, the debt figures areconverted into U.S. dollars from currencies of repay-ment at end-of-year official exchange rates. GNP isconverted from national currencies to U.S. dollarsby applying the conversion procedure described inthe technical note for Tables 2 and 3.
Total debt service as a percentage of exports of goodsand services is the sum of principal repayments andinterest payments on total external debt (as definedin the note for Table 20). It is one of several conven-tional measures used to assess a country's ability toservice debt.
Interest payments as a percentage of exports of goodsand services are actual payments made on total exter-nal debt.
Concessional debt as a percentage of total externaldebt conveys information about the borrower's re-ceipt of aid from official lenders at concessionalterms as defined by the DAC, that is, loans with anoriginal grant element of 25 percent or more.
Multilateral debt as a percentage of total external debtconveys information about the borrower's receipt ofaid from the World Bank, regional developmentbanks, and other multilateral and intergovernmen-tal agencies. Excluded are loans from funds admin-istered by an international organization on behalf ofa single donor government.
The summary measures are weighted by exportsof goods and services in current dollars and by GNPin current dollars, respectively.
Table 24. Terms of external public borrowing
Commitments refer to the public and publicly guar-anteed loans for which contracts were signed in theyear specified. They are reported in currencies of re-payment and converted into U.S. dollars at averageannual official exchange rates.
Figures for interest rates, maturities, and grace peri-ods are averages weighted by the amounts of theloans. Interest is the major charge levied on a loanand is usually computed on the amount of principaldrawn and outstanding. The maturity of a loan isthe interval between the agreement date, when aloan agreement is signed or bonds are issued, andthe date of final repayment of principal. The graceperiod is the interval between the agreement dateand the date of the first repayment of principal.
Public loans with variable interest rates, as a percent-age of public debt refer to interest rates that float withmovements in a key market rate; for example, theLondon interbank offered rate (LIBOR) or the U.S.
prime rate. This column shows the borrower's ex-posure to changes in international interest rates.
The summary measures in this table areweighted by the amounts of the loans.
Table 25. Population and labor force
Population and labor force growth rates are expo-nential period averages calculated from midyearpopulations and total labor force estimates. (See theKey for survey and census information.)
Population estimates for mid-1992 are made bythe World Bank from data provided by the U.N.Population Division, the U.N. Statistical Office, andcountry statistical offices. Estimates take into ac-count the results of the latest population censuses,which in some cases are neither recent nor accurate.Note that refugees not permanently settled in thecountry of asylum are generally considered to bepart of the population of their country of origin.
The projections of population for 2000, 2025, andthe year in which the population will eventually be-come stationary (see definition below) are made foreach economy separately. Information on total pop-ulation by age and sex, fertility, mortality, and inter-national migration is projected on the basis of gener-alized assumptions until the population becomesstationary
A stationary population is one in which age- andsex-specific mortality rates have not changed over along period, during which fertility rates have re-mained at replacement level; that is, the net repro-duction rate (defined in the note for Table 26) equals1. In such a population, the birth rate is constant andequal to the death rate, the age structure is constant,and the growth rate is zero.
Population projections are made by age cohort.Mortality fertility, and migration are projected sepa-rately, and the results are applied iteratively to the1990 base-year age structure. For the projection pe-riod 1990 to 2005, the changes in mortality are coun-try specific: increments in life expectancy and decre-ments in infant mortality are based on previoustrends for each country. When female secondaryschool enrollment is high, mortality is assumed todecline more quickly. Infant mortality is projectedseparately from adult mortality. Note that the pro-jections incorporate the impact of acquired immunedeficiency syndrome (AIDS) on mortality.
Projected fertility rates are also based on previ-ous trends. For countries in which fertility hasstarted to decline (termed "fertility transition"), thistrend is assumed to continue. It has been observedthat no country in which the population has a life
241
expectancy of less than 50 years has experienced afertility transition; for these countries, fertility tran-sition is delayed, and the average decline of thegroup of countries in fertility transition is applied.Countries with below-replacement fertility are as-sumed to have constant total fertility rates until2005 and to regain replacement level by 2030.
International migration rates are based on pastand present trends in migration flows and migra-tion policy. Among the sources consulted are esti-mates and projections made by national statisticaloffices, international agencies, and research institu-tions. Because of the uncertainty of future migrationtrends, it is assumed in the projections that net mi-gration rates will reach zero by 2025.
The estimates of the size of the stationary popu-lation are very long-term projections. They are in-cluded only to show the implications of recent fertil-ity and mortality trends on the basis of generalizedassumptions. A fuller description of the methodsand assumptions used to calculate the estimates iscontained in World Population Projections, 1994-95Edition (forthcoming).
Total labor force is the "economically active" popu-lation; a restrictive concept that includes the armedforces and the unemployed but excludes home-makers and others unpaid caregivers. Labor forcenumbers in several developing countries reflect asignificant underestimation of female participationrates. Labor force growth rates are derived from In-ternational Labour Organisation (ILO) data.
Table 26. Demography and fertility
The crude birth rate and crude death rate indicate, re-spectively, the number of live births and deaths oc-curring per thousand population in a year. Theycome from the sources mentioned in the note toTable 25. (See the Key for survey and census infor-mation.)
The total fertility rate represents the number ofchildren that would be born to a woman if she wereto live to the end of her childbearing years and bearchildren at each age in accordance with prevailingage-specific fertility rates. The rates given are fromthe sources mentioned in the note for Table 25. (Seethe Key for survey and census information.)
Births to women under age 20 and over age 35 areshown as a percentage of all births. These births areoften high risk because of the greater risk of compli-cations during pregnancy and childbirth. Childrenborn to very young or to older women are also morevulnerable.
The net reproduction rate (NRR), which measuresthe number of daughters a newborn girl will bear
242
during her lifetime (assuming fixed age-specific fer-tility and mortality rates) reflects the extent to whicha cohort of newborn girls will reproduce them-selves. An NRR of I indicates that fertility is at re-placement level: at this rate women will bear, on av-erage, only enough daughters to replace themselvesin the population.
Married women of childbearing age using con trucep-tion are women who are practicing, or whose hus-bands are practicing, any form of contraception.Contraceptive usage is generally measured for mar-ried women age 15 to 49. A few countries use mea-sures relating to other age groups, especially 15to 44.
Data are mainly derived from demographic andhealth surveys, contraceptive prevalence surveys,and World Bank country data. For a few countriesfor which no survey data are available and for sev-eral African countries, program statistics are used.Program statistics may understate contraceptiveprevalence because they do not measure use ofmethods such as rhythm, withdrawal, or absti-nence, or use of contraceptives not obtainedthrough the official family planning program. Thedata refer to rates prevailing in a variety of years,generally not more than three years before and oneyear after the year specified in the table.
All summary measures are country dataweighted by each country's share in the appropriatepopulation subgroup. Thus the crude birth (death)rate is weighted by the number of births (deaths) ineach country, and the total fertility rate and births towomen under 20 and over 35 are weighted by therelevant population subgroups.
Table 27. Health and nutrition
The estimates of population per physician and pernursing person are derived from World Health Orga-nization (WHO) data and are supplemented by dataobtained directly by the World Bank from nationalsources. The data refer to a variety of years, gener-ally no more than two years before the year speci-fied. Nursing persons include auxiliary nurses, aswell as paraprofessional personnel such as tradi-tional birth attendants. The inclusion of auxiliaryand paraprofessional personnel provides more real-istic estimates of available nursing care. Because de-finitions of doctors and nursing personnel varyand because the data shown are for a variety ofyearsthe data for these two indicators are notstrictly comparable across countries.
Low birthweight babies are children born weighingless than 2,500 grams. Low birthweight is frequentlyassociated with maternal malnutrition. It tends to
raise the risk of infant mortality and to lead to poorgrowth in infancy and childhood, thus increasingthe incidence of other forms of retarded develop-ment. The figures are derived from both WHO andU.N. Children's Fund (UNICEF) sources and arebased on national data. The data are not strictlycomparable across countries because they are com-piled from a combination of surveys and adminis-trative records that may not have representative na-tional coverage.
The infant mortality rate is the number of infantswho die before reaching one year of age, per thou-sand live births in a given year. The data are fromthe sources mentioned in the note to Table 25. (Seethe Key for survey and census information.)
Prevalence of malnutrition measures the percent-age of children under 5 with a deficiency or an ex-cess of nutrients that interfere with their health andgenetic potential for growth. Methods of assessmentvary, but the most commonly used are the follow-ing: less than 80 percent of the standard weight forage; less than minus 2 standard deviation from the50th percentile of the weight-for-age reference pop-ulation; and the Gomez scale of malnutrition. Notethat for a few countries the figures are for children 3or 4 years of age and younger.
The under-5 mortality rate shows the probabilitythat a newborn baby will die before reaching age 5.The rates are derived from life tables based on esti-mated current life expectancy at birth and on infantmortality rates. In general, throughout the worldmore males are born than females. Under good nu-tritional and health conditions and in times ofpeace, male children under 5 have a higher deathrate than females. These columns show thatfemalemale differences in the risk of dying by age5 vary substantially. In industrial market econo-mies, female babies have a 23 percent lower risk ofdying by age 5 than male babies, but the risk ofdying by age 5 is actually higher for females thanfor males in some lower-income economies. Thispattern is not uniformly associated with develop-ment. There are low- and middle-income countries(and regions within countries) where for example,the risk of dying by age 5 for females relative tomales approximates the pattern found in industrialcountries.
The summary measures in this table are countrydata weighted by the relevant population subgroup.
Table 28. Education
The data in this table refer to a variety of years, gen-erally not more than two years distant from thosespecified. Figures for females, however, sometimes
refer to a year earlier than that for overall totals. Thedata are mostly from UNESCO.
Primary school enrollment data are estimates ofthe ratio of children of all ages enrolled in primaryschool to the country's population of school-agechildren. Although many countries consider pri-mary school age to be 6 to 11 years, others do not.For some countries with universal primary educa-tion, the gross enrollment ratios may exceed 100percent because some pupils are younger or olderthan the country's standard primary school age.
The data on secondary school enrollment are cal-culated in the same manner, and again the defini-tion of secondary school age differs among coun-tries. It is most commonly considered to be 12 to 17years. Late entry of more mature students as well asrepetition and the phenomenon of "bunching" infinal grades can influence these ratios.
The tertiary enrollment ratio is calculated by di-viding the number of pupils enrolled in all post-secondary schools and universities by the popula-tion in the 20-24 age group. Pupils attending voca-tional schools, adult education programs, two-yearcommunity colleges, and distant education centers(primarily correspondence courses) are included.The distribution of pupils across these differenttypes of institutions varies among countries. Theyouth populationthat is, 20 to 24 yearshas beenadopted by UNESCO as the denominator, since itrepresents an average tertiary level cohort eventhough people above and below this age group maybe registered in tertiary institutions.
Primary net enrollment is the percentage of school-age children who are enrolled in school. Unlikegross enrollment, the net ratios correspond to thecountry's primary-school age group. This indicatorgives a much clearer idea of how many children inthe age group are actually enrolled in school with-out the numbers being inflated by over- or under-age children.
The primary pupilteacher ratio is the number ofpupils enrolled in school in a country divided by thenumber of teachers in the education system.
The summary measures in this table are countryenrollment rates weighted by each country's sharein the aggregate population.
Table 29. Gender comparisons
This table provides selected basic indicators disag-gregated to show differences between the sexes thatillustrate the condition of women in society. Themeasures reflect the demographic status of womenand their access to health and education services.Statistical anomalies become even more apparent
243
when social indicators are analyzed by gender be-cause reporting systems are often weak in areas re-lated specifically to women. Indicators drawn fromcensuses and surveys, such as those on population,tend to be about as reliable for women as for men;but indicators based largely on administrativerecords, such as maternal and infant mortality, areless reliable. More resources are now being devotedto developing better information on these topics,but the reliability of data, even in the series shown,still varies significantly.
The health and welfare indicators in Table 27 andin the maternal mortality column of Table 29 drawattention, in particular, to discrimination affectingwomen, especially very young girls, and to the con-ditions associated with childbearing. Childbearingstill carries the highest risk of death for women ofreproductive age in developing countries. The indi-cators reflect, but do not measure, both the availabil-ity of health services for women and the generalwelfare and nutritional status of mothers.
Life expectancy at birth is defined in the note toTable 1.
Maternal mortality refers to the number of femaledeaths that occur during childbirth per 100,000 livebirths. Because deaths during childbirth are definedmore widely in some countries to include complica-tions of pregnancy or the period after childbirth, orof abortion, and because many pregnant women diefrom lack of suitable health care, maternal mortalityis difficult to measure consistently and reliablyacross countries. The data are drawn from diversenational sources and collected by the World HealthOrganization (WHO), although many national ad-ministrative systems are weak and do not recordvital events in a systematic way. The data are de-rived mostly from official community reports andhospital records, and some reflect only deaths inhospitals and other medical institutions. Sometimessmaller private and rural hospitals are excluded,and sometimes even relatively primitive local facili-ties are included. The coverage is therefore not al-ways comprehensive, and the figures should betreated with extreme caution.
Clearly, many maternal deaths go unrecorded,particularly in countries with remote rural popula-tions. This accounts for some of the very low num-bers shown in the table, especially for severalAfrican countries. Moreover, it is not clear whetheran increase in the number of mothers in hospital re-flects more extensive medical care for women ormore complications in pregnancy and childbirth be-cause of poor nutrition, for instance. (Table 27shows data on low birth weight.)
244
These time series attempt to bring together read-ily available information not always presented in in-ternational publications. WHO warns that there areinevitably gaps in the series, and it has invited coun-tries to provide more comprehensive figures. Theyare reproduced here, from the 1991 WHO publica-tion Maternal Mortality: A Global Factbook. The datarefer to any year from 1983 to 1991.
The education indicators, based on UNESCOsources, show the extent to which females haveequal access to schooling.
Percentage of cohort persisting to grade 4 is the per-centage of children starting primary school in 1970and 1987, respectively, who continued to the fourthgrade by 1973 and 1990. Figures in italics representearlier or later cohorts. The data are based on en-rollment records. The slightly higher persistenceratios for females in some African countries mayindicate male participation in activities such as an-imal herding.
All things being equal, and opportunities beingthe same, the ratios for females per 100 males shouldbe close to 100. However, inequalities may cause theratios to move in different directions. For example,the number of females per 100 males will rise at sec-ondary school level if male attendance declinesmore rapidly in the final grades because of males'greater job opportunities, conscription into thearmy, or migration in search of work. In addition,since the numbers in these columns refer mainly togeneral secondary education, they do not capturethose (mostly males) enrolled in technical and voca-tional schools or in full-time apprenticeships, as inEastern Europe.
Females as a percentage of total labor force, based onILO data, shows the extent to which women are"gainfully employed" in the formal sector. Thesenumbers exclude homemakers and other unpaidcaregivers and in several developing countries re-flect a significant underestimate of female participa-tion rates.
All summary measures are country dataweighted by each country's share in the aggregatepopulation or population subgroup.
Table 30. Income distribution and PPP estimatesof GNP
The first columns report distribution of income orexpenditure accruing to percentile groups of house-holds ranked by total household income, per capitaincome, or expenditure. The last four columns con-tain estimates of per capita GNP based on purchas-ing power parities (PPPs) rather than exchange rates(see below for the definition of the PPP).
Columns 2 through 7 give the shares of popula-tion quintiles and the top decile in total income orconsumption expenditure for 45 low- and middle-income countries and 20 high-income countries. Thedata sets for these countries refer to different yearsbetween 1978 and 1992 and are drawn mostly fromnationally representative household surveys.
The data sets for the low- and middle-incomecountries have been compiled from two mainsources: government statistical agencies (oftenusing published reports) and the World Bank(mostly data originating from the Living StandardsMeasurement Study and the Social Dimensions ofAdjustment Project for Sub-Saharan Africa). Wherethe original unit record data from the householdsurvey were available, these have been used to cal-culate directly the income (or expenditure) shares ofdifferent quantiles; otherwise, the latter have beenestimated from the best available grouped data. Forfurther details on both the data and the estimationmethodology, see Chen, Datt, and Ravallion 1993.The data for the high-income OECD economies arebased on information from the Statistical Office ofthe European Union (Eurostat), The Luxembourg In-come Study, and the OECD. Those for other high-in-come countries come from national sources.
There are several comparability problems acrosscountries in the underlying household surveys.These problems are diminishing over time as surveymethodologies are both improving and becomingmore standardized, particularly under the initia-tives of the United Nations (under the HouseholdSurvey Capability Program) and the World Bank(under the Living Standard Measurement Studyand the Social Dimensions of Adjustment Project forSubSaharan Africa). The data presented hereshould nevertheless be interpreted with caution. Inparticular, the following three sources of noncompa-rability ought to be noted. First, the surveys differ inusing income or consumption expenditure as theliving standard indicator. For 28 of the 45 low- andmiddle-income countries, the data refer to con-sumption expenditure. Typically, income is moreunequally distributed than consumption. Second,the surveys differ in using the household or the in-dividual as their unit of observation; in the firstcase, the quantiles refer to percentage of householdsor per capita, rather than percentage of persons.Third, the surveys differ according to whether theunits of observation are ranked by household or percapita income (or consumption). The footnotes tothe table identify these differences for each country.
The 1987 indexed figures on PPPbased GNP percapita (US = 100) are presented in column 8. Note
two changes from previous editions: GDP has beenreplaced by GNP; and PPC (purchasing power ofcurrencies) by PPP. PPP is the term commonlyused to refer to the parities computed for a fixedbasket of products, even though theoretically theseare more appropriately labeled PPC. The data in-clude (a) results of the International ComparisonProgramme (ICP) Phase VI for 1990 for OECD coun-tries extrapolated backward to 1987; (b) results ofICP Phase V for 1985 for non-OECD countries ex-trapolated to 1987; (c) the latest available resultsfrom either Phase IV for 1980 or Phase III for 1975extrapolated to 1987 for countries that participatedin the earlier phases only; (d) a World Bank estimatefor China and the economies of the former SovietUnion; and (e) ICP estimates obtained by regressionfor the remaining countries that did not participatein any of the phases. Economies whose 1987 figuresare extrapolated from regression estimates are foot-noted.
The blend of extrapolated and regression-based1987 figures underlying column 8 is extrapolated to1992 using Bank estimates of real per capita GNPgrowth rates and expressed as an index (US = 100)in column 9. For countries that have ever partici-pated in the ICP as well as for China and theeconomies of the FSU, the latest available PPP-based values are extrapolated to 1992 by Bank esti-mates of growth rates and converted to current "in-ternational dollars" by scaling all results up by theU.S. inflation rates. The blend of extrapolated andregression-based 1992 estimates is presented in col-umn 10. Economies whose 1987 figures are extrapo-lated from another year or imputed by regressionare footnoted accordingly. The adjustments do nottake account of changes in the terms of trade.
The ICP recasts traditional national accountsthrough special price collections and disaggregationof GDP by expenditure components. ICP details areprepared by national statistical offices, and the re-sults are coordinated by the U.N. Statistical Division(UNSTAT) with support from other internationalagencies, particularly Eurostat and the OECD. TheWorld Bank, the Economic Commission for Europe,and the Economic and Social Commission for Asiaand the Pacific (ESCAP) also contribute to this exer-cise. For Nepal, which participated in the 1985 ex-ercise, total GDP data were not available, and com-parisons were made for consumption only.Luxembourg and Swaziland are the only twoeconomies with populations under 1 million thathave participated in the ICP; their 1987 results, as apercentage of the U.S. results, are 83.1 and 15.0, re-spectively. The next round of ICP surveys, for 1993,
245
is expected to cover more than 80 countries, includ-ing China and several FSU economies.
The "international dollar" (1$) has the same pur-chasing power over total GNP as the U.S. dollar in agiven year, but purchasing power over subaggre-gates is determined by average international pricesat that level rather than by U.S. relative prices.These dollar values, which are different from thedollar values of GNP or GDP shown in Tables I and3 (see the technical notes for these tables), are ob-tained by special conversion factors designed toequalize the purchasing powers of currencies in therespective countries. This conversion factor, the pur-chasing power parity (PPP), is defined as the num-ber of units of a country's currency required to buythe same amounts of goods and services in the do-mestic market as one dollar would buy in theUnited States. The computation involves derivingimplicit quantities from national accounts expendi-ture data and specially collected price data and thenrevaluing the implicit quantities in each country at asingle set of average prices. The average price indexthus equalizes dollar prices in every country so thatcross-country comparisons of GNP based on themreflect differences in quantities of goods and ser-vices free of price-level differentials. This procedureis designed to bring cross-country comparisons inline with cross-time real value comparisons that arebased on constant price series.
The ICP figures presented here are the results of atwo-step exercise. Countries within a region orgroup such as the OECD are first compared usingtheir own group average prices. Next, since groupaverage prices may differ from each other, makingthe countries in different groups not comparable,the group prices are adjusted to make them compa-rable at the world level. The adjustments, done byUNSTAT and Eurostat, are based on price differen-tials observed in a network of "link" countries rep-resenting each group. However, the linking is donein a manner that retains in the world comparisonthe relative levels of GDP observed in the groupcomparisons, called "fixity."
The two-step process was adopted because therelative GDP levels and rankings of two countriesmay change when more countries are brought intothe comparison. It was felt that this should not be al-lowed to happen within geographic regions; that is,that the relationship of, say, Ghana and Senegalshould not be affected by the prices prevailing in theUnited States. Thus overall GDP per capita levelsare calculated at "regional" prices and then linkedtogether. The linking is done by revaluing GDPs ofall the countries at average "world" prices and real-
246
locating the new regional totals on the basis of eachcountry's share in the original comparison.
Such a method does not permit the comparisonof more detailed quantities (such as food consump-tion). Hence these subaggregates and more detailedexpenditure categories are calculated using worldprices. These quantities are indeed comparable in-ternationally, but they do not add up to the indi-cated GDPs because they are calculated at a differ-ent set of prices.
Some countries belong to several regionalgroups. A few of the group have priority; others areequal. Thus fixity is maintained between membersof the European Union, even within the OECD andworld comparisons. For Austria and Finland, how-ever, the bilateral relationship that prevails withinthe OECD comparison is also the one used withinthe global comparison. But a significantly differentrelationship (based on Central European prices)prevails in the comparison within that group, andthis is the relationship presented in the separatepublication of the European comparison.
To derive ICP-based 1987 figures for countriesthat are yet to participate in any ICP survey, an esti-mating equation is first obtained by fitting the fol-lowing regression to 1987 data:
where all variables and estimated values are ex-pressed as US = 100 and where
r = ICP estimates of per capita GDP converted toU.S. dollars by PPP, the array of r consisting of ex-trapolations of the most recent actual ICP valuesavailable for countries that ever participated in ICP
ATLAS = per capita GNP estimated by the Atlasmethod
ENROL = secondary school enrollment ratio
RMSE = root mean squared error.
ATLAS and ENROL are used as rough proxies ofintercountry wage differentials for unskilled andskilled human capital, respectively. Following Isen-man 1980, the rationale adopted here is that ICPand conventional estimates of GDP differ mainlybecause wage differences persist among nationsdue to constraints on the international mobility oflabor, A technical paper (Ahmad 1992) providingfuller explanation is available on request. For fur-ther details on ICP procedures, readers may consult
the ICP Phase IV report, World Comparisons of Pur-chasing Power and Real Product for TI 980 (New York:United Nations, 1986). Readers interested in de-tailed ICP survey data for 1975, 1980, 1985, and1990 may refer to Purchasing Power of Currencies:Com paring National Incomes Using ICP Data (WorldBank 1993).
Table 31. Urbanization
Data on urban population and agglomeration inlarge cities are from the U.N.'s World UrbanizationProspects, supplemented by data from the WorldBank. The growth rates of urban population are cal-culated from the World Bank's population esti-mates; the estimates of urban population shares arecalculated from both sources just cited.
Because the estimates in this table are based ondifferent national definitions of what is urban, cross-country comparisons should be made with caution.
The summary measures for urban population asa percentage of total population are calculated fromcountry percentages weighted by each country'sshare in the aggregate population. The other sum-mary measures in this table are weighted in thesame fashion, using urban population.
Table 32. Infrastructure
This table provides selected basic indicators of thecoverage and performance of infrastructure sectors.Coverage. Indicators of coverage are based on the in-frastructure data most widely available across coun-tries which measure the extent, type, and sometimescondition of physical facilities in each infrastructuresector (examples are provided in the Appendix ta-bles). Such data are divided by national populationtotals to derive indicators of coverage or availability(as in telephone main lines per thousand persons orroad kilometers per million persons). More directmeasures of coverage are based on household sur-veys of actual access, reported as percentage ofhouseholds with electricity or access to safe water.For roads and railways, physical proximity (such asshare of population within 1 kilometer of a pavedroad) would be a good measure of coverage, but it israrely available.
Performance. Performance quality should be as-sessed from the perspectives of both the infrastruc-ture providers and the users. Indicators from theproviders' perspective measure operating efficiency(such as power system losses, unaccounted-forwater, and locomotive availability), capacity utiliza-tion, or financial efficiency (such as cost recovery).
Indicators from the users' perspectives would mea-sure the effectiveness of the service ultimately deliv-ered. Service quality indicators (such as faults per100 main lines per year) are the most difficult datato obtain on a comparable and recurrent basis for alarge sample of countries. Some indicators representboth system efficiency and service quality, such asthe share of paved roads in good condition.
Although the data reported here are drawn fromthe most authoritative sources available, compara-bility may be limited by variation in data collection,statistical methods, and definitions.
Electric power. Coverage is measured by the per-centage of households with access to electricity suf-ficient for at least electric lighting. This indicator isfrom Kurian 1991 and is available only for 1984. Sys-tem losses, which are obtained from the "PowerData Sheets" compiled by the Industry and EnergyDepartment of the World Bank and lEA Energy Sta-tistics, combine technical and nontechnical losses.Technical losses, due to the physical characteristicsof the power system, consist mainly of resistancelosses in transmission and distribution. Nontechni-cal losses consist mainly of illegal connection toelectricity and other sources of theft. System lossesare expressed as percentage of total output (net gen-eration).
Telecommunications. The measure of coverage isthe number of telephone exchange mainlines perthousand persons. A telephone mainline connectsthe subscriber's equipment to the switched net-work and has a dedicated port in the telephone ex-change. This term is synonymous with "main sta-tion," also commonly used in telecommunicationdocuments. Faults per 100 main lines per year referto the number of reported faults per 100 main tele-phone lines for the year indicated. Some operatorsinclude malfunctioning customer premises equip-ment as faults, while others include only technicalfaults. Data on main lines and faults per 100 mainlines are from the International TelecommunicationUnion database.
Roads. Indicators used to represent coverage inthis sector include spatial road density (a country'sroad length divided by land area) and per capitaroad density (length of the road network per popu-lation size). The latter measure (kilometers of pavedroads per million population) is used here as an ap-proximate indicator of coverage. As the measure ofperformance, paved roads in good condition is de-fined as roads substantially free of defects and re-quiring only routine maintenance. Data for pavedroads are from Queiroz and Gautam 1992 and areavailable for 1988 only.
247
Water. For most countries, the percentage of thepopulation with access to safe water either bystandpipe or house connection is the measure ofcoverage and is drawn primarily from the WorldHealth Organization's The International DrinkingWater Supply and Sanitation Decade series, variousyears. For the economies of the FSU, the percentageof public housing equipped with running water isthe measure of coverage, and the source is HousingConditions in the USSR, published by the State Com-mittee on Statistics for the USSR. Data for waterlosses are from Garn 1987 and are for metropolitanarea systems. Where 1986 data were not available,the closest available year was taken. Water losses in-clude physical losses (pipe breaks and overflows)and commercial losses (meter underregistration, il-legal use including fraudulent or unregistered con-nections, and legal, but usually not metered, usessuch as firefighting).
Railways. The coverage indicator is the number ofrail traffic units per million U.S. dollars GDP. Railtraffic units are the sum of passenger-kilometers andton-kilometers and were obtained from the databasemaintained by the Transport Division of the Trans-port, Water, and Urban Development Department,World Bank. Diesel locomotive availability is one ofthe better measures of technical and managerial per-formance because locomotives are the most expen-sive rolling stock the railways own. Data for diesellocomotive availability as a percentage of diesel in-ventory are from the same World Bank database.GDP figures are from Summers and Heston, ThePenn World Tables (Mark 5.5), forthcoming.
Table 33. Natural resources
This table represents a step toward including envi-ronmental data in the assessment of developmentand the planning of economic strategies. It providesa partial picture of the status of forests, the extent ofareas protected for conservation or other environ-mentally related purposes, and the availability anduse of fresh water. The data reported here are drawnfrom the most authoritative sources available,which are cited in World Resources Institute, WorldResources 1994-95. Perhaps even more than otherdata in this Report, however, these data should beused with caution. Although they accurately charac-terize major differences in resources and usesamong countries, true comparability is limited be-cause of variation in data collection, statistical meth-ods, definitions, and government resources.
No conceptual framework that integrates naturalresource and traditional economic data has yet been
248
agreed on. Nor are the measures shown in this tableintended to be final indicators of natural resourcewealth, environmental health, or resource deple-tion. They have been chosen because they are avail-able for most countries, are testable, and reflectsome general conditions of the environment.
The natural forest total area refers to natural standsof woody vegetation in which trees predominate.These estimates are derived from country statisticsassembled by the Food and Agriculture Organiza-tion (FAQ) and the United Nations Economic Com-mission for Europe (UNECE). New assessmentswere published in 1993 for tropical countries (FAQ)and temperate zones (UNECE/FAO). The FAO andthe UNECE/FAO use different definitions in theirassessments. The FAQ defines natural forest in trop-ical countries as either a closed forest where treescover a high proportion of the ground and there isno continuous grass cover or an open forest, definedas mixed forest/grasslands with at least 10 percenttree cover and a continuous grass layer on the forestfloor. A tropical forest encompasses all stands ex-cept plantations and includes stands that have beendegraded to some degree by agriculture, fire, log-ging, or acid precipitation.
UNECE/FAO defines a forest as land where treecrowns cover more than 20 percent of the area. Alsoincluded are open forest formations; forest roadsand firebreaks; small, temporarily cleared areas;young stands expected to achieve at least 20 percentcrown cover on maturity; and windbreaks and shel-terbelts. Plantation area is included under temper-ate country estimates of natural forest area. Somecountries in this table also include other woodedland, defined as open woodland and scrub, shrub,and brushland.
Deforestation refers to the permanent conversionof forest land to other uses, including shifting culti-vation, permanent agriculture, ranching, settle-ments, or infrastructure development. Deforestedareas do not include areas logged but intended forregeneration or areas degraded by fuelwood gather-ing, acid precipitation, or forest fires. The extent andpercentage of total area shown refer to the averageannual deforestation of natural forest area.
Some countries also conduct independent assess-ments using satellite data or extensive ground data.A 1991 country-wide assessment using Landsat im-agery estimated India's forest cover at 639,000square kilometers. An inventory based on 1990LANDSAT TM imagery estimated Mexico's forestcover at 496,000 square kilometers, with a deforesta-tion rate of 4.06 square kilometers per year between1980 and 1990. In Brazil two recent satellite-
imagery-based assessments of deforestation in theBrazilian Amazon have resulted in different defor-estation rate estimates for this region. A study bythe U.S. National Space and Aeronautics Adminis-tration (NASA) and the University of New Hamp-shire estimated forest loss at 15,000 square kilome-ters per year during 1978-88. Brazil's NationalInstitute for Space Research (INPE) and National In-stitute for Research in the Amazon (INPA) esti-mated deforestation at 20,300 square kilometers peryear for the same period. Deforestation in sec-ondary forest areas and dry scrub areas were not in-cluded in either study. The FAQ data presented inthis table include forestation in all Brazil, includingsecondary forest areas and other forested areas.Note also that according to the FAQ Brazil has an es-timated 70,000 square kilometers of plantation land,defined as forest stands established artificially by af-forestation and reforestation for industrial and non-industrial usage. India has an estimated 189,000square kilometers of plantation land and Indonesiaan estimated 87,500 square kilometers.
Nationally protected areas are areas of at least 1,000hectares that fall into one of five management cate-gories: scientific reserves and strict nature reserves;national parks of national or international signifi-cance (not materially affected by human activity);natural monuments and natural landscapes withsome unique aspects; managed nature reserves andwildlife sanctuaries; and protected landscapes andseascapes (which may include cultural landscapes).This table does not include sites protected underlocal or provincial law or areas where consumptiveuses of wildlife are allowed. These data are subjectto variations in definition and in reporting to the or-ganizations, such as the World Conservation Moni-toring Centre, that compile and disseminate them.Total surface area is used to calculate the percentageof total area protected.
Freshwater resources: annual withdrawal data aresubject to variation in collection and estimationmethods but accurately show the magnitude ofwater use in both total and per capita terms. Thesedata, however, also hide what can be significantvariation in total renewable water resources fromone year to another. They also fail to distinguish theseasonal and geographic variations in water avail-ability within a country. Because freshwater re-sources are based on long-term averages, their esti-
mation explicitly excludes decade-long cycles of wetand dry. The Département Hydrogeologie in Qr-leans, France, compiles water resource and with-drawal data from published documents, includingnational, United Nations, and professional litera-ture. The Institute of Geography at the NationalAcademy of Sciences in Moscow also compilesglobal water data on the basis of published workand, where necessary, estimates water resources andconsumption from models that use other data, suchas area under irrigation, livestock populations, andprecipitation. These and other sources have beencombined by the World Resources Institute to gen-erate data for this table. Withdrawal data are for sin-gle years and vary from country to country between1970 and 1992. Data for small countries and coun-tries in arid and semiarid zones are less reliable thanthose for larger countries and countries with higherrainfall.
Total water resources include both internal renew-able resources and, where noted, river flows fromother countries. Estimates are from 1992. Annual in-ternal renewable water resources refer to the aver-age annual flow of rivers and aquifers generatedfrom rainfall within the country. The total with-drawn and the percentage withdrawn of the totalrenewable resource are both reported in this table.Withdrawals include those from nonrenewableaquifers and desalting plants but do not includelosses from evaporation. Withdrawals can exceed100 percent of renewable supplies when extractionsfrom nonrenewable aquifers or desalting plants areconsiderable or if there is significant water reuse.Total per capita water withdrawal is calculated bydividing a country's total withdrawal by its popula-tion in the year for which withdrawal estimates areavailable. For most countries, sectoral per capitawithdrawal data are calculated using sectoral with-drawal percentages estimated for 1987. Domesticuse includes drinking water, municipal use or sup-ply, and use for public services, commercial estab-lishments, and homes. Direct withdrawals for in-dustrial use, including withdrawals for coolingthermoelectric plants, are combined in the final col-umn of this table with withdrawals for agriculture(irrigation and livestock production). Numbers maynot sum to the total per capita figure because ofrounding.
U.N. Department of International Economic and Social Affairs. Various years. Statistical Yearbook. NewYork.
Various years. Energy Statistics Yearbook. Statistical Papers, series J. New York.U.N. International Comparison Program Phases IV (1980), V (1985), and VI (1990) reports, and data
from ECE, ESCAP, Eurostat, OECD, and U.N.FAO, IMF, UNIDO, and World Bank data; national sources.
International Monetary Fund. Government Finance Statistics Yearbook. Vol. 11. Washington, D.C.Various years. International Financial Statistics. Washington, D.C.
U.N. Department of International Economic and Social Affairs. Various years. World Energy Supplies.Statistical Papers, series J. New York.
IMF data.
International Monetary Fund. Various years. International Financial Statistics. Washington, D.C.U.N. Conference on Trade and Development. Various years. Handbook of International Trade and
Development Statistics. Geneva.U.N. Department of International Economic and Social Affairs. Various years. Monthly Bulletin of
Statistics. New York.Various years. Yearbook of International Trade Statistics. New York.
FAQ, IMF, U.N., and World Bank data.
Bos, Eduard, My T. Vu, Ernest Massiah, and Rodolfo A. Bulatao. World Population Projections, 1994-95Edition (forthcoming). Baltimore, Md.: Johns Hopkins University Press.
Garn, Harvey. 1987. "Patterns in the Data Reported on Completed Water Supply Projects." WorldBank, Transport, Water, and Urban Development Department, Washington, D.C.
Heiderian, J., and Wu, Gary. 1993. "Power Sector: Statistics of Developing Countries (1987-1991)."World Bank, Industry and Energy Department, Washington, D.C.
Institute for Resource Development/Westinghouse. 1987. Child Survival: Risks and the Road to Health.Columbia, Md.
International Energy Agency. 1993. lEA Statistics: Energy Prices and Taxes. Paris: OECD.International Road Transport Union. 1990. World Transport Data.International Telecommunication Union. 1994 World Telecommunications Development Report. Geneva.Kurian, G. T. 1991. The New Book of World Rankings. New York: Facts on File.Querioz, Caesar, and Surhid Gautam. 1992. "Road Infrastructure and Economic Development." Policy
Research Working Paper 921. World Bank, Washington, D.C.Ross, John, and others. 1993. Family Planning and Population: A Compendium of International Statistics.
New York: The Population Council.Sivard, Ruth. 1985. Women-A World Survey. Washington, D.C.: World Priorities.U.N. Department of Economic and Social Information and Policy Analysis (formerly U.N. Department
of International Economic and Social Affairs). Various years. Demographic Yearbook. New York.Various years. Population and Vital Statistics Report. New York.Various years. Statistical Yearbook. New York.1989. Levels and Trends of Contraceptive Use as Assessed in 1988. New York.1988. Mortality of Children under Age 5: Projections 1950-2025. New York.1986. World Comparisons of Purchasing Power and Real Product for 1980. New York.1991. World Urbanization Prospects: 1991. New York.1991. World Population Prospects: 1990. New York.1993. World Population Prospects: 1993 Revision. New York.1993. World Urbanization Prospects: 1992 Revision. New York.
U.N. Educational Scientific and Cultural Organization. Various years. Statistical Yearbook. Paris.1990. Compendium of Statistics on Illiteracy. Paris.
UNICEF 1989. The State of the World's Children 1989. Oxford: Oxford University Press.World Bank. 1993. Purchasing Power of Currencies: Comparing National Incomes Using ICP Data.
Washington, D.C.World Health Organization. Various years. World Health Statistics Annual. Geneva.
1986. Maternal Mortality Rates: A Tabulation of Available Information, 2nd edition. Geneva.1991. Maternal Mortality: A Global Fact book. Geneva.Various years. World Health Statistics Report. Geneva.Various years. The International Drinking Water Supply and Sanitation Decade. Geneva.
World Resources Institute. 1994. World Resources 1994-95. New York.FAQ, ILO, U.N., and World Bank data; national sources.
External Organization for Economic Cooperation and Development. Various years. Development Co-operation.finance Paris.
1988. Geographical Distribution of Financial Flows to Developing Countries. Paris.IMF, OECD, and World Bank data; World Bank Debtor Reporting System.
Part 1 Classification of economies by income and region
(Table continues on thefo lowing page)
251
Incomegroup Subgroup
Sub-Saharan Africa Asia Europe and Central Asia Middle East and North Africa
Americas
East and EasternSouthern East Asia Europe and Rest of Middle NorthAfrica West Africa and Pacific South Asia Central Asia Europe East Africa
Former Yugoslav Republic of Macedonia.Other Asian economiesTaiwan, China.
Definitions of groups
These tables classify all World Bank member economies, and all othereconomies with populations of more than 30,000.
Income group: Economies are divided according to 1992 GNP percapita, calculated using the WorldBankAtlas method. The groups are:low-income, $675 or less; lower-middle-income, $676-2,695;
252
upper-middle-income, $2,696$8,355; and high-income, $8,356 ormore.
The estimates for the republics of the former Soviet Union arepreliminary and their classification will be kept under review.
Incomegroup Subgroup
Sub-Saharan Africa Asia Europe and Central Asia Middle East and North Africa
Amencas
East and EasternSouthern East Asia Europe and Rest of Middle NorthAfrica West Africa and Pacific SouthAsis Central Asia Europe East Africa
Other Asian economiesTaiwan, China.Economies in which no single export category accounts for more than 50 percent of total exports.Former Yugoslav Republic of Macedonia.
Definitions of groups
These tables classify all World Bank member economies, plus all othereconomies with populations of more than 30,000.
Major export category: Major exports are those that account for 50percent or more of total exports of goods and services from onecategory, in the period 1987-91. The categories are: nonfuel primary(SITC 0,1,2, 4, plus 68), fuels (SITC 3), manufactures (SITC 5 to 9,less 68), and services (factor and nonfactor service receipts plusworkers' remittances). If no single category accounts for 50 percentor more of total exports, the economy is classified as diversified.
Indebtedness: Standard World Bank definitions of severe andmoderate indebtedness, averaged over three years (1990-92) are usedto classify economies in this table. Severely indebted means either ofthe two key ratios is above critical levels: present value of debt service
254
to GNP (80 percent) and present value of debt service to exports (220percent). Moderately indebted means either of the two key ratiosexceeds 60 percent of, but does not reach, the critical levels. Foreconomies that do not report detailed debt statistics to the World BankDebtor Reporting System, present-value calculation is not possible.Instead the following methodology is used to classify the non-DRSeconomies. Severely indebted means three of four key ratios (averagedover 1990-92) are above critical levels: debt to GNP (50 percent); debtto exports (275 percent), debt service to exports (30 percent); andinterest to exports (20 percent). Moderately indebted means three offour key ratios exceed 60 percent of, but do not reach, the criticallevels. All other classified low- and middle-income economies arelisted as less-indebted.
ECONOMIC INFRASTRUCTUREincluding transport, electric power, telecommunications,water and sanitation, and waste disposalproduces services that are vital to development. Theavailability of infrastructure has increased significantly in developing countries over recentdecades. Yet performance is often poor, inadequate maintenance leads to premature deteriora-tion of facilities, and services frequently do not match users' needs and willingness to pay. Thepotential for infrastructure services to contribute to poverty reduction and to environmentalimprovement is often neglected. The time is ripe to broaden the focus of policy beyond increas-ing the quantity of infrastructure facilities to improving the quality and efficiency of services.
This seventeenth annual World Development Report examines the record of both successes andfailures in infrastructure provision and concludes that obtaining more effective and efficientdelivery of services requires important changes in incentives. The Report identifies three waysof reforming institutions to improve incentives:
Apply principles of commercial managementoperating infrastructure more like a busi-ness than a bureaucracy, with clear objectives, with managerial and financial autonomy; and withcustomer satisfaction a key measure of performance.
Introduce competitiondirectly where feasible, indirectly where not. By giving users moreoptions, competition makes providers more efficient and more accountable to users.
Give users and other stakeholders a strong voice and real responsibility in planning, oper-ating, regulating, and financing services.
Governments will need to act more often as facilitator, coordinator, and guardian of the pub-lic interest (especially to ensure access for the poor)and less often as direct operator. The pri-vate sector will need to provide management expertise and a significantly greater share of financethan in the past. From a broad menu of specific institutional arrangements for infrastructure pro-vision, the Report assesses four options by sector and type of country. The Report notes thatimproved performance of infrastructure can enable countries to achieve substantial efficiencygains, a higher level of service from existing infrastructure resources, and improved access toessential services for the poor.
This Report includes the World Development Indicators, which offer comprehensive, currentdata on social and economic development in more than 200 countries and territories. TheIndicators are also available on diskette. An appendix to the Report provides selected statisticson infrastructure.
COVER DESIGN BY BRIAN NOYES / TI-4E MAGAZINE GROUP