a)~~~C a)~C l 0 gf ~E'c* . CL) 'io Recent World Bank Technical Papers No. 370 Dejene, Shishira, Yanda, and Johnsen, Land Degradation in Tanzania: Perception from the Village No. 371 Essama-Nssah, Analyse d'une repartition du niveau de vie No. 372 Cleaver and Schreiber, Inverser la spriale: Les interactions entre la population, l'agriculture et l'environnement en Afrique subsaharienne No. 373 Onursal and Gautam, Vehicular Air Pollution: Experiencesfrom Seven Latin American Urban Centers No. 374 Jones, Sector Investment Programs in Africa: Issues and Experiences No. 375 Francis, Milimo, Njobvo, and Tembo, Listening to Farmers: Participatory Assessment of Policy Reform in Zambia's Agriculture Sector No. 376 Tsunokawa and Hoban, Roads and the Environment: A Handbook No. 377 Walsh and Shah, Clean Fuelsfor Asia: Technical Options for Moving toward Unleaded Gasoline and Low-Sulfur Diesel No. 378 Shah and Nagpal, eds., Urban Air Quality Management Strategy in Asia: Kathmandu Valley Report No. 379 Shah and Nagpal, eds., Urban Air Quality Management Strategy in Asia: Jakarta Report No. 380 Shah and Nagpal, eds., Urban Air Quality Management Strategy in Asia: Metro Manila Report No. 381 Shah and Nagpal, eds., Urban Air Quality Management Strategy in Asia: Greater Mumbai Report No. 382 Barker, Tenenbaum, and Woolf, Governance and Regulation of Power Pools and System Operators: An International Comparison No. 383 Goldman, Ergas, Ralph, and Felker, Technology Institutions and Policies: Their Role in Developing Technological Capability in Industry No. 384 Kojima and Okada, Catching Up to Leadership: The Role of Technology Support Institutions in Japan's Casting Sector No. 385 Rowat, Lubrano, and Porrata, Competition Policy and MERCOSUR No. 386 Dinar and Subramanian, Water Pricing Experiences: An International Perspective No. 387 Oskarsson, Berglund, Seling, Snellman, Stenback, and Fritz, A Planner's Guidefor Selecting Clean-Coal Technologiesfor Power Plants No. 388 Sanjayan, Shen, and Jansen, Experiences with Integrated-Conservation Development Projects in Asia No. 389 International Commission on Irrigation and Drainage (ICID), Planning the Management, Operation, and Maintenance of Irrigation and Drainage Systems: A Guide for the Preparation of Strategies and Manuals No. 390 Foster, Lawrence, and Morris, Groundwater in Urban Development: Assessing Management Needs and Formulating Policy Strategies No. 391 Lovei and Weiss, Jr., Environmental Management and Institutions in OECD Countries" Lessonsfrom Experience No. 392 Felker, Chaudhuri, Gy6rgy, and Goldman, The Pharmaceutical Industry in India and Hungary: Policies, Institutions, and Technological Development No. 393 Mohan, ed., Bibliography of Publications: Africa Region, 1990-97 No. 394 Hill and Shields, Incentivesfor Joint Forest Management in India: Analytical Methods and Case Studies No. 395 Saleth and Dinar, Satisfying Urban Thirst: Water Supply Augmentation and Pricing Policy in Hyderabad City, India No. 396 Kikeri, Privatization and Labor: What Happens to Workers When Governments Divest? No. 397 Lovei, Phasing Out Leadfrom Gasoline: Worldwide Experience and Policy Implications No. 398 Ayres, Anderson, and Hanrahan, Setting Prioritiesfor Environmental MAnagement: An Application to the Mining Sector in Bolivia No. 399 Kerf, Gray, Irwin, Levesque, Taylor, and Klein, Concessionsfor Infrastructure: A Guide to Their Design and Award No. 401 Benson and Clay, The Impact of Drought on Sub-Saharan African Economies: A Preliminary Examination No. 402 Dinar, Mendelsohn, Evenson, Parikh, Sanghi, Kumar, McKinsey, and Lonergan, Measuring the Impact of Climate Change on Indian Agriculture No. 403 Welch and Fr6mond, The Case-by-Case Approach to Privatization: Techniques and Examples No. 404 Stephenson, Donnay, Frolova, Melnick, and Worzala, Improving Women's Health Services in the Russian Federation: Results of a Pilot Project (List continues on the inside back cover) WORLD BANK TECHNICAL PAPER NO. 446 Eastern Europe and CentralAsia Pollution Management Series Economic Reform and Environmental Performance in Transition Economies Gordon Hughes Magda Lovei The World Bank Washington, D.C. Copyright © 1999 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing September 1999 Technical Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown 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. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission promptly. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use, is granted by the World Bank, provided that the appropriate fee is paid directly to Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, U.S.A., telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the World Bank at the address above or faxed to 202-522-2422. ISBN: 08213-4564-8 ISSN: 0253-7494 Cover photo by Curt Carnemark, the World Bank, Estonia, 1993. Gordon Hughes is senior advisor and Magda Lovei is an environmental economist in the Environment Department at the World Bank. Library of Congress Cataloging-in-Publication Data has been applied for. Contents Foreword vii Abstract ix Acknowledgments xi Abbreviations and Acronyms xiii Executive Summary xv Chapter 1 Overview of Progress in Economic Transition I Reform, economic growth, and restructuring 2 Trends in industrial output and energy intensity 5 Fuel use: Composition and trends 7 Chapter 2 Air Pollution 11 Trends in total emissions 11 Emissions intensities 13 Ambient air quality in urban areas 14 Air pollution trends in hot spots 15 Chapter 3 Water Pollution 19 Trends in emissions and water quality 19 Water pollution trends in hot spots 21 Chapter 4 The Health Impacts and Costs of Environmental Damage 23 Air pollution 23 Water-related health impacts 25 The overall burden of environmental damage 28 Annexes A. Analyzing Trends in Environmental Quality, Emissions and their Determinants 31 B. Estimating the Health Impacts of Air Pollution 35 C. Estimating the Health Impacts of Access to Water and Sanitation 43 D. Tables 49 iii iv Economic Reform and Environmental Performance in Transition Economies Notes 55 References 59 Boxes 1. Reform and environmental benefits 2 2. Perverse incentives against industrial restructuring in "virtual economies" 3 3. Pan-European strategy to phase out leaded gasoline 13 4. Industrial production and air quality trends in two hot spots 16 5. Demand management and water consumption in Central and Eastern Europe 20 6. No improvement in some hot spots despite declining industrial emissions 22 7. Economic damage from fuel combustion in Krakow, Poland 25 B.1 Focus on airborne particulate matter 37 B.2 DALYs as a measure of the burden of disease 41 Figures 1. Average GDP growth rates between 1989 and 1997/98 4 2. GDP and industrial output by country group, 1989-98 5 3. Energy consumption and GDP in transition economies compared with international trend, 1990 6 4. "Excess" energy consumption by country group, 1990 and 1995 6 5. "Excess" energy consumption in selected advanced- and slower-reforming countries, 1990 and 1995 6 6. Trends in primary energy consumption, 1989-95 7 7. Indices of fuel use by small sources in advanced reform countries, 1989-95 8 8. Use of fuels by small sources in advanced reform countries, 1989-95 8 9. Indices of emissions of air pollutants in advanced reform countries, 1989-96 11 10. Emissions of air pollutants in slower-reforming countries, 1989-95 11 11. Emissions intensities for air pollutants in advanced reform countries, 1989-95 14 12. Emissions intensities for air pollutants in slower-reforming countries, 1989-95 14 13. Trends in urban air quality by country group, 1989-96 14 14. Trends in emissions of air pollutants in hot spots, 1990-96 15 15. Trends in urban air quality in hot spots, 1990-96 15 16. Indicators of water pollution, 1990-96 19 17. Emissions of water pollutants in hot spots, 1990-96 21 18. Surface water quality in hot spots, 1990-96 21 19. Excess mortality due to air pollution by country group, 1990 and 1995 24 20. Excess cases of chronic bronchitis due to air pollution by country group, 1990 and 1995 24 21. Total costs of air pollution by country group, 1990 and 1995 25 22. Infant mortality rates by country group, 1990 and 1995 26 23. Actual versus expected rates of infant mortality by country group, 1995 27 24. Estimated impact of improvement in water and sanitation on infant mortality by country group, 1995 27 25. Total burden of mortality and illness associated with environmental factors by country group, 1995 28 26. Relative burden of mortality and illness associated with environmental factors by country group, 1995 28 27. Relative burden of mortality and illness, selected countries 29 28. Relative burden of mortality and illness for the NIS Northwest countries 29 29. Relative burden of mortality and illness for the NIS Southeast countries 29 Contents v Tables 1. Total lead emissions in selected transition economies, 1990 and 1996 12 A.1 Estimated coefficients for emissions of air pollutants 34 B.1 Dose response coefficients used in the study 39 C.1 Variables used in the analysis 46 C.2 Equations for mortality of infants and children under age 5 47 C.3 Impact of water and sanitation on mortality rates in the ECA Region 48 D.1 GDP growth and GDP per capita in Central and Eastern Europe and the Newly Independent States 49 D.2 Inflation in Central and Eastern Europe and the Newly Independent States, 1989/93 and 1993/97 50 D.3 Motor vehicle use and change in Central and Eastern Europe and the Newly Independent States, 1990-96 51 D.4 Mortality and life expectancy in the Newly Independent States, Central and Eastern Europe, and the European Union 52 D.5 Dissolved oxygen levels in selected rivers 53 Foreword he transition to a market economy is expect- economies and to the less encouraging trends in slower- ed to lead to environmental as well as eco- reforming countries. In the latter group, pressures on nomic improvement. In practice, realization the environment decreased somewhat during transi- of the expected gains may come slowly, especially for tion only as a consequence of declining economic ac- countries where economic reform and growth have tivity, while persistent problems remained in lagged. This Technical Paper reviews progress in en- environmental "hot spots" around high-polluting in- vironmental trends since transition began. It looks at dustrial centers. air and water pollution and health indicators in the This report continues and builds on the World region, over time, in comparison with worldwide Bank's work in analyzing the environmental effects trends and in light of the environmental issues iden- of transition, restructuring, and privatization with a tified in the Environmental Action Programme for view to identifying priority areas for investment and Central and Eastern Europe. policy initiatives. As this study emphasizes, economic reform and commitment to introducing environmental regula- Kein Cleaver tions and programs are vital for reaching lasting en- Director vironmental improvements. The paper points to the Europe and Central Asia solid progress achieved in the advanced reform Environmentally and Socially Sustainable Development vii Abstract conomic transition to a market economy is ex- was used, along with other indicators, to assess the pected to yield environmental benefits as health effects of changes in environmental conditions changing incentives foster more efficient pro- under the transition. duction, better use of resources, and increased com- Damages from air pollution have declined in the munity input. A review of environmental and health CEE group but have increased in the NIS, partly be- conditions in the countries of Central and Eastern cause of the general worsening of the health status of Europe (CEE) and the newly industrialized states the population. Water quality depends greatly on lo- (NIS) shows that pollution has declined across the cal infrastructure, which has deteriorated in the area, but for different reasons. In the slower-reform- economically hard-hit NIS. Nevertheless, infant ing countries, reduced pollution may be simply the mortality rates do not indicate any increase in water- temporary effect of an economic downturn and could related deaths and disease, except in some countries. be erased when growth resumes. By contrast, the ad- Although use of agricultural fertilizers and pesticides vanced reformers, mainly in the CEE group, have might be expected to lead to a decrease in nitrate-lad- made solid gains in improving energy efficiency and en runoff, nitiate levels in rural drinking water do not reducing emissions intensity of pollutants. Both appear to have improved significantly. groups of countries inherited environmental "hot Air pollution and dirty water are more serious spots" where highly polluting activities cause poor problems in the NIS than in the CEE group. Within environmental quality. the NIS, air pollution is the more important problem The report looks at the key environmental issues in the northwestem part, and clean water is a high identified in the 1993 Environmental Action Program priority in the southeastem countries, especially in for Central and Eastern Europe, including air and rural areas. International and country efforts should water pollution and the associated health burdens. reflect these differing concerns and needs. The concept of a disability-adjusted life-year (DALY) ix Acknowledgments T his report was prepared as a contribution to Steedman for data analysis, and Milen Dyoulgerov a forthcoming report by the Organisation for for research assistance. The report greatly benefited Economic Co-operation and Development from constructive and helpful comments by Michele (OECD), Environment in the Transition to a Market Econ- de Nevers, David Hanrahan, Kseniya Lvovsky, and omy. The preparation of the report was supported by Paavo Eliste from the World Bank; Brendan Gillespie the Environmentally and Socially Sustainable Devel- and others from OECD; and experts from Central and opment Sector Unit of the Europe and Central Asia Eastern Europe and the newly independent states who Region, and the Environment Department of the reviewed the paper. The authors would like to ex- World Bank. press their thanks to Nancy Levine for editorial assis- The authors thank Tony Zamparutti and Glen tance, Jim Cantrell for desktop publishing, and to Anderson for contributions and comments, Jennifer Sriyani Cumine for administrative support. xi Abbreviations and Acronyms CB Chronic bronchitis CEE Central and Eastern Europe COI Cost of illness CO2 Carbon dioxide CVM Contingent valuation method DALY Disability-adjusted life year EAP Environmental Action Programme EBRD European Bank for Reconstruction and Development ECA Europe and Central Asia Region EU European Union GDP Gross domestic product IEA International Energy Agency MPC Maximum permitted concentration NIS Newly independent states NOx Nitrogen oxides OMR Operation, maintenance, and repair PM Particulate matter PM2s 9 Particulate matter smaller than 2.5 microns in diameter PMIO Particulate matter smaller than 10 microns in diameter QWB Quality of well being SO2 Sulfur dioxide TSP Total suspended particulates VOSL Valuation of a statistical life UNECE United Nations Economic Commission for Europe USEPA United States Environmental Protection Agency VOC Volatile organic compound WHO World Health Organization WTO World Trade Organization WTP Willingness to pay xiii Executive Summary T he transition from centrally planned to market region. However, in advanced reform countries, total economies started almost a decade ago in Cen- emissions of key air pollutants such as particulates, tral and Eastern Europe (CEE) and the newly sulfur dioxide, and nitrogen oxides have fallen more independent states (NIS). As emphasized by the than gross domestic product (GDP), while in the slow- Environmental Action Programme for Central and er-reforming countries they have fallen approximate- Eastern Europe-a document endorsed by the envi- ly in line with GDP. The analysis presented in the ronmental ministers of 50 countries in 1993-the eco- paper indicates that advanced reform countries have nomic reform and restructuring associated with the managed to achieve a clear and sustained reduction transition were expected to eliminate the perverse in the emissions intensities of key air pollutants; that incentives that underlay many of the environmental is, the changes associated with economic reform have problems of centrally planned economies. It was also produced real and lasting environmental benefits in recognized, however, that economic reform had to be these countries. By contrast, there has been little evi- harnessed by effective environmental policies, dence of any significant improvement in emissions institutions, and investments to achieve lasting envi- intensities in slower-reforming countries. ronmental improvements. This study attempts to eval- While the reductions in emissions have generally uate how the transition affected the environmental improved air quality, the improvements have not been performance of faster- and slower-reforming econo- proportionate to emissions reductions because of pol- mies, tracing the combined effects of economic lution from small sources such as households and growth, reform, and environmental interventions on growing traffic in urban areas. In addition, emissions environmental trends in the region. of pollutants have fallen less rapidly and air quality All countries in the region experienced more or has improved less markedly in environmental "hot less severe economic recession after their transition spots" than in the respective country groups on aver- started. However, the depth of recession and the speed age. These results suggest that progress toward trans- of economic recovery showed significant differences, lating overall improvements in emissions into better strongly influenced by progress in implementing eco- environmental quality, especially for those living in the nomic reform. Advanced reform countries managed most polluted regions, has been relatively slow. to turn around their economies faster than slower- While trends in water pollution from industrial reforming countries. sources have generally followed air pollution trends, Economic recession, which led to declines in eco- the impacts of other sources, such as municipal waste, nomic output, industrial production, and energy use, on water pollution have often been more important contributed to reduced pollution loads in the entire than changes in industrial activity. Maintenance and xv xvi Economic Reform and Environmental Performance in Transition Economies investments in municipal water, sanitation, and at about $6 billion a year -as much as 5 percent of wastewater treatment infrastructure seem to be at least total urban income in the mid-1990s.1 Economic dam- as importantin determining water qualityaseconomic ages associated with air pollution have declined in restructuring. While several countries, especially in CEE but have increased in the NIS, where small im- CEE, have achieved significant improvements in mu- provements in urban air quality could not compen- nicipal infrastructure and wastewater treatment, sate for the general worsening of the health status of water infrastructure has generally deteriorated dur- the population. ing the transition in the slower-reforming countries Water-related problems are a key determinant of the NIS group. of the levels of infant mortality and waterborne dis- The use of fertilizers and pesticides has declined ease. Historically, the economies now in transition had in both advanced reform and slower-reforming coun- relatively low infant mortality rates in comparison tries. As a result, a decline in agricultural fertilizer with countries at similar income levels, especially in runoff to water bodies must have occurred, especial- the CEE and NIS Northwest group.2 During transi- ly in slower-reforming countries, where the decline tion, there has been no evidence of increasing water- in fertilizer use has been greater. However, nitrate related ill health except for the NIS Northwest country levels in rural drinking water, a significant environ- group, where average infant mortality rates increased mental health problem in many countries in the re- during the early 1990s. The proper maintenance of the gion, do not appear to have improved significantly. existing water and sanitation infrastructure should Recent data indicate that this has remained a serious receive special attention in these areas. Infant mortal- concern in, for example, Estonia, Kazakhstan, and ity rates have generally declined in the region, with Romania. the largest declines occurring in the NIS Southeast Water quality in some rivers has improved some- country group. what - in the Czech Republic, for example -but oth- Both air pollution and dirty water are more seri- er countries (Armenia, Hungary, Poland, and the ous problems throughout the NIS than in CEE coun- Slovak Republic) have recorded little significant tries. The overall environmental burden on health and change. In some hot spots, modest improvements the relative importance of air and water pollution are have been measured, but in several others, typically also distributed unevenly between the northwestern in slower-reforming countries, levels of industrial and southeastern parts of the NIS. Air pollution causes pollutants have actually increased. Examples include over three times more damage than dirty water in the Bulgaria's Maritsa Basin and the Volga-Urals region Russian Federation and Ukraine, whereas in Azer- in the Russian Federation. baijan and Central Asia the ratio is reversed. Among Environmental problems are associated with se- CEE countries, the largest relative burdens are asso- rious health damage. The two largest components of ciated with dirty water in Albania and air pollution overall health damage associated with air pollution in Bulgaria. The heaviest costs of environmental dam- are premature deaths and chronic bronchitis, which age are found in the NIS Southeast countries: air pol- result primarily from exposure to particulates. Using lution in Azerbaijan and dirty water in Azerbaijan, economic valuation techniques, this study estimates Kazakhstan, Tajikistan, Turkmenistan, and Uzbeki- the monetary costs of damage associated with air pol- stan. Future environmental interventions should tar- lution in a group of 57 cities in transition economies get these problem areas. Chapter 1 Overview of Progress in Economic Transition T ransition to a market economy, for all its long- Contamination of drinking water, mainly by ni- term rewards, is not an easy process. In addi- trates originating from agricultural enterprises and tion to the necessary but often difficult social, by heavy metals and toxic chemicals originating political, and economic adjustments, the countries of from industries. Central and Eastern Europe (CEE) and the newly in- A central principle of the EAP was that the pro- dependent states (NIS) have had to cope with a legacy cess of economic reform and restructuring would of inefficient industries, obsolete and polluting tech- eliminate the perverse incentives that underlay many nologies, and weak environmental management and of the environmental problems of centrally planned regulation. In addition, past limitations on civic ac- economies (Box 1). At the same time, it was recog- tivities and on dissemination of information meant nized that economic reform alone was no panacea. that mechanisms for bringing public pressure to bear Effective environmental policies, institutions, and in- on environmental issues were stunted or lacking. This vestments would be needed to harness the positive inherited burden led to serious environmental and forces of market reform and match the environmen- health problems and to dangerous environmental "hot tal performance of more developed economies. This, spots.' in turn, would require economic stability and the pros- The Environmental Action Programme (EAP) for pect of sustained economic growth, to encourage Central and Eastern Europe, which was endorsed by governments and industrial enterprises to take the the environmental ministers of 50 countries during steps required to make more efficient use of energy the 1993 Environment for Europe Conference in Luc- and natural resources, mitigate pollution, and enhance erne, Switzerland, identified a number of serious the positive environmental effects generated by eco- environmental problems requiring immediate and ur- nomic reform. gent attention. They included: More than five years have passed since the EAP * High levels of airborne particulates from coal burn- was endorsed. Significant progress has been made in ing by domestic sources, small-scale enterprises, many countries introducing environmental initiatives power and heating plants, and metallurgical plants at the national, regional, and local levels. Internation- * High levels of sulfur dioxide and other gases, es- al cooperation has supported many of these efforts. pecially in combination with particulate matter, However, transition has proved to be a much longer emitted from power and industrial plants and from and more difficult process than most had anticipat- household burning of high-sulfur coal and fuel oil ed, and progress has varied. By 1998 only one country, * Lead in air and soil stemming from emissions from Poland (which embarked on economic reform before industry and transport using leaded gasoline the rest of the region), had reestablished sustained 2 Economic Reform and Environmental Performance in Transition Economies Box 1. Reform and envirounmental benefits chapter first examines how advanced reform coun- tries3 have managed to improve their environmental Economic reform measures are expected to lead performance in comparison with slower-reforming to significant environmental benefits: transition economies and then assesses the impacts * The elimination of price controls and subsidies, of economic reform and additional environmental especially for fuels, generally leads to an ini- measures undertaken at the local, regional, national, tial increase in the costs of production and and international levels. thus promotes adjustment of production pro- Many enviromnental problems remain to be ad- cesses, energy and resource savings, and dressed, even in the most successful reformers. The restructuring of industry toward less study sets out to quantify the cost of environmental resource-intensive sectors. . damage in different countries as a basis for compar- • The imposition of a hard budget constraint, g coupled with changes in the incentive struc- ing current priorities which appear to present ture of enterprises, encourages managers to significant threats to human health to determine improve the efficiency of their operations, re- whether the priorities identified in the EAP remain duce waste, and improve the overall manage- valid for the new millennium. This overview covers ment of resources. all countries in the region including NIS Southeast * Privatization andfavorable conditionsforforeign countries which were not included in the EAP. The investments are expected to improve corporate governance, efficiency, profitability, and ac- study does not cover issues that have a less pervasive cess to financing-making possible the re- effect on human health in the region. More general- cess to financing -making possible the re- newal of outdated, highly inefficient, and ly, this study does not attempt to review the state of polluting capital stock. the environment in the region which is covered, for . Trade and market liberalization increases the ex- example, by the European Envirorment Agency's re- posure of enterprises to international market cent report: Europe's Environment: The Second requirements, management and environmen- Assessment, presented at the 1998 Environment for tal practices, and enhances their access to Europe Conference in Arhus, Denmark. cleaner production technologies. Implementation of the above reforms has begun Reform, economic growth, and in most CEE countries, but the reforms have not restructring been undertaken in a substantial way in the NIS. The countries of the region started the transition with widely differing capabilities and have progressed at economic growth and surpassed the pretransition lev- very different rates. The NIS countries that emerged el of real gross domestic product (GDP). Several NIS following the breakup of the Soviet Union have had countries have still not progressed very far in their to cope with the disintegration of economic and po- economic reform programs or achieved even a modi- litical links, as well as the difficulties of establishing cum of economic stability; their real GDP levels are new monetary and fiscal systems. Many of them face little more than one-half of those before transition. . significant barriers in their efforts to develop new Lower production has meant reduced volumes of pol- markets, and they lack the commercial traditions, lution, especially air and water pollution from large human resources, and access to expatriate communi- enterprises, but unless efficiency and environmental ties with capital that have eased the transition for performance are improved, any short-term reductions may easily be reversed when production levels recover. m n CeE count This study attempts to evaluate the combined On average, GDP per capita for the NIS coun- effects of economic reform and other measures on tries is much lower than for the CEE countries (Table environmental performance in the transition econo- D.1). The severe drop in oil and commodity prices mies since 1989-90. It examines several underlying during 1998, combined with lower real exchange rates, factors - economic growth, industrial structure, ener- will mean a substantial decrease in GDP per capita gy intensity, and the composition of fuel use -that measured in U.S. dollars. GDP per capita for Russia affect trends in volumes of emissions, emissions in- and Belarus, for example, will have fallen substan- tensities, and ambient environmental quality. This tially in 1998, thus widening the gap between the two Overview of Progress in Economic Transition 3 country groups. The differences in real living stan- Box 2. Perverse incentives against industrial dards are not quite as large as the figures imply, since restructuring in 'virtual economies" the estimates rely on current or weighted average ex- change rates rather than purchasing power parity Governments in many slower-reforming NIS (PPP) rates. Recent estimates of PPP rates for the CEE the traditional ties between enterprises and the countries suggest that the better-off countries such as state, introduce hard budget constraints and in- the Czech Republic, Hungary, and Slovenia have real dustrial reform, and insist on the bankruptcy of incomes per capita that are not far behind those in unviable firms. Subsidies have been maintained, Greece and Portugal, the poorest member states of the largely in the form of tax arrears. The result is European Union (EU). what Gaddy and Ickes (1998) call a "virtual economy" in which enterprises do not add value and operate by nonmonetary means such as bar- The pace of reform ter. Russia's largest companies, for example, Progress in implementing economic reform has been have been reported to conduct more than 70 per- uneven among CEE/NIS countries. In advanced re- cent of their business by nonmonetary means. uneve amog CE/NIS ounties.In adance re- Bankruptcies and the entr of new firms are very form countries, state-owned enterprises have been rare, tamering the gro of a real c ery v ~~~~~rare, hampering the growth of a real competi- privatized or at least subjected to much stricter bud- tive environment. Pure survival and rent-seek- get constraints than before. Subsidies for energy and ing rather than profit-maximizing behavior natural resources have been reduced, although usu- become the main characteristics of enterprises in ally not entirely eliminated. Tax systems have been the virtual economy. restructured, and efforts have been made to improve No significant industrial restructuring has taken both the collection of taxes and the enforcement of place in these slow-reforming NIS countries. regulations, including those relating to environmen- Enterprises are more interested in maintaining tal performance. Overall, the cumulative impact of employment, avoiding taxes, and keeping up these reforms in CEE countries has been to bring them good relationships with business partners and much closer to the market economies of Western Eu- government officials than in producing profits rope, increasing the gap between the CEE and NIS that would be heavily taxed. Exports are fre- rcoopie, .quently pursued only to obtain the cash neces- sary to pay wages in the otherwise cashless By contrast, many NIS countries have continued economy. Economic performance is therefore to shield their enterprises from market forces through extremely weak - often much weaker than offi- indirect subsidies, including loans from government cial statistics indicate. As Gaddy and Ickes show, banks, and by tolerating large tax and payment ar- in Russia, for example, the volume of real spend- rears (Box 2). As a result, industrial restructuring has ing on plant and equipment in the manufactur- proceed yin these countries, and the limited ing sector in 1997 was about 5 percent of the 1990 proceeded slowly in these counhries, and the limited level, and investment in all productive sectors progress in privatization has not yielded the expect- was about 17 percent of the 1990 level. ed benefits. Unlike the situation in advanced reformers, where privatized enterprises improved The ability of the public sector to collect and ef- their economic performance, progress has been dis- ficiently reallocate tax revenues throughout the appointing in the NIS. There, ownership was often economy is severely limited in virtual economies: , ~~~~~~because taxes are paid in kind, the supply of transferred to former employees or managers with- goods and sre rate tn effiieny on goods and services, rather than efficiency con- out changes in economic incentives, management siderations, determines public spending. practices, traditional ties to government, or depen- dence on budget support. The lack of economic reform has also severely Throughout the region, there is growing recog- limited improvements in environmental perfor- nition that the relatively rapid pace of liberalization mance: many firms have done little to improve apittntdevel- their productivity or to invest in environmental and pitianhanobemacprotection. Instead, soft credits and waivers for opment of the institutions necessary to support a pollution charges become part of the system of well-functioning market economy. The advanced re- government support to enterprises. form countries have made some progress, but 4 Economic Reform and Environmental Performance in Transition Economies inadequate institutional arrangements in slower-re- icantly higher than in 1989, most of the other CEE forming countries have resulted in serious distortions countries could expect to achieve this goal within five that vested interests have been able to exploit, often years, given reasonable external circumstances (in- at significant social and environmental cost. The sub- cluding strong support from the EU). Two of the CEE sequent high levels of poverty and inequality have countries, Bulgaria and Romania, have renewed eco- fueled disillusion with market reforms and have fur- nomic reform efforts in recent years, which should ther undermined the authority of governments to bring sustained growth. On the other hand, political establish effective institutions. instability in the Balkans, including the wars in the former Yugoslavia, has caused severe economic prob- Economic trends lems and uncertainty for these countries. It is likely to take at least 10 years for the NIS countries to sur- All countries in the region experienced a more or less pass their real GDP levels at the end of the 1980s. severe recession at the beginning of the transition, but The liberalization of controlled prices, combined there have been large differences in the depth of the with weak monetary discipline, led to severe episodes recession and the speed of recovery. Success in rein- of inflation in almost all countries. As shown in Table stating economic stability and growth has been D.2, by 1993 the general price level had risen to more strongly influenced by progress in implementing than 10 times its 1989 level in all of the CEE countries market reforms (EBRD 1997). Most countries had neg- except the Czech Republic, Hungary, and the Slovak ative average growth rates during 1990-96 (Figure 1). Republic. Most NIS countries experienced some form However, the countries of the former Soviet Union of hyperinflation, with 1993 prices soaring to more (including the Baltic states) experienced larger falls than 200 times their 1989 levels. A few countries, no- in real GDP than the CEE countries (other than those tably the Kyrgyz Republic and Moldova, managed to affected by war).' Despite occasional setbacks, ad- establish some kind of price stability after 1993, but vanced reform countries have succeeded in turning severe inflation remains a major barrier to the rees- tablishment of economic stability in many NIS their economies around much sooner than slower re- ctrises Fo exampe betwen193 and avr formes. Tey ahievd a easoabledegre of countries. For example, between 1993 and 1997 aver- formers. They achieved a reasonable degree of age annual price increases were about 350 percent in macroeconomic stability,5 with nominal interest rates Belarus (even higher in 1998) and 250 percent in Az- in the range 10-20 percent per year, implying posi- erbaijan. Worse still, inflation rates have varied greatly tive real interest rates that provide a genuine incentive from year to year as a result of erratic economic poli- for saving and investment, and they had firmly rees- cies, and there are no mechanisms for monetary tablished economic growth by 1994-95 (Table D.1). correction to protect savers and investors from the con- Although by 1998 only Poland's real GDP was signif- fiscatory effects of inflation. Figure 1. Average GDP growth rates between 1989 and 1997/98 2- -2- -14 -16 Note: Averages refer to growth rates between 1989-98 except for the following: for Azerbaijan, Bulgaria, Georgia, Latvia, Moldova, Slovenia, Tajikistan, Turkmenistan, Ukraine, and Uzbekistan, they refer to averages between 1989-97, and for Macedonia FYR, 1990-97. Source: OECD data and international Finance Statistics. Overview of Progress in Economic Transition 5 Under such circumstances it makes no sense for Figure 2. GDP and industrial output by any enterprise to invest time, labor, and money in any- country group, 1989-98 thing that will not produce an almost immediate 140 return. Even simple measures to improve output and 120 pro- ductivity or to reduce the waste of energy or raw materials are of little significance in comparison with ~80 efforts to take advantage of the fiscal and monetary distortions created by rapid and fluctuating rates of 60 inflation. In addition, high inflation and political, le- & 40 gal, and other uncertainties increase the risks and costs 1 20 of long-term financing, further delaying investments 0 in the replacement of outdated, highly polluting cap- 1989 1990 1991 1992 1993 19 1995 1996 1997 1998 ital stock and in other environmentally beneficial - GDP: advanced U Industry: advanc improvements. Thus, it is unrealistic to expect any A GDP: other Indutry: ower substantial progress toward better environmental Source: World Bank estimates based on OECD data. performance or, specifically, an increase in environ- mental investment until the core elements of macro- some of which have faced renewed economic mnsta- economic stability are in place. This means that spe- bility and crises in recent years. cial environmental financing mechanisms such as environmental or pollution abatement funds are like- *nutra ou*u haalnmresapyta GDP in both country groups, but the overall shift in ly to have only a marginal impact on the overall environmental performance of the enterprise sector. he setoralcmoin of advaway frm inst Suchfund canaddrss svereenvionmetal rob as been greater in the advanced reform economies. eSuthat ponds can imminent thvreat tovumane halth pIndeed, in the slower reformers, the industrial sector lems that pose an imminent threat to human health satdt eoe n19,atog grgt D or critical ecosystems, but they cannot correct the un- started to recover in 1995. although aggregate GDP derlying weaknesses of the macroecononmic and policy was still falling. In 1998 industrial output fel as GDP framework. recovered slightly. Within the industrial sector, the initial decline in output from heavy industry, includ- Trends in industrial output and energy ing chemicals and metallurgy, was particularly intensity marked. These sectors have recovered somewhat since -1994 but still account for a significantly lower indus- The EAP predicted that structural changes generated trial output than in the late 1980s. by the process of economic reform during transition would be beneficial to the environment Reform was expected to promote a shift from heavy industries Ener,.,y intensity toward less resource- and pollution-intensive sectors, Absolute levels of energy intensities have tradition- improvements in the efficiency of resource use, and a ally been high in all transition economies, but switch from dirtier fossil fuels such as coal toward especially in the slower-reforming countries. Even in cleaner fuels such as gas. This section reviews actual 1995 the average energy intensity of GDP in those changes in these areas. countries was nearly twice its value in the advanced reform countries. Allowing for differences in climate Industnal production and other factors, the slower-reforming countries have Figure 2 shows average trends for GDP and industri- much greater opportunities for improving their ener- al output between 1989 and 1996 for the advanced gy efficiency; thus, the only modest improvements reform economies and for slower reformers. As not- made so far represent an important loss of opportu- ed above, the transition process precipitated a collapse nity. in GDP and industrial output in all countries, but the International experience shows that the amount advanced reformers stabilized and turned their econ- of energy consumed to produce a unit of GDP tends omies around far earlier than did the slower reformers, to decrease with increasing levels of GDP. From this 6 Economic Reform and Environmental Performance in Transition Economies viewpoint, CEE/NIS countries have consumed exces- Figure 4. "Excess" energy consumption by sive amounts of energy per unit of GDP at the start of country group, 1990 and 1995 the transition. In 1990 "excess" energy consumption- 90.0 the difference between expected consumption based 80.0 77.8 76.3 on GDP per capita and actual consumption, expressed 70 ° 57.3 as a percentage of total consumption-reached more s.o than 80 percent in Kazakhstan and Ukraine, more than 40.0 70 percent in Azerbaijan, Bulgaria, the Russian Fed- 30.0 eration, and the Slovak Republic, more than 60 percent l 0.0 in Bulgaria and Poland, and almost 50 percent in 0.0a Hungary, resulting in high pollution intensities. Fig- 1990 1995 ure 3 presents the 1990 energy consumption figures El Advanced reformers U Slower reforners for five transition countries compared with the inter- Source: Worid Bank estimates. national trend.6 During the transition, energy use decreased mea- sumption declined by 12,13, and 17 percentage points, surably in advanced reform countries, and it fell respectively (Figure 5). drastically in slower reforming countries, where 1996 By contrast, most slower reformers have hesitat- energy use stood at half its 1989 level. However, the ed to embark on reforms or have failed to follow response of the transition economies to the dual chal- through on initial commitments. In some cases, as in lenge of modernizing their energy and industrial Belarus, initial reform was halted. In others, such as sectors and improving efficiency has been far from Ukraine, implementation of reforms has been slow uniform. Most advanced reformers managed to put and uneven. As a result, in the slower-reforming coun- in place the institutional foundations for a lasting try group, energy consumption remained nearly as transformation from a centrally planned to a market- "excessive" in 1995 as at the beginning of the transi- based system. Rapid adjustment of energy prices to tion (Figure 4). reflect costs, the enforcement of payment discipline, There were, however, significant differences and large-scale privatization of industrial firms have among individual countries. Energy intensities wors- decreased the energy intensity of the economy; "ex- ened and "excess" energy consumption increased in cess" energy consumption declined in the advanced several countries, including Azerbaijan, Bulgaria, and reform country group from 70 percent in 1990 to 57 the Kyrgyz Republic, where no significant reform took percent in 1995 (Figure 4). Especially significant im- place during this period. Little change in "excess provements have taken place in the Czech Republic, Figure 5. "Excess" energy consumption in selected Hungary, and Poland, where "excess" energy con- advanced- and slower-reforming countries, 1990 and 1995 Figure 3. Energy consumption and GDP in transition loo- 1990 economies compared with international trend, 1990 01995 7 6 60 0~ ~ ~~~~~~~~4 ""n 1~~~~~ r.~ 20- (o 0~~~~~~~~~~~~~~~~~~~~~ L 0 1 2 3 4 5 6 7 8 9 10 GDP per capita (thosand US$) Source: World Bank estimates on World Development Indicators. Source: World Bank estimates based on World Development Indicators. Overview of Progress in Economic Transition 7 energy consumption occurred in the Russian Federa- portunity costs of operating nuclear and hydroelec- tion and Ukraine, whereas in some cases (for example, tric power plants are small relative to those of thermal Kazakhstan and Romania) there was noticeable power plants, so that it makes sense to utilize as much improvement. While the data show significant im- as possible the capacity of existing facilities of these provements in some slower-reforming countries such sources while placing the burden of adjustment to as Albania, Armenia, and Georgia, these improve- lower demand on plants that use fossil fuels. On the ments were mostly due to the impact of war and demand side, even limited adjustments. in relative severe restrictions on household energy consumption prices have favored a shift away from coal and heavy rather than to the effects of structural changes and fuel oil toward gas, and institutional factors have re- improvements in efficiency. Overall, the advanced re- inforced this trend. form countries have started to narrow the gap between The switch from coal and, in the slower-reform- their energy efficiency and the energy efficiency of ing countries, petroleum products toward gas was market economies. Slower reformers failed to achieve even stronger for final consumption (i.e., excluding similar improvement in this area in the first half of power generation and other types of energy conver- the decade. sion) than for primary consumption of energy in both Fuel use: Composition and trends groups of countries, indicating that the decline in coal Changes in composition use outside the power and heating sectors has been Changes ln cvnnposition larger than its overall decline. Price adjustments and The composition of total energy use has changed in convenience both suggest that the share of coal in fi- many countries but most significantly in the slower- nal consumption is likely to continue to fall. reforming economies. Figure 6 shows the average decrease in the consumption of different categories Coal combustion by small sources of fuel between 1989 and 1995. In both the advanced and slower-reforming From an environmental perspective, any shift from country groups, consumption of solid fuels has fallen coal toward the use of gas, but especially outside the by more than total energy consumption, while con- power sector, is beneficial because of the resulting sumption of gas has declined by about the same as reductions in emissions of air pollutants such as par- total energy. The shifts in the composition of energy ticulates and sulfur dioxide. The most damaging type consumption have been small in the advanced reform of fuel use is the burning of coal in stoves and small countries but large in the slower-reforming countries. boilers operated by households, small businesses, and The consumption of petroleum products has also fall- service establishments. Emissions from these sources en more rapidly in the slower-reforming countries typically represent a disproportionately larger source than has that of all energy. These shifts reflect a of damage to human health than emissions from large combination of demand and supply factors. The op- stationary sources, such as power and industrial plants, because the small sources are closer to people Figure 6. Trends in primary energy consumption, and it is more difficult to control their emissions. I 989-95 Historically, as noted by the EAP, the combus- Other energy tion of coal by small sources such as households, Gas _ _ municipal buildings, and small commercial boilers Petroleum prodwts represented a serious air pollution problem in many Petroleum products transition economies, including Armenia, the Czech Solid els Republic, Estonia, Georgia, Latvia, Poland, Russia, Ukraine, and parts of Kazakhstan and the Kyrgyz 0 5 10 15 20 25 30 Republic. Merage decline, % per year It is not possible to measure fuel use in small sources directly, but the consumption of coal by sec- DAdvanced reform countries * Slower-reforming coUflnes tors other than energy conversion (including power Source: World Bank estimates based on OECD and UN data. and heating plants) and manufacturing industry pro- 8 Economic Reform and Environmental Performance in Transition Economies vides a reasonable indication. Figures 7 and 8 show was especially marked for gasoline in the advanced trends in such use for the two country groups, along reform countries. Consumption of motor fuels has with the other major category of fuel consumption by been rising strongly since 1993-94 in these countries, small sources, motor gasoline. The strong downward and there are signs of a similar recovery in the slow- trend in the use of coal by small sources is clear for er-reforming countries from 1995 onward. The reason both groups of countries. for the recovery has to do with a substantial growth The downward trend in coal consumption is like- in vehicle fleets in almost all the CEE countries and in ly to continue for some time because in many countries several NIS countries (Table D.3). Data indicate a the relationship between the prices of coal and sub- median increase of 33 percent over six years (nearly 5 stitute fuels, especially gas, remains more favorable percent per year) even though GDP declined, on av- to coal than would be the case if these prices fully -erage, nearly 20 percent. In Belarus, Russia, and reflected world market prices (appropriately adjust- Ukraine, vehicle numbers have grown even more rap- ed for fuel quality and heating value). For example, idly, despite larger falls in GDP. Such growth may the price of coal in Poland is now above export pari- reflect the extent to which the ownership of motor ty, but coal prices are still implicitly subsidized in the vehicles was held down under the previous econom- other main producing and consuming countries, in- ic system. cluding the Czech Republic, Romania, Russia, and As one would expect, the number of vehicles per Ukraine. By contrast, average prices for natural gas 1,000 people is closely correlated with GDP per capi- and petroleum products have typically exceeded im- ta. (The figures for Armenia and Tajikistan seem to port parity, except in Russia, where gas prices have be anomalies.) Some countries, such as Estonia, the been controlled to subsidize both household and in- Czech Republic, and Slovenia, however, have mark- dustrial consumers. As these pricing distortions are edly higher levels of automobile ownership than gradually corrected, it is likely that the share of coal would be predicted by their levels of GDP per capita in the energy consumption of households and other (OECD/EEA 1997). small sources will continue to decline. In time, coal Most of the increase in motorization has been will be used almost entirely for power generation and concentrated in urban and metropolitan areas. The by certain large industrial consumers, as is the pat- average annual growth rate of passenger car numbers tern in Western Europe today. reached almost 7 percent in Warsaw and 5 percent in Budapest between 1985 and 1994. The growth in ve- Motor fuels consumption hicle numbers appears to be especially high in some Another aspect of the energy transition is that the use NIS cities. In Moscow, for example, the fleet size al- of petroleum products for transport fell by significant- most tripled during 1985-94, with an average annual ly less than GDP in both groups of countries. This gap growth rate of about 11 percent, rising to more than Figure 7. Indices of fuel use by small sources in Figure 8. Use of fuels by small sources in advanced advanced reform countries, 1989-95 reform countries, 1989-95 g_120_ 100 - a 120 _ Ambient air quality in urban areas 9 80 Although the relationship between aggregate emis- xs 70 X sions and air quality in specific cities is not a simple 60 _____________________________________ one, large reductions in emissions of key air pollut- 1989 1990 1991 1992 1993 1994 1995 ants should mean that urban air quality has improved in the transition economies. Because emissions from large stationary sources such as power plants and Source: World Bank estimates based on OECD and national data. large industries are farther away from urban popula- Figure 12. Emissions intensities for air pollutants in tions and disperse over a wide geographic area, their slower-reforming countries, 1989-95 relative impact on local air quality and human health 120 is smaller than the impact of emissions from small 115 sources such as households, small commercial boil- o105 ers, and vehicles. Therefore, if the reduction of 11 105 \ \ / ~_ emissions has extended to small sources, as the anal- a 100 ysis of the pattern of coal consumption suggests, this 95 should translate into better urban air quality. 90 Figure 13 presents the results of an analysis of 85 air quality trends for a sample of 44 urban areas in 80 - Western Europe, Central and Eastern Europe, and the 1989 1990 1991 1992 1993 1994 1995 former Soviet Union covering the period 1989-96. + PM -- S02 A NOx -- vocs Since many factors influence the nature and extent of Source: World Bank estimates based on OECD and national data. air pollution in different cities, the analysis focused on estimating average time trends for the improve- For the slower-reforming countries, there is lit- ment or worsening of air quality indicators in urban tle evidence of any significant improvement in areas in different groups of countries. For example, emissions intensities over the period; they have fluc- the average improvement of 4.3 percent per year in tuated within a range of 10 percent on either side of thel991baselinevalues. (Itispossiblethatemissions Figure 13. Trends in urban air quality by country intensities for PM have started to fall in a sustained group, 1989-96 manner, but more recent observations are needed be- Alr pollutant fore the improvement can be regarded as convincing.) Overall, this analysis suggests that the changes PM 4 .7 associated with economic reform and the transition 95 have produced real environmental benefits in the ad- so2 vanced reform countries which go significantly ll7.9 beyond the reductions in emissions resulting from 3. NO 3__ _5_ declines in income and output. Similar gains are not 2 . observable in the slower-reforming countries. This -.o -4.0 0.0 4.0 8.0 1Z may simply be a matter of Timing; output and energy Average annual Improvement In air quality (% pa) use were still falling in 1995, and additional improve- O Western Europe E Advanced reforn countries E Othercountries ments in emissions -intensities may have been Source: World Bank estimates based on OECD and national data. Air Pollution 15 PM in Western Europe means that ambient concen- Figure 14. Trends in emissions of air pollutants in trations in the typical Western European city in the hot spots, 1990-96 sample were falling by this amount each year over the eight years.8 The outcomes for individual cities Dor -Kh-fkov show a fair degree of dispersion, but there are statis- tically significant differences between the time trends M&fta BW .l . for the three groups of countries. _ ] The advanced reform countries experienced a UPPW Sle . -: .. : ... significantly greater average improvement in ambi- ent concentrations of PM than did the other two . . ... ... country groups. Despite the substantial reduction in dust emissions in the slower-reforming countries, their o s 10 15 20 proportional improvement in average concentration Average reducfion in emissions, % per year flDust OSO of PM was no better than that achieved in Western Source: World Bank estimates based on Electrowatt 1998. Europe, although their much higher initial levels I_______________________ meant a greater absolute decline in concentrations. For SO2, the improvement in Western Europe outstripped with major air pollution problems.9 The comparison that in both the advanced reform and slower-reform- suggests that emissions of both pollutants have fall- ing countries, though the latter did better than the en much less rapidly in these hot spots than in the former. Finally, the improvements in ambient con- relevant reform group as a whole. The shortfall is centration of NO are almost the same in Western particularly marked for the Maritsa Basin and the Europe and in the advanced reform countries, whereas Donetsk-Kharkov region. Only in the case of dust there was a clear deterioration in the slower-reform- emissions in Upper Silesia does the average rate of ing countries, particularly those in the CEE. This reduction come close to the general average for ad- worsening seems to reflect the rapid growth in the vanced reform countries. number and use of motor vehicles without any im- When improvements in urban air quality in hot spots (shown in Figure 15), are compared with the provement in their average emissions per kilometer. genra tshown in Figure 1 3, the rsts are Furthermore, nitrogen oxides are products of all fuel general mrends shown in Figure 13, the results are combustion, and fuel switching therefore had less ef- ment in PM concentratons for the Baltics and Upper fect onNOx emissions than on2PM and 502 . Silesia (6.5 percent and 7.3 percent per year, respec- tively) are very similar to the average improvements (6.7 percent per year) in urban areas in the advanced Although national trends are important, the extent to whichenvironmentalconditionshaveimprovedinthe Figure IS. Trends in urban air quality in hot spots, most polluted regions should be given equal weight 1990-96 in assessing the environmental benefits from the tran- sition. For this reason data have been collected on five hot spots in the region: the Baltic countries D - (primarily with respect to water pollution), the Do- s netsk-Kharkov region in Ukraine, the Maritsa Basin in Bulgaria, the Upper Silesia region in Poland, and UPP&5Sb the Volga-Urals region in the Russian Federation. The data cover both emissions of key pollutants and indi- l8 5 5 10 IS 20 cators of environmental quality from 1990 to 1997. Average annual improvement in air quality (% pa) Figure 14 shows the average annual reduction in to- c: PM so fl LeO tal emissions of dust and SO2 for the four hot spots Source: World Bank estimates based on Electrowatt 1998. 16 Economic Reform and Environmental Performance in Transition Economies Box 4. Industrial production and air quality trends in two hot spots Hot spots in many countries show simnilarities because of past economic development paths. For example, Upper Silesia in Poland and Donetsk-Kharkov in the industrial belt of Ukraine are both centers of heavy industry with complex engineering infrastructure. Industrial production and, in particular, coal-based power and heat generation have been identified as the major sources of air pollution in both regions. As a result, airborne particulates are a major air pollution problem in both areas. During the transition, marked differences between Upper Silesia and Donetsk-Kharkov have developed. In both areas, industrial output has declined, but in Upper Silesia the decline did not exceed 20 percent, and by 1997 industrial production had surpassed its 1990 level by 32 percent. In Donetsk-Kharkov indus- trial output in 1996 was less than 50 percent of the 1990 level. Ambient concentrations of PM increased temporarily in both areas during the early 1990s even though industrial output declined. However, the increase was reversed much sooner in Upper Silesia than in Donetsk-Kharkov. Whereas in Upper Silesia ambient concentrations have declined steadily since 1991 despite increasing industrial production, in Donetsk-Kharkov the relative improvement in ambient con- centrations has been smaller, as particulate emissions per unit of output have increased continuously. The underlying reasons for this phenomenon can be traced to differences in the overall economic conditions of the two countries, in economic restructuring, and in environmental protection measures. In Upper Silesia air quality was improved through a switch to cleaner fuels, investment in new capital stock, and air pollution abatement measures. No improvements have taken place in the Ukraine, where, because of serious macroeconomic difficulties, priority was given to maintaining the production capabili- ties of existing capital stock and technologies and even planned investments in pollution abatement were delayed or abandoned. Of the 47 pollution abatement facilities planned for 1994, only 20 became opera- tional, and their capacity reached only 15 percent of the target. The share of out-of-order equipment ex- ceeded 20 percent in 1994 and has continued to grow due to shortages in operation and maintenance (O&M) funds, personnel, and materials. Energy intensities remained high or rose. The result has been an actual increase in emissions per unit of output, especially in power and heat generation. Deterioration in air pollution control has also been linked with inadequate environmental management practices at all levels. Environmental regulations have often been violated, with no response by the regulatory bodies. Ambient particulate concentrations and industrial output indices in Upper Silesia and Donetsk- Kharkov, 1990-97 Upper Silesia Donetsk-Kharkov 120 ,,,/ 100 T 100--~~~~~~~~~~0 80 60 40 ~~~~~~~~~~40 20o- -l | | | | 111 11 20 0 ~~~~~~~~~~~~0- 90 91 92 93 94 95 96 97 90 91 92 93 94 95 96 97 Ambient Air Quality / Particles -+- Industrial Production Ambient Air Quality/ Particles 4 Industrial Production Source: Electrowatt 1998. Air Pollution 17 reform countries. The average improvements in am- Throughout the region, and even in advanced reform bient concentrations of PM in the Maritsa Basin and countries, progress in privatizing and restructuring the Donetsk-Kharkov region outstrip the overall av- large heavy industries has been slower than in other erage for the slower-reforming countries. However, sectors of the economy. Without budgetary discipline the relative improvement in the Kharkov-Donetsk and measures to enforce better management, these plants region is considerably smaller than in Upper Silesia tend to continue business as usual, although at a lower (see Box 4 for a more detailed discussion). A cause of level of output, so that there is little or no improvement concern is the deterioration of urban air quality in in environmental performance in terms of pollution in- Russia's Volga-Urals region. tensity of output It is unrealistic to expect significant The pattern for SO2 is less encouraging than that progress in the hot spots in the slower-reforming coun- for PM. Only in the Donetsk-Kharkov region did the tries until the painful decisions involved in restructuring rate of improvement outstrip the average rates for the or dosing plants have been taken. relevant reform group.10 These results suggest that, to achieve the high- Overall, these results suggest that progress to- est returns in terms of improving human health and ward translating overall improvements in emissions welfare, priority should be given to: and urban air quality into better environmental qual- * Improve access to piped rural water supply and ity for those living in the most polluted regions has the coverage of urban sanitation in Albania been relatively slow. This is clearly one of the major and the NIS Southeast countries, especially in challenges facing envirorunental policy in the slow- Tajikistan, Turkmenistan, and Uzbekistan er-reforming countries, especially in the NIS, and it * Improve urban air quality in Armenia, Bulgaria, highlights the difficulty of achieving sustained im- Georgia, Moldova, the Russian Federation and provements in environmental performance in the Ukraine absence of effective economic reform. Hot spots, by * Improve the state of existing water and sanitation their very nature, contain large concentrations of in- infrastructure throughout the NIS, especially in the efficient and polluting heavy industrial plants. NIS Northwest group. Chapter 3 Water Pollution Trends in emissions and water quality pended solids decreased by almost two-thirds be- T n he EAP predicted that economic reforms by tween 1990 and 1996. In some other CEE countries, themselves would not have a strong influence water pollution discharges have also fallen. Howev- on water pollution: although restructuring er, comparable data that could be used to examine should reduce industrial discharges, and a reduction the environmental effects of investment in wastewa- in agricultural chemicals and thus their runoff into ter treatment are not available. rivers and streams, in most countries household and Given the paucity of direct evidence of aggre- other municipal sources generate the largest share of gate discharges, indirect indicators can be used. Figure water pollution, and their discharges would not be 16 shows the trends in variables associated with dif- directly effected by economic changes. This chapter ferent types of water pollution. Use of fertilizers and looks at indicators of water pollution trends during pesticides has declined by much more than crop pro- transition. duction in both groups of countries. Runoff of Data that can provide direct insights into the fertilizers into rivers, lakes, and groundwater should impact of the transition on water pollution are sparse. have fallen dramatically in the slower-reforming coun- Most water pollution is highly localized and is often tries and to a lesser extent in the advanced reform linked to discharges of human wastes rather than to countries. The same should be true for pesticide run- industrial or other economic activities. Arrangements off in advanced reform countries. This may not regarding municipal finance and infrastructure that affect investments in and the operation of sewerage Figure 16. Indicators of water pollution, 1990-96 and wastewater treatment systems are at least as im- Etl¶uentdischarges portant as economic restructuring. Many of the l advanced reform economies, notably the Czech Re- Indwtnal water uset public and the Baltic countries, have made or are beginning to make large investments in municipal LUvestock producfion wastewater treatment, in connection with their goal Cr0p producion of becoming members of the EU within a few years. Pestcide use = - - - In the Czech Republic, some point-source water pol- Fertilizeruse lution has fallen significantly: according to data from 0 5 10 15 20 25 30 the Organisation for Economic Co-operation and De- Avage dedine, % per year velopment (OECD), discharges of biological oxygen 0 Advanced reform countries * Slower reforming countries demand (measured over five days, BOD5) and sus- Source: World Bank estimates. 19 20 Economic Reform and Environmental Performance in Transition Economies translate into immediate improvements in water qual- er contain nitrate levels far exceeding acceptable ity, but people who rely on shallow groundwater limits. sources for their drinking water should benefit from The overall volume of water abstraction has de- these changes. Similarly, the decline in livestock pro- clined, but by much less than the use of water by the duction should have reduced the runoff of nutrients, industrial sector. The decline in total water use in the especially from intensive livestock operations, which industrial sector was about 5 percent per year less than were important sources of water pollution in the Bal- the decline in industrial production in the slower-re- tic countries. forming countries, while in the advanced reform The EAP identified high nitrate levels in rural countries it was about 5 percent per year greater. drinking water as a significant environmental health Water losses by municipal water utilities have been problem in several countries (Belarus, Bulgaria. Hun- traditionally high in the region but were typically con- gary, Lithuania, Romania, and the Slovak Republic); cealed by generous water consumption norms and by high nitrate levels cause "blue-baby syndrome" the lack of consumption metering, which resulted in (methemoglobinemia) in newborns. Updated data on false consumption records. Demand management, the nitrate levels have been scarce, but in Romania indi- installation of water meters, consumption-based bill- cationsarethatelevatednitratelevelsarestillcommon ing, and increased water tariffs have changed in shallow wells and in nearly 12 percent of deep wells consumer behavior and reduced water consumption (OECD 1998) and that mortality and morbidity among (see Box 5). Reduced consumption, in turn, resulted newborns from methemoglobinemia have not de- in lower utility operating costs and less need for new dined. New evidence of high nitrate levels in wells investments, making it possible to shift resources to has also been found in Estonia, where 3 to 5 percent badly needed investments in rehabilitation and effi- of the 4,000 wells tested during 1993-96 showed ni- ciency improvements. trate levels above 45 milligrams per liter (mg/1). In As with water use, effluent discharges in the Kazakhstan, too, artesian wells and the Syrdaria Riv- slower-reforming countries have decreased at about Box 5. Demand management and water consumption in Central and Eastern Europe The introduction of metering and consumption-based billing has in many cases resulted in significant reductions in billed water consumption in advanced reform countries. * In several cities of Saxony, Germany, including the capital, Dresden, total domestic water consumption was reduced from about 300 liters per capita per day (I/c/d) to about 90 l/c/d in just five years. * In Rostock, under a concession run by Lyonnaise des Eaux, water production diminished by more than 60 percent. • In some neighborhoods of Bydgoszcs, Poland, metering led to a 30 percent reduction in water consump- tion over three years. * In Gdansk, Poland, metering and higher tariffs introduced during 1992-95 led to a 33 percent decrease in total water consumption (see figure below). Effect of metering on annual water production in Gdansk Morena District 2500-. 100 " 2000 - - 7 Mte 80 a °1500 60 / - ) C oX 50004 0 0 1992 1993 1994 1995 Source: World Bank 1 998b. Water Pollution 21 the same rate as industrial production. This reinforc- Figure 17. Emissions of water pollutants in hot es the earlier conclusion that up to 1995-96, economic spots, 1990-96 reform and industrial restructuring had little or no s effect on the overall environmental performance of the industrial sector in the slower-reforming countries. Dak,r Almost all of the environmental improvements were the result of declines in industrial output rather than in the emissions intensity of production. V04-g-a s Identifying overall trends of water quality can s lo Is 20 be difficult as these are often linked to local factors. Average reduc,Ion ehom, % pe yaw As a result, data on water quality are limited and of ns I AmorJumsat uncertain reliability. Reports to the OECD on water Source: Electrowatt 1998. quality data for major rivers in four advanced reform countries suggest that water quality in the Czech Re- Figure 18 relies on monitoring of surface water public improved somewhat during 1989-95.11 The quality in rivers, lakes, and reservoirs in hot spots. other reporting countries -Hungary, Poland, and the Overall, the data suggest a pattern of modest improve- Slovak Republic-recorded little significant change ment but with some deterioration in the levels of over the period. In the Baltic countries, too, water certain industrial pollutants, notably oil products in pollution discharges from point sources have fallen the Maritsa Basin and metals in the Volga-Urals re- significantly, yet aggregate data for these countries gion. The recorded levels of oil products (phenols) in show no improvement in river water quality. In ad- rivers in the Maritsa Basin are mostly quite low but dition, the EAP noted that discharges of saline waters are subject to erratic variations. Thus, not too much from coal mines was a serious problem in several can be read into this result. The pattern of change in countries, in particular, the Czech Republic, Poland, the levels of metals in the Volga-Urals area is also un- and Ukraine. While Poland has constructed plants to even, but the trends for particular rivers are more remove saline from mine waters, discharges to the Odra consistent, suggesting that discharges from certain River (though not yet from the Vistula River) the prob- metallurgical plants may have increased substantial- lem remains serious in the Ukraine. Similarly, data for ly. This would be consistent with previous obser- rivers in some NIS countries such as Armenia show no vations showing lack of progress in improving envi- clear trend, although there have been substantial chang- ronmental performance in this region (see Box 6). es at particular monitoring points. There are indications of some improvement in the quality of drinking water in the three hot spots; how- Water pollution trends in hot spots ever, the data are rather scarce and are based on the More data are available on water pollution and qual- proportion of samples that fail to meet either chemi- ity in the hot spots, though coverage is still uneven cal or biological standards. and the pattern of change is complex and difficult to interpret. Figure 17 suggests that total emissions of Figure 18. Surface water quality in hot spots, 1990-96 BOD and nutrients have declined substantially in three of the hot spots even though discharges from Baltia municipal sources have probably remained relative- Donetk-Kharkov ly constant. Any improvement in the Donetsk- Maran _.......... Kharkov region has been small and erratic from year to year. Again, after allowing for changes in industri- Upper Slimia al production, it seems likely that there has been a Volg-UrI._ significant improvement in environmental perfor- mance in the two hot spots analyzed in the advanced Avea annual improvement in water qualiy (pa) reform countries but little if any progress in the slower- * BOD El Nitrats 13 Metai Oil products reform countries. Source: Electrowatt 1998. 22 Economic Reform and Environmental Performance in Transition Economies Box 6. No improvement in some hot spots despite declining industrial emissions Sverdlovsk Oblast is one of the major centers of heavy industry in the Russian Federation, with a strong concentration of ferrous and nonferrous metallurgy, machinery, and electric power generation. The most pressing environmental problem facing the oblast is the contamination of water bodies with heavy metals and toxic chemicals. For the past 10 years, drinking water has been polluted, and an estimated 90 percent of the population has been using substandard, often hazardous, drinking water. The problem is further aggravated by inadequate drinking water treatment and poor water supply facilities. The rivers in the oblast are among the most polluted in the Urals region, with practically all water bodies dassified as "very dirty" or "exceptionally dirty." Despite a 30 percent decline in industrial production during 1990-94 and related decreases in emissions, water quality failed to improve and in some cases further deteriorated. The Neyva River, for example, had to be reclassified from "dirty" to "very dirty." Pollutants originate from point sources and include oil products, ammonium salts, and various heavy metals. The concentration of these substances often exceeds the maximum permitted concentration (MPC) for fishing purposes by several hundred times. Samples taken from the Salda River, for example, showed concentrations of copper 200 times the MPC; of manganese, 50 times; of zinc, 23 times; of iron, 15 times; and of oil products, 7 times. Groundwater pollution from point sources is also prevalent. Within the complex factors affecting pollution and water quality, two stand out: * The presence of already accumulated pollutants, which diminishes the effects of relatively short-term decrease in emissions * The negative effects of declining capital stock, deferred operations and maintenance measures, and lack of pollution ibatement as a result of the general economic crisis and weak environmental management and regulatory efforts. Pollutants accumulated in bottom sediments of water bodies from past discharges are a major source of continuous contamination. Additional pollution comes from contaminated soils, which are washed into adjacent water bodies, and from contaminated biota. These residues are expected to continue to largely offset decreased pollution discharges in the oblast. Unless lasting pollution abatement measures are taken, revival of economic activity in the oblast will mean a further buildup of contaminants. Despite the efforts to conserve water resources, the negative impact of the general economic crisis largely offsets any benefits from the decrease in pollutant discharges. Capital investments in the oblast are dwin- dling, and investment priorities are focused on maintaining production capacity. Not surprisingly, the implementation of scheduled investments in industrial pretreatment and sewage treatment facilities has fallen substantially behind schedule. The worst impact of the economic crisis stems from the decline in operation, maintenance, and repair (OMR) expenditures and activities, which has led to significant dete- rioration in water treatment. Most of the water treatment installations are outdated and overloaded. Be- cause of lagging OMR, more than half of them are in a state of emergency. Source: Electrowatt 1998. Chapter 4 The Health Impacts and Costs of Environmental Damage educing the adverse environmental impacts form countries. As guidance for policymakers in set- on health was the EAPs key criterion for set- ting environmental priorities, the following discussion ting environmental priorities. Environmen- attempts to (a) provide an overall picture of how the tal health concerns were reinforced by the disturbing human health effects and costs of air pollution and trend of diverging life expectancy between CEE/NIS water-related health damages have changed since and western countries. The divergence, which start- transition began and (b) compare the magnitude of ed during the 1960-80s, increased sharply after 1989, these environmental problems. especially among males. The largest declines in life expectancy occurred Air pollution in the NIS -most markedly in Russia, where the av- Significant reductions in air pollution have been erage life expectancy for males fell from 63.8 years in achieved in many transition economies, especially 1990 to 57.3 years in 1994, with the largest declines with respect to particulates and sulfur dioxide. There occurring in 1992 and 1993. Similar trends have been is evidence that reductions in total emissions have observed in other republics of the NIS. In Kazakhstan, been translated into real improvements in the levels for example, average life expectancy dropped by 3.7 of exposure, at least in the CEE countries.12 years during 1990-96 (Table D.4). The following analysis draws on an extensive The key factors of these trends are deteriorating sample of cities, on dose-response relationships es- health services and the socioeconomic stress brought tablished in a number of studies, and on economic about by difficult economic and social changes for valuation techniques to estimate the impacts of chang- large segments of the population during transition. es in exposure to key air pollutants on mortality and In many transition economies, the economic crises - morbidity in the exposed populations, and the costs. accompanied by declining GDP, hyperinflation, rising (See Annexes B and D for a description of the meth- unemployment, deteriorating social services-have odology.) Nonhealth costs associated with air brought large segments of the population into pover- pollution are not taken into account in this analysis. ty. Although environmental factors are not the According to most studies that have attempted to primary reason for worsening health conditions in the quantify the full range of costs of air pollution, non- region, they exacerbate existing trends. health costs (excess mortality and illness) account for The analysis in chapters 1-3 showed that eco- less than 20 percent of total costs. nomic reform in the transition economies has led to The following analysis covers 57 CEE and NIS significant reductions in pollution and improvements cities for which comparable data for 1990 and 1995 in environmental quality, especially in advanced re- were available. The results of the analysis are based 23 24 Economic Reform and Environmental Performance in Transition Economies on the reductions in mortality and illness that would suggests that since exposure to high levels of airborne be achieved by reducing annual average levels of ex- particulates increases the risk of mortality at all lev- posure to not more than 50 gg/m3 PM10 (particulate els of mortality,13 even constant or improving levels matter 10 microns or less in aerodynamic diameter). of ambient air quality lead to higher numbers of ex- This represents a realistic target; it is the middle of cess deaths because of the deteriorating health of NIS the range recommended in the pre-1997 World Health populations. Most of the increase in mortality from Organization guidelines and is similar to pre-1997 other causes, as well as from air pollution, is concen- standards for the United States and the EU. trated among males. There was also a small increase The two largest components of overall health in the average mortality rate for CEE countries, from damage are premature mortality and excess cases of 11 to 12 per 1,000 population, but it was not sufficient chronic bronchitis, primarily resulting from exposure to offset the gains from decreased exposure to airborne to particulates. With the 50 mg/m3 target as a basis, it particulates. was possible to estimate excess mortality (the num- Figure 20 shows the numbers of excess new cas- ber of premature deaths) attributable to exposure to es of chronic bronchitis, one of the most serious key air pollutants in the 57 sample cities, which have morbidity impacts associated with exposure to par- a combined population of 61 million. As Figure 19 ticulates. (These numbers are not affected by the shows, the estimated number of excess deaths in the increase in mortality rates and thus reflect changes in 57 cities was 32,000 in 1990, increasing to 40,000 in average levels of exposure.) The total number of ex- 1995. There was a large difference between CEE and cess new cases was estimated to be constant at about NIS cities. In the CEE cities, with a combined popula- 110,000 per year, but differences between country tion of 15 million, the number of estimated excess groups were significant. In the CEE countries, the es- deaths fell sharply, from 5,700 in 1990 to 3,700 in 1995. timated number of new cases fell by nearly one-half, By contrast, the NIS cities, with a combined popula- while the NIS experienced only a small increase. tion of 46 million, experienced an estimated increase Using economic valuation techniques, it is pos- from 26,000 in 1990 to 36,000 in 1995. sible to estimate the monetary cost of the damage to The underlying factors in these differences are health caused by air pollution. (See Annex B for a de- complex; they include changes in the levels of ambi- scription of the methodology and Box 7 for an ent concentrations, the size of the populations application to Krakow.)14 The results of this analysis exposed, and the overall health status of the popula- are shown in Figure 21. For the 57 cities as a group, tion. For example, the average mortality rate for NIS the estimated cost per year increased from $5.3 bil- countries increased from 11 per 1,000 population in lion in 1990 to $6.5 billion in 1995. These costs 1990 to 14 per 1,000 in 1995. Epidemiological evidence represented nearly 5 percent of total urban income Figure 19. Excess mortality due to air Figure 20. Excess cases of chronic bronchitis pollution by country group, 1990 and 1995 due to air pollution by country group, 1990 and 1995 50 - 120- 40 40 _ioo 0 _- - - l - (0~~~~~~~~~~~~~~~~~~6 30 -_ _ -8 0)~ ~ ~ ~ ~ ~ ~~~~~~~~4 10 ~ ~ ~ ~ ~ ~ ~ ~ ~~a6 20 ~~~~~~~~~~~~~~0~ CEEINIS CEE NIS 1990 CEE/NIS CEE NIS 019 1995 * 1995 Source: World Bank estimates using data for 57 cities with Source: World Bank estimates using data for 57 cities with comparable data for 1990 and 1995. comparable data for 1990 and 1995. The Health Impacts and Costs of Environmental Damage 25 Box 7. Economic damage from fuel combustion in Krakow, Poland A World Bank study (Lvovsky et al. forthcoming) examined the economic costs of pollution from fuel combustion in a number of cities around the world. In Krakow, Upper Silesia, as in other cities, the largest environmental damage is associated with fuel combustion in small sources, especially coal burning by households. Of all environmental costs, health costs dominate. (The shares of health, climate change, and other costs differ across sources, with health costs dominating in small sources, climate change costs in large sources.) There are significant differences between environmental costs across combustion sources. Large power and district heating boilers usually have medium- or even high-efficiency dust controls - typically with 80-90 percent removal efficiency -and their emissions are discharged from tall stacks and are widely dispersed. The most cost-effective measures are therefore those that target sources with no or low-efficiency controls and low or medium dispersion. High priority should be given to reducing emis- sions from small and medium-size sources, which include households as well as small and medium-size boilers in industrial and commercial premises. Comparison of the costs and benefits of pollution abatement alternatives leads to the following list of priorities (ranked by the cost per unit of improvement in average exposure to MO: * Installation of medium- or high-efficiency dust controls (with greater than 80 percentremoval efficiency) for industrial facilities and small district heating plants * Use of smokeless coal in small boilers and stoves * A shift from coal to gas in boilers in multifamily buildings and industrial or commercial premises. The balance between fuel switching - whether the move is to gas or to district heating - and the installa- tion of controls is largely a function of the size and age of existing equipment. The cost advantage of coal for large industrial boilers will justify the installation of dust (but usually not sulfur) controls. For older, medium-size boilers, the most cost-effective option will often be to switch to gas; for newer, more effi- Environmental cost of coal combustion in Krakow cient boilers, it may be cheaper to fit controls. In w 200 - the case of small industrial, commercial, or resi- lS0 -_ dential boilers, fuel switching will almost always lOo - be the cheapest option because controls are diffi- t 5 cult and expensive to install and operate, while m modern gas boilers are both more efficient and less labor-intensive to operate than coal boilers. X Source: Lvovsky and others (Forthcoming). _ for the sample cities in 1995. As would be expected countre 2g.rTotal costs9 o air pollution5by from the earlier results, the total estimated cost for country group, 1990 and 1995 the CEE cities fell from $1.2 billion (2.9 percent of ur- ban income) in 1990 to $0.7 billion (1.6 percent of urban e) 7 - income) in 1995. For the NIS cities, the burden im- 6 - posed by air pollution is much heavier, and it has c 5 increased by 40 percent-from $4 billion (4.6 percent == E 4 ~ _ f : fof urban income) in 1990 to $5.7 billion (6.6 percent of .C _ urban income) in 1995. -6 2- s 1 ~ _ Water-related health impacts CEE/NIS CEE NIS EI 990 The earlier analysis of trends in water quality and dis- 01995 charges of water pollutants in selected hot spots showed that both industrial discharges and agricul- Source: World Bank estimates using data for 57 cities with tural runoff have declined as a consequence of changes comparable data for 1990 and 1995. in economic activity and in the use of inputs to pro- 26 Economic Reform and Environmental Performance in Transition Economies duction. However, the effects of water pollution as Figure 22. Infant mortality rates by country group, such on the health of the population may be relative- 1990 and 1995 ly small. Individuals or small communities may suffer severe problems as a result of exposure to high levels of nitrates, heavy metals, and toxic chemicals in the 30 water supply or through the food chain. But in most o countries such problems are not a large threat to the 20 health of the general population, so long as people - have access to safe piped supplies of drinking water. 10 It is relatively easy to ensure that piped water sup- o IV 0- plies are protected from hazardous pollutants or are c All tranSoon CEE NiS Northwest NIS Southwast a 1 *Z ecanO,T1esEl19 treated to remove microbiological and other contam- z*995 inants. Source: World Bank estimates. The key questions are whether dwellings have piped water supplies and whether water quality is compromised by lack of proper treatment at the treat- country group. Within that group the deterioration ment plant or by contamination resulting from occurred primarily in Moldova and Ukraine. The sta- deterioration of the distribution network. In most CEE tistics for Moldova are difficult to interpret because and NIS countries, more than 90 percent of urban of internal divisions, but the worsening in Ukraine's households have access to piped water supplies.'" For infant mortality rate, from 13 per 1,000 live births in rural households, the situation is much worse. 1990 to 15 in 1995, suggests a significant problem. In- Rural or urban households that rely on surface fant mortality rates have also worsened in Azerbaijan water or shallow wells for drinking water are most (from 23 to 25) and Tajikistan (from 40 to 42); in those severely affected by cross-contamination from poor countries capacity to deal with public health problems or no sanitation. Even septic tanks may cause high has been seriously affected by war and the resulting levels of nitrates in water drawn from nearby wells. movements of population. Infant mortality rates de- More generally, poor management of sewage may teriorated in several other countries, including result in people being exposed to a range of infectious Kazakhstan, Russia, and Uzbekistan, immediately diseases, via contamination of drinking water supplies after the breakup of the former Soviet Union, but by where distribution networks are in poor condition, or 1995 they were no worse than in 1990 or had even via recreational activities such as boating and swim- improved. ming that bring them into contact with polluted water. ipoed. Water-related problems associated with lack of There have been notable instances of such problems access to water and sanitation are not the only influ- in many NIS countries since 1990.16 ences on infant mortality. Around, the world, infant The simplest and best indicator of the burden of mortality rates are also strongly affec*od by such ill health resulting from water-related problems is the level of infant mortality.17 As Figure 22 shows, aver- urbanization. Therefore, it is helpful to contrast actu- age infant mortality rates for all GEE and NIS countries declined slightly between 1990 and 1995, from 22 to al levels of infant mortality in the region with the rates 21 per 1,000 live births. The average for CEE coun- that would be expected holding the other relevant fac- tries declined from 18 to 16, while that for the NIS tors constant. Figure 23 shows that actual infant Southeast country group (Armenia, Azerbaijan, Geor- mortality rates, especially for the CEE and NIS North- gia, Kazakhstan, the Kyrgyz Republic, Tajikistan, west country groups, are significantly lower than Turkmenistan, and Uzbekistan) decreased from 32 to would be expected for their income levels and other 31. Only for the NIS Northwest group (Belarus, Mold- non-water-related factors. Most of the CEE and NIS ova, the Russian Federation, and Ukraine) was there countries have achieved relatively high levels of edu- an increase, from 16 to 17 deaths per 1,000 live births. cation, as well as high levels of access to water and Thus, there is no evidence of any clear increase sanitation, by comparison with other middle-income in water-related ill health outside the NIS Northwest countries with similar levels of GDP per capita. The The Health Impacts and Costs of Environmental Damage 27 Figure 23. Actual versus expected rates of infant the NIS Northwest countries has access to piped wa- mortality by country group, 1995 ter supplies, but reaching 95 percent coverage would be expected to decrease the average infant mortality so 50 rate only from 17 to 14, reducing infant deaths by i a 40 _ about 5,400 per year. (The much smaller change re- t M 30 * flects the success of these countries in achieving 9- c 20 [ i * ] . ]relatively low levels of infant mortality through pub- lic health and other measures even through the water iso ° |-- U , - * and sanitation infrastructure is in a poor state, espe- All transition CEE NIS Northwest NIS Southeast cially in rural areas.) economies rActual The additional effect on infant mortality of en- U Expected Source: World Bank estimates using expected values from cross-sectional suring 100 percent access to sewer connections for regression. urban populations is much smaller than the impact of access to piped water. Even in the NIS Southeast differences within the CEE/NIS group indicate the countries, where the existing average level of access success of countries in lowering infant mortality to sewer connections is less than 35 percent and the through effective public health and other measures. urban population is 50 percent of the total, reaching This achievement appears to be under some threat both the water supply and the sanitation targets in Ukraine but is being sustained in most other would be expected to bring average infant mortality countries. down ontly slightly, from 24 (with achievement of the As this analysis shows, improved access to wa- water supply target) to 22 deaths per 1,000 live births. ter and sanitation can have a significant impact on In addition to increased infant mortality, children infant mortality rates, especially where these levels and adults are affected to a greater or lesser extent by are relatively high. Because it is unrealistic, at least in diarrheal illnesses or by other water-related parasitic the medium term, to expect that countries with large and infectious diseases. Health analysts have devel- rural populations can reach 100 percent access to oped the concept of a disability-adjusted life year piped water supplies or sewer networks, it is neces- (DALY) to calculate the overall burden caused by dis- sary to set achievable targets. The targets used here eases and injuries of varying duration and severity are (a) access to piped water supplies for at least 95 (Murray and Lopez 1996).19 On the basis of an inter- percent of the population and (b) access to sewer con- nationally established ratio of DALYs attributable to nections for all urban households.'18 water-related mortality and illnesses, it is estimated Figure 24 shows the reductions in infant mortal- ity that would be expected given achievement of the Figure 24. Estimated impact of improvement in water target alone or the water and sanitation targets water and sanitation on infant mortality by country together. Ensuring that 95 percent of the population group,1995 in the NIS Southeast countries has access to piped 40 water supplies would be expected to reduce the aver- age rate of infant mortality by over 20 percent, from so 31 per 1,000 live births to 24. This represents an esti- mated reduction of over 11,600 in the annual toll of _ infant mortality in this group of countries. The effect cD is much smaller in the CEE countries as a whole be- cause the proportion of the population with access to O piped water exceeds 90 percent. (The exception is Al- z All tranms0on CEE NIS Northwest NIS SouthIst bania, which, with little more than 60 percent access uActual oWith water o With water& urban sanitation to piped water, is very similar to the NIS Southeast Source: World Bank estimates using expected values from cross-sectional countries.) Just over 70 percent of the population in regression. 28 Economic Reform and Environmental Performance in Transition Economies that the NIS Southeast and NIS Northwest countries each year is used in order to compare the costs of air lose 1.1 million and 0.45 million DALYs each year, pollution and waterborne diseases.20 respectively, as a result of limited access to piped A sample of 77 cities with actual or interpolated water. The total number of DALYs lost each year as a data on ambient air quality for 1995 was used to cal- result of water-related illnesses over the whole region culate the costs of urban air pollution. The cost figures would fall by 1.6 million if the target of 95 percent were scaled up by the ratio of total urban population access to piped water were achieved and by 1.8 mil- to the population of the sample cities for each coun- lion per year if both the water and urban sanitation try.21 targets were met. Figures 25 and 26 show the estimated total and It is rural populations that have the lowest levels relative burdens of premature deaths and illness as- of access to piped water and that experience more sociated with environmental factors, measured in total mortality and sickness from water-related diseases. The DALYs lost per year and in DALYs lost per 1,000 pop- rural population numbers about 35 million in the NIS ulation per year. The estimated total burden amounts Southeast countries and 60 million in the NIS North- to 3.9 million DALYs per year for the whole region. west countries. Thus, the above estimates imply a loss The largest components of the total burden are attrib- of about.31 DALYs per year per 1,000 rural inhabitants intheNISSoutheastcountriesandabout7DALYsper Figure 25. Total burden of mortality and illness year per 1,000 rural inhabitants in the NIS Northwest associated with environmental factors by country countries. Although the situation is much better than group, 1995 in other parts of the world, for example in East Asia 2500 (where the estimated average loss is nearly 90 DALYs per year per 1,000 rural inhabitants), nonetheless, the 22 burden of water-related disease for rural inhabitants 15______ ________ of the NIS Southeast countries implies a reduction of A about 2 years in average life expectancy. -_ The overall burden of environmental damage s ° - AU bansitfon CEE NIS Norimest NIS Southeast It is extremely difficult to devise satisfactory ways of I ecorrnies weighing the relative costs caused by air pollution and o A r pouludon a Water supply * Urban sanvation water-related health impacts. Air pollution in the Source: World Bank estimates. CEE/NIS region is almost entirely an urban problem, whereas rural populations bear most of the costs of Figure 26. Relative burden of mortality and illness dirty water. Most of those who die or fall sick as a associated with environmental factors by country result of air pollution are over 40 years old, while the group, 1995 deaths and illnesses resulting from dirty water are 0 20 concentrated among children, especially infants and 0a 0. children under age 5. Estimates of the costs of dirty 8 15 water cover each country as a whole, but the reliabil- ity of much of the key data is either unknown or poor. 10- Information on urban air pollution outside large cit- o * ies is very scarce or nonexistent. o .a 5 Despite these caveats, it is important to get a o sense of which environmental problems are likely to >; o-I1T I _ be most important in different groups of countries or a AN transiWn CEE NE Nydwrst NS Souhew even individual countries. The following analysis corpTies therefore makes a number of heroic assumptions to OAFrpoluton U Water & uba saito permit comparisons. The total number of DALYs lost Source: World Bank estimates. The Health Impacts and Costs of Environmental Damage 29 utable to air pollution in the NIS Northwest countries Figure 27. Relative burden of mortality and (1.6 million DALYs per year lost) and to poor water illness, selected countries supplies in the NIS Southeast countries (1.1 million DALYs per year lost). The CEE countries account for Ukraine e____ less than 10 percent of the total burden, even though SlovakPqp.. they have nearly 29 percent of the total population of na 3 the region. Thus, their relative position is much bet- Pland ter than that of the NIS countries. Although hot spots Lkhuania of air pollution remain a significant problem in the Latvia CEE countries, the improvements in air quality that Hungary have occurred since 1990 mean that air pollution is Abania no longer as dominant an environmental issue as it Abani once was. 0 2 4 6 8 10 12 Both air pollution and dirty water are much more DALYs lost per year per 1,000 people serious problems throughout the NIS than they are E3Air polktlon U Water & urban sanitation for the CEE countries, but the overall burden of dam- Source: World Bank estimates. age is distributed very differently between the NIS Northwest and NIS Southeast groups. In the Russian Federation and Ukraine, air pollution causes over Figure 28. Relative burden of mortality and illness three times as much damage as does dirty water, for the NIS Northwest countries whereas in Azerbaijan and Central Asia that ratio is reversed. Poland Figures 27 to 29 provide a more detailed break- Uraine down of the estimates for individual countries. Again, Russian Fed. it must be emphasized that these estimates are very Mold l crude and should be used only as a rough guide to Belarus the relative importance of different problems in the various countries. Still, certain features stand out. 0 1 2 3 4 5 6 7 8 9 10 Among the CEE countries, the largest relative bur- DALYs lost per year pe 1,000 people dens are associated with dirty water in Albania and oAirpollution *Water& urban sanitation poor air quality in Bulgaria. Among the NIS North- Source: World Bank estimates. west countries, air pollution imposes the heaviest burdens in Moldova, the Russian Federation, and Ukraine. Ukraine also has the highest relative cost Figure 29. Relative burden of mortality and illness associated with dirty water, which reinforces the ear- for the NIS Southeast countries lier point about the deterioration in its infant mortality rate since 1990. Improvements in the state of water lRusia _ infrastructure should receive high priority here. Finally, the heaviest costs of environmental dam- Turknenistan age are found in the NIS Southeast countries. Air Tajikistan pollution in Armenia and dirty water in Azerbaijan, Kyrgyz Wep. Kazakhstan, Tajikistan, Turkmenistan, and Uzbeki- Kazakrhstan stan account for significant health damage (a loss of Georgia more than 10 DALYs per 1,000 population per year). Azerbaijan Within this group, the highest priority should be to Annenia improve access to water and urban sanitation in Tajiki- 0 10 20 3 4 stan, Turkmenistan, and Uzbekistan, since in each of , , , z~~~~~~~~~~~~~ ~ ~~~~ Air pollution * Water & urban sanitation these countries the cost of dirty water exceeds 20 DALYs per 1,000 population per year. As noted, it is Source: World Bank estimates. 30 Economic Reform and Environmental Performance in Transition Economies rural populations that suffer most from lack of access the NIS Southeast countries, especially in to adequate piped water supplies. Efforts to rectify Tajikistan, Turkmenistan, and Uzbekistan this deficiency will therefore generate the highest re- * Improving urban air quality in Armenia, Bulgaria, turns in terms of improvements in human welfare and Georgia, Moldova, the Russian Federation, and prospects. Overall, these results suggest that, to Ukraine achieve the highest returns in terms of improving * Improving the state of existing water and sanita- human health and welfare in the region, priority tion infrastructure throughout the NIS, especially should be given to: in the NIS Northwest group. * Increasing access to piped rural water supply and the coverage of urban sanitation in Albania and Annex A Analyzing Trends in Environmental Quality, Emissions and their Determinants Data notes Data on air pollution and ambient air quality Information on emissions and environmental quality ., ny analysis of the links between econonmic is even more erratic than economic and energy data. change and the environment in the transi- Most countries estimate and report emissions of sul- . s tion economies of the Europe and Central fur dioxide, nitrogen oxides, and VOCs using the Aisa (ECA) Region must deal with the problems posed methods recommended by the UN Economic Com- by the very uneven quality of the data available. In mission for Europe under the terms of international particular, the breakup'of the former Soviet Union in conventions relating to transboundary air pollution. 1991 and the progressive dissolution of the former Yu- However, data on emissions of particulates is not stan- goslavia presents difficulties for any attempt to dardized and are, in any case, difficult to compile compile consistent data that span the full period of because of the diversity of sources-ranging from the transition from 1988 onwards. The results report- large power plants to the multitude of small boilers ed in this document rely upon economic and and coal-burning stoves. (Of course, the same sourc- environmental data from a wide range of sources. es also emit sulfur dioxide and nitrogen oxides, but for these pollutants it is easier to rely upon standard Economic and energy data emission coefficients per ton of fuel burned that are based on the sulfur content of the fuel and the com- Wherever possible, the primary sources of data used bustion technology.) As a consequence, coverage of in the report were the databases compiled by the small source emissions of particulates may be incon- OECD (including the International Energy Agency) sistent across countries since the data for many NIS and the World Bank, because some attempt is made countries relate only to large industrial sources. to ensure consistency both over time and across coun- For information on air quality, the report relied tries in the information that is recorded in these on the databases compiled by WHO (for their AMIS databases. Even so, there are many missing observa- system) and OECD, since these are the most consis- tions, often as a result of the collapse of statistical tent and reliable data on urban air quality for CEE systems in some of the NIS countries. In other cases, countries. However, these cover no more than one or recorded values for some series-particularly for two cities in each country, so that the data have been energy consumption, both in total and by compo- supplemented by other information from a variety of nent-appear to be so erratic that we decided to drop national and international sources. Many of the coun- such data from our analysis. tries of CEE received considerable assistance from 31 32 Economic Reform and Environmental Performance in Transition Economies donors to upgrade their equipment for collecting and sults are kept in the form of paper records which are analyzing data on ambient air quality in the early easily lost or can only be compiled with a consider- 1990s. As a result, recent data on air quality from these able effort. countries are usually more reliable than older data. There are also some large discontinuities in data se- Analytical methods ries-especially for dust or particulates-where the Trend analysis location of monitoring equipment or the procedures For the reasons outlined above, little weight can be followed were substantially altered. attached to particular observations or even trends for For the NIS, the major problem with air quality individual countries. Instead, the compilation of eco- data for many cities arises from the manner in which nomic and environmental data must be regarded from the figures were reported. Under the Soviet system it a statistical point of view as a set of observations with was standard to express observations as multiples of substantial errors in measurement. It is, therefore, the maximum allowable concentration-e.g. 100 gg/m3 necessary to be very cautious about relying upon con- for dust. In many cases, the published averages were ventional regression analysis to establish the nature rounded to a single significant figure -such as one, of the relationships between the various variables, two, or three times the maximum allowable concen- since the regression coefficients will be biased towards tration. This accounts for the frequency with which zero by an amount that is difficult to estimate. concentrations of 100 or 200 gg/m3 occur in the pub- Instead, we have relied upon simple analysis of lished statistics. It has been sometimes possible to variance techniques to identify differences between obtain more detailed information from a variety of groups of countries and over time, since these are other sources, for example, data collected either by more robust with respect to the impact of measure- local or regional environmental agencies or, in some ment errors. Thus, the basic equation used for each cases, by Sanitary-Epidemiology Stations. series is: log( Xi,) = as + br, Di, + b,t Di, + ui, (1) Data on water pollution and ambient water qity where: It is even more difficult to obtain consistent data on emissions of water pollutants and on ambient water i is the country quality than for air pollution and ambient air quality. t is the year Tlhis is especially true for large cities where ambient r stands for countries that are dassified as ad- air quality data are more readily available than data vanced reformers on rivers which pass through them. Most NIS countries continue the former Soviet s stands for countries dassified as slower reformers. practice of recording the total volume of wastewater The coefficients: discharged with separate figures for wastewater that a. is the base value of the variables is discharged with and without some form of treat- I ment. Some regional environmental agencies compile brt and b5t are trend variables for advanced and data on discharges of BOD, metals, and other pollut- slower reform groups of countries ants but-like the wastewater discharge data-this u.t is the error term seems to be based on information supplied by large D. and D. are du industrial plants, which usually account for no more y vas than 20-30 percent of all discharges. ues 0 or 1 according to the group of reformers in The lack of reliable data on water quality is i which country i is classified. part an issue of reporting, especially in the NIS. The Since the variance components model is speci- successors of the Soviet State Committee on Meteo- fied in levels, it is not necessary to weight the rology and Hydrology (usually known as Hydro-Met) observations when estimating equation (1). In a few continue to collect and analyze water samples for cases the analysis focuses on shares rather than lev- many sampling points all over the NIS. However, els, for which levels of total consumption were used sampling frequencies are erratic and most of the re- as the weighting variable. The year 1991 was used as Annex A - Analyzing Trends in Environmental Quality, Emissions and their Determinants 33 the base year for the analysis largely for reasons of Fpit is total consumption of petroleum products. data availability. This was the year in which the re- formprocss bgan n may CE contris, ad itwas This specification implies that emissions are the prod- form process began in many CEE countries, and it was ucof the transitional year prior to the dissolution of the uct of: Soviet nion an the fomer Yugslavia.(a) a country-specific term (A.) which is constant over Rather thionand displayn ymear bygo yearvialues.time and which is normalized by setting the value of A. = 1 for one country (the Czech Republic), so the time components, some of the figures show aver- that'the values of this term reflect the intensity of age trends in energy consumption or environmental emissions in each country relative to the level of quality over the full period. In these cases, the trends the base country; are estimated as the coefficients br and b5 in the equa- (b) two sets of trend variables which capture changes tion: in emission intensities in the advanced and slower- log( Xi,) = as + br Dif t + b, Di, t + ut (2) reform countries (B1t, BF) so that B1t represents the relative emission intensity in advanced reform Changes in emission intensities countries in period t with both B =1 and BA = 1 for t = 1991; and All of the transitional economies in the ECA region (c) a baseline measure of emissions of the pollutant are experiencing traumatic macroeconomic and struc- in country i given the country's economic activ- tural changes. The scenarios of emissions examined ity, industrial output, and fuel use in year t, where in the EAP combined the short term effects of macro- the coefficients a, b, g. and gp are used to predict economic and industrial adjustment with the longer average emissions for the whole sample. term impact of economic reform on the efficiency with The equation for estimation that is implied by this which resources are used and on environmental per- specification is: formance. In analyzing the actual changes in emissions since transition began, it is important to separate these effects. The following outlines a frame- log( E1 ) = aj + b4t Dir + b5t Di5 + (3b) work for estimating the contribution of various factors log( cY,1 + , + yCFCs1 + 'yFp) + ui 3 to changes in overall emission loads and to identify whether the fall in total emissions of air pollutants is This is non-linear in the coefficients and was, there- simply the consequence of lower levels of GDP, in- fore, estimated by non-linear least squares. A linear dustrial production, and energy use or whether there specification may be obtained by redefining the coef- have been a real improvement in emissions intensi- ficients a, b, .. as elasticities which yields an alternative ties, i.e. in emissions per unit of output and energy equation for estimation: use. In order to disentangle the joint effects of declin- log( Et ) = a, + bn D1, + bst Di2 + a log(Y1) + (4) ing economic activity and changes in energy use from j3 log(Zi1) + y. log(F1ii) + yp log(Fpi,) + uit improvements in environmental performance associ- The disadvantage of this formulation is that it can only ated with economic reform we have assumed that total be applied when all of the values of Y, Z, and F are emissions of each pollutant in country i in period t strictly positive, whereas (3b) can be estimated even can be expressed as: if the consumption of one fuel is zero or not recorded. Eit = Ai * (Brt Di, + B~t Di) The results from using the alternative specifications (aY1t + I Zit + yc Fci + yp Fpt) (3a) are quite similar with respect to the evolution of emis- sion intensities, so that we will focus on the results Where: from the non-linear specification (3b). Eit is emissions of the pollutant in country i in year t Table Al gives the estimated coefficients for the Yit is GDP four air pollutants examined - particulates/ dust, sul- Z.t is industrial output fur dioxide, nitrogen oxides, and VOCs. The number Fit is total consumption of coal and other solid of observations varies from 54 on 9 countries for par- fuels ticulates to 93 on 14 countries for sulfur dioxide. In 34 Economic Reform and Environmental Performance in Transition Economies all cases the data for 1996 onwards are so limited that The results show that there were clear, and sta- these years were excluded from the analysis, so that tistically significant, differences between the trends the estimation is based on data for as many years as in emissions intensity for the advanced and slower- were available in the period 1989-95. Note that the reforming groups of countries for 3 out of the 4 relative emission intensities for each year are obtained pollutants -not VOCs for which the difference is not by taking the exponential of the brt and b5 coefficients. statistically significant. For all of the pollutants there Thus, the estimated coefficient of -0.222 for bft in 1994 was a substantial decline in emissions intensities from for particulates means that the average emissions 1989 to 1995 in the advanced reform group, whereas intensity in advanced reform countries in 1994 was the changes for the slow reform group were much exp(-0.222) or 0.8 times that in 1991. smaller and inconsistent over time. Table A. I Estimated coefficients for emissions of air pollutants Particulates S02 NOx VOCs Advanced reform ( br) Year 1989 0.028 0.051 0.139 0.335 (0.098) (0.034) (0.057) (0.115) Year1990 -0.016 0.016 0.047 0.118 (0.091) (0.032) (0.045) (0.072) Year 1992 -0.065 -0.045 -0.008 0.048 (0.039) (0.030) (0.041) (0.069) Year 1993 -0.166 -0.117 -0.073 -0.027 (0.036) (0.029) (0.040) (0.067) Year 1994 -0.222 -0.168 -0.160 -0.008 (0.041) (0.032) (0.041) 0.065) Year 1995 -0.315 -0.276 -0.199 -0.106 (0.044) (0.035) (0.047) (0.064) Slower reform (bn ) Year 1989 0.102 0.143 0.102 (0.082) (0.095) (0.068) Year 1990 0.020 -0.025 0.007 0.072 (0.013) (0.020) (0.020) (0.018) Year 1992 0.013 -0.021 0.002 0.070 (0.016) (0.023) (0.023) (0.020) Year 1993 -0.046 -0.009 0.074 0.036 (0.030) (0.035) (0.036) (0.034) Year 1994 -0.043 -0.011 0.080 0.040 (0.048) (0.059) (0.056) (0.051) Year 1995 -0.087 -0.004 0.073 -0.023 (0.055) (0.059) (0.079) (0.063) GDP ( a ) -0.0004 0.0220 0.0209 0.0199 (0.0025) (0.0069) (0.0044) (0.0032) Industrial output ( f ) 0.0144 0.01 12 -0.0128 -0.0228 (0.0025) (0.0113) (0.0062) (0.0033) Total coal use (Yc) 0.0109 0.0250 0.0005 -0.0015 (0.0042) (0.0116) (0.0062) (0.0035) Total petroleum use 0.0045 -0.0013 0.0078 0.0005 ( Y) (0.0026) (0.0023) (0.0042) (0.0011) Note: The figures in brackets are the asymptotic standard errors of the coefficients. Annex B Estimating the Health Impacts of Air Pollution I n this annex (which draws on Lvovsky and oth- of ill-health which are associated with exposure to air ers 1999, which describes the methodology in pollution. Both of these topics have been the subject more detail) we outline the methodology that was of extensive research and debate because they have used to estimate the health damage caused by urban enormous implications for the setting of priorities with air pollution in the transitional economies of CEE/ respect to environmental and other policies. Interest- NIS. It is based upon an approach that has been widely ed readers should refer to Ostro (1996) and Lvovsky used in recent years to estimate the benefits of im- and others (1999) for fuller descriptions of the basis provements in air quality resulting from the adoption for the methodology, reviews of the relevant litera- of stricter ambient standards or the imposition of ture, and examples of how the methodology has been stricter limits on emissions-for example, in the U.S., applied in developing countries. the European Union, and the U.K. There are minor variations between different applications of the ap- Applying dose-response relationships proach according to the exact questions that are addressed and the nature of the data available. The W s results presented in this report are based on a rela- to pollutant j in an urban area that we are studying is tively simple analysis because we are interested in the currently A. while the relevant air quality standard damage associated with the exposure of urban popu- for the pollutant is S.. We will define: lations to levels of air pollution in excess of particular DA = max (A - S., 0) (5) air quality standards. Thus, it is not necessary to mod- - m ( el the dispersion of pollutants which is required only if one needs to estimate how reductions in emissions If the population exposed to the air pollutant in the of one or several pollutants will affect ambient con- urban area is P and we are interested in the incidence centrations in a particular town or city. of a particular health effect denoted as H., then the First we focus on the dose-response relationships reduction in this incidence as a result of the improve- between levels of exposure to various air pollutants ment DA. in exposure to pollutant j will be: and the incidence of premature deaths or of more or less severe illness due to a range of diseases. To trans- DHi = b.. * DA. * P (6) late such measures of ill-health into monetary terms - for comparisons with the costs of mitigation or with where: overall levels of income -it is also necessary to de- b.. is the slope of the dose-response function of health ef- velop a consistent basis for valuing the various kinds fect i with respect to changes in exposure to pollutant j. 35 36 Economic Reform and Environmental Performance in Transition Economies Acute vs. chronic impacts by air pollution from acute effects. The reason is that The interpretation of this formula depends upon the this establishes a lower bound for the impact of air pol- nature of the data available as well as the epidemio- lution on health. A similar approach has been adopted logical model from which the slope of the dose- in choosing dose response coefficients for various ill- response function is derived. In the case of air pollu- nesses. The most important of these is the incidence tion -in particular, exposure to particulates - there is of chronic bronchitis which is associated with substan- tial medical costs and impairment of capacity to an important distinction between the many epidemi- ta eia ot n mareto aaiyt anogimpotaltudiestinctio between the manyt epidems engage in a wide variety of activities. For this illness ologial sudie thatexamne te acue efects the lower bound value of the dose response coefficient associated with day to day variations in levels of air use her ione-half of the cenrestimae fromet pollution and the much smaller number of studies best available study. which examine the chronic effects associated with dif- Readers who wish to et a sense of the scale of ferences in average exposure over many years. The damage caused by air gton a one acceptssthe of problem in relying studies that deal with acute effects idemiological evidence of chronic effects should is that these may amount to no more than a shifting multiply the figures presented by a factor of 4 or 5. in the time at which the affected individuals fall ill or Thus, the reservations about the quality of the data die. There is a very large body of evidence which on exposure to air pollution in cities throughout the shows that a sharp increase in exposure to either PM25 region discussed in Annex A are swamped by the or PM1O will be followed within a short period of 1 to deliberate downward bias in the estimates of dam- 5 days by an increase in both deaths and various ill- age resulting from the use of lower bound values for nesses in the population affected. However, those who the dose response coefficients. die are typically people who were already in poor health, so that the effect of the higher air pollution Exposure to airborne particulate matter may simply be to advance their date of death by a A related set of questions concerns the way in levels few months at most. This is, of course, not a trivial o xoueadteepsdppltosaeetmt matter, but it is very different from an increase in the ed ese are link ed t our sting of the motlt aefryugridvdaswihlast ed. These are linked to our understanding of the mortality rate for younger individuals which leads to mechanisms through which exposure to air pollution telso maeny yearsdof aali continue totakeaccount affects health, which has gradually changed since the Why,then do nalyts ontiue t tak accunt middle of the 1980s. At that time it was accepted that of the dose-response parameters derived from stud- health e should be analyz ed bfsg ies of the acute effects of air pollution? The reasondamage should be analyzed by focusmg on lies in the fact that there are only a very small number exposure to PM1O and that measures of dust (TSP) of highe quality studiet the re lonly averymal conibefr were often poorly correlated with PMI, because the of high quality studies of the long term, chronic, ef- share of natural dust, which is dominated by large fects, because it is extremely difficult and costly to particles, collected by TSP monitors varies greatly by mount studies that monitor the health of large panels season and location (see Box B.1). Since then, biologi- of individuals over many years. Both of the two clas- cal and epidemiological research has built up a strong sic studies (Dockney and others 1993, and Pope and case for the view that it is fine particles - as measured others 1995) rely upon data that were collected for by PM25- which are the primary source of damage completely different reasons -primarily to study the to health. It is this research which underpins the deci- causes of heart disease and cancer. Further, the re- sion of the U.S. EPA to adopt new standards for sults of these studies suggest that long term exposure exposure to particulates that are based on PM2.5 rath- to even moderate levels of certain air pollutants has a er than PM10. very large effect on the incidence of mortality and The crucial point of PM2.5 is that fine particles, some types of illness. The b.. coefficients derived from because they are so small, remain suspended for long these two studies are 4-5 times the consensus value periods of time. As a result, levels of PM2.5 tend to be of the coefficients derived from studies of acute ef- much more uniform across urban areas including fects. metropolitan regions than PM10 or, especially, TSP.22 In this report we have used the consensus value Further, fine particles penetrate indoors through nor- of the dose-response coefficient for mortality caused mal air flows, so that comparisons of monitors placed Annex B - Estimating the Health Impacts of Air Pollution 37 Box B.1. Focus on Airborne Particulate Matter Exposure to sulfur dioxide and nitrogen oxides The situation is somewhat different for sulfur diox- Airborne particulate matter represents one of the ide and nitrogen oxides. Again, there has been a largest environmental heath threats associated with air pollution. Chronic exposure to particu- change in the scientific assessment of their contribu- lates can lead to premature death by exacerbat- tion to the health damage caused by air pollution over ing respiratory illness, pulmonary disease or the past 10-15 years. Most of the current biological cardiovascular disease. Acute exposure (short- and epidemiological evidence suggests that the long term peaks in the levels of particulates) can in- term damage caused by direct exposure to these pol- crease a chance that a person in a weakened state lutants is minimal. It is certainly true that very high or an especially susceptible person will die (some levels of both sulfur dioxide and nitrogen dioxide can studies have also found a correlation between . . . mortality and other pollutants like sulfur diox- provoke breathing difficulties, especially for people ieortality one) other pollutants like sulfurdiox- with asthma. However, such concentrations tend to be highly localized and relatively infrequent in CEE/ As monitoring methods and data analysis have NIS countries, so that the period of exposure is rela- become more sophisticated, the focus of atten- tively short and the overall magnitude of the damage tion has shifted gradually from total suspended is small, though an attempt has been made to allow particulates (TSP) to inhalable particles below 10 f microns in diameter (usually measured as PM,d for these effects in the analysis. and to fine particles below 2.5 microns (PM25). More seriously, sulfur dioxide and nitrogen ox- The share of PM10 larger than 2.5 microns is ides are converted in the atmosphere into sulfates and termed coarse particles. Evidence from studies nitrates respectively. These form fine particles and, completed in the last 5-8 years suggests that it is thereby, contribute to the average level of exposure fine particulates which are most likely respon- to PM2 5*3 Hence, controlling emissions of sulfur di- sible for the excess mortality and morbidity as- oxide and nitrogen oxides may be a significant sociated with exposure to particulates. component of any strategy to reduce exposure to fine particles. But, for this study it is sufficient to focus on inside and outside houses show that people are ex- observed levels of particulates without knowing the posed to very similar levels of PM2.5 wherever they extent to which these are comprised of primary parti- are. This means that the issue of differences in the level cles emitted by fuel combustion and other sources or and length of exposure is much less important for secondary particles generated in the atmosphere. PM25 than for measures with a substantial fraction of large particles. Exposure to ozone Thus, our analysis is based on using levels of A final issue concerns the effects of exposure to ozone exposure to PMIO as a proxy for exposure to fine par- produced by photochemical processes in the atmo- ticles. The dose response coefficients have normalized sphere. Again, short term exposure to high levels of on the assumption that the average ratio of PM10 to ozone can provoke breathing difficulties and acute TSP is 0.65 which is consistent with data for the US discomfort for asthmatics. It is much more difficult to and Europe and is likely to be reasonable for coun- establish whether differences in the average level of tries in CEE/NIS, since the primary sources of exposure to ozone are associated with significant im- PM10-in particular, combustion of fuels-are quite pacts on health, in part because most of the data relates similar. Further, it is reasonable to assume that the to peak levels of exposure. There is little doubt that entire population of large towns, cities, and metro- such differences do affect the growth and yields of politan regions are exposed to a level of PM2.5 that is trees and certain crops, so that the agricultural losses equivalent to that derived from the annual average caused by ozone can be significant, but we have not level of PM10 recorded at the monitor or monitors from attempted to estimate the non-health damages caused which the data were obtained. This assumption re- by air pollution. Further, exposure to high peak lev- flects the basis on which the dose response coefficients els of ozone is not a significant problem for most of were obtained in the studies of the chronic effects of the countries covered by this study. Two countries air pollution. -the Czech Republic and Hungary -do have inter- 38 Economic Reform and Environmental Performance in Transition Economies mittent episodes during the summer months, but all mortality rate in Russia increased from 11 per 1,000 these are largely the consequence of the transport of population in 1990 to 15 in 1995 and from 12 to 14 in pollutants from Austria and southern Germany. Ukraine. In contrast, the mortality rate in Uzbekistan -with rapid population growth and fewer of the Dose-response coefficients problems which led to the deterioration in health sta- In the case of mortality, the b.. dose response coeffi- tus in Russia and Ukraine-remained constant at 6 cients are country-specific. This is because epidemio- per 1,000 population. Other NIS countries such as logical evidence suggests that the effect of exposure Kazakhstan fall between the two extremes in the lev- to air pollution on mortality is multiplicative rather el and trends in mortality rates represented by Russia than additive. This means that a given increase in and Uzbekistan. In the case of Kazakhstan, the mor- average exposure to particulates-for example, 10 gg/m3 tality rate increased from 8 to 9 from 1990 to 1995 -for a population of 1 million people will cause a which was not sufficient to offset the decline in aver- larger absolute increase in mortality when the death age exposure to particulates in large cities. rate for the exposed population is high than when it The inclusion of the baseline mortality rate in the is low. Hence, the dose response coefficient for mor- specification of the dose response coefficients address- tality, i.e., i = m, in city k can be written as: es concerns about the transfer of relationships from rich industrial countries, where most of the epidemi- bmj(k) = pj * M(k) (7) ological studies of the health effects of air pollution have been conducted, to poorer and less industrial- where: ized countries. While some uncertainty remains, M(k) is the baseline mortality rate for the exposed recent studies undertaken in less developed countries, population such as Mexico City, Santiago, Chile and Bangkok, j is the proportionate increase in mortality asso- Thailand, have provided support for the transfer of ciated with a one unit increase in average expo- relationships specified in the form of equation (7). sure to pollutant j. A related concern focuses on the linearity of the It is usual to use the current level of mortality relationship, especially at high levels of exposure, for the whole country in which city k is located as the because the average levels of exposure to PMIO and baseline mortality rate, but a case can be made for sulfur dioxide in some developing countries as well using specific regional or urban data in countries as in various countries in the ECA Region are much where there are large differences between regions or higher than were observed in the studies that provide between urban and rural areas. the basis for the coefficients used here. Unfortunate- Note that aggregate mortality rates are heavily ly, it is not possible to settle this issue. The studies influenced by the age structure and general health of above covered cities with high average levels of ex- the population. Thus, the relative impact of lower posure, but they focused primarily upon mortality exposure to air pollution will be smaller in countries with more limited analysis of other health effects. whose population is relatively young or whose over- Their result suggest that it is reasonable to transfer all health is good. This observation is particularly coefficients, even for high levels of exposure, until important in understanding what has happened to the better evidence is available. Nonetheless, more com- estimated damage caused by air pollution in the CEE/ prehensive studies are required before it is possible NIS countries. to provide any kind of definitive assessment of wheth- In several NIS countries-notably Russia and er the response to changes in exposure to air pollution Ukraine -the burden of ill-health caused by air pol- is similar at low and high levels of exposure. Table lution increased from 1990 to 1995, even though B.1 provides the details of the dose response coeffi- average levels of exposure did not increase in general cient used for this study. and clearly decreased in some large cities. The reason The air quality standards, Sj, used in the calcu- for the increase was the general deterioration in the lations were annual average values of 50 gg/m3 for health status of the populations exposed combined PMW sulfur dioxide, and nitrogen dioxide. These val- with the ageing of the relevant populations. The over- ues were chosen for consistency with air quality Annex B - Estimating the Health Impacts of Air Pollution 39 Table B. I Dose response coefficients used in the study (per I lpg/r3change in the annual average level of exposure) Health effect PM,0 S02 Mortality, percent change in all-cause mortality rate 0.084 Chronic bronchitis I 100,000 adults 3.06 Respiratory hospital admissions / 100,000 people 1.2 Asthma attacks / 100,000 asthmatics 3,260 Emergency room visits / 100,000 people 23.54 Restricted activity days / 100,000 adults 5,750 Lower respiratory illnesses / 100,000 children 169 Respiratory symptoms!/ 00,000 adults 18,300 Cough days! 100,000 children 1.81 Chest discomfort days / 100,000 adults 1000 Source: Lvovsky and others 1999. standards or recommended targets adopted by WHO, States, it is more usual to employ the COI approach the European Union and the U.S. EPA in the first half to value morbidity in different countries. The COI of the 1990s. Since 1997 the European Union and the approach relies upon estimates of the economic costs U.S. EPA have adopted stricter standards for ambi- of health care and the loss of output preceding recov- ent air quality that will gradually come into force over ery or death. The total cost is equal to the sum of direct the period up to 2008. In the case of the European costs (such as hospital treatment, medical care, drugs) Union, the ultimate standards will be 20 jig/m3 for and indirect costs (such as the value of output lost). PM10 and 30 gg/m3 for sulfur dioxide, though the lat- However, the COI approach fails to account for the ter was adopted for the protection of ecosystems rather disutility of illness, while subsidized or inadequate than for health reasons. The U.S. EPA's new standard medical services complicate the calculation of econom- for particulates, which is still the subject of legal chal- ic costs of health care. As a result, the analysis here lenges, is 15 gg/m3 for PM2. which is equivalent to relies upon the WTP approach when such estimates 23 gg/m3 for PM1O using the conversion ratio cited were available and uses the COI approach only when above. For the calculation of the health damage asso- they were not. ciated with urban air pollution we have used the early 1990s air quality standards to ensure that our esti- C broflCit?s mates are biased downwards. Further, there is little Chronic bronchitis may last from the beginning of the realistic prospect that most cities in the region will be illness through the rest of the individual's life, so that able to meet the stricter European Union and US. stan- it is not possible to use the COI approach. The esti- dards within the next 10 years, so that policy-makers mate of WTP to reduce the risk of contracting chronic may reasonably conclude that damage estimates bronchitis takes account of differences in the frequency based on unattainable standards are of little relevance and severity of the disease associated with different to the practical issues which they have to address. causes, since severe cases of chronic bronchitis are extremely disabling. Based on U.S. evidence, the Valuing morbidity caused by air pollution average willingness to pay to avoid a typical pollu- Several techniques are known to estimate the econom- tion-related case of chronic bronchitis was estimated ic value of the damage associated with illness and at $195,000 at 1990 U.S. incomes and prices. disability caused by exposure to air pollution. The approaches which are most frequently used are (a) Accute morbidity willingness to pay to reduce the risk of illness or dis- To overcome the limited evidence relating to either ability (WTP), and (b) cost of illness (COI). Since the WTP or COI valuations of many of the illnesses pro- literature on WTP to avoid the morbidity effects is voked or exacerbated by air pollution, it is useful to rather limited and based entirely within the United combine them with indicators which attempt to mea- 40 Economic Reform and Environmental Performance in Transition Economies sure individuals' perceptions of the Quality of Well- damage from 2 per million people per year to 1 per Being (QWB) on a cardinal scale from 0 (death) to 1 million people per year with a total population ex- (perfect health).24 This facilitates the extrapolation of posed to the earthquakes of 10 million people. The valuations for a small number of specific conditions annualized cost of meeting the stricter standards will to a much wider range of symptoms. It is important amount to $20 million per year, after allowing for the to distinguish between acute effects and chronic ef- net saving on reconstruction in the aftermath of earth- fects, because the irreversibility of a poor health state quakes. adds a significant component to the WTP for avoid- The question then is whether the government or ing this health state, which will not be captured by the population considers that the expenditure of $20 WTP estimates for the avoidance of temporary acute million per year is justified in relation to the reduc- disorders. Once the relationship between the health- tion in the risk of mortality equivalent to the loss of status index and WTP is found, WTP for any condition 10 statistical lives per year averaged over a period of which can be described using the QWB score, can be 10-20 years. No one can be certain when or even estimated. whether the reduction in lives lost to earthquakes will occur or who will turn out to be the beneficiaries of Private and social costs of illness the stricter building standards. That is why we refer The use of individual WTP to avoid particular health to the value of a statistical life, because the focus is on problems is problematic if the costs of health care are willingness to pay to reduce certain kinds of risk to borne by the public sector. The social costs of illness which a particular population is exposed. If this will- comprise the private WTP plus the publicly borne ingness to pay exceeds a value of $2 million per costs. On the other hand, COI valuation cover treat- statistical life, then the benefits of imposing the strict- ment cost plus loss of earnings, but not the disutility er building standards may be judged to exceed the from illness. It is clear that some components of COI costs that will be incurred. On the other hand, a much estimates may be used to supplement WTP estimates, lower value per statistical life would imply that the to reflect the full costs to society. In this study, we benefits of stricter building standards do not exceed have relied upon the extrapolated WTP estimates that the costs involved, unless there are other consider- correspond to the privately borne costs of morbidity ations that have not been taken into account. effects, as these can be viewed as a lower bound to A variety of valuation techniques have been used the full social costs. to estimate the value of a statistical life. These include the analysis of occupational risks and wage differen- Valuing mortality: The value of a statistical tials, contingent valuation models that rely upon life detailed survey questions, and various types of mar- The effects of air pollution on mortality can be best ket studies. The literature on this issue is relatively assessed by relying on the value of a statistical life well developed for rich industrial countries. Surveys (VOSL). It is important to be clear what this does not of the existing literature suggest a mean VOSL for the attempt to measure, because the notion of valuing U.S. that lies in the range from $3.6 million to $4.8 human life is so controversial in public discussion. million, again at 1990 incomes and prices. We have An estimate of the value of a statistical life does not used the lower value of $3.6 million for the willing- purport to measure the compensation that would be ness to pay to avoid a statistical premature death as required by or should be paid to an average person the starting point for estimating the value of a statis- who dies in a road accident or a plane crash. Rather, tical life for the risks of mortality associated with it is derived by considering a different problem. exposure to air pollution. As an example, suppose that a government is trying to assess what kind of building standards ransferrng valuations to transition economies should be satisfied in an area that is prone to inter- There are several uncertainties which complicate the mittent but usually moderate earthquakes. It is transfer of WTP estimates for rich market economies estimated that a particular set of standards will, on for use in valuing premature deaths caused by air average, halve the number of deaths from earthquake pollution in transitional economies. The most impor- Annex B - Estimating the Health Impacts of Air Pollution 41 tant of these arise from (a) differences in income Applying (8) by using average per capita GNP levels, and (b) differences in age profiles of the popu- for Yk gives ratios for (VOSL/Yk) of 322 for Russia in lations affected by air pollution and the risks that were 1997,429 for Ukraine, and 524 for Armenia.25 The im- used in deriving the VOSL estimates. plication that the relative willingness to pay for The most general way of adjusting for differenc- reductions in the risk of mortality in Armenia is more es in income levels is to calculate country-specific than 3 times the U.S. level seems an extremely strong valuations for country k using the formula: and probably counter-intuitive assumption. It is more reasonable to infer that estimates of the income elas- 0log ( Vk) p log ( Yk / Yus ) + log ( Vus ) (8) ticity of willingness to pay to avoid risks of death or ill-health are not robust to large income differentials where: on the grounds that they refer to cross-sectional dif- Vk and Vusare estimates for the valuation of pre- ferences in income within a country but should not mature deaths caused by air pollution in country be applied to differences across countries. We have k and the U.S., respectively thus chosen to assume that the income elasticity of Yk and are the per capita incomes of country k willingness to pay is 1, so that attention focuses pure- and the U.S., respectively p is the income elasticity of the relevant willing- o Eencso in income. ness to pay. ~~~~~~~Even so, it may be argued that income should ness to pay. not be measured by income per head in U.S. dollars The literature on the income elasticity of the converted using either the current or some average VOSL and other valuations of health effects is quite exchange rate. Instead, the conversion could be based limited, but it suggests an average value of 0.7. How- on purchasing power parity (PPP) exchange rates ever, there must be some question about the which tend to yield substantially higher estimates of plausibility of such an estimate when applied to the income per head in low income and transitional econ- very large income differences between the U.S. and omies - for example $2480 for Uzbekistan in 1995 relatively poor transitional economies. For example, using the PPP exchange rate and $980 using the World the VOSL for the U.S. that we have used is equivalent Bank Atlas method based on an average exchange to about 160 times the average per capita GNP of the rate. One practical difficulty is that PPP exchange rates country in 1990. are not available for all of the countries covered by Box B.2. DALYs as a measure of the burden of disease DALYs (disability-adjusted life years) are a standard measure of the burden of disease. The concept of DALYs combines life years lost due to premature death and fractions of years of healthy life lost as a result of illness or disability. A weighting function that incorporates discounting is used for years of life lost at each age to reflect the different social weights that are usually given to the illness and premature mortality at different ages. The combination of discounting and age weights produces the pattern of DALY lost by a death at each age. For example, the death of a baby girl represents a loss of 32.5 DALYs, and a female death at age 60 represents 12 lost DALYs (values are slightly lower for males). The use of DALYs as a measure of the burden of disease has provided a consistent basis for systematic comparisons of the cost-effectiveness of alternative interventions designed to improve health. When com- bined with the results of large epidemiological studies it enables public health specialists to identify priori- ties and focus attention upon developing programs which have the potential to generate large improvements in the health status of poor households in the developing world. In Africa and Asia, lack of access to clean water and sanitation and high levels of exposure to indoor air pollution resulting from the use of dirty cooking fuels account jointly account for more than 10% of the total burden of ill-health, mostly as a consequence of high levels of mortality among infants and young children under the age of 5. Source: Murray and Lopez (1996). 42 Economic Reform and Environmental Performance in Transition Economies our analysis. Even if they were, it is important to con- set of weights representing the number of DALYs lost sider the reason for attempting to value the costs of per case, it is possible to compute an overall figure of damage caused by air pollution or other forms of en- the damage caused by air pollution which can be com- vironmental damage. The absolute magnitude of the pared with the overall burden of ill-health from all values is much less important than their relationship causes. to total income or to the costs of measures that could This approach provides a method of adjusting reduce the damage. For such comparisons it is easier the value of a statistical life for the age profile of those to rely upon an average exchange rate rather than a who die as a result of exposure to air pollution. The PPP exchange rate. At the same time this method is average age at death of such is assumed, on the basis consistent with our general approach of seeking to of evidence from U.S. and other studies, to be 65, establish a lower bound for the magnitude of the costs which corresponds to a loss of 10 DALYs per death. of environmental damage. By comparison the populations at risk in the VOSL For many purposes, it may be sufficient to rely studies reviewed to obtain the estimate of $3.6 mil- upon an overall measure of the burden of ill-health lion had average ages in the range 40-45 which would caused by air pollution. In this analysis we have used imply an average loss of 22.5 DALYs per death. This the approach of calculating disability-adjusted life ratio yields an adjusted VOSL of $1.6 million per death years (DALYs) as our measure of ffil-health. This mea- at 1990 U.S. prices and incomes that was used for the sure has been adopted by the WHO and other present analysis. international organizations as the basis for making Estimates of the number of DALYs lost as a re- cross-country comparisons of the burden of disease - sult of non-fatal illnesses including chronic bronchitis see Murray and Lopez (1996) for a full description and were based on the figures used in the Murray & Lo- justificationofthemethodsusedinthesecalculations. pez (1996) analysis or by applying the ratio of By the incidence of premature deaths and the various willingness to pay estimates to the figure of 10 DALYs other forms of ill-health caused by air pollution by a per death from exposure to air pollution. Annex C Estimating the Health Impacts of Access to Water and Sanitation T here is ample evidence from intervention and stan and Turkmenistan. Nonetheless, as Chapter 4 cross-community studies to establish that ac- shows, infant mortality rates are relatively low for cess to improved water supply and sanita- most countries in the ECA region when compared tion can have a significant impact in reducing the with countries elsewhere with similar levels of income incidence of both morbidity and mortality associated and urbanization. with diarrhea, intestinal nematodes, and other tropi- Casual observation supported by detailed com- cal diseases -see, for example, the review of 144 munity studies show that income and education are studies inEsrey and others (1991). However, few stud- -both critical factors in understanding the impact of ies have attempted to use these epidemiological water supply and sanitation on health. Income mat- results as the basis for (a) estimating the total burden ters both as a determinant of, and possible proxy for, of ill-health caused by lack of access to clean water nutritional status and as an indicator of the resources and sanitation, and (b) comparing the cost of this ill- available for avoidance or mitigation expenditures. health with other types of environmental damage in Infants who are well-nourished are less likely to die order to establish priorities for environmental poli- during episodes of diarrhea. A higher average income cy.26 This annex will provide details of how the results means that parents are more likely to be able to buy obtained from a cross-country study of the impact of either bottled water or other clean water, thus reduc- water and sanitation on infant and under-5 mortality ing the frequency of such episodes, or to pay for were used to estimate the burden of ill-health associ- medical attention to mitigate their severity. Education, ated with lack of water and sanitation. especially of women, plays a critical role in dissemi- The primary indicators of the health impacts of nating the importance of personal and domestic water supply and sanitation that were used are (i) the hygiene in reducing the frequency and severity of infant mortality rate, defined as the number of deaths waterborne disease. Thus, the influence of these fac- of babies less than 12 months old per 1,000 live births, tors must be taken into account when estimating the and (ii) the under-5 mortality rate, defined as the num- impact of water and sanitation improvements. ber of deaths of babies and young children under 5 Diarrheal diseases are a-often, the-major years old per 1,000 children under 5. In the ECA Re- cause of mortality among infants and young children gion, infant mortality accounts for about 75 percent in Africa, Asia (except China), and Latin America, typ- of under-5 mortality. The average infant mortality rate ically accounting for 20-25 percent of deaths of in 1995 for the countries covered by this study was 21 children from 0 to 4 years old as well as about 8 per- per 1,000 live births, but this varied from under 10 in cent of the total loss of DALYs in the developing world the Czech Republic and Slovenia to over 40 in Tajiki- as a whole (ust behind lower respiratory infections). 43 44 Economic Reform and Environmental Performance in Transition Economies Other water-bome diseases impose substantial costs water. In most cases, the collection of sewage and the as a result of disability and loss of productivity by discharge of wastewater to more distant or less harm- those who suffer from them. The magnitude of such ful locations is more effective in achieving a reduction costs seems to be strongly correlated with levels of in e-coli or fecal coliform counts of, for example, bath- mortality for infants and young children. Thus, it is ing or shellfish waters. reasonable to use changes in these mortality variables Of course, there have been widely publicized in response to access to water supply and sanitation episodes such as the cholera epidemic in Lima, out- as a general indicator of the impact of water and san- breaks of hepatitis and other diseases in Kharkov, itation on the overall burden of water-bome diseases. Odessa, Riga and Rostov in the NIS, and deaths as a Small scale studies have shown that water sup- result of e-coli in the UK. In the case of Lima, the real ply and sanitation are not equally important as problem was the spread of the disease as a result of influences on the incidence of different diseases. Wa- poor water supply and sanitation which encouraged ter supply seems to be the critical variable for guinea its transmission within the large but very poor com- worm (dracunculiasis), schistosomiasis, and trachoma munities of recent migrants to the city. In Riga, the in the sense that access to ample water of reasonable problem was caused by a failure to operate the water quality for drinking and hygiene substantially reduc- treatment plant properly because of a lack of money es the transmission of these diseases. For hookworrn to buy supplies of the chemicals required to ensure infection (ancylostomiasis and necatoriasis) it is proper disinfection of raw water, which was contam- sanitation that is critical in breaking the cycle of trans- inated by untreated sewage. Similarly, the episodes mission, though its impact on incidence as opposed in Kharkov, Odessa, Rostov, and elsewhere in the NIS to severity of the disease is uncertain. The incidence have usually been traced to failures in maintaining and severity of diarrhea diseases and ascariasis seem infrastructure or in operating treatment facilities, com- to be influenced by both water supply and sanitation, bined with a willingness of some people to expose though it is not possible to conclude which of the two themselves to risk by swimming in contaminated riv- is more important. ers or lakes in spite of warnings. In such cases, the Since important questions of priorities may arise issue becomes one of how much the population is in allocating resources to the development of water willing to pay to avoid small risks of exposure to and sanitation services in poor communities, the study waterbome diseases which may be unpleasant and was designed to identify the separate influences of inconvenient but are rarely seriously disabling or fa- water and sanitation on mortality and, by inference, tal except for those who are frail or old. on morbidity. No attempt has been made to examine the impact of quality variables -in particular, the Epidemiological analysis quality of water supply and the extent and nature of A standard grouped logit model has been used in or- sewage treatment. der to allow for the S-shaped relationship between M., The overall burden of ill-health associated with -the number of babies who die before the age of 1 transmission through contact with untreated or par- year (or 5 years in the case of the under-5 mortality tially treated sewage outside the local neighborhood rate) per 1,000 live births in country i-and the inde- is relatively small. Children and adults may develop pendent variables X.. for j = 1, ..., J. diarrhea, cholera, typhoid, or hepatitis as a result of bathing in contaminated waters or eating contaminat- ed shellfish, but both mortality and the overall loss of g =i log ~~ao +XIaj Xi (9) DALYs from such causes are usually small. Moreover, (1000 - M) v =, conventional sewage treatment on its own yields min- imal-or zero health benefits. Unless the treatment plant This was estimated by a maximum likelihood proce- chlorinates its effluent - an unusual measure - or in- dure for grouped data using the number of infant cludes maturation ponds, land treatment or other deaths and the number of live births-both measured measures for bacterial removal, sewage treatment has in thousands -as the dependent variables. Since the little impact on the bacteriological quality of waste- group sizes are very large for some countries, the as- Annex C - Estimating Estimating the Health Impacts of Access to Water and Sanitation 45 ymptotic t-values for many of the parameters are participation in the labor force and urbanization could large. C)n the other hand, this will not cause the coef- go either way. A higher share of women in the labor ficients of variables - which have little or no influence force may mean that infants and children receive less on mortality rates -to be statistically different from attention, but the women may be better informed zero, because this is determined by the number of about the benefits of good hygiene practices. Similar- countries in the sample analyzed rather than the size ly, infants and children in urban areas may be more of the samples in each country. exposed to the diseases associated with squalid liv- Differentiating the logit equation yields: ing conditions, but equally their parents may have better access to information and services which reduce A Mj = aj Mj (I - - ) A Xjj (10) the chances of contracting these diseases or mitigate 1000 their effect. As for the educational variables, atten- If mortality rates M. are either very low or very high, tion has focused on female education but it is not clear then the corresponding slope coefficients for a given whether female illiteracy is a better indicator than to- change in X.. will be small, while the maximum slope tal iliteracy Curretteroi nto r in to - coefficients occur when M = 500. In practice, infant tal iltera Current enrollment of girls in primary mortality rates are all less than 170 per 1,000 live births educathon may be important as a channel by which and under-5 mortality rates are less than 320, so that health educaton reaches mothers. all of the observations fall in the lower part of the S- lare numberoof ntiv setin were shaped relationship. ~~~~estimated using the cross-country dataset in order to Twope deataonseshave benue o hesaitcl obtain the best basic model. Because of missing val- Two datasets have been used for the statistical analysis: ues it was impossible to test all of the key variables (1) The main dataset is a cross-country sample of 133 on the same subset of observations, since the sample countries based on the World Development Indi- size would have been so small. Thus, a series of pair- cators 1996 dataset.27This was supplemented with wise comparisons including the main alternative additional data on educational, health, and infra- specifications was examined. The final set of main structure variables prepared for the World Devel- independent variables was selected by choosing those opment Reports published in 1993-1995. which consistently yielded the lowest root mean (2) A second dataset was constructed by combining square error in these comparisons. This procedure the cross country dataset with pooled data for identified female illiteracy as the best educational 1965-90 taken from World Tables 1995.28 variable for inclusion in the model. Additional tests Since most of the data series are incomplete, five- suggested that none of the infrastructure variables yearly averages were constructed for 1965-69,1970-74, apart from those relating to access to water and sani- [... ],,1985-89,1990-93 using whatever data were avail- tation have a significant influence in determining able. This provides a quasi-panel dataset covering infant or child mortality rates. Similarly, the variables nearly 30 years, though the time series for each coun- which attempt to measure access to health services try are still too small to attempt a full cross-section, did not contribute to the explanatory power of the time series analysis. The World Tables data do not model and were, therefore, dropped. Finally, the anal- contaiin any information on access to sanitation, so the ysis compared linear and logarithmic specifications second dataset was used simply in order to cross- for the main variables with the conclusion that the check whether the results from a strict cross-country linear specification yielded the best results. The coef- analysis were robust to the inclusion of data from past ficients for the main variables in the equation and their time ]periods. The variables used in the analysis are asymptotic t-values are given in Table C.2. presented in Table C.1. It is possible that water supply and sanitation A priori one would expect that infant mortality have different effects for urban and rural households, (or under-5 mortality) will decline with increasing so various alternative specifications were estimated values of (a) income per person, (b) the level of fe- to examine whether this is case. The model decom- male education, (c) the availability of health services, poses the influence of the water variable by including (d) access to safe water and sanitation, and (e) indica- (a) access to piped water in urban areas multiplied by tors of infrastructure provision. The effects of female the urban share of the population, and (b) access to 46 Economic Reform and Environmental Performance in Transition Economies Table C. I Variables used in the analysis Dependent variables INF_MORT Infant mortality rate * no per 1,000 live births US_MORT Under-5 mortality rate: no per 1,000 live births Independent variables Income GNP-CAP GNP per person in US dollars (for 1994 in the cross-country dataset, deflated to 1990 prices for the pooled dataset) PPP_INC GNP per person at purchasing power parity in US dollars at 1994 international prices INCLOW20 Percentage share of income/consumption for the poorest 20% of the population Population and labor force PCT_URB Percentage of population living in urban areas PCT_RUR Percentage of population living in rural areas LABF_FEM Females as percentage of the labour force Education and health ILIT FEM Percentage of females aged 15 or greater who are illiterate ILIT TOT Percentage of total population aged 15 or greater who are illiterate PENR_FEM Percentage of females in age group who are enrolled in primary education POP_DOC No of people per doctor ACC HLTH Percentage of total population with access to health services Infrastructure SWAT_TOT Percentage of total population with access to safe water SWAT_URB Percentage of urban population with access to safe water SWAT_RUR Percentage of rural population with access to safe water SAN_TOT Percentage of total population vwith access to sanitation SAN_URB Percentage of urban population with access to sanitation SAN_RUR Percentage of rural population with access to sanitation ROAD_CAP Length of roads (km) per million people TEL_CAP Main line telephone connections per million people ELEC_CAP Electricity production (MWh) per million people piped water in rural areas multiplied by the rural supply is more important than urban water supply in share of the population plus a similar decomposition determining mortality rates. Indeed, the coefficient for of the sanitation variable. The results for both infant urban water supply is not statistically significant in and under-5 mortality suggest that the rural water any of the models examined, though its value is Annex C - Estimating Estimating the Health Impacts of Access to Water and Sanitation 47 Table C.2 Equations for mortality of infants and children under age 5 (using cross-country dataset) Infant mortality Under-S mortality Intercept -3.74 -3.69 -2.81 -2.76 (0.16) (0.17) (0.13) (0.14) GNP-CAP -0.00006 -0.00005 -0.00017 0.00022 (0.00002) (0.00002) (0.00002) (0.00002) PCT_URB 0.0127 0.0142 0.0137 0.0048 (0.0016) (0.0035) (0.0014) (0.0029) LABF_FEM 0.0153 0.0161 0.0105 0.0145 (0.0026) (0.0028) (0.0022) (0.0023) ILIT_FEM 0.0156 0.0145 0.0165 0.0143 (0.0010) (0.0009) (0.0008) (0.0007) SWAT_TOT -0.0078 -0.0120 (0.0008) (0.0006) SWAT_URB * PCT_URB -0.0029 -0.0028 (0.0026) (0.0021) SWAT_RUR* PCT_RUR -0.0090 -0.0114 (0.0009) (0.0007) SAN_TOT -0.0044 -0.0024 (0.0008) (0.0006) SAN_URB * PCT_URB -0.0124 -0.0015 (0.0030) (0.0025) SAN_RUR * PCT_RUR -0.0039 -0.0059 (0.001 1) (0.0009) -2IlogL 1108.2 1144.6 2399.7 2416.2 Sample of countries 71 69 68 66 Note: The figures in brackets are the asymptotic standard errors for the reported coefficients. consistently negative. The pattern of coefficients for shown in Table C.3. To put these figures in context, urban and rural sanitation is more puzzling. Again, the average levels of access to water and sanitation they are consistently negative, but their relative mag- for all countries in the region are 76 percent for water nitudes reverse in the equations for infant and under-5 and 51 percent for sanitation, so that the potential re- mortality. Urban sanitation has a much bigger effect duction in infant and under-5 mortality if these on infant mortality, whereas it is rural sanitation average levels were to be increased to 90 percent or which has the larger influence for under-5 mortality, more could exceed 25 percent. Other equations not reported here suggest that the As a working hypothesis, it will be assumed that pattern in the equation for infant mortality results is morbidity and mortality of those aged 5 or above will more consistent over alternative specifications. decline as a result of increased coverage of urban The relative importance of water and sanitation as influences on infant and under-5 mortality may be pipediwatertandtsewrseinrproportiono t the reduc- inferred from the equations by calculating the impact to nmraiyrts h cl fterdcini inferedfro th equtios b caculaingtheiinact the overall burden of disease has been estimated from on mortality rates of a 10 percentage point increases in the water and sanitation variables. To do this it is the data compiled for the Global Burden of Disease necessary to assume a level of urbanization since the study sponsored by the World Health Organization access variables are multiplied by the percentage of and the World Bank (Murray and Lopez 1996). This households affected. For this purpose we have used a gives estimates of the total burden of various diseas- value of 65 percent which is the average share of the es linked to water and sanitation in terms of DALYs urban population in total population for the transi- for former centrally-planned economies of the ECA tion countries in the ECA Region. The results are region as a group. 48 Economic Reform and Environmental Performance in Transition Economies Table C.3 Impact of water and sanitation on mortality rates in the ECA Region Percentage reduction in mortality if access were increased by 10 percentage points Infant mortality Under-S mortality All water supply 8.0 12.0 Urban water supply 2.0 2.0 Rural water supply 3.2 3.9 All sanitation 4.0 2.0 Urban sanitation 7.8 0.7 Rural sanitation 1.4 2.1 To scale changes in mortality, the total number the current level of access to water and sanitation in of DALYs lost each year to (a) diarrhea diseases, (b) country i are denoted by W. and S,, while the target hepatitis B & C, (c) tropical cluster diseases excluding levels of access are W.* and Si*. The predicted level of trypanosomiasis and Chagas disease, and (d) intesti- mortality at the target levels of access is Mi*. Then, nal nematode infections were summed to provide an the logit equation in (9) implies that aggregate estimate of the burden of water- and sani- tation-related diseases for the region.29 This was [Mi*/(1000 - MMI = [Mi /(1000 - M)] * normalized by the total number of deaths of children exp[aw(Wi* - Wi) + a1(Ss* - Si)] under 5 years old each year from the same group of Defining diseases.30 Thus, the scaling factor represents the av- erage number of DALYs lost as a result of water- and * = expl aw(Wi* - Wi) + a1(S1* - S,) ] (12) sanitation-related diseases per under-5 death associ- and ated with the same group of diseases. Applying it to R; M, /(1000 - Mj) (13) the estimated reduction in under-5 mortality as a re- sult of improvements in water and sanitation yields an estimate of the total number of DALYs that may this implies that be saved by such measures. The aggregate scaling Mi* = RjZj* / (1 + RjZi*) (14) factor is 48 DALYs per under-5 death from water- and sanitation-related diseases for the standard assump- which is the equation used to calculate the improve- tions, of which 34 DALYs represents the loss of DALYs ment in mortality rates. Note that this derivation holds due to the under-5 death and 14 DALYs represents constant the influence of unobserved variables which the associated burden of morbidity and of mortality lead the actual mortality rate for the country to be for people over 5 years old. Thus, under-5 deaths ac- higher or lower than that predicted from the estimat- count for 70-75 percent of the total burden of disease ed equation. affected by water and sanitation. The estimates of reductions in mortality rates due to improved access to water and sanitation that are The benefits of improving access to water presented here are based on two assumptions, which and sanitation are used both separately and in combination: We can use these results to estimate the reduction in * At least 95 percent of the entire population has ac- infant or under-5 mortality rates that would be ex- cess to piped water supplies pected if access to water and/or sanitation were * All urban households have access to sanitation. improved. The data on levels of access to water and The total number of DALYs saved as a result of sanitation for urban and rural areas separately is only the resulting decline in mortality rates was obtained available for some of the countries studied. Thus, we by using the current number of births annually in each have used the coefficients for SWAT_TOT and country and applying the estimates of the DALYs lost SAN_TOT in carrying out this analysis. Suppose that per under-5 death that are discussed above. Annex D Tables Table D. I GDP growth and GDP per capita in Central and Eastern Europe and the Newly Independent States, 1993 and 1997 Real GDP as % of 1989 value GDP per capita, US $ Central and Eastern Europe 1993 1997 1997 Albania 65 83 750 Bulgaria 68 60 1140 Croatia 53 60 4610 Czech Republic 79 91 5200 Estonia 63 74 3330 Hungary 82 88 4430 Latvia 49 52 2430 Lithuania 37 42 2230 Macedonia, FYR 1090 Poland 88 110 3590 Romania 72 81 1420 Slovak Republic 77 97 3700 Slovenia 83 97 9680 Newly Independent States Armenia 35 42 530 Azerbaijan 53 41 510 Belarus 78 ' 57 2150 Georgia 29 25 840 Kazakhstan 70 1340 Kyrgyz Republic 440 Moldova 540 Russian Federation 69 55 2740 Tajikistan 55 330 Turkmenistan 82 630 Ukraine 42 42 1040 Uzbekistan 88 90 1010 Source: World Bank, World Economic Indicators, 1998 CD-Rom; EBRD, OECD. 49 50 Economic Reform and Environmental Performance in Transition Economies Table D.2 Inflation in Central and Eastern Europe and the Newly Independent States, 1989/93 and 1993/97 Percent increase in GDP deflator Central and Eastern Europe 1989-93 1993-97 Albania 910 120 Bulgaria 890 6430 Croatia Czech Republic 120 Estonia 5470 150 Hungary 140 Latvia 5380 100 Lithuania 13780 200 Macedonia, FYR 180 Poland 1400 120 Romania 3190 1050 Slovak Republic 90 40 Slovenia 70 Newly Independent States Armenia 26090 Azerbaijan 23120 13820 Belarus 30680 39940 Georgia 463950 Kazakhstan 52160 Kyrgyz Republic 20760 530 Moldova 650 Russian Federation 41530 1760 Tajikistan 17030 Turkmenistan 85550 Ukraine 146230 10480 Uzbekistan 18700 19000 Source: World Bank, World Economk Indicators 1998 on CD-ROM; EBRD; OECD. Annex D - Tables 51 Table D.3 Motor vehicle use and change in Central and Eastern Europe and the Newly Independent States, 1990-96 Vehicles per 1,000 population Percentage increase in number of CEE Countries (1996) vehicles, 1990-96 Albania 32 106 Bulgaria 225 33 Croatia 195 10 Czech Republic 350 39 Estonia 323 40 -Hungary 273 27 Latvia 188 30 Lithuania 238 49 Macedonia, FYR 142 15 Poland 249 50 Romania 125 69 Slovak Republic 220 9 Slovenia 385 26 NIS countries Armenia 2 -65 Azerbaijan 49 -3 Belarus 102 69 Georgia 87 -19 Kazakhstan 81 -2 Kyrgyz Republic 32 -30 Moldova 54 1 Russian Federation 132 49 Tajikistan 2 -36 Turkmenistan Ukraine 92 45 Uzbekistan Source: World Bank, World Economic Indicators 1998 on CD-ROM; EBRD, OECD. 52 Economic Reform and Environmental Performance in Transition Economies Table D.4 Mortality and life expectancy in the Newly Independent States, Central and Eastern Europe, and the European Union, 1994 and 1995 Infant deaths from Infant deaths from intestinal infectious Total deaths from all couses per disease per infectious disease 100,000 live births, 100,000 live births, per 100,000 Life expectancy, Newly independent states 1994 1994 people, 1994 (year) 1995 Armenia 1,850 284 15 71.1 Azerbaijan 2,520 283 29 70.2 Belarus 1,2S0 26 8 69.6 Georgia 1,590 51 13 73.3 Kazakhstan 2,740 289 30 68.8 Kyrgyz Republic 2,910 325 33 67.6 Moldova 2,260 148 16 68.6 Russia 1,870 47 20 64.8 Tajikistan 4,060 1,146 84 67.4 Turkmenistan 4,640 910 76 66.7 Ukraine 1,400 31 13 68.6 Uzbekistan 3,200 360 38 70.1 Central and Eastern Europe Bulgaria 1,630 49 9 70.9 Croatia 990 6 9 73.7 Czech Republic 8S0 2 3 73.2 Estonia 1,450 30 11 70.4 Hungary 1,160 5 8 70.1 Latvia 1,550 46 20 68.7 Lithuania 1,410 17 15 69.2 Poland 1,510 2 7 70.2 Romania 2,330 80 12 69.9 Slovenia 650 0 3 73.9 European Union Denmark 540 0 10 7S.2 France 640 6 13 77.8 Note: Bold, italicized numbers indicate that the data are for a year other than that listed. In all cases these are the most recent years (since 1990) for which data are available. Source: WHO, World Health Statistics Annual, various years; World Bank World Development Indicators 1997. Annex D - Tables 53 Table D.5 Dissolved oxygen levels in selected rivers Annual mean concentrations (mg/litre) 1989 1990 1991 1992 1993 1994 1995 Czech Republic Labe 7.3 7.7 7.9 8.4 8.5 8.9 10.2 Odra 7.3 6.4 7.9 8.3 8.3 8.5 9 Morova 10.2 10.1 10.7 10.9 10.8 11.5 11.7 Hungary Druna 10 9.8 10.3 9.7 9.7 9.3 9.6 Drava 9.9 10.3 10.3 10.7 10.9 10.5 9.8 Tisza 12.1 11.6 11.7 11.9 12.4 12.1 11.3 Poland Wisla 10.2 10.1 10.2 10.7 10.9 10.9 Odra 10.2 10.6 10.8 11.1 10.8 10.6 Nysa Luzycka 9.5 9.9 10.4 9.3 Slovak Republic Maly Dunaj 10.2 9.5 8.9 9.9 10.1 7.9 Vah 10.2 10.5 10.5 10.5 10.3 10.5 Hron 10 11.1 10.9 10.8 10.8 11.1 Hornad 8.2 7.6 8.6 8.2 8.8 9.2 Armenia Pambak 6.1 8.3 Debet 7.4 7.4 Araks 6.1 6.5 Kasakh 7.8 7.1 Sevjur 8.2 6 Hrazdan 5.5 6.1 Azat 7.4 7.5 Arpa 8.9 8.3 Goris 3.2 6.4 Source: OECD. Notes 1. Dollar amounts are U.S. dollars. Denmark, Finland, France, Ireland, Germany, Ja- 2. NIS Northwest group includes: Belarus, Moldova, pan, the Republic of Korea, the Netherlands, New the Russian Federation, and Ukraine. Zealand, Portugal, South Africa, Spain, Sweden, NIS Southeast group includes: Armenia, Switzerland, Turkey, the United Kingdom, the Azerbaijan, Georgia, Kazakhstan, the Kyrgyz Re- United States, and Uruguay. public, Tajikistan, Turkmenistan, and Uzbekistan. 7. Two different statistical models have been used: 3. The group of advanced reform countries, as de- (a) a linear specification in which total emissions fined in EBRD (1996), consists of Croatia, the Czech of a pollutant in each country are expressed as the Republic, Estonia, Hungary, Latvia, Lithuania, sum of appropriate emissions coefficients multi- Poland, the Slovak Republic, and Slovenia. Slower- plied by GDP, industrial output, and consumption reforming countries are, in the CEE group, of different fuels and (b) a log-linear specification Albania, Bulgaria, Macedonia FYR, and Romania that relies on emissions elasticities rather than on and, among the NIS, Armenia, Azerbaijan, Belarus, emissions coefficients. In both cases, the statisti- Georgia, Kazakhstan, the Kyrgyz Republic, cal model allows for differences across countries Moldova, the Russian Federation, Tajikistan, and over time, with different time parameters for Turkmenistan, Ukraine, and Uzbekistan. the advanced reform and slower-reforming coun- 4. Of course, it may be argued that GDP estimates tries. The two models produce broadly similar for the countries of the former Soviet Union be- results. The discussion in this chapter focuses on fore the transition are substantially overstated the linear specification (see detailed discussion in because of distorted prices and the inclusion of a Annex A). significant fraction of output that was effectively 8. The findings are the same regardless of the spe- unsalable or of negligible value. cific measurement used-black smoke, PM10 5. Bulgaria's hyperinflation of 1996-97 was stopped (particulate matter less than 10 microns in diam- by the establishment of a currency board in July eter), or total suspended particulates. 1997. 9. For comparison, the average annual reduction in 6. The trend in the chart is the result of a linear re- total dust emissions during 1990-96 was 16.8 per- gression of the logarithm of energy consumption cent per year in the advanced reform countries and per capita in 22 market economies with similar cli- 12.8 percent per year in the slower-reforming coun- mates, using the logarithm of their GDP per capita tries, while for S02 the equivalent figures were 9.3 figures as the explanatory variable. The 22 coun- percent in the first group and 18.8 percent per year tries are Argentina, Austria, Belgium, Chile, in the second. 55 56 Economic Reform and Environmental Performance in Transition Economies 10. Because of the small number of observations, the dents. Most CEE and NIS countries had relatively change in S02 concentrations in the Baltics is not effective mechanisms, based on networks of insti- included in Figure 14. Overall concentrations of tutes of sanitary health and epidemiology, for S02 have increased slightly, but from a very low dealing with outbreaks of infectious disease. The level, (less than 10 ug/m3), so that the changes breakup of the former Soviet Union has caused have little significance. particular problems in this sphere for some NIS countries because so much of the expertise in this 11. The level of dissolved oxygen in the Labe (Elbe), field was concentrated in Moscow. Morova, and Odra rivers increased about 15 per- cent. The largest improvement was in the Labe, 17. Reported rates of infant mortality show substan- where average BOD increased from 7.3 mg/l in tial differences. In particular, United Nations 1989 to 10.2 mg/l in 1995. This would be consis- estimates of infant mortality for the Central Asian tent with a substantial reduction in organic countries are much higher than the figures used pollution loads from industrial sources within the here. For consistency over time, the figures used river basin. in this analysis are World Bank estimates, which correspond in most cases to national estimates. 12. The link is less obvious than it may seem. Much of the reduction is accounted for by large sources - 18. Access to urban sanitation has been shown to have power plants and heavy industrial activities - a significant impact on infant mortality, whereas emissions from which are widely dispersed and the influence of rural sanitation is slight or zero. account for a relatively minor share of ambient air 19. Estimates of DALYs take account of the finding that pollution in large urban areas. It is emissions from people care more about the effects of ill health that small sources, including households and vehicles, occurs during the prime years of life; thus, a year of that are typically the primary determinants of ur- life lost as a result of the death of aninfantis weighted ban air quality. While economic reform may be less heavily than a year of life lost as a result of the expected to influence emissions from small sources death of an adult in the age group 30-45 years. (via fuel prices and the adoption of less polluting vehicles or heating equipment), this may be a 20. Restating the costs of air pollution in terms of the much slower process than adjustments in the in- total number of DALYs lost each year, it is as- dustrial and power sectors. sumed that each death caused by air pollution results in the loss of 10 DALYs, based on the aver- 13. This is because such exposure increases stress on re numbe oss of lf lYst ed onath ave the respiratory and cardiac systems, so that people tomcardovacuar an respiaor illneses, d wh ar aled,ikaemrefkl odea to cardiovascular and respiratory illnesses, and whsutof arearineay s are more likely tod that each new case of chronic bronchitis results in result of their illness and others are more likely to tels f12DLs h egt o te yp become sick. ~~~~~~~~the loss of 1.2 DALYs. The weights for other symp- become sick. toms were scaled to these values using the ratios 14. The analysis was based on standard values for the of the relevant estimates of willingness to pay. willingness to pay to reduce the risks of prema- 21 ture mortality, chronic bronchitis, and other 21. The implicit assumption is that, on average, air illnesses or symptoms provoked by exposure to polluto smal an l erities A ura arasi particulates and other pollutants, using 1995 in- is similar to that in larger cities. Although this is comes as the basis for the calculation. probably a reasonable assumption overall, it does lead to anomalies in particular countries because 15. The main exceptions are cities such as Baku and of the threshold values used to compute the costs Dushanbe where the water supply infrastructure of airepoll For ex in Hungary both has been overwhelmed by large influxes of people apestiand ior had annualara leel displaced b or fleein from war.Budapest and Miskolc had annual average levels displaced by or fleeing from war, of particulates below the 50 ug/m3 threshold in 16. Exposure to contaminated water is an intermittent 1995, so that the estimated cost of air pollution for problem; the actual burden of ill health that it these cities, and thus for the country as a whole, causes depends greatly on the response of the was zero. However, some Hungarian sources public health system. Prompt and effective action suggest that levels of particulates in smaller Hun- to prevent exposure and transmission and to treat garian towns such as Bekescsaba, Nyiregyhaza, those affected can minimize the cost of such inci- and Szeged may be quite high. Therefore, the ap- Notes 57 proach used here probably underestimates the 25. These figures were calculated by using the GDP overall cost of urban air pollution. per capita figures given in Annex Table Al with 22. Studies in the U.S. suggest that levels of fine par- an average per capita GNP of $28,600 for the U.S. ticulates are relatively uniform over metropolitan in 1997. regions with a radius of up to 50 km provided that 26. The original report on the study undertaken by the terrain is not too uneven. These observations Esrey and others included an analysis of the ex- have prompted investigators to examine data on pected net benefits from water and sanitation average visibility or haze indices -which tend to interventions for six diseases for Africa, Asia, and be collected at airports on the outskirts of large Latin America. This analysis takes account of the cities-as a way of supplementing data from di- costs of both mortality and morbidity, but the fig- rect monitoring which is usually collected nearer ures are inevitably very broad-brush since the city centers. The results suggest that the correla- authors had to rely upon continent-scale data. tion between visibility measures and monitored 27. A more recent version of this dataset is published levels of fine particles is typically very high. as World Bank 1999 World Development Indicators 23. For example, U.S. studies have found that sulfates 1999, Washington, DC: World Bank. account for as much as 40 percent of fine particles 28. The most recent version of the complete World in the eastern half of the U.S., while nitrates are Tables databank is available on the World Devel- equally important in the western half of the coun- opment Indicators CD-ROM, Washington, DC: try. The reason for the difference is the relative World Bank. importance of vehicles and coal combustion (es- pecially of coal with a relatively high sulfur 29. Trachoma might also have been included but it is content) in contributing to overall levels of fine of negligible importance in Latin America and the particles in the two regions. evidence for the impact of water and sanitation on its incidence is rather mixed. The survey by 24. Any health state can be evaluated by considering Esrey et al suggests that it may be personal hy- its iimpact upon various symptoms, its effect upon giene practices rather than the availability of water social activity, physical activity and mobility, and that is the critical factor. its duration. By these means, the appropriate WTP values can be obtained for each acute morbidity 30. Diarrheal diseases accounted for 99 percent of the impact that is described in the health-status index total deaths and 85 percent of total DALYs lost as literature and investigated in the air pollution lit- a result of this group of diseases. erature, given the established correlation between WTP values and QWB scores. References DEPA (Danish Environmental Protection Agency). Lovei, Magda, and Barry S. Levy. 1997. "Lead Expo- 1998. "UN/ECE Task Force to Phase Out Leaded sure and Health in Central and Eastern Europe: Petrol in Europe: Country Assessment Report." Evidence from Hungary, Poland, and Bulgaria." Danish Ministry of Environment and Energy, In Magda Lovei, ed., Phasing Out Leadfrom Gaso- Copenhagen. line in Central and Eastern Europe: Health Issues, Dockery, D., and others. 1993. "An association be- Feasibility, and Policies. Implementing the Envi- tween air pollution and mortality in six cities." . he ronmental Action Program for Central and East- New England Journal of Medicine 329 (24): 1753- ern Europe series. Washington, D.C.: World Bank. 1759g Lvovsky, Kseniya, Gordon Hughes, David Maddison, 1759. Bart Ostro, and David Pearce. Forthcoming. 1999. EBRD (European Bank for Reconstruction and Devel- Air Pollution and the Social Costs of Fuels: Evidence opmnent). 1996. Transition Report. London. from Major Cities in Developing Countries. World .1997. Transition Report. London. Bank Technical Paper. Washington, D.C. _- 1998. Transition Report Update. London. Murray, Christopher J. L., and Alan D. Lopez, eds. Electrowatt. 1998. "Trends in Exposure to Air Pollu- 1996. The Global Burden of Disease. The Global Bur- tion and Contamination of Drinking Water: Five den of Disease and Injury Series, 1. Harvard School Case Studies." Draft update prepared for the of Public Health. Cambridge, Mass.: Harvard Uni- World Bank. Zurich. versity Press. Esrey, S.A., J.B. Potash, L. Roberts, and C. Shiff. 1991. OECD (Organisation for Economic Co-operation and "Effects of Improved Water Supply and Sanitation Development). 1997. Environmental Data Compen- on Ascariasis, Diarrhoea, Dracunculiasis, Hook- dium 1997. Paris. worm Infection, Schistosomiasis, and Trachoma." -. 1998. Public Exposure to Priority Pollutants Iden- Bulletin of the World Heath Organization 69(5): 609- tified in the Environmental Action Programme: Five Years On. Paris. 621. OECD/IEA (International Energy Agency). 1997. En- Gaddy, C. G., and B. W. Ickes.1998. 'Russia's Virtual ergy Efficiency Initiative. Paris. Economy." Foreign Affairs (September-October). Ostro, Bart. 1996 A Methodologyfor Estimating Air Pol- Lovei, Magda, ed. 1997. Phasing Out Lead from Gaso- lution Health Effects. Geneva: WHO. line in Central and Eastern Europe: Health Issues, Pope, C.A., and others. 1995. "Particulate Air Pollu- Feasibility, and Policies. Implementing the Environ- tion as a Predictor of Mortality in a Prospective mental Action Program for Central and Eastern Study of U.S. Adults." American Journal of Respira- Europe series. Washington, D.C.: World Bank. tory Critical Care Medicine 151: 669-674. 59 60 Economic Reform and Environmental Performance in Transition Economies TME (Institute for Applied Environmental Econom- _ . 1993. World Development Report 1993. Invest- ics). 1995. "Urban Transport and Air Pollution: ing in Health. New York: Oxford University Press. Trends and Policy Options." Draft paper prepared . 1997a. World Development Indicators 1997. for the World Bank. The Hague. Washington, D.C. WHO (World Health Organization). Various years. . 1997b. World Development Report 1997: The World Health Statistics Annual. Vienna. State in a Changing World. New York: Oxford Uni- World Bank/OECD. 1998. "Environmental Action versity Press. Programme for Central and Eastern Europe." .1998a. World Development Indicators 1998 on Abridged version of the document endorsed by CD-ROM. Washington, D.C. the Ministerial Conference Lucerne, Switzerland, World Bank. 1998b. "Russian Federation Joint Munici- April 28-30,1993. World Bank, Washington, D.C. pal Water and Wastewater Sector Study." World World Bank. 1992. World Development Report 1992: Bank, Washington, D.C. Development and the Environment. New York: Ox- ford University Press. Distributors of World Bank Group Publications Prices and credit terms vary Irom CZECH REPUBUC INDIA Eulyos Putilisthing Co., Ltd. PERU SWEDEN country to country. Consult yeur USIS, NIS Prodejna Allied Publishers Ltd. 46-1, Susong-Dong Editonal Desarrollo SA Wennergren-Williams AB local distributor betore placing an HFavelkova 22 751 Mount Road Jongro-Gu A,artado 3824, Ica 242 OF. 106 P 0. Box 1305 order. 130 00 Prague 3 Madras - 600 002 Seoul Lma 1 S-171 25 Solna Tel: (420 2) 2423 1486 Tel: (91 44) 852-3938 Tel: 82 2) 734-3515 Tel: (51 14) 285380 Tel: (46 8) 705-97-50 ARGENTINA Fax: (420 2)2423 1114 Fax: (91 44) 852-0649 Fax: (822)732-9154 Fax: (51 14) 286628 Fax: (46 8) 27-00-7t World Publications SA URL: http:/Mwww.nis.czl INDONESIA LEBANON PHIUPPINES E-mail: mail@wwi.se Acv Cordoba 1877 1120O Ciudad de Buenos Aires DENMARK Pt. Indira Limited Librairie du Liban International Benksnurce Center Inc. SWITZRLAND Tel: dden11 4815-8156 SamfundsLitteratur Jalan Borobudur 20 P.O. Ben 11-9232 1127-A Anpole St, Baralngay, Librairie Payot Service Institutionnel Tel: (54 ) ~ RosenDerns AJId 11 P.O. Bun 181 Beirut Venezuela C(tm)tes-de-Montbenon 30 F-axl: (54 o li infov5- com.6 r OK-1 970 Frederiksberg C Jakarta 10320 Tel: (001 9) 217 944 Makati City 1002 Lausanne E-mail: wpoeke@infeia.com.ar Tel: (45 35) 351942 Tel: (02 21I) 390-4200 Fax: (961 9) 217 434 Tel::(63 2) 896 6501; 6505; 6507 Tel::(41 21) 341-3229 AUSTRALA. FUI, PAPUA NEW Fax: (4535) 357822 Fan:(62 2)30-4289 E-mail: hsayegh7Iibrairiewdu- Fax: (63 2) 896 1741 Pan (41 21) 341-3235 GUINEA, SOLOMON ISLANDS, URL: httpjAwww.sl.cbs.dk IRiN hban.com.lb POLAND ADECO Van Diermen VANUATUI, AND SAMOA ECUADOR Ketab Sara Co. Publishers URL: httpJ/wwaw.Iibrairie-du- International Publishing Service EditionsTechniques D.A. Infermatio n Senices Libri Mundi Khaled Eslamboli Ave., 6th Street libancem .lib Ul. Piekna 31/37 Ch. de Lacuen 41 648 Whitehorse Read Libreria Internacional Delatrosz Alley No. 8 MALAYSIA 00-677 Warzawa CH1 807 Bonney Mitcham 3132, Victoria P.O. Box 17-01-3029 P.O. Box 15745-733 Universty ot Malaya Cooperative Tel: (482)628-6089 TFa: (41 21' Tel: (61) 3 9210 7777 Juan Leon Mera 851 Tehran 15117 B hokhp, Limited Fan:,(48 2) 621-7255 Tel: (41 21) 94323673 Fan: (61(392107108 JuanLeon8Quite Tel: (98211 8717819:8716104 P.O. Bon 1127 E-ail:bouks%ipskp.atm.com.pl THAILAND URL: tt:Awddrtcmau Tel: (1593 2) 521-606; (593 2) 544- Fan: (002 ()8712479 Jalan Pastel Bars UFRL: CeoTHmiBunksDisrbt URL htpi wmw.dadirect.corn.au T S3)26;5 )4185 E-mail: ketab-sara@neda.net.ir 59700 Kuala Lumpur htltp:/ wwwipscg.waw.plfips/export Centra3 Books Oistdbution AUSTRIA Fan: (593 2) 594-209 ~~~~~~~~~~~~Kowekab Publishers Tel: (60 3) 756-000 PORTUGAL Bangkok 10500 Gerold and Cn. E-mail: lihrimal0Iibrimundi.com.ec P e 97-1 Fan: (60 a) 755-4424LirraPtul Weiurgasse 26Teran E-mail: umkoop5tm.net.my Apartado 288 , Rua Do Carm B 0500 A-1011 ien ~~CODEU Tel F8a10658373fEXIO 727 Tel: (43t) 512-47-31-29 Expo rTel (9821 258-3723 MEXICO 0LsoTRINIDAD & TOBAGO Fa:(4 ) 512-47-31-29 Ruiz de Castilla 763, Edit.Epcee Fan: (00 21( 258-3723 INFOTEC 1200 Llubon AND THE CARRIBBEAN URL'htp:/ww.grol.coatonlne Primer pine, Of. #2 IEADAv. San Fernanidso.3 Tel: (l 347-4982 URL: hop:f~.gerol.ce/ut.online iDad IRELANDlo ueraFa 1( Syntematics Studies Ltd. BANGLADESH Tel/Fan: (593 2) 507-383; 253-091 Govemment Supplies Agency C13. TrEl0s Guerra FDsutnS 347-02 I MicroIndutrie Devlopmnt Email codugimsaLnt.ec Olflg as tSolithrair 14050 Mexico, DRFEsenMi od St. A uga stineShpigCnr Micrn Industries Dnelopment E-mail:codes impeaLnetec 4-5 Harcoart Read Tel: (52 56420 Assistance Society (MIDAS) EGYPT, ARAB REPUBLIC OF Dublin 2 Fax , )64-2822 ROMANIA Trinirdad & Tobago, Want Indies HouseaS, Road 16 E A P Fn 1E-madl: Intotec@rtn.net.mx Compani De Uibrarl Bucuresli S.A: Tel: (868) 645-8466 DHanouses RAead 6Al Abram Distdbutuon Agency Tel: (353 11661-3111 man nte opc eLbriBcrtiS. Fan: (886)645-8467 Dhakmandi R/Area Al Galaa Street Fax: (353 1) 475-2670 URL: htfpJ/iln.net.mx SIr. Lipeani no. 26, secter 3 E-mail: tobe5trinidad.nel Tel: (880 2) 326427 Cairn 1SAL ud PeUaMeGoS..dNDA.Bchrs Fan: (880~) 811188 Tel: (2021 578-6083 ISRRE Moni-Peanuo Me1uico onA.dCV Tel::(40 1)313 9645SGAO TelF(80a2x32647Te (20 2i 578 60833 Yozamot Uterature Ltd. c/R anc 141-Clonia0 ) 12 4000 G D BELGIUM P.O. Box 56055 Cuauhtemoc ~~~~~~~~~~~~~RSSIN EDEATONPD Bee 9997, Madhvani Building Jean De Lannoy The Middle East Observer 3 Yara Haadia Street 00500 Mesics, DRERSINFDRTO Punta14JnjaLtd. Ror 202 ~~~~~~41, Obese Street Tel Acmv 61560 Tel: (62 5)533-5658 lsdatelstvo cVes Mir> Ac. du Caio T 7 Fan: (525 ) 514-6799 Ba, Kopacbni7 Pereulok Kampala 1060 Brussels CMioscl:ow7210183197 Tel: (256 41) 251 467 Tel:r(32 538-169 Tel: (20 2) 393-9732 Fax: (972 ) 5285-397 NEPAL Tel 101831 Fax: (256 41) 251468 Fax: (32 )538-0841 (20 -97 ROY Intemational Everest Media International Services Tel (7 095 917 87 49 FN:TED6INGD 580841 FINAN P Box 13 056 U Ltd Fax: (7 7\) 917 925 k -al ufwtuado BRAZIL FIGNDP ee105 DBox 5443 ozimarin@glasneL.ru UNITED KINGDOM Pshlicacdes Tecnicas Internac'ionais Akeateeminen Kiriakauppa Tel Aviv 61130 KtmnuSNAOE AWN HN Microinio Ltd. Ltda. P.O. Bsox 128 Tel: (972 31 649 9469 Kalsh97manda 02 SINGMARE; TAIWAN, CHIAOR. Boo 3, Omega Park, Altos, Rona Peixuto Gomide, 209 FIN-00101 Helsinki Fan: (972 3) 1648 60139 Fax,:(99771), 224 431 Hiemisphere Publi'casinSrie Hampshire GU34 2PG 01409 See Paulo, OP. Tel: (358 0 121 441(8 E-mail: royi 3nev isionnet0 .l a t onneNces Tengatn Tei: (55 11 -6) 447Fax: (358 w) 1214E435 URL:9 hffp:/vww.royinLoiI 41 KaaPadding Read #04-03 Tel: a4 o FaN: 25860Fan: (358 (121-S443 E-aR: hopskcc.id NETHERLANDS GOUlden eel Building Fan: (44 142 9 Tel: (55 1) 259-6949-al0urtlu@sokanI Palestinian AuthsritylMiddle East De LindeboomAlntemationale Singapore 349316 E-ail- a4 4) 69889 E-mail: postmantEonpti.uoml.br URL: http:/Avww.ukateeninen.cm Indes Information bervices Publicaties h.ers Tel: (65) 741-516 E-a wbankmicrninfo.couk URL: http:/AwwwuaI.br FRANCE ROB. 19502 Jerusalem P.O. Ben 202, 7480 AE Haaksbergen Fan: (60) 742-9356 URL: http://ww.microinfo.co.ak CFaADA Editions Eska; DBJ Tel: (9722) 6271219 Tel: (31 53574-0004 E-mail: anhgnte@asianconnect.com The Stationery Office Resat Puhlishing Co. Ltd. 48, rae Say Lussac Fan: (972 2( 6271634 Fax: (31 53) 572-9296 Fax: 51 Nine Elms Lane 526oaa Oanotak Roa Tel:5 (33 55araiStr. ITALY, LIBE8iA E-mail: lindeoeBuworidonline.nrl LoVan 2ndAn B B 5369iCang tek RoadpTel: '3301 3642-82 UL htpAwki ves2nik Publishing8Tel: 44o1711a873- Otlawa, OntafiFan (331KI4-29-1-67Licosa Commissioneaia Sansoni SPA UR:htIwwwdslo.I-i- GroupTe:44118340 Tel: (6131745-2665Fx3 32 9-91i67 Via D-a4 Di Calabria, Fax deboo Du2taisa) cesta Fan: (4417)6873-8242 Fan: (81 3r 745-7820 GERMANY Casella Pu stale 552 NWusND1000 Ljubljana URL: httpJAmw.Ctbe-stationery- E-mail: UNO-Velag 50125 Firenze NEnELN 3 23 ffice.co.uk& order.depl@±innsatbook.csm PoppeludorferAlle 55 Tel: (3955)645-415EBSCO HZ Ltd. Tel: (386 611133 8347:1213 UrLe r(dpt7 rennutb28 ks,c7 m 531pennf Fale2 58 Tn: (39255)2)1- 645415 P e MiB91 Fax: (386 6) t33 8020 VENEZUELA URL: htt:tl wwwrnoufbook.com 5315 Bonn ax: (39 ) 641-27 Privaew Market 991 E-mail: repansekjggvestnik.ei Tecni-Ciencia Libros, S.A. CHINA Tel: (49 228) 940020 E-mail: lico2a0fttcc.k Ne Max: Castro Cuidad Comercial Tamanco China Financial & Economic Fan: (4922) 217492 URL: http:/5wwwtbccitflcosa Auckland SOUTH AFRICA, BOTSWANA Hivel C2, Caracas Publishing House IURL: http:/Mwww.uno-verg.de JAMAICA Tel: (6419) 524-81195 0 Fnr sin ers. Tel: (58 2109595547:65036 001 6 8, Do Fo Si Dung Jie E-mail: unoveragg@a3l.com Ian Randle Publishers Ltd.4 Atca Fan: (58 4) 95965636 B oe Uig GHANA 206 Old Hope Road, Kingston 6 Oasis Oficial Vance Boulevard, Gooo dood ZAMBIA Tel: (0610 6401-7365Epp Benks ServicY Tel: 876-927-2085 RPO. Boe 3627 R.O. Be 12119, Hi City 7463 University Bookshop, University of Fan: (86 0) 6401 -7360 RI. SB n 44 Fan: 876-977-0243 WAllitegon Cape Tuwn Zambia China Book Import Centre TUC E-mail: i npl Dcobs.com Tel: (64 4299 1551 TeL (2721 595 4400 mat East Read Campus RDO. Boe 2825 Tel: 232 JAPAN Fa: (64 ) 499 1972 Fan: (7 2?) 595 44305 Beijing Fan: 223 21 779099 Enarem Bunk Service E-mnail: oasisactrix.gen.nz E-mail: oxordo@sup.co.za Lusaka 3-13R Hseama3-chome,iank.-ne122t7 Fax: 223 1 7790993-13 Hong 3-chome BunlWku UROL: htrtpJ/wwvr.easitbouks.cs.nzJ For subscription orders: Tel: (260 1 5 7 Chinane Cortporation for Promotion GREECE Tnkyo 113 NIEIIntemational Subscription Service Fan: (260 i( 263 952 52, YoumFangtHesTng Paprasotidou S.A. Tel: (81 3)381"-861 NIGRIOR. Boe 41095 ZIMBABWE X2Yu angHe o u Ji ng 35, Stournara Str. Fax: (Bi 3) 3818-0864 University Press Limited Xuan Nei Da162 thnsE-ai: rer@Ji-becnp Three Crowns Building Jericho Craighall Academic and Baobab Books (Pvt.) B,e,i(ingTl:(06 12 364-182 URmal: odSsvebcop Private Mail Bag 5895 Johannesburg 2024 LId. Tel(86 10660072 494 Tl (01) -86UL lbadsn Tel: (27 11) 880-1446 4 Conald Road, GraniHeside Fax: (86 M ~~~Fax:(30 1) 364-8254 http:/vwmwhekdkoame.or.jp/-svt- aelFan:32711) 880-6248FRO. Bee 667 FCOL(81M (00749 HAITI eshi Fax.(34 22)( 41-12086 E-mait: issOiscosea Haraem CnoelOBac ta Culture Dittus'iun KENYA . 4-06SPAIN Tel: 263 4 755035 Carrera 6 No. 51-21 5, Rue Capois Africa Bank Service (E.A.) Ltd. PAKISTAN Mundi-Prensa Libros, SA. Fan: 263 4781913 Apartado Aereo 34270 C.P. 257 Guaran Hoante, Mfanganoa Street Mirza Bank Agenicy Castefon 37 Santalti de Bogotd, D.C. Port-au-Pdnce RDO. Ben 45245 60, Shahrah-e-Duaid-e-Azam 28001 Madrid Tel: (57 1? 285-2798 Tel: (509 23 9260 Nairohi Lahore 54000Te:349 4670 Fax: (57 ()285-2700 Fan: (5) 234858 Tel: (254 2' 223 641 Tel:: 9 2 735 3601Fa:349)5798 HONG KONG, CHINA; MACAO Fax: (254 2) 330 272 Fan: (92 42 (576 3714 E-mail: libreria@mundiprensa.es COTE D'IVOIRE is20LtLicyBoasdUnvrtyPesURL: http://www.mundrprensa.csm/ Ceranrer d CEdiina eDiA i Sales & Circulation Department Loita House S Bangalore Town Monsedi-Prenta Baceon At4iBane 5 CEDA 302 Seabird House Mezzanine 1 Sharae Faisal Cun9n arelo eCna,39 04bidja 54122-28 Wyndham Street, Central RDO. Bon 60077 PO Ben 13833 08009 Barcelona349 Tel: (225) 24 651 0 248651 1 Hong Kung, China Nairobi TelarachFaii(37(87-60 Fee: (225)25 o5e7 Tel. (85212630-1409 Tel: (254) 2-330853, 221426 aTel i-9751443507 Fax: (34 1)488-76492 Fan: (BSO) 2026-1107 Far (25) 2-330854, 561654 437E-maiL: bsrcelonagmendiprensa.es CYPRUS E-mail: saIes@asma2D00.com.hk E-mail: Legacy@orsn-net.com Fax:'(92 2 4547640 Center for Applied Research URL: hlrp://icww.nsia2000.com.hk KRA EULCO E-mail: ouppak@TheOffice.nel SRI LANKA, THE MALDIVES CyomsClleseRA EPSCO Lake House Bookshop 0 Dosgenes Street, Engami HUNGARY Dayang Books Trading Ce. Pale Book Corporation 108, Sir Chrittampalam Gardiner RDO. Ben 2006 Earn Info Service Intereaotisal Division Aziz Chambers 21, Qseen's Read Mawathe Hicosia Mritsngeli Europa Haz 783-20, Paogba Ben-Doug, Lahore Colombo 2 Tel: (357 21 59-0730 H-iT138 Budapest Socho-ku Tel: (92 42) 636 3222: 636 0885 Tel: (94 1) 32105 Fan: (357 ~) 66-2051 Tel: (36 1135000024, 36080 25 Seoul Fan: (92 42( 6362328 Fan: (941(i)432104 Fan: (36 1) 35090032 Tel: (82 2) 536-9555 E-mail: pbc@brain.net.pk E-mail: LHL@sri.lanka.net E-mail: euroinfo@mail.matav.hu Fan: (82 2) 536-0025 E-marl: seamnap@cholfian.net RecerLt World Bank Technical Papers (continued) No. 405 Onorato, Fox, and Strongman, World Bank Group Assistancefor Minerals Sector Development and Reform in Member Countries No. 406 Milazzo, Subsidies in World Fisheries: A Reexamination No. 407 Wiens and Guadagni, Designing Rulesfor Demand-Driven Rural Investment Funds: The Latin American Experience No. 408 Donovan and Frank, Soil Fertility Management in Sub-Saharan Africa No. 409 Heggie and Vickers, Commercial Management and Financing of Roads No. 410 Sayeg, Successful Conversion to Unleaded Gasoline in Thailand No. 411 Calvo, Optionsfor Managing and Financing Rural Transport Infrastructure No. 413 Langford, Forster, and Malcolm, Toward a Financially Sustainable Irrigation System: Lessonsfrom the State of Victoria, Australia, 1984-1994 No. 414 Salman and Boisson de Chazournes, International Watercourses: Enhancing Cooperation and Managing Conflict, Proceedings of a World Bank Seminar No. 415 Feitelson and Haddad, Identification of Joint Management Structuresfor Shared Aquifers: A Cooperative Palestinian-Israeli Effort No. 416 Miller and Reidinger, eds., Comprehensive River Basin Development: The Tennessee Valley Authority No. 417 Rutkowski, Welfare and the Labor Market in Poland: Social Policy during Economic Transition No. 418 Okidegbe and Associates, Agriculture Sector Programs: Sourcebook No. 420 Francis and others, Hard Lessons: Primary Schools, Community, and Social Capital in Nigeria No. 421 Gert Jan Bom, Robert Foster, Ebel Dijkstra, and Marja Tummers, Evaporative Air-Conditioning: Applications for Environmentally Friendly Cooling No. 422 Peter Quaak, harrie Knoef, and Huber Stassen, Energyfrom Biomass: A Review of Combusion and Gasification Technologies No. 423 Energy Sector Unit, Europe and Central Asia Region, World Bank, Non-Payment in the Electricity Sector in Eastern Europe and the Former Soviet Union No. 424 Jaffee, ed., Southern African Agribusiness: Gaining through Regional Collaboration No. 425 Mohan, ed., Bibliography of Publications: Africa Region, 1993-98 No. 426 Rushbrook and Pugh, Solid Waste Landfills in Middle- and Lower-Income Countries: A Technical Guide to Planning, Design, and Operation No. 427 MarinXo and Kemper, Institutional Frameworks in Successful Water Markets: Brazil, Spain, and Colorado, US.A No. 428 C. Mark Blackden and Chitra Bhanu, Gender, Growth, and Poverty Reduction: Special Program of Assistance for Africa, 1998 Status Report on Poverty in Sub-Saharan Afrtica No. 429 Gary McMahon, Jose Luis Evia, Alberto Pasc6-Font, and Jose Miguel Sanchez, An Environmental Study of Artisanal, Small, and Medium Mining in Bolivia, Chile, and Peru No. 430 Maria Dakolias, Court Performance around the World: A Comparative Perspective No. 431 Severin Kodderitzsch, Reforms in Albanian Agriculture: Assessing a Sector in Transition No. 432 Luiz Gabriel Azevedo, Musa Asad, and Larry D. Simpson, Management of Water Resources: Bulk Water Pricing in Brazil No. 433 Malcolm Rowat and Jos6 Astigarraga, Latin American Insolvency Systems: A Comparative Assessment No. 434 Csaba Csaki and John Nash, editors, Regional and International Trade Policy: Lessons for the EU Accession in the Rural Sector-World Bank/FAO Workshop, June 20-23, 1998 No. 436 Roy Prosterman and TIme Hanstad, ed., Legal Impediments to Effective Rural Land Relations in Eastern Europe and Central Asia: A Comparative Perpective No. 437 Csaba Csaki, Michel Dabatisse, and Oskar Honisch, Food and Agriculture in the Czech Republic: From a "Velvet" Transition to the Challenges of EU Accession No. 443 Luc Oecuit, John Elder, Charistian Hurtado, Francois Rantrua, Kamal Siblini, and Maurizia Tovo, DeMIStifying MIS: Guidelinesfor Management Information Systems in Social Funds No. 444 Robert F. Townsend, Agricultural Incentives in Sub-Saharan Africa: Policy Challenges No. 445 Ian Hill, Forest Management in Nepal: Economics of Ecology No. 447 R. Maria Saleth and Ariel Dinar, Evaluating Water Institutions and Water Sector Performance THE WORLD BANK 1818 [1 Street, NAN. Washington. 1).C. 20433 lJSA Telephone: 202-477-1234 Facsimile: 202-477-6391 telex: MCI 64145 W)RIA)BANK M\CIA 248423 WORDI)BANK Internet: www.worldhank.org l.-mail: hooks(iovorldhank.org litES ISBN 0-8213-4564-8