Integrated national energy planning and management : methodology and application to Sri Lanka

WORLD BANK TECHNICAL PAPER NUMBER 86 WTP86 INDUSTRY AND ENERGY SERIES October 1988 Integrated National Energy Planning and Management Methodology and Application to Sri Lanka Mohan Munasinghe 'ad~~ FILE COPys  RECENT WORLD BANK TECHNICAL PAPERS No. 40. Plusquellec and Wickham, Irrigation Design and Management: Experience in Thailand and Its General Applicability No. 41. Bord na M6na, Fuel Peat in Developing Countries No. 42. Campbell, Administrative and Operational Procedures for Programs for Sites and Services and Area Upgrading No. 43. Simmonds, Farming Systems Research: A Review No. 44. de Haan and Nissen, Animal Health Services in Sub-Saharan Africa: Alternative Approaches (also in French, 44F) No. 45 Sayers, Gillespie, and Queiroz, The International Road Roughness Experiment: Establishing Correlation and a Calibration Standardfor Measurements No. 46. Sayers, Gillespie, and Paterson, Guidelines for Conducting and Calibrating Road Roughness Measurements No. 47. Segura, Guidelines for Evaluating the Management Information Systems of Industrial Enterprises No. 48. Tschannerl and Bryan, Handpumps Testing and Development: Proceedings of a Workshop in China No. 49. Gunnerson and Stuckey, Anerobic Digestion: Principles and Practices for Biogas Systems No. 50. Turtiainen and Von Pischke, Investment and Finance in Agrcultural Service Cooperatives No. 51. Shuval and others, Wastewater Irrigation in Developing Countries: Health Effects and Technical Solutions No. 52. Armstrong-Wright, Urban Transit Systems: Guidelines for Examining Options (also in Spanish, 52S) No. 53. Bamberger and Hewitt, Monitoring and Evaluating Urban Development Programs: A Handbookfor Program Managers and Researchers (also in French, 53F) No. 54. Bamberger and Hewitt, A Manager's Guide to "Monitoring and Evaluating Urban Development Programs: A Handbook for Program Managers and Researchers" (also in French, 53F; Spanish, 53S; and Chinese, 53C) No. 55. Technica, Ltd.. Techniques for Assessing Industrial Hazards: A Manual No. 56. Silverman, Kettering, and Schrnidt, Action-Planning Workshopsfor Development Management: Guidelines No. 5 7. Obeng and Wright, The Co-composting of Domestic Solid and Human Wastes No. 58. Levitsky and Prasad, Credit Guarantee Schemes for Small and Medium Enterprises No. 59. Sheldrick, World Nitrogen Survey No. 60. Okun and Emst, Community Piped Water Supply Systems in Developing Countries: A Planning Manual No. 61. Gorse and Steeds, Desertification in the Sahelian and Sudanian Zones of West Africa No. 62. Goodland and Webb, The Management of Cultural Property in World Bank-Assisted Projects: Archaeological, Historical, Religious, and Natural Unique Sites No. 63. Mould, Financial Information for Management of a Development Finance Institution: Some Guidelines No. 64. Hillel, The Efficient Use of Water in Irrigation: Principles and Practices for Improving Irrigation in Arid and Semiarid Regions (List continues on the inside back cover)  Integrated National Energy Planning and Management Methodology and Application to Sri Lanka  INDUSTRY AND ENERGY SERIES This series is sponsored by the Industry and Energy Department of the World Bank's Policy, Planning, and Research Staff to provide guidance on technical issues to government officials and World Bank staff and consultants involved in the various industrial and energy sectors and subsectors. Other volumes in this series are: o The Forest Industries Sector: An Operational Strategy for Developing Countries (World Bank Technical Paper Number 83) o The New Face of the World Petrochemical Sector: Implications for Developing Countries (World Bank Technical Paper Number 84 o Proposals for Monitoring the Performance of Electric Utilities (World Bank Technical Paper Number 85)  WORLD BANK TECHNICAL PAPER NUMBER 86 INDUSTRY AND ENERGY SERIES Integrated National Energy Planning e and Management Methodology and Application to Sri Lanka Mohan Munasinghe with Peter Meier REcONS0ETINN The World Bank Washington, D.C.  Copyright (C 1988 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 October 1988 Technical Papers are not formal publications of the World Bank, and are circulated to encourage discussion and coniment and to communicate the results of the Bank's work quickly to the development community; citation and the use of these papers should take account of their provisional character. 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. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries conceming the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitation of its boundaries or its national affiliation. Because of the informality and to present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to Director, Publications Department at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to photocopy portions for classroom use is not required, though notification of such use having been made will be appreciated. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list and indexes of subjects, authors, and countries and regions; it is of value principally to libraries and institutional purchasers. The latest edition of each of these is available free of charge from the Publications Sales Unit, Department F, The World Bank, 1818 H Street, NW, Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'lena, 75116 Paris, France. Mohan Munasinghe is chief of the Energy and Infrastructure Division, Country Department I, in the World Bank's Regional Office for Latin America and the Caribbean. During the period of the study (1982-86), while on a leave of absence from the World Bank, he served as senior energy adviser to the President of Sri Lanka. Library of Congress Cataloging-in-Publication Data Munasinghe, Mohan, 1945- Integrated national energy planning and management methodology and application to Sri Lanka / Mohan Munasinghe with Peter Meier. p. cm. -- (World Bank technical paper, ISSN 0253-7494 ; no. 86. Industry and energy series) Includes bibliographies. ISBN 0-8213-1136-0 1. Energy policy--Sri Lanka. I. Meier, Peter, 1942- II. International Bank for Reconstruction and Development. III. Title. IV. Series: World Bank technical paper ; no. 86. V. Series: World Bank technical paper. Industry and energy series. HD9502.S722M86 1988 333.79'09549'3--dcl9 88-27872 CIP  ABSTRACT Given the importance of energy in modern economies, the first part of this volume is devoted to examining some of the key conceptual and analytical tools available for energy policy analysis and planning. Policy tools and institutional frameworks that will facilitate better energy management are also discussed. These will ensure that major energy policy initiatives, programs and projects are effectively implemented. More specifically, energy planning is broadly interpreted here to denote a series of steps or procedures by which the myriad of interactions involved in the production and use of energy may be studied and understood within an explicit analytical framework. Energy policy analysis is explained, while effective energy management techniques are discussed to achieve desirable national objectives, using a selected set of policies and policy instruments. While the conceptual framework for policy analysis and planning is integrated (to facilitate the formulation of broad energy strategies), the implementation process must involve maximum use of decentralized policy instruments and market forces to improve effectiveness. This would include an increased role for the private sector where appropriate. In the second part of this volume, the actual application of the principles set out earlier is explained through a case study of Sri Lanka. The success of the approach is demonstrated by the significant progress already made over the past few years toward implementation of a National Energy Strategy (NES) in Sri Lanka. Specifically, in late 1982, the Government of Sri Lanka sought to strengthen energy policy formulation and coordination. Several important programs with obvious short-term benefits were quickly identified and launched in late 1982 and early 1983. These programs, while clearly yielding early results, were essentially of a one-time nature. They reduco'd energy consumption over the short to medium run, but even if completely adopted throughout the economy, could not arrest the longer term trends. The case study shows that over the medium to long term, therefore, some of the more fundamental issues must be addressed. In resource-poor Sri Lanka, not only are the energy options more limited, but also the impact of not adopting a comprehensive long-term strategy would be much more severe. Thus, this monograph integrates the many aspects of the short-term programs already begun with the options for the medium to long term, and ends with the outline of a long-term strategy for Sri Lanka.   - vii - PREFACE The contents of this volume constitute primarily a record of the work and experiences of the author while serving the Government of Sri Lanka, on a leave-of-absence from the World Bank. Nevertheless, the analytical methodology, institutional framework, and Sri Lanka case study provide helpful insights into the process of formulating and implementing a practical energy strategy in the presence of real world problems and constraints to be found in a typical developing country. The results of the case study (conducted in Sri Lanka during the period 1982-1985) became the foundation for the National Energy Strategy (NES) of the Sri Lanka Ministry of Power and Energy which was completed in late 1985. Although this volume also includes some information acquired after the publication of the NES in 1985, the presentation of the original analytical work has been left essentially untouched so that the reader may follow the process by which policy decisions were made, based on the technical studies available at the time. For example, the world oil price scenarios that are presented in Chapter 3 were developed in early 1985, well before the price decline of early 1986. Thus the approach taken in this volume is not to revise these scenarios with the benefit of hindsight, but rather to assess the degree to which the decisions taken on the basis of the best forecasts available in 1985 have proven to be sufficiently robust. The work described here could not have been carried out without the invaluable assistance of the Energy Coordinating Team Task Forces and Staff, as well as other associated government officials. Peter Meier, as consultant to the Office of the Presidential Senior Energy Advisor (through a technical assistancy financed by the Asian Development Bank), coordinated and led the Sri Lankan group responsible for developing and implementing the computer models. Dr. Meier is presently Chairman of the Energy Program at the State University of New York, Stonybrook, NY. Mark Bernstein, Chitrupa Fernando, Mostefa Ouki, Tom Wilbanks, and several anonymous referees gave helpful comments. The encouragement and support provided by His Excellency J.R. Jayewardene, President of Sri Lanka; Honourable P. Dayaratne, the present Minister of Power and Energy; and Professor K.K.Y.W. Perera, Secretary of the Ministry of Power and Energy, are also gratefully acknowledged. Haijiang Wang and Alex McPhail helped to prepare the manuscript, while William DeCarlo assisted with the word processing. Mohan Munasinghe January 1988   - ix - TABLE OF CONTENTS ABSTRACT v PREFACE vii LIST OF FIGURES xi LIST OF TABLES xii 1. A FRAMEWORK FOR INTEGRATED NATIONAL ENERGY PLANNING AND POLICY ANALYSIS 1.1 Introduction and Outline. 1 1.2 Overview of Integrated National Energy Planning 4 1.3 Tools for Policy Implementation. 9 1.4 Constraints on Policy .10 References .14 2. ENERGY MODELING METHODOLOGY 1.1 Introduction .15 2.2 Evaluation Criteria .16 2.3 Classification and Evaluation of Energy Models .26 2.4 Conclusions .32 References .37 3. SRI LANKA CASE STUDY 3.1 Overview of the Macroeconomy and Energy Situation 38 3.2 Energy Policy Coordinating Framework .54 3.3 Hierarchical INEP Modeling Framework ........................... 61 Appendix 3.1 .72 References .76 4. FUELWOOD 4.1 Introduction ..77 4.2 The National Fuelwood Conservation Program (NFCP) 78 4.3 The Analytical Design ..83 4.4 Analytical Results ..89 4.5 Conclusions ..98 Appendix 4.1 Assessment of Improved Fuelwood Stove .......... 101 Appendix 4.2 Technical Assessment of Fuelwood Stoves .106 References ..110  x 5. ELECTRIC POWER 5.1 Introduction ...............................................111 5.2 Analytical Approach .............................................. 112 5.3 Demand Projections .............................................. 112 5.4 The Impact of Conservation .............................................. 118 5.5 The Impact of Drought Years ........................................... 120 5.6 The System Load Factor (SLF) ......................................... 123 5.7 Delayed Construction Schedule ........................................... 123 5.8 Institutional Reform ............................................... 125 Notes ............................................... 129 References .............................................. 130 6. OIL IMPORTS AND ENERGY-ECONOMIC ISSUES 6.1 Introduction .............................................. 131 6.2 Methodological Issues .............................................. 135 6.3 Organization of the Macroeconomic Accounting Framework ............................... ............... 138 6.4 Energy Sector Investment and Debt Service .................... 147 References .............................................. 148 7. ENERGY DEMAND MANAGEMENT AND CONSERVATION 7.1 Background and Policy Options ........................................ 149 7.2 Industrial and Commercial Energy Conservation Program (IECP) .......................................... 154 7.3 Transportation Energy Conservation Program (TECP) .............................................. 160 7.4 Energy Supplying Institutions ............................................ 160 7.5 Household and Other .............................................. 161 7.6 Agriculture .............................................. 161 References ............ 163 8. A NATIONAL ENERGY STRATEGY FOR SRI LANKA 8.1 The Major Issues ............ 164 8.2 The Basic Directions of Policy ........................................... 165 8.3 Short Term Options .............................................. 167 8.4 Options for the Medium to Long Term ............................ 172 References ............................................... 175 BIBLIOGRAPHY .177  - xi - LIST OF FIGURES 1.1 Integrated National Energy Planning Framework ............................ 7 3.1 Sri Lanka Energy Supply Pattern (1983) ........................................ 48 3.2 Sri Lanka Energy Consumption Pattern (1983) .............................. 51 3.3 Sri Lanka Energy Sector Institutional Framework (Mid-1982) ....... 56 3.4 Energy Sector Institutional Framework (End-1982) ......................... 57 3.5 Energy Sector Institutional Framework (Mid-1984) ......................... 60 3.6 Energy-Econiomy Modelling Framework ............................................ 63 3.7 GDP Scenarios ...................................................... 67 3.8 World Oil Price Scenarios ...................................................... 68 3.9 The Design for Model Runs ...................................................... 70 A.3.1. The RESGEN Concept .75 4.1 Institutional Framework for NFCP ................................................... 84 4.2 Sample Household Location and Zonal Definition in WiJesinghe (1984) ..................................................... 86 4.3 Percentage Composition of Fuelwood by Region ............................ 88 4.4 Anticipated Trends in Total Fuelwood Consumption ..................... 90 4.5 Harvesting Patterns for Eucalyptus Camuldulensis ......................... 93 4.6 Movement of Natural Forest Area ................................................... 94 4.7 Natural Forest Area with Plantation and Cookstove Programs .... 95 4.8 Foreign Exchange Value of Fuelwood Plantation ........................... 97 4.9 Annual Foreign Exchange Flows with Fuelwood Plantation Prograrme. .99 5.1 Overall Scenario Design ................................................. 115 5.2 Evolution of Electricity Demand ................................................. 116 5.3 Impact of Conservation Program on CEB Oil Consurnption ......... 119 5.4 Impact of a Drought Year ................................................. 121 5.5 Impact of a Drought Year with Conservation ................................ 122 5.6 Impact of Delayed Construction ................................................. 126 6.1 External Resource Gap ................................................. 132 6.2 The Analytical Approach ................................................. 134 6.3 Net Foreign Exchange Saving due to the Conservation Program ........................................ 146 7.1 Electricity Generation ........................................ 150 7.2 Net Oil Consumption (Imports) ........................................ 151 7.3 Industrial and Commercial Conservation Program .......................... 156 8.1 Impact of Conservation in a Drought Year .................................... 169 8.2 Fuelwood Consumption ..................: 170  - xii - LIST OF TABLES 1.1 Annual Energy Investment as a Percentage of Total Annual Public Investment (Early 1980's) .............................. 1 2.1 Summary of Evaluation Criteria ............................................. 17 2.2 Classification of Models ................................................ 27 3.1 Consumption Trends for Major Petroleum Products ............ 41 3.2 Petroleum Import Bill ................................................ 42 3.3 Evolution of GDP, Energy Demand and Prices for the Period 1970-1986 ..................................................... 43 3.4 Imports of Oil and Petroleum Products ................................ 46 3.5 Consumption of Electricity and Petroleum Products ............ 47 3.6 Basic Energy Data from the Energy Balance Tables ........... 49 3.7 Energy from Hydroelectric Plants .......................................... 69 4.1 Benefits of Using Improved Wood Stoves ............................. 82 4.2 The Current Plantation Program ........................................... 89 4.3 Sample Fuelwood Balance ................................................ 92 A.4.1. Present Value of Economic Costs for Woodstove and Open Hearth ................................................ 102 A.4.2. Preseiit Value of Financial Costs for Woodstove and Open Hearth ................................................ 103 5.1 CEB Generation Capacity ................................................ 111 5.2 Typical RESGEN Output ................................................ 113 5.3 C.E.B. Capacity Expansion Plan ............................................ 114 5.4 Generation Requirements Forecasts ........................................ 117 5.5 Conservation Program Goals ................................................ 118 5.6 Summary of Scenario Results ................................................ 120 5.7 Delayed Construction Scenario ................................................ 124 6.1 The Data Tables of the Model .............................................. 139 6.2 Sample Balance of Payments: ................................................ 141 6.3 Sample Resource Gap Projection ............................................ 143 6.4 NEDMCP Conservation Goals ................................................ 145 6.5 Foreign Exchange Imipact of the Deferred Investment Scenario ................................................ 147 8.1 NEDMCP Conservation Goals ................................................ 167 8.2 Oil Import Savings Due to the Conservation Program ....... 168 8.3 Impact of Petroleum Product Pricing Policy ........................ 172 8.4 Maximum Renewables Case ................................................ 173  CHAPTER 1 A FRAMEWORK FOR INTEGRATED NATIONAL ENERGY PLANNING AND POLICY ANALYSIS 1.1 INTRODUCTION AND OUTLINE The pervasive and critically important role of energy in national economies indicates that the identification of energy issues and energy policy development are important areas of study by governments, researchers, and the development community. While the softening of world oil prices in 1986 provided relief to oil importing nations, the availability of adequate energy resources at a reasonable cost is still a vital precondition for continued economic progress. At the same time, most of the key energy issues identified during the past decade have not disappeared. Thus, developing country energy investments still average about 25 percent of total public investments, oil importers are spending an average of 15-20 percent of export earnings on petroleum imports, and fuelwood shortages and deforestation problems continue unabated (especially in Africa and Asia). Table 1.1 shows recent energy investment requirements, with the electric power sector accounting for about three fourths of these amounts. If oil prices remain relatively low and the world economy expands, there is likely to be increasing economic growth in and demand for exports from, the developing counltries. This will, in turn, entail greater industrial activity and demand for einergy in developing countries, with a consequent need for further investments in supply. Table 1.1 Annual Energy Investment as a Percentage of Total Annual Public Investment (Early 1980's) Over 40% 30-40% 20-30% Below 20% Argentina Ecuador Botswana Egypt Brazil India China Ethiopia Colombia Pakistan Costa Rica Ghana Korea Philippines Liberia Nigeria Mexico Turkey Nepal Sudan  - 2 - Given the importance of energy in modern economies, the first part of this volume is devoted to examining some of the key conceptual and analytical tools available for energy policy analysis and planning. Policy tools and institutional frameworks that will facilitate better energy management are also discussed. These will ensure that major energy policy initiatives, programmes and projects are effectively implemented. More specifically, energy planning is broadly interpreted here to denote a series of steps or procedures by which the myriad of interactions involved in the production and use of energy may be studied and understood within an explicit analytical framework. Planning techniques ranging from basic manual methods to sophisticated computer modeling are reviewed in this volume. The complexity of energy problems has forced increasing reliance on the latter approach. Energy policy analysis involving the systematic investigation of the impact of specific energy strategy or policy packages on the economy and society, at all levels, is explained. Effective energy management techniques (which includes both supply and demand management) are discussed using a selected set of policies and policy instruments, to achieve desirable energy and economic objectives. While the conceptual framework for policy analysis and planning is integrated (to facilitate the formulation of broad energy strategies) the implementation process must involve maximum use of decentralized policy instruments and market forces to improve effectiveness. This would include an increased role for the private sector where appropriate. In the second part of this volume, the actual application of the principles set out earlier is explained through a case study of Sri Lanka. The success of the approach is demonstrated by the significant progress already made over the past few years towards implementation of a National Energy Strategy (NES) in Sri Lanka. Specifically, in late 1982, the President of Sri Lanka created within the Ministry of Power and Energy, an Energy Coordinating Team (ECT) consisting of three Task Forces to strengthen energy policy formulation and coordination. Several important programmes with obvious short term benefits were quickly identified by the ECT, and launched in late 1982 and early 1983. These included: (1) the National Energy Demand Management and Conservation Programme (NEDMCP), aimed at promoting greater efficiency in the production and use of the major commercial fuels (oil, electricity and gas) in all sectors of the economy; and (2) the National Fuelwood Conservation Programme (NFCP), to manage the use of biomass more efficiently, including the introduction of more efficient domestic cookstoves. Among other recent initiatives, the power utility company (CEB) has embarked on a loss reduction programme, and the national oil company (CPC) has started work on a facility that will decrease the cost of oil imports by reducing transportation expenses.  - 3 - It is shown that these programmes, while clearly yielding short-term benefits, are essentially of a one-time nature. They will reduce energy consumption over the short to medium run, but even if completely adopted throughout the economy, cannot arrest the longer term trends--as development proceeds, and both the economy and population expand. This is well illustrated by the impact of the improved cookstove programme on fuelwood consumption. As shown later (Figure 8.2), even if all households adopt the more efficient stoves by 1989, total consumption will increase after 1990, and consumption will again be at or above present levels by the turn of the century. In brief, the NEDMCP and NFCP are essential to immiediately reduce energy costs, but they must be followed up with supporting initiatives, within the framework of a systematic and consistent long run national energy strategy (NES). The case study shows that over the medium to long-term, therefore, some of the more fundamental issues must be addressed. In contrast to many other countries in South Asia that nature has endowed with significant fossil resources (natural gas in Pakistan and Bangladesh; coal, oil and gas in India), the prospects for discovering similar resources in Sri Lanka remain poor. In such a situation, not only are the energy options more limited, but also the impact of a failure to adopt a comprehensive long-term strategy would be much more severe. Otherwise, the ever increasing oil import needs will result in severe foreign exchange difficulties, which in turn would dampen economic growth. Moreover, continuing deforestation will also have severe adverse consequences. In recognition of these longer-term problems, the Energy Planning and Policy Analysis Task Force (EPPAN) of the ECT was charged with the identification of the major objectives and issues of national energy policy and the definition of a national energy strategy. This monograph integrates the many aspects of the short-term programmes already begun with the options for the medium- to long-term, and ends with the outline of a long-term strategy for Sri Lanka.  - 4 - 1.2 OVERVIEW OF INTEGRATED NATIONAL ENERGY PLANNING Because of the many interactions and non-market forces that shape and affect the energy sectors of every economy, decision makers in an increasing number of countries have realized that energy sector investment planning, pricing and management should be carried out on an integrated basis, e.g. within a integrated national energy planning (INEP) framework which helps analyze a whole range of energy policy options over a long period of time (Munasinghe, 1980). We emphasize again that INEP provides primarily a conceptual framework for policy analysis and energy strategy formulation, while policy imiplementation should rely mainly on market incentives and decentralized competitive forces. The development of the concepts and methodology of INEP and its subsequent application can be traced to the energy crisis of the 1970's. Before this period, energy was relatively cheap, and any imbalance between supply and demand was invariably dealt with by augmenting supply. The emphasis was more on the engineering and technological aspects. Furthermore, planning was confined to the various energy subsectors such as electricity, oil, coal, etc., with little coordination among them. From the mid-1970's onwards, the rapidly increasing cost of all forms of energy, led by the world oil price, stimulated the development of new analytical tools and policies (Munasinghe, 1980). First, the need became apparent for greater coordination between energy supply and demand options, and for the more effective use of demand management and conservation. Second, energy-macroeconomic links began to be explored more systematically. Third, the more disaggregate analysis of both supply and demand within the energy sector offered greater opportutnities for inter-fuel substitution (especially away from oil). Fourth, the analytical and modelling tools for energy subsector planning became more sophisticated. Fifth, in the developing countries, greater reliance was placed on economic principles, including the techniques of shadow pricing. INEP makes use of all these separate threads. Some early attempts at comprehensive energy planning were made, particularly in the second half of the 1970's. However, it was soon recognized that the constraints imposed by limiited data, skilled manpower and time posed formidable pronblems, especially in third world countries. Many of the early planning models were designed to be all-encompassinig, and therefore proved to be too large and unwieldy. On the other hand, more specific models tended to overlook important energy sector or macroeconomic linkages. Finally, most models were not policy- oriented, and were often treated as mere academic exercises. This initial learning process led to a more hierarchical analytical framework that recognizes  - 5 - at least three distinct levels of analysis, viz: energy-macroeconomic, energy sector, and energy subsector, as well as the interactions among them. This approach also gives a better policy focus. More recently, the availability of microcomputers has provided third world analysts with a relatively cheap, powerful and flexible tool to develop and apply some of these ideas (Meier, 1985 and Munasinghe, 1986). We will set out the conceptual approach to INEP below, and then discuss in Chapter 2 the appropriateness of presently available microcomputer-based models to implement part or all of such a framework. Coordinated energy planning and pricing require detailed analyses of the interrelationships between the various economic sectors, and their potential \ energy requirements, versus the capabilities and advantages/disadvantages of the various forms of energy such as electric power, petroleum, natural gas, coal and traditional fuels (e.g., firewood, crop residues and dung) to satisfy these requirements. Non-conventional sources, whenever they turn out to be viable alternatives, must also be fitted into this framework. The discussion applies both to the industrial world and developing countries. In the former, the complex and intricate relationships between the various economic sectors, and the prevalence of private market decisions on both the energy demand and supply sides, make analysis and forecasting of policy consequences a difficult task. In the latter, substantial market distortion, shortages of foreign exchange as well as human and financial resources for development, larger numbers of poor households whose basic needs somehow have to be met, greater reliance on traditional fuels, and relative paucity of energy as well as other data, add to the already complicated problems faced by energy planners everywhere. An important objective of developing countries must be to upgrade the quality of energy planning, policy analysis and management. In particular, such efforts must also focus on methods of enhancing the effectiveness of energy policy implementation. In order to better understand the role of planning, we begin by identifying below some of the broad goals of energy policy from the national perspective. The broad rationale underlying planning at the national level and V policymaking of all kinds in the developing countries, is the need to ensure the best use of scarce resources in order to further overall socio-economic development and improve the welfare and quality of life of citizens. Energy planning in particular is therefore an essential part of national economic planning, and should be carried out and implemented in close coordination with the latter. However, the word "planning", whether applied to the national economy or tile energy sector in particular, need not imply some rigid framework along the lines of centralized and fully planned economies. Planning, whether by design or deliberate default, takes place in every economy, even where so-called free market forces reign supreme. In energy  - 6 - planning and policy analysis, the principal emphasis is on the detailed and disaggregated analysis of the energy sector, its interactions with the rest of the economy, and the main interactions within the various energy subsectors themselves. Energy policy analysis and planning must be developed to meet the many interrelated and often conflicting overall national objectives as effectively as possible. Specific goals usually include: (a) determining the detailed energy needs of the economy and meeting them, to achieve growth and development targets; (b) choosing the mix of energy sources to meet future energy requirements at lowest costs; (c) minimizing unemployment; (d) conserving energy resources and eliminating wasteful consumption; (e) diversifying supply and reducing dependence on foreign sources; (f) meeting national security and defense requirements; (g) supplying the basic energy needs of the poor; (h) saving scarce foreign exchange; (i) identifying specific energy demand/supply measures to contribute to possible priority development of special regions or sectors of the economy; (j) raising sufficient revenues from energy sales to finance energy sector development; (k) price stability; (l) preserving the environment; and so on. Scope of INEP The scope of INEP, policy analysis, and supply-demand management may be clarified by examining the hierarchical framework depicted in Figure 1.1 (Munasinghe, 1980). At the highest and most aggregate level, it must be clearly recognized that the energy sector is a part of the whole economy. Therefore, energy planning requires analysis of the links between the energy sector and the rest of the economy. Such links include the input requirements of the energy sector such as capital, labour, raw material and environmental resources (such as clean air, water or space), as well as, in relation to national objectives, energy outputs such as electricity, petroleum products, wood fuel and so on, and the impact on the economy of policies concerning availability, prices, taxes, etc. While some of these relationships are at the macro-level--such as foreign exchange requirements for energy imports, or investment capital requirements for the energy sector--others are more directly linked with or limited to specific y user sectors. For example, policies affecting the transport sector such as subsidies to public transport, construction or non-construction of super-highways or airports, the level of license fees for vehicles or relative excise taxes on diesel versus gasoline vehicles, tax credits for energy conservation, pollution control legislation or specific end-use planning policies may have as profound  MACROECONOMY MACRO LEVEL Transport Industry Energy Sector Agriculture Other MACRO LeVeL Interaction Between Energy Sector & the Rest of the Economy Resource Availability Resource Requiremnents (Inputs K.LM) Enmergy Outputs J Energy Demand Energy Sector Constraints (Disaggregate) National Objectives J ) / ENERGY SECTOP \s{ \ & Constraints INTERMEDIATE LEVEL Gas Oil Electricity Wood Coal Other a Energy Subsector Interactions ELECTRICIlY SUBSECTOR MICRO LEVEL Supply Demand a Subsector Planning & Management - Investment Planning - Pricing Policy (LR) - Operations - Physical Controls (SR) - Loss Optimisation - Technological Methods - Reliability Optimisation - Education & Propaganda Figure 1.1. Integrated National Energy Planning Framework  - 8 - an impact on energy demands as more broad-based energy pricing, allocation or supply management policies. The second level of INEP treats the energy sector as a separate entity composed of sub-sectors such as electricity, petroleum and so on. This permits detailed analysis of each sector with special emphasis on interaction among the different energy sub-sectors substitution possibilities, and the resolution of any resulting policy conflicts such as competition between natural gas, bunker oil or coal for electricity production, diesel or gasoline for transport, kerosene and electricity for lighting, or woodfuel and kerosene for cooking. The third and most disaggregate level pertains to planning within each of the energy sub-sectors. Thus, for example, the electricity subsector must determine its own demand forecast and long-term investment program; the petroleum sub-sector, its supply sources, refinery outputs, distribution networks and likely demands for oil products; the woodfuel sub-sector its consumption projections and detailed plans for rotation or reforestation, harvesting of timber, and so on. In practice, the three levels of INEP merge and overlap considerably. For example, a class of demand management issues that affects both macro and micro aspects of energy planning are those related to energy substitution or conservation. These issues may need to be analyzed at all three levels of the INEP hierarchy. Within certain limits many energy resources are substitutes for each other although convenience in use and overall systems cost may vary widely. Hence appropriate supply and pricing policies may bring about significant shifts in energy demand for specific energy resources, at least in the long run. Similarly, individual actions or deliberate policies aimed at bringing about energy conservation--e.g. reductions in energy usage relative to prevailing levels- -may significantly affect energy consumption. Such conservation may simply be achieved at the expense of some loss in personal comfort or convenience (like reducing thermostat settings, driving within mandated speed limits or switching off lights in unoccupied rooms). Other means may include substitution of energy by capital or labor, replacement of pilot lights by electronic switches, reduction in the curb weight of automobiles, recirculation of process heat in industrial plants through better engineering or lighter materials, and installation of insulating materials in buildings.  - 9 - 1.3 TOOLS FOR POLICY IMPLEMENTATION INEP should result in the development of a flexible and contemporary energy strategy which can meet the national goals discussed earlier. Such a national energy strategy may be implemented through a set of energy supply and demand management policies and programmes. To achieve the desired national goals, the policy instruments available to third world governments for optimal energy management include: (a) physical X controls; (b) technical methods; (c) direct investments or investment-influencing policies; (d) education and promotion; and (e) pricing, taxes, subsidies and other financial incentives. Since these tools are interrelated, their use should be closely coordinated for maximum effect. Physical controls are most useful in the short-run when there are unforeseen shortages of energy. All methods of limiting consumption by physical means such as load shedding or rotating power cuts in the electricity sub-sector, reducing or rationing the supply of gasoline or banning the use of motor cars during specified periods, are included in this category. However, physical controls can also be used as long-run policy tools. Technical means used to manage the supply of energy include the determination of the most efficient means of producing a given form of energy, choice of the least-cost mIix of fuels, research and development of substitute fuels such as oil from shale, coal or natural gas, the substitution of alcohol for gasoline, and so on. Technology may also be used to influence energy demand, such as introducing more fuel-efficient automobiles or better wood- stoves, and by research into and promotion of solar heating devices. Investment policies have a ir.ajor effect on both energy supply and consumption patterns in the long run. The extension of natural gas distribution networks, the building of new power plants based on more readily available fuels such as coal, or the development of public urban transport networks are just some of these policies. It should be noted that while many of these may well be undertaken by sectors other than energy--examples are investments in transportation facilities or the systematic installation and/or electrification of deep-well irrigation pumps--close cooperation between the energy administration and planning authorities of these other sectors are obviously called for.  - 10 - The policy tool of education and promotion can help to improve the energy supply situation by making citizens aware of cost-effective ways to reduce energy consumption, of the energy use implications of specific appliances or vehicles, and of the potential for substitution of energy by capital (e.g. proper insulation). Taxation and subsidies are useful policy instruments that can also profoundly affect energy consumption patterns in the long run. For example, countries which have imposed high taxes on gasoline have generally had significant results in terms of reduced automobile use, more efficient vehicle fleets, and so on. Subsidies have similarly encouraged energy saving capital investments. Pricing is politically sensitive, but, as discussed later, it is a most effective means of demand management, especially in the medium- and long-run. However, pricing has limitations and must be skillfully combined with other non-price policy instruments for best results. For example, in the traditional fuels sector (e.g., fuelwood), where there is no well developed market, pricing policy may not be effective. Reduction of fuelwood use may require other non-price measures such as disseminating a more effective domestic cooking stove. In the case of very high income consumers, who are willing to pay an extremely high price for their comfort (e.g., air conditioning), high prices might not reduce demand significantly, but could be used instead as a revenue generating mechanism. At the other end of the scale, raising prices to low income energy consumers may simply cause them to spend a greater share of their income for energy purchases without affecting energy use, because these consumers are already at the basic needs level and could not do without a minimum amount of energy (e.g., kerosene for lighting in rural homes). In such cases, energy pricing policy would cause hardship by limiting the income available to poor households to purchase essential non-energy items. In summary, controlling the use of energy through coor(linated use of the various policy tools is the principal goal of energy supply and demand management. Practical aspects of such policy coordination are discussed later. 1.4 CONSTRAINTS ON POLICY In a recent review of the national energy planning experience in developing countries, Wilbanks (1987) echoes the consensus that formal national energy planning has seldom made much of a difference in important energy decisions, despite wide recognition of its importance. This clearly reflects the formidable constraints that come into play in implementing such a process.  - 11 - The chief constraints that limit effective policy formulation and implementation are: (a) a poor institutional framework; (b) insufficient manpower and other resources; (c) weak analytical tools; and (d) inadequate policy instruments. (a) Institutional Framework A properly functioning institutional framework is a vital pre-condition for the success of any planning process. Indeed, as pointed out by Wilbanks, the greatest challenge to implementing integrated national energy plans comes from the institutional, rather than the methodological, side. An adequate institutional framework for energy policy formulation and implementation should have at least three well defined and balanced elements: policymaking, implementation, and research and development. Most often, to X, the detriment of all, all three aspects are mixed together. For example, senior energy decisionmakers, who should be devoting a significant part of their time to major policy issues, are invariably embroiled in day-to-day crisis management. The academic and research community frequently focus their interest on energy problems that are either insignificant or irrelevant from the national policy viewpoint. Finally, even when separate organizations exist for policy, operations, and R&D, their efforts are often uncoordinated. Studies of institutional frameworks that have been successful should be of considerable value to many developing countries seeking to restructure their energy organizations. Usually, the existing line agencies, such as electric utilities and oil companies, are able to attend to daily operational needs, while institutions such as universities and research centres are geared to perform R&D functions. The weakest area is that of policy analysis and formulation; unfortunately, it lies at the heart of INEP and hence, needs considerable strengthening. (b) Manpower and Other Resources The lack of skilled manpower for policy analysis and development, as well as energy management and project implementation, are severe handicaps. p Education and training programmes have a vital role. Furthermore, specialists with broad skills in energy planning and policymaking are likely to be more scarce, and therefore more valuable, than those having expertise in narrower areas such as electricity generation or refinery operations. One approach to maximize the use of scarce skills may be to build up multidisciplinary teams (consisting of technologists, economists, financial analysts and social scientists, among others). Adequate remuneration, autonomy, recognition and provision of facilities for research, are other means by which specialists may be retained within a country.  - 12 - (c) Analytical Tools A broad range of models recently used for energy planning are reviewed in Chapter 2. As mentioned earlier however, work done during the last decade has led to the conclusion that a hierarchical modelling framework might be best suited to more effective policy analysis and application. The hierarchy of models closely follows the INEP analytical outline shown in Figure 1.1. The operation of the modelling framework is described in greater detail in Chapter 3, which also discusses one such set of models developed and used in Sri Lanka in 1984-85. The Sri Lanka model used a MS-DOS compatible, 16-bit microcomputer, with 256 kilobytes of main memory and a 10 megabyte hard disc--i.e. computer technology readily available at that time. It is important to note tllat all the computerized activities may be first carried out manually, at a relatively simple level, and only later, as data and local analytical skills improve, should more sophisticated computer modelling be pursued. Furthermore, while the computer models are being developed, energy policy formulation should not be neglected--key options may be examined using more conventional techniques, and appropriate decisions taken and implemented (e.g., re-forestation, efficient cookstoves, energy pricing, demand management and conservation, and choice of renewable supply technologies). These early decisions will facilitate timely action on important issues, and may be coherently incorporated into a subsequent longer-term national energy strategy when more comprehensive computer runs can also be carried out. Great care must be taken to ensure that the models represent reality as closely as possible while remaining tractable, and they should be designed to provide answers to typical questions that senior decisionmakers might be expected to ask. (d) Policy Implementation Tools The constraints and research issues concerning policy implementation tools have been discussed earlier in Section 1.3. While in theory many tools are available at the disposal of the policymaker, it was observed that they are frequently either limited in effectiveness or impose considerable social and economic costs. In practice, policy implementation calls for the coordinated use of several tools in order to maximize impact. (e) Other Constraints In the context of developing countries we generally face additional constraints on energy policies. There may be severe market distortions due to taxes, import duties, subsidies, or a externalities which cause market (or financial) prices to diverge substantially from the true economic opportunity costs (or shadow prices). Therefore, on the grounds of economic efficiency  - 13 - alone we may have to make (second-best) shadow pricing adjustments. However, these again may have to be modified in anticipation of energy user reactions that will be based on market prices rather than underlying economic cost considerations. Furthermore, severe income disparities and social considerations need to be addressed by subsidized energy prices or rationing to meet the basic energy needs of poor consumers. Finally, there are usually many additional considerations that affect policy requirements such as financial viability and autonomy of the energy sector, regional development needs, as well as socio-political, legal and other constraints. Often the distinction between policy objectives (described earlier in Section 1.2) and constraints is blurred.  - 14 - REFERENCES Meier, Peter, "Energy Planning in Developing Countries: The Role of Microcomputers," Natural Resources Forum, Vol 1, January 1985, pp. 41-52. Munasinghe, Mohan, "Integrated National Energy Planning in Developing Countries," Natural Resources Forum, Vol. 4, October 1980, pp. 359-73; also available as Reprint No. 165, The World Bank, Washington, D.C. Munasinghe, Mohan, "Practical Application of Integrated Natural Energy Planning (INEP) Using Microcomputers,' Natural Resources Forum, Vol. 10, February 1986, pp. 17-38; also available as Reprint No. 374, The World Bank, Washington, D.C. Munasinghe, Mohan, "Energy Economics in Developing Countries: Analytical Framework and Problems of Application," The Energy Journal, Vol. 9 January 1988, pp. 1-17; also available as Reprint No. 421, The World Bank, Washington, D.C. Wilbanks, Thomas J., "Lessons from the National Energy Planning Experience in Developing Countries," The Energy Journal, Vol. 8 (special LDC issue), July 1987, pp. 169-182.  CHAPTER 2 ENERGY MODELLING METHODOLOGY 2.1 INTRODUCTION As described in Chapter 1, successful energy management depends on both good analysis and effective implementation. In this chapter, we explore some of the analytical tools for modelling and policy development. The implementation of policy is described later in the context of the Sri Lanka case study. There is an unquestioned need for analytical tools that have the ability to integrate project and policy sector information across the different energy sub-sectors, and also incorporate sector-wide planning considerations and energy-related linkages into the overall macroeconomic picture. First, energy analysts, both in lending agencies and in country planning entities, are having to deal with more complex problems at a higher level of sophistication. Second, the macroeconomic and sectoral outlook in many countries is much worse than before, and new approaches and initiatives need to be explored. Third, the policy package associated with new lending instruments such as sector restructuring loans and structural adjustment loans must be more clearly defined and elaborated. Finally, the uncertainty regarding future energy prices, world economic growth, interest rates, access to foreign capital etc., has put a greater premium on a more flexible approach to policy analysis and decision- making. The need for improved analytical approaches has, of course, been recognized both by national governments and their often newly created energy planning entities, as well as by multi- and bi-lateral assistance agencies that, over the past 10 years, have provided guidance and technical assistance to such bodies. Indeed, particularly over the past few years, there has been a dramatic increase in the number of countries that have introduced microcomputer-based energy planning tools. However, despite the great interest in the topic, and the proliferation of energy models over the past few years, it is unclear to what extent the analytical frameworks and modelling tools are in fact useful, be it from the country perspective (e.g. in terms of the contribution made to the definition of national energy strategies), or from the perspective of development agencies (e.g. in terms of clarifying the impacts and benefits of policy packages associated with structural adjustment loans). Several factors have impeded the wider use of such tools (Wilbanks, 1987). On the methodological side these include data constraints, lack of integration with other sectors, lack of uncertainty considerations, geographic differenices and disregard for the demand side. Perhaps even more important, according to Wilbanks, are the  - 16 - institutional problems that prevail in many developing countries. Despite the general lack of information, it seems clear from what is available, and from the experience of the UNDP/World Bank energy assessments, that the scope, appropriateness and sophistication of energy models varies quite considerably. Sometimes, even when the models developed are useful, the unavailability of trained local staff or good documentation can be a significant hurdle. In this chapter the specific analytical characteristics necessary to support an integrated national energy planning process are presented. This takes the form of a discussion of criteria by which energy models are evaluated. A classification scheme is then developed for energy models together with a discussion of the leading models that fit each category. The chapter concludes with a discussion of the practical implementation of a modular, hierarchical approach, which implies a set of small, interlinked, well focussed models that adhere to some common data transfer protocol to facilitate interaction among them. 2.2 EVALUATION CRITERIA In this section we develop the criteria by which energy models could be evaluated. These criteria represent those attributes of analysis that we feel are important to a successful modelling effort in support of policy-making and investment planning. Obviously, given the broad scope of such efforts, this could be a very long list indeed. However, the experience in a large number of small- to medium-sized developing countries indicates that a major part of the analytical needs of energy planning entities are (or should be) focussed on a relatively limited set of short- to medium-term issues: integration of subsectoral planning efforts to achieve sector-wide consistency and integration with the overall macroeconomy, pricing and demand management, project evaluation and-medium term investment planning. The criteria discussed below address such concerns, with the qualification that smaller and less complex economies can often get by using a relatively unsophisticated set of criteria. A summary of the evaluation criteria is presented in Table 2.1. Since we are reviewing the many different types of models that constitute the typical energy modelling system, it is obvious that not all criteria apply to all models; some are specific to macroeconomic models, others are specific to integrated energy sector or energy pricing models.  - 17 - Table 2.1 Summary of Evaluation Criteria General Criteria 1. Ability to respond to key questions of policy and priorities. 2. Degree of integration with and replication of the planning process. 3. Approach to treatment of uncertainties and the types of uncertainties that can be handled. Macroeconomic Model 1. Integration and consistency with government accounts and forecasts. 2. Disaggregation between energy and non-energy sectors. 3. Geographic disaggregation. 4. Alignment of model with available data. Economic Model 1. Incorporation of project level information. 2. Formulation of demand structures. 3. Technical representation of the electric power sector. 4. Does output format follow standard conventions? 5. Representation of institutional structure. 6. Representation of petroleum sector. 7. Treatment of non-commercial fuels. 8. Extent of financial detail including tariff structures. Subsectoral Models 1. Extent to which model accepts data from models at subsectoral level.  - 18 - Integration with the Planning Process Perhaps the most important attribute of a modelling system from the standpoint of the energy planning agency is the degree to which it replicates the institutional realities. If one accepts the premise that a key function of sector level policy analysis and plaining is to integrate, rather than duplicate, the planning efforts at the subsector level into a consistent, sector-wide picture, and into the overall macroeconomic planning framework, then it follows that the modelling framework must also be consistent with this function. This requirement is reflected both in the overall analytical and modelling approach as well as in matters of data and technical detail. Philosophically, given the premise of a hierarchical planning process itself, there follows also a hierarchical modelling framework in which the individual modules can be exercised independently. Ideally, the models at the third, subsectoral level are run jointly by both the subsectoral institution and the energy sector level planning institution; or at the very least the energy model must have the ability to accept information at the same level of detail as that offered by the subsectoral entity. For example, in the case of the electric sector, the model should have the ability to accept the capacity expansion plan of the electric utility (a subject examined further below). Treatment of Uncertainties Energy and macroeconomic policy analysis and planning in the typical oil- importing country is beset with a multitude of uncertainties, and the ability to identify energy and investment strategies that are robust under a variety of future outcomes is perhaps the key element in a successful planning process. Several fundamentally different types of uncertainty must be faced: (1) the stochastic character of all natural phenomena (such as rainfall and streamflow, or the distribution of oil reservoir sizes); (2) the uncertainties of the international economic and energy environment, over which the decision- makers in typical developing countries have no (or at best only marginal) control; and (3) the uncertainties in the functioning of the domestic economic system (and hence uncertainty in the responsiveness of the system to policy initiatives). There is a well established body of analytical techniques that deal with the first type of uncertainty, usually referred to as decision analysis; and indeed these techniques are widely used in subsectoral eniwrgy models for petroleum exploration, probabilistic simulation of electric generating systems, technology choice, siting of facilities, and the management of multi-purpose  - 19 - water resource systems. The successful application of such techniques depends on the ability to define the underlying probability distributions, which in turn demands an adequate database. However, despite the usual problems of data in developing countries, experience shows it to be much easier to extrapolate probability distributions concerning natural systems (from other countries, or similar climate and geologic regimes) than to extrapolate socio-economic phenomena. Thus these techniques have been applied fairly successfully to developing countries. Application of such techniques to the features of the international economic system are much more difficult, because physical laws do not apply to political, economic and social behaviour, and hence a derivation of the necessary probability distributions becomes an almost impossible task. For example, all manner of game-theoretical models were developed in the late 1970's to model OPEC behaviour and the future trajectory of the world oil price. Yet few, if any, came even close to projecting the oil price downturn in the 1981-1983 period, much less the sudden collapse that followed in 1986. Indeed, such models have not been used anywhere in developing countries as a basis for decision-making. In response to these difficulties, the standard technique has been the use of scenario simulation, in which values of such exogenous factors are hypothesized without any initial judgement on the likelihood of occurrence. In effect, the judgement about uncertainty is shifted fronm the analyst to the decision-maker, since it is the latter who must make the trade-off between scenario and policy response, on the basis of the projected impact. However, if such an approach is to be practical, it poses some rather severe limitations on a prospective modelling system. First, since there may be several exogenous scenarios as well as policy preferences of decision makers for each of several parameters, the. possible permutations of scenarios quickly proliferates. The analyst must be able to quickly test over a multitude of scenarios, discard those with duplicative or uninteresting outcomes, and focus on a more manageable subset. Too many scenarios will simply overwhelm the typical decision-making body. If such a model is to be used in practice, a simple rule of thumb might be the requirement that a complete model run consume a period of some minutes, rather than hours. Integration with Macroeconomic Plans At the macroeconomic level, perhaps the most important concern is to develop a base case scenario that is consistent with the official government  - 20 - forecast. In most countries there exists an official plan, typically prepared in an annual or 5-year cycle by the Ministry of Planning or Finance. Whilst such forecasts may well be in the nature of a set of goals rather than a forecast, and therefore subject to some controversy, it is nevertheless true that as a matter of practice this forecast usually has official standing, and therefore provides the basis for the participation of the macroeconomic planning authorities in sectoral planning efforts. It follows that the macroeconomic model used for the short- to medium- term covered by this forecast should have the ability to replicate the official forecast as the baseline for analysis. Ideally, this baseline should then be capable of perturbation to reflect other assumptions in a consistent fashion, particularly if the official baseline is considered unrealistic and could be replaced with a more responsible alternative. Moreover, the presentation should be in a format that closely follows the official guidelines (for example, using the same sectoral disaggregation, even if this poses some difficulties elsewhere in the analysis, e.g. in the demand projections). By the same token, given that some of the more important impacts of energy sector decisions are on the external accounts, the format and presentation should follow closely that of the central bank and its reporting requirements to the international financial community. All of this implies a high degree of disaggregation: however, it is easier to produce an aggregated measure from the disaggregated categories, than it is to disaggregate a measure in a manner consistent with official statistics. For example, instead of a single variable for external debt in an energy-macroeconomic model, it is helpful to use the same disaggregation as the central bank (short, medium, long term, government, private, etc), and then assign the energy sector debt service to the appropriate category. Disaggregation of Energy and Non-energy Sectors There are two basic approaches to the reconciliation of energy models and highly aggregated macroeconomic models (in which energy and non-energy sectors are either not distinguished at all, or at a level of disaggregation unsuited to energy work), particularly with respect to the evaluation of the macroeconomic impacts of the energy sector. The simple expedient is to run the two models independently, with a comparison of the energy sector impacts derived from the energy model (say the level of energy-derived government revenues, or foreign debt service) against the overall aggregates projected by the macro model. Of course, since the aggregate includes the energy sector, this approach is both imprecise and inconsistent, although perhaps adequate for a first order estimate.  - 21 - A more detailed, but also rnore difficult approach is to attempt some explicit disaggregation of the macro model. Where formal models are concerned this may require extensive changes to the equation structure. On the other hand, where the basis is a macroeconomic accounting framework rooted in the official government plan, it is usually quite easy to subtract out energy sector investment (and debt service) from the aggregate investment projections. Moreover, since the energy sector often contributes a relatively small amount to GDP, its inclusion in other value added sectors poses few problems and an explicit disaggregation here is not necessary (electricity is usually included in a sector such as municipal services that also includes water and other utilities, while refinery operations are frequently included in the industrial sector). Alignment of the Model to Available Data Many macroeconomic models developed for use in energy-economic studies prove to be poorly designed from the standpoint of possessing a level of cornplexity that is consonant with the available data. Even those models that make no claim of relevance for the short- to medium-term, but whose focus is the long-run structural adjustment process, frequently use theoretically convenient equations whose coefficients are difficult if not impossible to derive from actual, in-country data. Such models typically use log-log production functions involving constant substitution elasticities (among energy, labor, capital and other material inputs) that are quite different from the sort of production functions estimated in sectoral studies. This problem is perhaps most acute in that subset of countries whose economies are still largely agricultural, and whose export earnings depend heavily on a few export crops. Any reasonable and useful formulation of plantation sectors involving perennial crops would need to include in a sectoral production function--beyond only price and energy (and/or fertilizer) inputs- -such additional variables as the area under cultivation, and the age distribution of the planting stock. Since in many cases increases in the area under agricultural or plantation cultivation implies a decrease in the natural forest area, there may also exist an indirect link to the fuelwood supply; irI some cases (such as rubber), the replanting cycle itself contributes to the fuelwood supply (in Sri Lanka, for example, a significant portion of the urban fuelwood supply in the Colombo area is rubber wood, i.e. trees felled to make way for new plantings). Typically, production functions in macro-economic models are of the Cobb- Douglas form, which are the sinmplest that allow factor substitution. One such form is: X = aLb Ec Kd  - 22 - where X is incremental output, L is the incremental labor input, E is the incremental energy input, K is the previous year(s) investment, and a,b,c and d are technical coefficients. Yet production functions encountered in typical sectoral studies are frequently of a quite different form, especially for agricultural sectors, including at a minimum variables for land area, fertilizer inputs (of importance anyway to energy planning because of the energy intensiveness of their manufacture), and irrigation/rainfall inputs. The argument here is not that macroeconomic models should necessarily be encumbered with vast sectoral detail. However, to the extent that macroeconomic models are built to examine the longer-term issues of structural adjustment (and the impact of that adjustment on the energy sector), then maximum use should be made of the sectoral models built by others interested in the rehabilitation or adjustment of major producing or exporting sectors, especially where there are direct links to important energy supply issues as well. Project Level Information One of the most important questions concerning the suitability of a general modelling framework for both the national decision makers and international lending agencies is the ability to include project level information. The smaller the country in question, the more important this issue becomes, since a single electric sector project may in such cases account for a major share of public sector investment and debt service. In brief, the need to evaluate project level information within the broader context of the energy sector and the macroeconomy is not only important to financing institutions, but is also fundamental to energy sector investment planning on the part of the borrower. It should be noted that this criterion is not met by the simple availability of a project analysis package. Indeed there are many commercial packages now available for cost-benefit type calculations. Equally important are questions such as how alterinative investment plans are assembled, and, once assembled, how they are integrated into the overall analysis. Are project portfolios assembled on an ad hoc basis, by ranking of internal rates of return (IRR)--in fact a possibly misleading procedure if capital is constrained, or by some sophisticated capital budgeting model that maximizes net present value (NPV) subject to capital constraints? Such questions are especially important for the electric power sector, where the determination of the optimal capacity expansion path is a matter of considerable complexity. The main issue here is the degree to which the  - 23 - energy model can be integrated with typical electric utility capacity expansion plans. Demand Structures The adequacy of the demand structures in many models is in some doubt, particularly with respect to the ability to estimate the impact of policy initiatives. Perhaps the best example concerns the electric sector which, in many developing countries, is supply constrained--the major determinant of demand being the ability of the electric utility to extend the system into both previously unconnected rural regions and rapidly expanding urban centers. Yet the traditional econometric formulation of demand as a function of price and income frequently ignores this issue (quite aside from other assumptions such as constant own-price elasticities, and neglect of cross-price elasticities). To be useful for an energy plaining and policy exercise, the demand structure needs to consider, beyond just price and income variables, such factors as the extent of energy and peak load accounted for by auto-gerieration (which is highly sensitive to the level of reliability provided by the central system), the theft rate (as distinct from purely technical losses, which are subject to altogether different types of policy intervention), and the rates of new conxections (in turn also related to the considerable costs of distribution system expansion). Uncertainty must also be taken into account as described in the case study. To be sure, the level of available data, and the unique circumstances facing a particular country, make such generalizations hazardous. Nevertheless, it is the experience in a very wide set of countries that such questions are almost always present. Indeed, they lie at the heart of the controversy over demand projections that are encountered by, and hence merit the attention of, energy planning and policymaking bodies as part of their coordinating functions. Teclnical Representation of the Electric Sector Because of the importance of the electric sector from the standpoint of investment and debt service, and the inherent technical complexity of the operation of this subsector, an important issue concerns the degree to which an energy model has sufficient technical credibility to demonstrate the impact of multiple, and simultaneous policy and project initiatives. It is not at all uncommon for several multi- and bi-lateral agencies, and the country government itself, to be simultaneously contemplating the financing of additional generation capacity, the rehabilitation of the distribution system to  - 24 - reduce system losses, and an industrial sector energy efficiency initiative designed to improve power and load factors. It follows that if an energy model is to be useful even from a very narrow perspective, it has to have the ability to demonstrate the interactive impacts of such diverse but parallel initiatives. This in turn requires a level of technical detail that at a minimum iiicludes the ability to replicate the essential features of system load-duration curves and the optimal dispatch of specific generation units. Output Format and Data Conventions Since one of the major reasons for using models is to increase the productivity of the staff engaged in the support work for decision-making bodies, it follows that models should as far as possible adopt the usual reporting conventions in their presentation of key results. For example, energy balances should be generated in the form almost universally accepted by international institutions. This may perhaps be a somewhat obvious point, but several existing models not only require manual transcription of energy balances, but in some cases require additional calculations as well. Consideration must also be given to making outputs readily comprehensible to local decision makers. Institutional Structure Beyond the question of how the overall modelling system is aligned to the institutional realities as discussed above, is the degree to which an energy miodel itself is structured around the institutional framework. Since the revenue, tax and subsidy flows that are of major concern to government decision makers and to an economically efficient pricing system are related to the transactions between specific einergy-sector institutions, it follows that for the energy model to be useful for short- to medium-term analysis, it should have as its basic building blocks not subsectors (petroleum, electricity, gas, coal), but institutions (the refinery or oil company, the electric utility, the LPG company, etc.). Petroleum Sector Representation The representation of the petroleum sector in energy models is frequently unsatisfactory. Problems range from inadequate disaggregati(oin (which makes them difficult to use as a basis for pricing studies), to an inadequate treatment of refinery flexibility (future petroleum product output from a domestic refinery is estimated by mechanistic multiplication of the crude run by some set of yield coefficients). Indeed, nowhere is the ability of an energy  - 25 - model to interact with a more detailed subsectoral model more important than the petroleum sector. In that they can be structured around a linear programming representation of a refinery, this is one of the few areas where LP-based energy models have an inherent advantage. Simulation model representations of refineries must therefore have either an appropriate interface to a detailed refinery LP, or find some other scheme to deal with refinery capacity expansion and flexibility issues. Treating the refinery as a black box, with product outputs given as linear functions of the crude inputs is problematic even for simple hydroskimming refineries because crudes can be spiked over quite wide ranges. For example, in 1985 the Refinery of the Dominican Republic used as input four different Mexican and Venezuelan crudes, as well as a reconstituted crude consisting of a blend of butane, naptha, kerosene and gasoil. Non-commercial Fuels There are a number of alternative strategies to deal with non-commercial fuels. Many LP energy mnodels simply ignore them. Others include them in the overall scheme for computing energy balances, but with demand projections for fuelwood and charcoal not well integrated with the commercial fuel demand projections (the issue being the inadequate understanding of substitution effects among the commercial and non-commercial fuels). Reference energy system models generally do provide a structure to maintain a consistency in such substitutions, but are ill equipped to relate the pace of change to the price eiivironment. In hierarchical systems that include a detailed fuelwood model, there are special difficulties in forcing consistency among projections for electricity, petroleum products and non-commercial fuels. The unavailability of data is also a reason for not modelling non-commercial fuels. Data Integration with Subsectoral Models The proposition that energy models should be linked to more detailed subsectoral models is logically sound, and the claim that a given model is consistent with other models is almost always made. Indeed, as a matter of computer technique, it is not too difficult to transfer data from one model to another. In practice, however, the question is not one of modelling, but of institutional relationships. What matters is not only the technical capability of an energy model to accept subsectoral detail (as argued above in the case of the electric sector capacity expansion plan), but an institutional mechanism that ensures a sufficient dialogue for the transfer to actually take place. In some countries (such as Morocco and Sri Lanka), such a mechanism is present; in others it is notably absent, which makes any modelling effort quite suspect.  - 26 - Relationship to Staff Capabilities Finally, in terms of in-country application, mention must be made of the issue of staff capabilities and training. Even a modelling framework that possessed all of the above attributes would be without value unless the capability exists to use it. The working level staff must be sufficiently well trained to be able to understand and operate the analytical tools successfully. One might note that this is not only a matter of training in modelling and the particular software implementation--indeed, this is probably the least important aspect (if the software is well designed). Rather, the central issue is the technical background and experience of the staff. Often, a well trained engineer or economist, with some years of experience, can usually be trained successfully to use a modelling system, whereas programmers, modellers, and others without experience in a refinery, electric or gas utility, whatever their computer expertise, rarely have the technical judgement to successfully understand and operate a model ensemble. However, the participation of a programmer/systems analyst may well be desirable if present as part of a broader team. 2.3 CLASSIFICATION AND EVALUATION OF ENERGY MODELS There are four broad categories of models: 1. Macroeconomic models that stand at the top level of the hierarchy, including both stand-alone models as well as those designed as modules of a broader modelling system. II. Energy models that occuipy the middle level of the hierarchy, again including both those designed as modules in comprehensive systems, as well as stand-alone models that might, in principle, be used in a modular system. III. Integrated, non-hierarchical energy-macroeconomic models. IV. Modular, hierarchical systems--here the focus is not on the components (reviewed individually in Categories I and II), but on the interactions between the models. A summarized classification of the models is contained in Table 2.2. It should be noted that although models at the subsectoral level of the hierarchy (of which there may now be many hundred in use in developing  - 27 - Table 2.2 Clasification of Models Model Country Macroeconomic Model. Conventionai Macroeconomic Models SIAM-2, World Bank Thailand Purpose-built macroeconomic drivers ENERPLAN, Tokyo Energy Economics Group for UNDP Costa Rica Macroeconomic Accounting Frameworks ENMAC, Munasinghe and Meier for Govt. of Sri Lanka Sri Lanka Extended Macroeconomic Models Hill, Oak Ridge National Laboratory for USAID Liberia Vinh, World Bank Tunisia Blitzer, MIT for USAID Jordan Energy Model. Reference-Energy System Based EDIS, SUNY for USAID Dominican Republic RESGEN-3, IDEA, Inc. Indonesia Simultaneous Equation Models ENERPLAN, Tokyo Energy Economics Group for UNDP Thailand Argonne Energy Model Jamaica Linear programming models Thailand Energy Master Plan, Systems Europe Thailand Integrated Non-Hierarchical Model. Input-Output Models BEEAM, Brookline National Laboratory for USDOE Portugal Optimization Model. TERI IO/-LP model, Mubayi India Other ENVEST, DSI Inc. for USAID Morocco & Costa Rica Hierarchical Systems Sri Lanka Model, Munasinghe and Meier for Govt. of Sri Lanka Sri Lanka Bangladesh Model, DeLucia, Metasystems for ADB Bangladesh Morocco Energy Model, IDEA, Inc. for USAID Morocco & Pakistan  - 28 - countries) are not reviewed here, a great deal of attention is paid to how such subsector models could be linked to the energy sector, and macro level models. Indeed, such questions as how electric utility capacity expansion plans are integrated into energy models are central to the ultimate effectiveness of a modelling effort to support policy analysis. Category I: Macroeconomic Models Several quite distinct varieties of macroeconomic models have been used for energy policy work. The first are conventional macroeconomic models designed for general macroeconomic analysis rather than energy work. These typically include one or more highly aggregated energy variables included in production functions, whose substitution elasticities with respect to capital, labor and other inputs are econometrically estimated. Such models, by virtue of their inclusion of such aggregated energy variables, can be used to assess a certain set of energy policy questions. The best known in this category is the SIAM-2 model for Thailand, used by the World Bank for analysis of the impacts of alternative policy options in response to falling prices. SIAM-2 is a computable general equilibrium (CGE) model designed to address general questions of fiscal, exchange rate, wage, and pricing policies. It has 22 producing sectors, all of which (with the exception of gas and petroleum products) are characterised by sophisticated neo-classical production functions. The most critical part is the provision for policy simulation within the capital market structure. Markets clear, either via price adjustment or quantity adjustment. Where international markets exist for the commodities in question, prices are taken as exogenous; the same applies in the case of government set prices such as energy products. SIAM-2 is a very sophisticated model which limits its wide application in the developing world. The second group consists of much smaller macroeconomic models, built expressly to drive energy models. These are usually limited to a fairly small equation set for projecting GDP, population growth, etc., and generally lack the sophistication of the first group. The ENERPLAN macroeconomic driver developed for Costa Rica by the Tokyo Energy Analysis Group falls into this category. The strength of ENERPLAN is its extreme flexibility, since it enables the user to build whatever equation structure that is deemed to be appropriate. The user is provided with both macroeconomic and energy sector models, with the former designed as a driver for the input side of the energy model. The two models can also be exercised independently.  - 29 - The third group, macroeconomic accounting frameworks, is distinguished from the second group by two main features. They are based on accounting identities rather than econometrically estimated relationships, and encompass a much greater level of detail. The Sri Lanka Macroeconomic Accounting Framework (which is embedded in the INEP structure as further described in Chapter 3) is representative of this category. The basic approach was to replicate, in the equation structure of LOTUS 1-2-3, the macroeconomic and investment projections of the Ministry of Finance and Planning (MFP), and the reporting structure of the external accounts of the Central Bank of Ceylon. As such, sectoral growth rates, investment, savings etc., are all taken from the corresponding official planning documents. From this projection is subtracted the MFP estimates of energy sector investment and trade transactions. Finally, in the fourth group, are extended macroeconomic models, in which a substantive macroeconomic model is extended to include a number of energy variables. They are distinct from the first group in that they were built expressly for analysis of energy-macroeconomic issues. Typical examples are those of Hill (Oak Ridge National Laboratory) for Liberia, Dinh (World Bank) for Tunisia, and Blitzer (MIT) for Jordan. The objective of the Liberia model was to analyze the relationships between energy demand and the macroeconomy, without addressing the supply side of the energy system. As such, it has limited scope, and does not address questions of the macroeconomic impact of energy sector policies or investment decisions. Nevertheless, it typifies the sort of linkages used in many energy models of this type, in which the level of detail on the macroeconomic side is extremely limited. The Tunisia model was developed to analyze oil and gas policies in Tunisia. The model consists of five main blocks covering the macroeconomy, a petroleum production block, an oil balance block, a natural gas block and a fiscal block that simulates government revenues. While this is a powerful model, it is, of course, limited to the oil and gas subsectors. It has the potential to serve as an excellent starting point for other countries in which the government has many private sector partners engaged in fossil resource development. The Jordan model was developed to analyze energy-economy interactions. While much of the national accounts system is fairly standard, the niodel has a number of features that permit a variety of useful insights not available from others. It has seven sectors which produce all the domestic value-added. For each sector the model also calculates import requirements, energy subsidies and customs duties.  - 30 - Exports, international prices, sectoral investment plans, indirect taxes and subsidy rates, and government and private consumption growth are all exogenous. Category II: Energy Models Energy models fall into three main groups: (i) simulation models based on the Reference Energy System framework (developed originally at Brookhaven for use in the USA, but subsequently modified and adopted by many developing countries); (ii) simulation models based on a simultaneous equation structure, and (iii) optimization models, primarily linear programming. Although the two simulation approaches can be made to be mathematically identical, in practice they reflect rather important differences in conceptual approach. The best known implementation of the Reference Energy System approach is the RESGEN model, used by the national energy planning agencies in Thailand, Indonesia, the Dominican Republic and Sri Lanka, among others. The RESGEN energy model was designed specifically for microcomputer implementation, and for developing countries. It has become one of the more widely used software packages for energy planning in developing countries. As in the case of ENERPLAN, RESGEN is more accurately described as a software package to generate models, than as a model per se. The key innovation int RESGEN is that the fixed network structure is abandoned in favor of a totally flexible structure in which the user can create networks of arbitrary structure. This flexibility is reflected in the diversity of applications, ranging from analysis of the rural energy sector in Rwanda, to the sophisticated multi-regional model used by the Ministry of Energy in Indonesia. RESGEN is designed to be used together with LOTUS 1-2-3; input information can be drawn from LOTUS spreadslheets and output information transferred to LOTUS for further graphics presentation. It is designed to be used as part of hierarchical modelling systems, and therefore requires some sort of macroeconomic driver. However, the unique feature of the program is that it allows three different types of demand structure: (1) econometric specification, (2) project demand specification, and (3) process models. Thus, it is a model of great flexibility, though this also imposes rather severe requirements on the user. Among simultaneous equation energy models, the ENERPLAN model for Thailand is again a good example. The energy model consists of some 60 equations, consisting of groups of equations for: (1) final demand; (2) energy conversion; (3) primary energy supply; and (4) energy prices. The equations are econometrically determined and major exogenous variables include real consumption, wholesale price index, and manufactured industry output. The  - 31 - general programming environment of ENERPLAN is extremely friendly and new models could be implemented rather easily. The Thailand Energy Master Plan model is representative of LP models. It consists of an extensive set of models adapted from the EEC models assembled by Systems Europe SA. The EEC models have a core of 15 energy networks. These networks can be run either in a simulation or optimization mode. The linear program selects that combination of resources, fuels and processes that minimizes the total discounted cost over the period 1981-2001, subdivided into five-year intervals reflecting the 5th to 8th National Development Plans. The major limitation of this model appears to be its sheer complexity. Furthermore the question always exists whether such institutional detail could be incorporated into a LP; however, the structure certainly seems warranted for simulation purposes. Category III: Integrated, Non-Hierarchical Energy-Economic Models This group of energy-economic models is to be distinguished from the hierarchical models of Category IV by virtue of their integration of the energy and macroeconomic components. In particular, energy and macroeconomic components cannot be run separately either as a matter of design (for example, the input-output model is embedded as a set of constraints in the BNL/TERI model for India), or as a matter of programming philosophy (for example in the ENVEST model for Morocco, the input/output driver is conceptually distinct, but the user cannot uncouple it from the energy sector simulation). The ENVEST model was built for the Ministry of Energy and Mines in Morocco by Decision Sciences, Inc., as part of a long term energy planning technical assistance project sponsored by USAID. It has the distinction of being among the first comprehensive microcomputer-based energy analysis systems implemented for developing countries. ENVEST's greatest strength is its close integration with the energy planning process in Morocco. In its design, the focus was on project analysis. The user can access a project analysis module, from which individual projects can be analyzed and returned to the database. As a completely menu-driven system, it is also user-friendly. Category IV: Hierarchical Models Finally, in Category IV, we have the hierarchical systens. These are modelling frameworks where not only are the macroeconomic and energy models modular in structure, but a great deal of emphasis is put on interaction with subsectoral models as well. Although too early for  - 32 - microcomputer implementation, the first such approach was by DeLucia and colleagues for the Bangladesh Energy Study of the mid 1970's. More recently, microcomputer versions of such systems have been implemented for Sri Lanka and Morocco. The hierarchical system developed for Sri Lanka, which is the analytical framework for the Sri Lanka case study in this book, is described in detail later, in Chapter 3. 2.4 CONCLUSIONS Since the issues, priorities and institutional structure will, of course, vary from place to place (and from time to time), definition of a universal model is neither possible nor desirable. However, based on the preceding model reviews, there is a great deal that can be said in the way of useful guidance for the technical design of a modelling framework. Hierarchical vs. Single Model Systems The foregoing discussion indicates that a hierarchical approach to energy modelling is likely to be the most effective, primarily on the grounds that it best captures the actual techno-economic linkages and institutional realities of the energy planning process. Moreover, our review of available models shows that many of the deficiencies of single models could be resolved by linkage to another model that provides some missing element of consistency (of which a good example is the Jordan Macromodel, which could be considerably enhanced by linkage to an energy sector model). Even in an integrated framework the question arises as to the desirable level of automation between models. In the Sri Lanka model ensemble, the information between subsectoral and(i energy models takes place manually, and the linkage between energy and macromodel, while using an intermediate data file of common format, requires manual intervention to implement. King (1985) has pointed out that if the model simulates the structure of the actual decision process, natural breaks would arise at points where decisions have to be made. This would be the case where energy sector plans affect an overall social and economic development plan and vice versa. Such interaction is frequently handled by meetings and committees. King has argued that this manual interface is far more desirable than an automatic linkage embedded in the model.  - 33 - He has also expressed the view that this modularity would be useful for progressively updating the model in a spiral fashion. Thus, each module could adjust to the need and ability of the agency responsible for its implementation, without disrupting the overall system. Just this approach is being followed in the new Morocco model, in which the LOTUS environment makes adherence to standard data protocols very easy (by definition of a common output range format used to transfer information). Staff training is, of course, also facilitated by a modular approach, in that training is not delayed towards near the end of the project, and can begin early in a technical assistance effort. Simulation Versus Optimization Much of what has been written about the relative merits of optimization and simulation approaches may be misleading. For example, three arguments are often advanced for avoiding the use of optimizing models for the overall energy sector. First are the difficulties in specifying objective functions, since least cost (and its discounted analogs in dynamic specifications) may not in fact reflect the actual decision process. However, even a simulation model must implicitly make such judgements. For example, in a network simulation of the electric or refinery sectors, certain rules inust be specified by which the allocations among fuels must be made. For example RESGEN assumes that electric plants are dispatched into the load duration curve in merit order, which produces identical results to those of an LP dispatch model. A second argument is that optimization tends to be expensive in terms of computer hardware, run time and software requirements. However, in today's microcomputer environment these are no longer valid objections, especially for linear programs (although for non-linear and mixed-integer programming the argument is perhaps still valid). The problem is more that the historical evolution of LP energy models in the U.S./European tradition has drifted toward larger and more complex LP's in which the real issue is the ability to interpret models having thousands of variables and constraints. A third argument is that optimization models tend to cut off debate because they present only one optimal solution, whereas simulation models permit a range of alternatives to be suggested. But this argument again is more a matter of how LP's have been misused, rather th1tn any intrinsic problem. The "what if" type simulation mode can just as well be applied to an LP in which the sensitivity of the solution can be examined as a function  - 34 - of a wide range of data assumptions (or for that matter alternative objective functions). Indeed, there is no reason why LP models might not be run in a highly constrained mode with respect to many of the key policy decisions (e.g. the rate at which a domestic natural gas resource should be exploited) in a what if simulation fashion, leaving the LP to make choices only in those subsectors to which it is well suited (e.g. the petroleum subsector). LP models also yield shadow (or dual) prices that are useful in economic analysis more readily than simulation models. The Macroeconomic Model The adage about what is worth doing is worth doing well, is particularly applicable to macroeconomic modelling efforts. This brief review suggests that there may be only limited value in simplistic macromodels that serve solely as drivers to the demand side of an energy model: exogenously specified GDP growth scenarios (an official government projection and one or two perturbations) would serve just as well. On the other hand, models of the sophistication of SIAM-2 are available only in a few countries, and the effort to develop a model of such detail cannot be justified as part of most technical assistance projects in the energy sector. In practice, then, in most cases a compromise must be found between these two extremes. Given the importance we attach in the smaller developing countries to project-level detail and the preparation of medium-term investment plans for the sector, it is likely that the first priority should be for a macroeconomic accountiilg framework of the type developed for Sri Lanka. This will at the very minimum permit a view of the energy sector consistent with official government projections, and permit the evaluation of the impact of the energy sector on the external accounts in a manner consistent with Central bank/IMF presentations. Although such a framework does not allow exploration of such issues as longer term structural adjustments, it does provide a means of forcing consistency between macroeconomic, sectoral, and subsectoral planning assumptions. As a second step, a model of the type represented by the Jordan and Tunisia models would be indicated. However, these should be designed in such a way as to exploit the advantages of being a component of a model system (once again we point to the advantages, say, of linking the Jordan macromodel to an energy sector model to provide consistency between electric sector investment and electricity demands). Moreover, if they are to be used in conjunction with perturbations of official government projections, they must  - 35 - be capable of replicating the latter to some reasonable degree. Energy taxes are important sources of revenue to developing country governments, both in countries that have significant domestic production as well as those that are purely oil-importers. The size of certain subsidy and tax flows are such that sharp changes in their magnitude--occasioned, for example, by a drop in the world oil price, or by a dramatic revision of the domestic tariff structure--may have a substantial impact on modelling of the budget deficit and the allocation of investment flows. It follows that an explicit representation of the government budget will be necessary to analyze the impact of alternative tarification strategies. The assumptions made about the structure of capital markets are critical to the outcome of policy simulations on such matters. The Energy Sector Representation To be suitable for analysis of most of the short- to medium-term issues faced by typical national energy planning bodies and the international financial agencies, the minimum standards required of the energy sector model can be reiterated as follows: 1. The model must have the capability to deal with project specific costs and energy impacts. 2. The model must have an adequate technical representation of the electric sector. At a minimum, specific plants (or groups of plants of roughly similar vintage and fuel type) should be dispatched into an annual load-duration curve with explicit representation of forced and scheduled outage rates. 3. The model should present the energy balance in the conventional form, and have sufficient flexibility to provide intermediate outputs for purposes of model and data validation. 4. Under no circumstances should the model have any data hard-wired into a compiled code. 5. The level of disaggregation must be sufficient to identify institution- specific transactions, and be consistent with the administrative structures used for pricing (e.g. if there exists a special subsidy for cement industry fuel oil consumption, then cement industry fuel consumption should be separately identified).  - 36 - 6. In most cases, the model must provide for a framework to examine the substitutions between commercial and non-commercial fuels (especially between fuelwood and petroleum products). 7. If the model is a compiled program, then it must provide for direct data exchange with spreadsheets. 8. The major unit operations of refineries must either be explicitly included in the energy model, or the energy model should interact with a detailed refinery/petroleum sector linear program. 9. The model should generate investment, debt service and trade data in a form compatible with overall macroeconomic models.  - 37 - REFERENCES Blitzer, Charles, Energy-Economy Interactions in Jordan, MIT Energy Laboratory Working Paper No. MIT-EL 82-038WP, June 1982. Dinh, H., Oil and Gas Policies in Tunisia: A Macroeconomic Analysis, World Bank Staff Working Paper No. 674, Washington, D.C. 1984. King, K., "Modelling the Energy Sector in Developing Countries," paper presented at the International Conference on Energy Planning in Bangladesh, Nov. 1985. Meier, Peter, Energy Planning for Developing Countries: An Introduction to Quantitative Methods, Westview Press Boulder, Colorado, 1986. Munasinghe, Mohan, "Practical Application of Integrated Natural Energy Planning (INEP) Using Microcomputers," Natural Resources Forum, Vol. 10, Febtuary 1986, pp. 17-38; also available as Reprint No. 374, The World Bank, Washington, D.C. Oak Ridge National Laboratory, The Liberian Macroeconomy and Simulation of Sectoral Energy Demand, Report No. ORNL/TM-9065, June 1984. Tokyo Energy Analysis Group, ENERPLAN User Manual, Report to the UNDP, Sept. 1985. Wilbanks, Thomas J., "Lessons from the National Energy Planning Experience in Developing Countries," The Eniergy Journal, Vol. 8 (special LDC issue), July 1987, pp. 169-182.  CHAPTER 3 SRI LANKA CASE STUDY 3.1 OVERVIEW OF THE MACROECONOMY AND ENERGY SITUATION Sri Lanka, situated in the Indian Ocean, is a small island of 65,610 sq. kilometers in area which includes 959 sq. kilometers of inland water. The land area is compact with a central hilly region. The rivers which spring up from this region (the most important of them is the Mahaweli Ganga), play an important economic role in providing hydroelectric power. Another major indigenous source of energy are the forests which have, however, declined from about 44% of total land area in 1955 to about 24% in 1980 mainly due to agricultural conversion and fuelwood use. Sri Lanka is essentially an agricultural economy dualistic in nature. It consists of an export-oriented plantation sector with primary crops of tea, rubber and coconut; and the domestic or rural sector with the major crop being paddy and other food crops--essentially for local consumption. Tea, which is one of the major exports, shows a steady decline, both as a percentage of total export value as well as volume, from 55.3% in 1977, to 27.2% in 1986. Rubber and coconut also show considerable declines; from 21.6% in 1970 to 7.7% in 1986 for rubber, and 11.7% in 1970 to 4.7% in 1986 for coconut. The policy of nationalization of the plantation sector in 1976 had adverse effects on production in the ensuing years, and the worst hit were the tea plantations. On the other hand, minor exports consisting of coconut by-products, spices, minor agricultural crops, precious and semi-precious stones, manufactured goods, minerals, and petroleum re-exports, had increased from 11.7% to 60.4% in value of total exports, between 1970 and 1986. Recent Economic Trends We briefly review below, as background information, the Sri Lankan economy--starting shortly before the first oil crisis of 1972-73. From 1970 to 1977, Sri Lanka was essentially a closed economy, with strict exchange control regulations and import controls. This period was marked by low investments and low growth patterns; GDP grew at only 2.9% per aniium on average; which was well below both the economy's previous performance of 4.4.% per annum in the 1960's and its inherent potential. One of the hardest hit  - 39 - sectors was the manufacturing industry--due to the sluggish economy and the poor investment climate, investments averaged less than 16% of GDP during this period. The oil price increase of 1973 had made petroleum, which supplies a third of the country's primary energy demands, a significant factor in the balance of payments. But strict import control measures and slow economic growth entsured that the burden of oil imports was manageable during the following years. At the same time the addition of new hydro capacity resulted in electricity supply exceeding demand, and thus helped to curtail the demand for oil. This period was marked by economic stagnation essentially due to the policies pursued. With the liberalization of the economy in 1977, and the adoption of a market oriented development strategy, Sri Lanka's economy has shown considerable change. Most significant has been the ability to break the low investment and low growth pattern of the 1970-1977 period. Investments have grown from 14% of GDP in 1977 to an average of 28% in 1980-86, with growth being shared by both public and private sectors. However, this growth has not been accompanied by any substantial increase in national savings, implying little shift in domestic resources from consumption to investment. Most of the increase in investment had to be financed by drawing down international reserves, and by resorting to commercial borrowing. The real rate of growth since 1977 has averaged about 5.8%, which is all the more significant since this was achieved in a period of considerable international economic turmoil with high international inflation, doubling of oil prices and recession in the developed countries (which reduced demand for exports and to somre extent constrained aid flows). Further, the increased investment programs were accompanied by gains in employment. Import liberalization and the decontrol of most prices provided an immediate boost to the manufacturing industry which was the hardest hit sector in the 1970-77 period. Growth in this sector jumped to 7.8% in 1978 and has thereafter averaged 5.2% from 1978-1985. The somewhat reduced growth in manufacturing since 1978 has been due to slow growth in capacity utilization and the low level of investment in the manufacturing sector due to easier, quicker and higher returns in other sectors like tourism, trade, and real estate. Economic Growth and Balance of Payments The bold economic reforms during the late 1970's resulted in an initially high growth rate of 8.25% in 1978. Against a background of increasing and unsustainable balance of payments and budgetary deficits, the initially high  - 40 - growth rate has shown a decline over time. During 1970-77 real GDP had grown at oiily 2.9% per annum; during 1977-80 growth averaged 6.1% while in 1980-85 growth had fallen to an average of 5.2%. The crux of these problems has been the high level of capital formation in relation to national savings and the slow growth of exports in relation to the import requirements. Between 1980 and 1983 capital formation remained at a high 30% of GDP largely because half of it was financed by foreign savings. The World Bank has estimated the decline of Sri Lanka's terms of trade to be over 30% between 1977 and 1981. This has been primarily due to the rapid increase in import prices of petroleum products in the international markets and decline of export prices in the tree crop sector. Though oil prices dropped in the international markets in 1982, it had no impact on the economy since the increases in the exchange rate offset the price decreases. Due to a combination of slow export volume growth relative to imports and a large deterioration in the terms of trade, total imports grew to twice the value of exports by 1980. The balance of payments showed a deficit of $579 million in 1986. The severe deterioration in the balance of payments in 1980 had reduced Sri Lanka's net international reserves by $220 million to $38 million. The current account balance deteriorated from a positive 2.4% of GDP in 1977 to a record deficit of 19.8% in 1980, subsequently levelling off at an average 10.9% deficit during 1981-86. The most serious domestic manifestation of Sri Lanka's financial difficulties has been higlh inflation. Inflation, as measured on all major price indices, accelerated sharply between 1978 and 1980, before subsiding between 1980 and 1981. The rate of inflation as measured by the Colombo cost of living index was 14%, on average in 1983 (over 1982), but reached 23% when measured between February 1983 and February 1984. It has declined somewhat since then, averaging 4.7% between 1984 and 1986. The crux of Sri Lanka's macro-economic problems remain: the high capital formation in relation to national savings; slow growth of exports relative to increasing imports, resulting in severe balance of payment problems; and depletion of external reserves. Energy-Macroeconomic Relationships During the Recent Past During the past decade international oil prices have twice risen sharply--in 1973-74 and 1979-80. The overall increase was from under US$2 per barrel in the early 1970's to US$34 per barrel in 1980. Since then, prices have declined  - 41 - to below US$20 per barrel by 1986/87. During both oil shocks, the economic growth rate declined in industrial as well as developing countries. The oil importing less developed countries (LDCs) continued to be highly vulnerable to developments in the world economy, not only because of high oil import costs, but also due to shrinkage of export markets, official aid flows, access to international credit etc. Each downturn in the developed countries had corresponding adverse effects on the LDCs. The impact of these two oil upheavals on the Sri Lanka economy is analyzed below. The impact of the first oil shock of 1973-74 was somewhat cushioned in the Sri Lanka economy due to a combination of factors. The strict exchange control regulations coupled with slow economic growth ensured that the burden of oil imports were manageable during the next few years. Though the oil import bill was a significant factor in the balance of payments, due to the prevailing economic climate oil imports fell. The turning points in the economy occurred in 1970 and 1977 with the change of governments and the economic policies adopted by them. From 1970-1977 the economy was highly protected from external shocks due to the closed economic policies that prevailed. As a result the first oil shock had no significant impact on GDP growth. Tables 3.1 and 3.2 give the key years in which significant chaniges took place in the consumption trends of petroleum products and the petroleum import bill. Table 3.1 shows that from 1973-77 the average growth rate of petroleum demand fell by 3.3%. This could be interpreted as a general response to doubling of petroleum prices, in a difficult economic period. Table 3.1 Consumption Trends for Major Petroleum Products a/ Average Annual Growth Product Consumption (1000 Tons) Rates (%) 1970 1977 1990 1983 1988 1970-77 1977-80 1980-86 Gasoline 148.4 111.6 107.7 117.5 130.6 -4.0 -1.2 21.3 Kerosene 272.6 213.1 188.7 169.1 154.2 -3.6 -4.0 -18.3 Autodiesel 264.5 261.4 399.5 464.3 487.3 4.0 15.2 22.0 Industrial Diesel 87.9 46.3 61.0 296.9 36.0 -8.8 9.6 -- Fuel Oil 208.8 134.7 247.2 253.1 129.0 -6.1 22.4 -- Power Sector Consumption 133.0 7.0 58.6 -- -- -34.3 102.9 -- Fuel Oil 133.0 7.0 45.0 -- -- -- -- -- Diesel -- -- 13.6 -- -- -- -- -- a/ Excludes Re-export and Bunker sales. Source: Coylon Petroleum Corporation  - 42 - Table 3.2 Petroleum Import Bill (USS Millons) 1970 1977 1980 1983 1986 A. Petroleum Imports 10 180 489 438 197 B. Petroleum Re-exports 5 84 181 106 84 C. Net Petroleum Imports 6 98 308 382 113 0. Non-Petroleum Exports 337 887 884 953 1132 E. C as Percentage of D 1.6 14.2 35.6 38.0 10.0 USS = Rs. 6.2 16.8 18.0 26.0 28.5 Source: Sri Lanka Customs The second oil shock of 1979-80 had much more severe repercussions than the first. Table 3.3 gives a clear picture of the impact of these two oil shocks. The periods are demarcated as shown since there were significant economic reactions and repercussions. First is 1970-73, the pre-embargo period; then the 1973-77 adjustment period for the first oil shock. 1977-80 includes tlle period of liberalization of the economy and the second oil shock; 1980-82 gives the short-run adjustment period after the second oil shock; and 1982-86 the long-run adjustment period. With the liberalization of the economy in 1977, the growth rate of GDP from 1977-80 averaged 6.8%. The rapidly improving economic growth during this period resulted in an accelerated demand for electricity (which previously, during 1973-77, grew at an average rate of 4.7%), due to the low tariff structure and untapped potential markets. The demand was met totally by hydro-power. During 1977-80, without taking into consideration the power cuts that prevailed, electricity growth averaged 10.2%; this had a significant impact on the economy since the supply had to be supplemented by thermal generation. During the same period, demand for petroleum products grew at 7.2% compared to -5.2% in 1973-77. In the short-run adjustment period (1980-82) after the second oil shock, demand growth for both electricity and petroleum stayed high, at an annual 10.1% and 14.7% respectively, as the economy  - 43 - Table 3.3 Evolution of GDP, Energy Demand and Prices for the Period 1970-1986 Adjustment Period Adjustment Period for the First Intermediate for the Second Pre-Embargo Oil Shock Period Oil Shock 1970-73 1973-77 1977-80 1980-82 1982-86 GDP Growth Rate 1/ 2.2X 3.4X 8.8! 5.4% 4.8! Growth Rate of Total Energy Demand 2/ -- 2.8X 2.3% 2.S% -- Electricity Demand 3/ 7.8e 4.7% 10.2X 10.1% 7.3X Non-Electricity Demand 4/ - Demand for Petroleum Production 0.9% -6.2% 7.2% 14.7% -8.7% - Demand for non-fuelwood energy 6/ -- 8.2X 0.4! -0.2! -- Changes in Energy Prices - Electricity Prices 6/ -3.12% -3.64% 32.6% 37.8! -- - Petroleum Product 11.2! 14.2! 30.0! -3.0% 7.1% Prices 7/ 1/ Annual average rate of change of real GDP for each specified period. 2/ Annual average rate of change of total energy demand (Commercial and non-commercial energy) 3/ Annual average growth rate of electricity deand (either in orig. unit or converted unit). 4/ Annual average growth rate of demand for each specified period. 6/ Includes all non-commercial energy sources such as firewood, crop residues, cow dung, etc., converted into a common unit (MTOE). S/ Changes in weighted average prices (or price index of all commercial energy sources. 7/ Weighted average of petroleum product prices.  - 44 - continued to expand. Since then growth rates have dropped significantly, particularly for petroleum (-6.7% in 1982-86) where domestic rupee prices followed the international price decline. The marginal increases in the demand for commercial energy in Sri Lanka up to 1985 were directly linked to the size of petroleum imports, regardless of whether these increases took the form of higher electricity consumption or direct consumption of petroleum products. Thus after 1977, the combination of increased consumption and a doubling of oil prices resulted in a rapidly growing oil import bill. By 1981, the net oil import bill more than tripled and the proportion of export earnings devoted to importing oil rose from 15% to 39%. As mentioned earlier, the first international oil crisis had no severe repercussions on the Sri Lanka economy due to the stagnant state of the domestic economy. The second oil crisis, however, not only effected the balance of payments, but also the terms of trade, which declined by over 30% between 1977 and 1980. This was due primarily to the increase in import prices of oil, and the decline in export prices of the tree crop sector. Though investments rose from 14% to 30% of GDP from 1977-1980, this was not accompanied by an increase in domestic savings. Increased investments have been financed by drawing down international reserves and commercial borrowings. As a result, Table 3.2 shows that up to 1977 the cost of oil imports was not a source of serious concern to Sri Lanka. Additions to international reserves declined from an inflow of 3% of GDP in 1978 to an outflow of 6% of GDP in 1980, while net use of commercial financing rose from -1% to 6% of GDP for the same period. The rupee, which was pegged to the dollar (floating exchange rate), depreciated in value from Rs. 15.608 in 1978 to Rs. 23.529 in 1983. The manifestation of these financial difficulties, along with international price increases, has been a high inflation rate. Before reducing significantly to an average of 4.7% between 1984 and 1986, inflation as measured on all major price indices accelerated sharply between 1978 and 1980. The recent decline in world oil prices had little effect on energy demand since this dollar price decrease was offset by the exchange rate fluctuations (with the rupee declining in value against the dollar), thus keeping domestic oil product prices relatively stable. Principal Energy Sources Sri Lanka has two major indigenous sources of energy--hydro electric power and fuelwood. The country's potential hydro electric power is estimated to be in the region of 2300 MW, of which 1122 MW has already been  - 45 - developed, with 533 MW of this being contributed by the recent Mahaweli scheme. By the end of the current decade, over 50% of the potential capabilities would have been realized. Hydro electric power is used primarily to meet base load energy generation, while gas turbines operate in the peaking mode and also provide spinning reserve. This pattern is expected to continue for at least the next five years. Though reliable data on fuelwood supply is difficult to obtain, the estimates clearly indicate its precarious and unsustainable nature. Over the past two decades, incremental wood production--from the natural regeneration of forests, agricultural residues and rubber replanting etc.--has fallen far behind consurmption, and today accounts for less than half the estimated annual consumption of around 5 million tons. The balance of wood supply has come mainly from the denudation of Sri Lanka's natural forest cover which has declined from about 44% of total land area in 1956 to about 24% by 1982. The balance of the major domestic demand is met by oil imports. The crude oil imported is refined into other petroleum products which are either used in the domestic market or re-exported. Table 3.4 shows the main types of oil imports which are crude oil, kerosene, and auto diesel. Table 3.5 gives the consumiption of electricity and petroleum products between 1975-1986. Figure 3.1 shows the typical Sri Lanka energy supply pattern (1983). Fuelwood (67%) dominates gross energy supply, with oil at 24%, thermal-based electricity at 4% and hydro electricity supplying 5%. Oil is the most important source in supplying useful energy or net supply (46%), with fuelwood at 38% and electricity at 16% (supplied by both hydro and oil) (useful energy was calculated by end-use efficiency estimates of the different energy using forms). Since 1978, there was a substantial increase in the demand for all forms of commercial energy, which rose in aggregate at 8.8% per annum in the 1978-80 period, as opposed to a slight decline in the preceding 7 years. Petroleum consumption has grown even more rapidly since the increase in demand for electricity in the past (1977 period) and had to be met by increased thermal generation. Table 3.6 shows that for 1983 oil for thermal generation in the form of diesel and fuel oil has increased very rapidly.  Table 3.4 Imports of Oil and Petroleum Products 1976 1976 1977 1978 1979 1980 1981 1902 1983 1984 1965 1966 af Valu, (US1 Millions) Crud. Oil 132.6 136.9 139.2 143.7 201.6 436.8 448.4 488.9 332.9 373.4 335.2 175.1 Gasoline -- -- 0.3 0.6 1.2 -- -- -- 4.4 4.4 -- -- Avtur 0.6 -- 3.3 8.2 22.3 22.5 16.4 1.8 3.2 -- -- -- Kerosone -- 1.3 4.9 3.6 12.1 -- -- 16.2 15.8 2.5 11.6 3.4 Automctivo Diesel -- 1.1 2.8 10.7 61.9 14.8 36.5 58.6 102.0 30.7 81.7 18.7 Other 8.8 5.8 6.4 8.3 8.6 20.3 14.9 11.4 10.2 13.5 -- -- Total 141.7 144.1 156.9 175.2 297.7 493.4 616.2 573.9 468.6 420.1 378.4 197.2 Volume ('000 Tons) Crude Oil 1,464.6 1,447.1 1,529.6 1,443.9 1,444.0 1,861.1 1,710.5 1,940.5 1,492.0 1,733.8 1,867.6 1,639.1 Gasoline -- -- 2.2 3.7 6.6 -- -- -- 15.0 -- -- -- Avtur 4.2 -- 16.8 66.7 66.3 58.4 46.0 5.4 10.9 -- -- -- Kerosene -- 9.8 32.2 25.4 41.9 -- -- 48.4 S5.8 8.8 44.6 15.4 Automotive Diesel -- 9.2 26.7 82.7 198.6 42.6 110.9 183.9 405.9 131.2 13S.8 131.6 n/ Provisional Soiare: Ceylon Petroleum Corporation  Table 3.5 Consumption of Petroleum and Eectricity Products 1975 1976 1977 1978 1979 1960 1981 1982 1983 1984 19s6 1966 (Est;-at.) Petroleum Products (Metric Tons) LPG C/ 582 807 3,108 2,432 6,404 7,110 6,683 8,197 7,1S0 8,178 11,278 16,476 Gasoline 96,067 101,065 111,491 129,994 116,146 107,691 109,028 114,217 117,477 118,831 121,678 130,624 Kerosene 209,764 206,598 212,593 244,832 229,918 188,268 166,266 174,098 159,146 150,926 153,665 164,182 Diesel, Automotive 245,516 267,667 261,9868 30,792 349,404 397,710 420,912 464,694 464,268 495,721 488,497 467,832 Diesel, Marine b/ 5,232 5,183 5,497 6,869 3,726 3,027 2,65s 5,616 7,829 3,902 778 513 Diesel, Industrial 37,314 38,e63 46,245 82,015 84,188 68,963 106,000 148,121 296,886 84,188 37,123 36,997 Furnace Oil, Domstic 143,664 126,578 135,630 162,6S6 183,639 269,731 240,326 247,138 263,098 218,913 142,783 129,048 Furnace Oil, Marine 20,108 20,088 18,762 21,233 16,099 12,887 22,884 26,974 26,881 9,397 85 706 Avtur 13,571 8,614 16,499 6,749 8,169 22,843 80,967 81,416 34,262 43,827 54,186 45,742 Lubricants c/ 15,848 19,696 14,933 17,346 18,899 21,312 20,430 20,614 20,715 21,039 20,635 20,516 Bitumen 22,444 26,023 26,162 26,190 24,286 10,269 16,477 21,116 24,423 33,099 32,661 43,690 Naphtha -- -- -- -- -- 38,642 66,063 98,021 75,044 78,431 7,486 N.A. Electricity Products (GUH): (Actual) Domestic 84.9 95.2 106.5 119.2 153.2 190.8 216.6 258.3 804.8 316.9 346.4 877.0 Industrial 622.6 616.3 619.4 692.0 629.9 626.6 647.6 739.2 762.0 790.9 8650.4 922.0 Commercial 122.6 137.4 147.9 158.9 208.0 223.2 219.9 262.6 291.8 299.6 380.0 877.0 Authorities 222.2 287.3 262.8 275.9 296.3 338.5 360.6 417.5 433.0 457.7 502.1 546.0 Street Lighting d/ 13.0 13.5 14.0 15.0 16.0 16.5 8.5 8.6 9.0 11.4 11.8 13.0 Total 965.2 999.7 1.040.8 1,161.0 1,298.3 1,391.6 1,503.1 1,666.1 1,790.6 1,676.5 2,060.7 2,2a4.0 a/ Since March 1977, reflects transfers to Colombo Gas A Water Cowpany. b/ Includes Marine Gas Oil, Marine Diosol Oil and Heavy Diesel. c/ Other than marine and aviation lubricants. d/ Estimated. Sourc Ceylon Petroleum Corporation Ceylon Electricity Board  GROSS ENERGY SUPPLY (1963) USERJL OR NEO SPPLY (1983) Electricity (Hyko + Theonal) EbddV 5S \ ~~~~67S TheimiV EIinctfidty ~~~~~~~~~~~~~~~~~~~~~~~~~~~00 Figure 3.1. Sri LankaEnergySupplyPattOil ern ( 195 (Nm.oc/ Figure 3.1. Sri Lanka Energy Supply Pattern (1 983)  - 49 - Table 3.6 Basc Energy Data from the Energy Balance Tables ('000 TOE) Pre-Em1ergo AdJustment Interadi ate AdJustment Period Period Period Second Oil Shock 1973 1974 1979 1968 Electricity Cons. 74 77 112 154 Imports DIell 0 10 208 426 Fuel Oil 0 0 0 0 Crude Oil 180 157 1497 1537 Fuel tor Therml 0s. Dilel 1.4 0.5 1.6 264 Fuel Oil 19 4.2 18 49.0 Non-conventional Energy Sources Sri Lanka has access to a variety of economically viable, non-conventional renewable energy resources--solar, wind, biomass and mini-hydro--which could be utilized to meet some of the couuitry's energy demand in the medium term. Solar Energy: Sri Lanka's location assures it of a relatively high and uniform level of insolation which could be harnessed for both water heating and crop drying. A very few households currently use hot water; the main market for these heaters will initially lie in the commercial and tourist sectors. The use of solar energy for crop drying is an important alternative to be developed since tea and other crop processing industries consume over one million tons of fuelwood (20% of total fuelwood consumption according to World Bank figures) per year.  - 50 - Mini-hydro: While emphasis in Sri Lanka has been on large hydro power schemes, about 10 MW of small schemes (5 KV to 250 KV range) have been operating in the tea estates of the central region since 1925. However, these have been abandoned because of the availability of cheap and reliable electricity from the national grid. In addition to the rehabilitation of these existing plants, potential sites for mini-hydro schemes exist in the central hilly areas and in irrigation systems in the north central part of the island. Other renewables: Apart from solar and mini-hydro, there are a number of promising renewable energy applications. These include generation of biogas from animal wastes, producer gas from rice husks and coir briquettes, wind energy for water pumping and electricity generation in isolated areas. At the outset, it was roughly estimated that all of these will contribute only about 2% of total energy requirements in 1990, and less than 5% by the year 2000. More detailed calculations regarding the role of new and renewable sources was made later, through the INEP computer modelling framework, but the orders of magnitude involved were not greatly effected. Nevertheless even this modest contribution will be useful since it will help to displace expensive imported oil and scarce domestic fuelwood at the margin. Energy Demand by Source Developing a detailed picture of sectoral energy consumption patterns in Sri Lanka has been hampered by the absence of reliable data at a sufficiently disaggregated level even for the commercial fuels; the inclusion of fuelwood increases the margin of error because total consumption figures vary widely according to the source. The estimate of fuelwood consumption for 1980 varies from 3.8 to 5.3 million tons. A major effort has been recently launched to identify household consumption at a disaggregate level i.e. Urban, Suburban, Rural, Estate, and Wet Zone and Dry Zone. Nevertheless preliminary analysis suggests a typical energy consumption pattern illustrated in Figure 3.2, where for commercial energy consumption (oil and electricity), transport (54%) and industry & commerce (26%) are the main users, with domestic demand amounting to 17% and other use 3%. These shares change dramatically when fuelwood is included in the analysis; domestic consumption then accounts for the major share of 68%, industry & commerce 18%, transport 13% and others 1%.  COMMERCLAL ENERGY ALL ENERGY ELECTRICITY OL E0R83) ELECTRICIlY, OIL & RJELWOOD (1983) Domestks Transport C"l hndushyv & /I Commenrcial Industry&1 Transport Commercial Figure 3.2. Sri Lanka Energy Cosumption Pattern (1983)  - 52 - 1. Fuelwood The absence of reliable data on fuelwood makes it difficult to identify the pattern in total energy demand and correlate it with economic variables. Estimates indicate no major shift from non-commercial to commercial energy sources over the last two decades. In fact the data points more towards the increase in the consumption of fuelwood given increased prices of commercial fuel like kerosene used for domestic cooking and lighting. 2. Commercial Energy The consumption trends of commercial energy have closely reflected the overall performance of the economy. Between 1970 and 1977, the near stagnation of economic activity and growth in real incomes was reflected in a largely constant level of commercial energy demand, with rising electricity consumption being offset by an average 3.3% per annum fall in the demand for petroleum products. With the liberalization of the economy, demand for all forms of commercial energy has shown a sharp increase, showing an average growth of 8.8% per annum. 3. Petroleum Products From 1970 to 1977, the consumption of all petroleum products fell, with the exception of LPG (used mainly for cooking) and auto diesel. This has been in part due to the virtual elimination of petroleum based electricity generation that followed the addition of new hydro electric capacity, and in part due to the response of demand to the doubling of prices in a difficult economic period. Since 1977, most products have shown increased growth in an improved economic climate. The oil demand for thermal power generation has shown more than average increases for diesel and fuel oil, dropping only with tlle recent addition of Mahaweli hydro capacity. Only the consumption of gasoline and kerosene has continued to fall after the post-1977 period. This is a reflection of the price sensitivity of demand for both these goods. 4. Electricity In contrast to petroleum consumption, electricity demand continued to grow at an average annual rate of 7% during the 1970-77 period. This has been due to: (a) the attractiveness of electricity as a source of energy since there were no tariff increases during this period, while petrodeum prices rose by more than 50% with the world energy crisis, and (b) the existence of a potential untapped market for electricity. Consequently, the demand for  - 53 - electricity has shown less responsiveness to price and changes in the overall economic performance than petroleum products. Since 1977 electricity demand accelerated--electricity sales grew at an average of 9.6% annually until 1980. This figure does not include the 3% reduction in demand due to prolonged power cuts in 1980 which was enforced due to the high growth rate of consumption and an unexpected and severe drought. The liberalized import of electric appliances for domestic use has also contributed to the increase in electricity consumption. The impact on the economy has been significant since this demand had to be met with increased import of petroleum products for thermal power generation. Energy Demand by Sector The energy demand sectors have been demarcated into four broad groups: 1. Industry/Agro-industry; 2. Transport; 3. Household/Agriculture; and 4. Government, Commercial and Others. 1. Industry/Agro Industry The industrial sector accounts for more than 50% of electricity consumption, about 1/3 of petroleum consumption and also 1/4 of the fuelwood used in the country. Fuelwood is an important source of energy for Agro-industry in the rural areas, (e.g. tree crop processing). Eighteen large industrial organizations in the public sector account for 35% of total electricity sales; while ten large private companies account for over half the sector's petroleum consumption. 2. Transport This sector depends entirely on petroleum products for energy, and transport accounts for more than half the total demand for petroleum. Diesel is the predominant transport fuel (75% of the sales), which reflects the extensive public transport network, and also the policy of pricing diesel well below gasoline. As a result of this price differential, the proportion of diesel car registration rose from 14% in 1978 to 35% in 1980. To reverse this trend, the government in 1980/81 raised the price of diesel to 60% of the price of gasoline, and increased the license fees of diesel cars to 3 times that of gasoline cars. Since then, the growth of auto-diesel consumption has been significantly reduced as seen in Table 3.5. 3. Household/Agriculture While per capita energy consumption in Sri Lanka is low by international standards and most households use energy for only cooking and lighting,  - 54 - the energy requirements in the household sector account for nearly half the country's primary energy consumption. The bulk of household energy requirements is met by fuelwood, which will continue to be the predominant source for the next decade. Only 13% of the households have access to electricity, and the consumption of kerosene, which is mainly used for lighting, has been declining over the past decade in response to higher prices. The demand for energy in this sector will continue to grow with increase in the population and rising standards of living. 4. Government, Commercial & Others This sector consumes about 1% of the total demand for energy in the economy. 3.2 ENERGY POLICY COORDINATING FRAMEWORK In late 1982, the President of Sri Lanka appointed a Senior Energy Advisor (SEA) to accomplish the following broad tasks: (a) Establish an effective organizational framework for overall energy coordination and integrated national energy policy analysis and planning. (b) Create a database, analytical procedures and systems to support activities under task (a) above. (c) Train and develop a team of energy specialists and other staff required to support activities under tasks (a) and (b) above. These objectives were largely achieved during the subsequent three years. In the process of carrying out this longer-run overall program, a number of urgently needed (and beneficial) specific activities were also carried out, such as the preparation of the national energy strategy (NES), the national energy demand management and conservation program (NEDMCP), and the national fuelwood conservation program (NFCP). In brief, the SEA's work encompassed regular policy advice to the President, overall staff training and institution building efforts, and direct advice and support to line agencies on energy management and policy planning aspects.  - 55 - Figure 3.3 shows the energy sector institutional framework in mid-1982. One major drawback was the large and varied number of Ministries and line agencies involved in the different energy subsectors, with inadequate coordination among them: (a) Electricity (Ministry of Power and Energy/Ceylon Electricity Board - CEB; Maliaweli Ministry/Mahaweli Authority - MA); (b) Petroleum (Ministry of Industries and Scientific Affairs/Ceylon Petroleum Corporation - CPC); (c) Fuelwood (Ministry of Lands and Land Development/Forestry Department - FD); and (d) Overall science and energy policy, and research and development (Natural Resources, Energy and Science Authority - NARESA). In general, energy policy up to 1982 tended to be subsector focussed, ad- X hoc, and supply oriented--with insufficient emphasis on demand management. In November 1982, President Jayewardene approved the setting up of an Energy Coordinating Team (ECT) to help remedy this situation. The initial X organization of the Energy Coordination Framework is shown in Figure 3.4. A special effort was made to ensure policy coherence in the commercial energy subsectors (oil and electricity), through improved links between ECT, CEB, CPC and MA. A similar strengthening of coordination with the fuelwood sector through the FD was also pursued. The rationale underlying the ECT concept was that the new framework, in the first instance, was not intended to be another bureaucracy that would seek to control the energy sector. The objective was to coordinate and facilitate the work of relevant Ministries and existing line agencies, prevent duplication of effort and policy conflicts, supplement weak or neglected areas in the energy sector, and provide direct advisory inputs to the President. In any case, major structural changes in the organization of the energy sector could not be undertaken in the short-run. The ECT framework was the most practical method of initiating and carrying out urgently needed tasks in the sector, without undue disruption of existing activities. As the ECT matured and experience was gained in energy coordination, some further organizational changes in the energy sector were gradually inmplemented, as described in the next section. ECT consisted primarily of 3 coordinating task forces (CTF's) that covered the following areas: (a) Energy Planning and Policy Analysis (EPPAN); (b) Energy Efficiency, Demand Management and Conservation (EDMAC); and (c) New, Renewable and Rural Sources of Energy (NERSE).  The Preskdent Ministfy h | Ministry o Ministry of Ministry of Other NARESA Power & Industries & Mahaweli Lands & Land Ministrbs Energy Scientific Affairs Development Development Com nittee CEB C i MA/ MDB _FD 2 ments O CEB = Ceylon Electricity Board NARESA = Natural Resources. Energy & Science Authority CFC - Ceylon Petroleum Corporotion MA/MDB Mahcweli Authority/Mahaoweli Development Board FD - Forest Department Figure 3.3. Sri Lanka Energy Sector Institutional Framework (Mid-1982)  Drect Link Liosan/Regular Conulation _ he Peident CEO - Cevion Electricity Board NARESA - Natural Resources. Energv c S cen Athoie tv tCabinet CPC - Ceyfon Petrolotn Cotpa:falbn Cdn Mdhwli Deve pmen Board FD Forest Depairtmrent p MinItr at Ministry at Ministry Of MWNOIstya te Power & Insrbis Mahawell Lands & Land Mnistes 8 Energy & Science MAlns Development Devepment Depten Celce of the Senkor Ener >sgy Adi3 _ _ ENERGY COORDONATNG TEAM Energ Coordintng Tearni Support Staff Noter Task forces (TF) and working -r__________________II groups are c_awn from relevant Enew Planning Energy Efficincyv Dernand Now,. Renewable & Rural energy and consumning rnWs- 8 POIICV Ardys Mangent & Conservatlon Sources of EnergV tries, governmnent agencies (EPPAN) (EDMAC) I I(NERSE) and prilvte sectr aogaS- zatios TFsandWGsalso llose wAth other coordinate efforts of energvgroupsin Fi eW O R K I N G G R E U P S F e (End-i 952) Figure 3.4. Energy Sector Instltutional Framework (End-1 982)  - 58 - Each coordinating task force (CTF) had as members 6 to 8 senior managers fromi the relevant organizations, who were seconded to help in ECT activities. The CTF's were chaired by the SEA, and initially met once a month. Major policy decisions agreed on at CTF meetings would normally be conveyed to the President through the SEA and if necessary, reach the appropriate line agency through the Cabinet. Agreed policy was also conveyed directly to concerned line agencies, through the relevant members. This facilitated speedy and accurate implementation of policy, and practically eliminated potential conflicts between the ECT and operating energy institutions. X CTF members were also obligated to carry out studies and/or provide data as requested at meetings. Meanwhile, each CTF also had a number of associated working groups (WG's) on specific topics (each responsible to a CTF member). WG's met quite frequently and normally submitted their recommendations to the appropriate CTF for approval and follow-up action. Several important WG's including those on Energy Modelling, Energy Economics, Industrial Energy Conservation, Transport Energy Conservation, Electricity Pricing, Electricity Loss Reduction, and Producer Gas were initially under the direct supervision of the Senior Energy Advisor, in order to build up rapid momentum. However, as the work progressed, the responsibility for the supervision and guidance of many of these WG's was delegated to others, thus relieving the workload of the SEA and building up management capability. While the ECT was originally set up as an inter-ministerial coordinating body, it became apparent that the CTF's would need more support staff to carry out various activities not covered by existing energy bodies (e.g., energy conservation)--thereby freeing the CTF's themselves to focus attention almost exclusively on their primarily policy-oriented role. Thus, during the first 18 months or so, the ECT support staff grew from one coordinating officer, one administrative officer, and one typist, to one experienced administrative official (Ministry assistant secretary for energy), 3 coordinating officers, over a dozen professional staff (some seconded on a part-time basis), two computer systems analysts, three clerks, three typists, and other minor employees. / All the ECT staff were either drawn from the Ministry of Power and Energy (MPE) or seconded there from other bodies. Therefore, the entire ECT gradually began to fall more under the Ministry of Power anid Energy, while retaining its inter-ministerial policy coordinating role. These developments essentially permitted MPE to take a greater responsibility and build more capability in the energy area--an aspect that had been largely neglected prior  - 59 - to 1983, due to the MPE's almost exclusive preoccupation with the CEB and electric power. This evolution is shown in Figure 3.5, which was the organizational structure by about mid-1984. As indicated earlier, the Ceylon Petroleum Corporation (the state oil company) was a part of the Ministry of Industries in 1982. Following strong arguments presented to the President from 1983 onwards, and considerable debate, this key institution was shifted over into the Ministry of Power and Energy in early 1985 (directly under the President). The Lanka Electricity Company (jointly owned by the CEB and Ministry of Local Government and Housing) was created in 1984 to gradually take over independent electricity distribution networks, hitherto operated very inefficiently by the local authorities or municipalities. These structural changes greatly strengtheiied policy coordination in the commercial fuel sector (i.e., electricity and oil products). The management of the ECT was also strengthened by the appointment of Deputy Chairmen for coordinating task forces. In early 1985 an additional secretary (energy) was brought into the MPE to understudy the SEA. When the SEA completed his 3-year assignment in late 1985, the ECT was placed directly under the control of the Secretary, MPE, with the Additional Secretary (Energy), Deputy Chairmen of the CTF's, and the Assistant Secretary (Energy) forming an executive group to help run the ECT. The energy coordinating framework was therefore well established, and functioned effectively. The principal activities of the three Task Forces are given below: 1. Energy Planning and Policy Analysis (EPPAN) EPPAN sought to meaningfully integrate all energy sector activities, and its time perspective was also basically medium-to long-term in nature. One of EPPAN's most important goals was to identify the overall objectives of national energy policy, and then attempt to define a national energy strategy (NES) that met these objectives (see Chapter 8 for details). A comprehensive computerized energy database was also set up. EPPAN reviewed the disaggregate supply and demand projections for oil, electricity and fuelwood, provided by other organizations. 2. Energy Demand Management and Conservation (EDMAC) EDMAC included a number of short- to medium-term activities aimed at increasing energy efficiency and eliminating waste in both the energy supplying and consulming sectors. The National Energy Demand Management and  - 60 - PR?ESIDENT (Minister ot Power & Energy) | CABINET OTHqER ________________ MINIT5RY OF MINISTRIES & POWER & DEPARTMENTS _OFRCE OF ENERGY TH E SENIOR _ __ ENERGY ADVISOR L= C B ECT ECT Executive Group Support Staff COORDINATING TASK FORCES Energy Planning Energy EfficiencyDernand & Policy Analysis Management & Conservation New,R. en & Rural (EPPAN) (EDMAC) Sources of Energy CPC = Ceylon Petroleurn Coparton WG = Working Group CEB - Ceybon Electricity Board LECO - Lanka Electricity Company - Direct Unk ECT Energy Coordinating Team - L-a-- asan/Consultation Flgure 3.5. Energy Sector Institutional Framework (Mrd-1 984)  - 61 - Conservation Program (NEDMCP) focussed on industry, commerce, transport, households, agriculture, and energy supplying institutions (see Chapter 6 for details). Coordinated application of policy tools, including rational energy pricing, was stressed. 3. New, Renewable and Rural Sources of Energy (NERSE) The principal focus of NERSE was to: (i) coordinate R&D activities in this area that are now being carried out by a large number of research organizations and universities; (ii) carry out technical, economic and financial reviews of these technologies, in order to identify the ones which would be most promising from the actual energy viewpoint; and (iii) promote, finance and encourage commercialization of these selected technologies. The highest priority of NERSE was qutickly identified as the National Fuelwood Conservation Program (NFCP), involving the speedy dissemination of improved cooking stoves to about 2.6 million Sri Lankan homes (see Chapter 4 for details). 3.3 HIERARCHICAL INEP MODELLING FRAMEWORK INEP Procedure Energy planning involves a number of analytical steps for which computer models can be useful: energy demand forecasting, supply-demand balancing, impact analysis, and so on. In Sri Lanka, such activities were first carried out at a relatively simple level as energy planning activity commenced, and only later, as data and staff capabilities improved, were more sophisticated techniques introduced. Indeed, it is one of the most important lessons of the energy planning experience of the last decade that the key to any successful modelling activity is to match the level of modelling detail and sophistication to the training and experience of the technical staff. A second criterion for the design of a modelling framework for Sri Lanka was to match it with the institutional structure. The essentially hierarchial nature of the planning process itself, as noted in Chapter 1, therefore  - 62 - demanded a corresponding approach to any modelling effort. Thus the modelling framework had the same three levels as the planning process itself. At the first level were the subsectoral models corresponding to the line agencies such as the CPC and CEB responsible for execution of policy and day-to-day operations. At the second level was the energy-sector-wide integration, the focus of the ECT. Finally, at the third level was the integration of the energy sector with the other sectors of the economy, institutionally the domain of the Ministry of Finance and Planning. The modelling system, whose main elements are shown on Figure 3.6, reflects the institutional structure. The Analytical Approach The process begins with an analysis of the macroeconomy in the first of three modules of the ENMAC macroeconomic accounting framework. The output of this step includes projections of the sectoral decomposition of GDP and non-energy merchandise trade. The energy-macroeconomic accounting framework is presented in more detail in Chapter 6. The RESGEN energy model, described in more detail below in Appendix 3.1 is run next. The demand equations in this model are driven by the GDP estimates passed from the macromodel. Other inputs are passed from the more detailed sectoral models: the refinery configuration and yield coefficients from the refinery optimization niodel, and the estimates of fuelwood demand from the fuelwood model. As part of obtaining the energy supply-demand balance, RESGEN estimates which energy facilities must be built to meet the demand, thereby providing a year-by-year estimate of investment requirements. The basis for the electric sector investments, which account for the bulk of total energy investment, is the generation plan of the CEB. Because different demand scenarios result in changes in the future electric load growth, the analysis includes an examination of the timing of capacity increments. An optimal dispatch model built into RESGEN also provides detailed estimates of CEB fuel consumption as a function of plant availabilities, hydro energy limitations, and the system load factor. The investment outlays, and their concomitant debt service obligations, are passed back to the macromodel together with the oil import bill (and the smaller earnings from bunkers and petroleum product exports). These energy sector transactions are merged with the non-energy transactions in a second module of the macroeconomic model. Finally the third module provides overall estimates of the balance of payments, the financing necessary to cover the external resource gap, the debt service ratio, and other macroeconomic impacts.  Macroeconomic Assumptions Balance of Paynments BP Debt Service Ratio Acrcounming oQResource Gap 4 Finance FrornewAork Energy Sector Investment, GOP Debt Service, Oil Import Bill Energy Sector RESGEN Price Assumptions Energy Model EFAM ' Assumptions | Subsector |l Models 1 lEnergy Institutional Balane i+ Financial Impacts Electric Refinery Fuetwoo | Demnd SectoF Figure 3.6. Energy-Economy Modelling Framework  - 64 - The RESGEN energy model passes the domestic energy transactions (e.g. sales from the CPC to the distributors, or CEB sales to consumers) to the EFAM Energy Finance Assessment Model, which assembles a picture of the energy sector financial flows among the key institutions and consumers in the sector. EFAM also provides estimates of the Government revenue from the sector (from petroleum product taxes, the Business Turnover Tax on energy institutions, customs duties, etc.). The financial flows are obtained by multiplying the physical energy flows (from RESGEN) by the appropriate set of prices and taxes. EFAM can therefore be used to assess the implications of alternative energy pricing and taxation schemes. As observed in Chapter 2, one of the unique features of RESGEN is that it allows three different types of demand structure (that can be used simultaneously): (i) econometric specifications (similar to ENERPLAN); (ii) project specific demands (that allows the user to modify demands according to the specific fuel inputs (or outputs or fuel substitutions or fuel use reductions from conservation); and (iii) process models (in which fuel demands are built up from a projection of end-use devices). The basis for the electric sector calculations is a plant-specific dispatch into a linearized load duration curve. The dispatch algorithm maps individual plants into the vertical blocks of the load duration curve based on the respective capacity availabilities. Again, with the institutional considerations in mind, the plant list takes the form of alternative capacity expansion plans as might be generated by a more detailed capacity expansion model such as WASP. There is explicit consideration of the energy constraints of hydro plants, and the algorithm can be shown to provide the optimum dispatch in mixed hydro-thermal systems. The Scenarios Since one of the most important aspects of the energy planning process is the treatment of uncertainties, an initial distinction is made between the policy options that lie under the control of the Sri Lanka Government, and those external factors over which the decision-maker has no control. Among the most important of these external factors are the world oil price, the overall state of the world economy that determines the external trade environment, and such factors as the level of interest rates on external debt. Because of the dependance of the electric sector on hydro electricity, rainfall variations represent another important source of uncertainty.  - 65 - It is also important to note that the analytical study reported here was carried out in early 1985. Although this volume also includes some recent information, the presentation of the original analytical work has been left essentially untouched so that the reader may follow the process by which decisions were made, based on the technical studies available at the time. For example, the world oil price scenarios that are presented below do not reflect the prices declines of early 1986. Thus, the approach taken in this volume is not to revise these scenarios with the benefit of perfect hindsight, but rather to assess the degree to which the decision taken on the basis of the best forecasts available in 1985 have proven to be sufficiently robust. Economic Growth. Ideally, one would wish to project the domestic economic growth rate as a function of domestic policy and the external environment. However, the present state of the art of macroeconomic modelling is such that great difficulties are still encountered in quantifying some of the key relationships. Indeed, the experience of Sri Lanka since the economic liberalization of 1977-1978 disproves much of the common wisdom concerning the short-term impact of external shocks, and oil price increases in particular, on the rate of economic growth. Due to these reasons, the domestic growth rate was used as one of the exogenous assumptions of the analysis: Three scenarios of GDP growth were posited over the period 1985-1995 ranging from a high case of 5.5% GDP growth through 1995 to a low case of 3.5% growth to 1990, falling to 2.5% in the early 1990's. The base case assumes a 5.5% growth from 1984-1988, which is consistent with the projection of the Ministry of Finance and Planning (MFP) in its last annual report on Public Investment, available at the time this study was carried out. From 1989 to 1995 it is assumed that the rate falls to 4% per year. This scenario implies a continuation of the growth trend over the 7 years since the economic reforms were introduced by the incoming administration in 1978. Even though this trend was maintained during a period of severe external shocks--the oil price increase of 1979-1980, sharp increases in the price of imported goods due to worldwide inflation--as well as falling world market shares of the international tea and rubber trades, growth rates in the late 1980's in excess of this trend seemed most improbable. However, a case in which the 5.5% rate is maintained through the mid 1990's is examined. There were, on the other hand, a number of reasons for the GDP growth rate to be substantially less. There was no guarantee that Sri Lanka's significant growth in manufactured exports would continue, which together with the increasing external payments burden may have begun to cause dislocations  200 (9 1coJ - b%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a (- I 0~~~~~~~~~~~~~~~~~ 0 I I a I- I I I I S- I I-- 73 75 77 79 81 83 85 87 89 91 93 95 Yeafs Figure 3.7. GDP Scenarios  - 67 - in the economy. The low scenario therefore posited a GDP growth rate of 3.5% to 1989, falling to 2.5% in the period 1990-1995. The GDP scenarios, in constant 1984 prices, are illustrated in Figure 3.7. It should, of course, be noted that these scenarios are not in the nature of forecasts or predictions, they were used in this study for the sole purpose of examining the consequences of energy policy options under different sets of conditions(4). Oil Price. The future condition of the world oil market is extremely difficult to forecast. Moreover, the complexity of the market, with the intricate interactions between the prices of different crudes and petroleum products, further complicates the situation for a small importing country such as Sri Lanka. As discussed in more detail in Chapter 5, even without some cataclysmic disruption to the world oil supply situation, significant near-term changes could occur in the petroleum product market in Southeast Asia as the new refineries in the Persian Gulf come on stream. Therefore, although the landed price of crude was used as the index of world oil price, it will be shown that the interplay of petroleum product prices around this level has- implications for Sri Lanka of the order of millions of dollars per year. In the high world oil price (WOP) case, constant real prices are assumed until 1988, followed by an annual increase of 4% thereafter (again in real terms). This brings the 1995 price to about the levels experienced at the very peak of the oil crisis in the early 1980s. (The 1984 average c.i.f. Colombo crude price was US$214/tonne(5)). The low WOP case assumes constant real prices until 1989, with a 1% per annum real price increase thereafter. Again it should be stressed that these scenarios were used (and are reported here) for indicative purposes only, rather than to imply a judgement as to the likely price path of the world oil price. Indeed, the entire thrust of the national energy strategy is the identification of policy initiatives that are robust under the expected uncertainties. The two scenarios examined are shown on Figure 3.8. Hydrological Uncertainties A national energy strategy must be concerned with short-term as well as long-term issues, of which temporary disruptions to the supply are among the most important. As Sri Lanka becomes more dependant on hydro electricity, the ability of the Ceylon Electricity Board to cope with drought years also becomes more important since most of the major iimpoundments are concentrated in a relatively small area of the hill country. The impact of a failure, delay or abnormal monsoon can be potentially quite serious, as  ~260 - 240 ~ ~ ~ ~ ~ ~ ~ ~~~~~~~4 ~220 n ~~~~~~~~~soo 200 _ WO0 1962 1964 1986 1968 1990 1992 1994 ye3rs Figure 3.8. World Oil Price Scenarios  - 69 - evidenced by the situation in 1983. Two cases of hydrological outcomes were examined: the current planning basis of the CEB that is based on firm plus 25% of secondary hydro energy (see chapter 5 for details of hydroelectric generation), and a case based on firm hydro only (that corresponds roughly to a 1 in 50 year hydrological event). It should, of course, be noted that there is a 1 in 5 chance that this event will in fact occur at some time over the next 10 year period. Table 3.7 presents the energy levels from hydro plants associated with these two cases. Table 3.7 Energy from Hydroelectric Plants Base Case Firm Hydro Only (Firm + 26S Secondary) (1 In 60 years) Non-Mahawoll hydro plants 1697 1509 Victoria 767 696 Kotmale 868 815 Randenigala 390 3ee Broadlands 95 86 Rantembe 18 166 Source: CEb The Analytical Design The analytical design for the model runs is shown on Figure 3.9. In constructing the overall scenarios, one obvious issue is the degree to which the world oil price and GDP assumptions are correlated. Increases in the oil price is only one of several types of external shocks that Sri Lanka has experienced (and will experience in the future), and the adjustment mechanisms to these external shocks can be quite complex. Thus, it is not possible to forecast a direct impact on GDP from a given oil price shock (as the experience of Sri Lanka in the years 1977-1982 reflects very well). Indeed, the impact of the relative ease with which Sri Lanka was able to adjust to the shocks of the 1970's, of which additional external financing was a major part, will only now be felt as the debt service obligations become due. For the same reason, the fact that oil prices are currently falling in real terms does not necessarily imply a faster domestic growth rate.  Base Conservation Hig/.5 % \h, /